100% found this document useful (4 votes)

7K views485 pages

Aij Recommendations For Loads On Buildings 2015

Q: What are the main distinctions between the 'ultimate limit state' and the 'serviceability limit state' in structural design?

The ultimate limit state in structural design refers to the maximum load-bearing capacity of a building, which is closely associated with safety concerns. It is synonymous with reaching the point where the structure may fail due to excessive loads. On the other hand, the serviceability limit state involves conditions under which a structure is considered usable and comfortable for occupation but may not involve a risk of collapse. This includes deformation, vibrations, or other service-related issues that affect the building's intended performance without risking structural failure .

Q: In what ways can the shape and slope of a roof influence the snow load calculations for structural design?

The shape and slope of a roof significantly influence snow load calculations through various factors. Roof slopes determine the sliding behavior of snow; a steeper slope promotes sliding, reducing snow load, while flatter roofs retain more snow . For slopes less than 10°, snow sliding is minimal, leading to increased snow weight on the roof, whereas slopes over 25° assist in snow sliding, decreasing the snow load . The shape coefficient, which adjusts the snow load from ground to roof, is contingent on the roof’s geometry and slope; irregular shapes can lead to uneven snow distribution and increased partial loads . Shape coefficients are dependent on the roof shape and can be influenced by meteorological conditions, such as wind and temperature, affecting how snow accumulates . M-shaped, multiple pitched, and multi-span roofs require specific coefficients because of their complex snow loading patterns . Thus, both roof shape and slope are critical in determining accurate snow load designs.

Q: What considerations are involved in evaluating torsional wind loads for buildings?

Evaluating torsional wind loads involves understanding how wind-induced vortices affect the structure's rotation around its vertical axis. Considerations include the building's aspect ratio, natural frequency, and the period of imposed vortices. High-rise buildings are particularly susceptible due to uniform vortex shedding, which may resonate with the building's natural modes. Accurate assessment requires accounting for both the structural stiffness and damping capacity to ensure that torsional loads do not lead to excessive vibrations or structural damage .

Q: Discuss how return periods influence the choice of basic load values in structural design.

Return periods are critical in determining basic load values as they represent the statistical likelihood of a particular loading event occurring within a given timeframe. For instance, a return period of 100 years is typically used to calculate the basic load value, meaning the load is expected to be exceeded once in 100 years. This expectation provides a balance between safety and practicality, ensuring structures are designed to handle events that are statistically likely to occur during their lifespan, thus aligning with the longevity and safety requirements of building standards. European and American codes sometimes use a 50-year return period, acknowledging that a 100-year load event may be rare but still probable during a building's lifecycle .

Q: How do load factors and resistance factors contribute to the reliability of structural designs?

Load factors and resistance factors are integral to ensuring the reliability of structural designs by accounting for uncertainties. Load factors are partial safety factors applied to cover the variability and uncertainty associated with the load estimations. They consider potential deviations in load magnitude. Resistance factors, similarly, account for uncertainties in the structure's resistance, including material strengths and construction quality. Together, these factors allow engineers to design structures to a target performance level, ensuring an acceptable probability of not exceeding the specified limits regardless of these uncertainties .

Q: What role do wind force coefficients play in the structural design of buildings with rectangular plans?

Wind force coefficients are crucial in determining how wind pressures are distributed across buildings with rectangular plans. These coefficients, defined based on aspect ratios rather than height, reflect how wind interacts with building surfaces under various wind conditions and structural geometries. Accurate determination of wind force coefficients ensures that the building's design can effectively counteract the lateral pressures imposed by wind, thus preventing structural failures related to wind loads .

Q: How does the occurrence of vortex-induced vibrations influence the structural design of high-rise buildings?

Vortex-induced vibrations occur when vortices generated in the wake of a building cause oscillations at frequencies nearing the building's natural frequency, leading to across-wind and torsional forces. This phenomenon is more pronounced in high-rise buildings due to their aspect ratios, contributing to increased periodicity and uniformity of vortices. Designing to mitigate these vibrations involves incorporating features that alter the flow patterns, such as aerodynamic modifications, as well as ensuring the structure can withstand such dynamic actions without compromising safety or performance .

Q: What methods are used for controlling snow loads on roofs, and how are they factored into design calculations?

Snow load control on roofs can be managed using mechanical removal, intentional snow sliding, and snow melting techniques. These methods aim to reduce the snow load by physically removing or altering accumulations. In structural design, the effectiveness of these control measures is factored into calculations by considering initial load expectations and reduction capabilities of each technique, with planners deciding on the design snow load based on these considerations and effectiveness. This ensures structures are efficiently designed to manage snow impacts while maintaining structural integrity .

Q: How are ground snow weights estimated using daily precipitation and mean air temperature data, and what are the implications for structural design?

Ground snow weights are estimated using daily precipitation and mean air temperature by applying equations that calculate the equivalent unit weight for ground snow. Specifically, methodologies such as those proposed by Joh & Sakurai involve integrating precipitation data with air temperatures below a certain threshold to determine the accumulated snow masses. This approach allows considerations of meteorological variations in snow load calculations, enabling more precise structural designs that account for real-time environmental conditions affecting snow accumulation and weight .

Q: Explain the significance of damping factors in the analysis of wind-induced vibrations in buildings.

Damping factors are crucial for analyzing wind-induced vibrations in buildings as they represent the reduction of vibration amplitude over time due to energy dissipation. Higher damping factors reduce the amplitude of vibrations, which is vital for ensuring structural stability and comfort. They influence the building's response to wind loads, particularly for structures with low natural frequency or large aspect ratio like tall buildings or chimneys, where vortex-induced vibrations can occur. The damping factor affects the non-dimensional critical wind speed used to evaluate vortex-induced vibration and aeroelastic instability. Accurate assessment of these factors ensures that structures can withstand dynamic wind effects without excessive movements that could affect structural integrity or habitability .

Loads on Buildings 2015

Uploaded by

sawmyo2007

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

100% found this document useful (4 votes)

7K views485 pages

Aij Recommendations For Loads On Buildings 2015

Loads on Buildings 2015

Uploaded by

sawmyo2007

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

ISBN978-4-8189-5003-0

AIJ Recommendations
for Loads on Buildings (2015)

AIJ-2019

Architectural Institute of Japan

Copyright © 2019 Architectural Institute of Japan. All Rights Reserved.
──
Recommendations for Loads on Buildings (2015)
──
Architectural Institute of Japan
5-26-20, Shiba, Minato-ku, Tokyo 108-8414, JAPAN
Tel: +81-3-3456-2051
Fax: +81-3-3456-2058
[Link]
──
Issued March 25, 2019
ISBN 978-4-8189-5003-0 C3052
──
Edited and published by Architectural Institute of Japan
──
Notice
Copyright of this publication is proprietary to the Architectural Institute of Japan and is
protected under the Copyright Act of Japan and other acts and conventions. Private use by the
individual or other use (including reproduction, modification, or distribution) of the publication
beyond the scope permitted by the Copyright Act is prohibited without the permission of the right
holder.
The Architectural Institute of Japan has made every effort to assure the accuracy of this
publication and shall not be liable for any decision or act that the User makes by using the
information contained in this publication.
Preface

This is the English version of the AIJ Recommendations for Loads on Buildings, which was revised in
February 2015. This document includes the entire text of the recommendations and a part of its
commentary. The first edition of the recommendations was published in 1975 and revisions regarding
wind loads and snow loads were made in 1981 and 1986, respectively. The third edition was published
in 1993, in which revisions considering the principles listed below were incorporated; the English
version of the third edition was published in 1996.

1. The design loads treated here are to be used in static structural analysis. Therefore, wind loads and
seismic loads are evaluated as equivalent static loads.
2. To consider various design methods, the design loads, estimated as objectively as possible
considering that design loads are common among various loads, must be provided. To ensure this,
a probabilistic/statistical method is applied uniformly and the concept of “basic load value” is
introduced, which indicates the characteristic value for each load based on statistical estimation.
3. Two design procedures are considered herein: A deterministic design procedure represented by
the allowable stress design, and a probabilistic design procedure represented by the limit state
design. The rationale in specifying design loads for each of these two procedures is presented.
4. For formulating or quantifying individual loads, “average” or average-like values with a specified
return period of each load or of its parameters are indicated, and the information regarding their
“variability” is provided.

In 2004, reparability was introduced as a performance indicator for buildings, in addition to safety and
serviceability, and the limit state design was treated as the main design procedure. In 2015, the fifth
edition of the recommendations was published with the following revisions.
1. Introduction of the concept of robustness as a performance indicator for buildings.
2. Introduction of the concept of accidental action.
3. Creation of a new chapter for tsunamis.
4. Creation of a new chapter for impulsive load.

Table 1 Publishing History of the Recommendations for Loads on Buildings

Japanese version English version
1st. edition 1975
2nd. edition 1981 (1986 for snow loads)
3rd. edition 1993 1996
4th. edition 2004 2006
5th. edition 2015 2018

It should be noted that a design load considered to be appropriate according to these recommendations
is not always legitimate under the current Japanese National Building Code.

March 2019
Architectural Institute of Japan
AIJ Recommendations for Loads on Buildings

MEMBERS OF COMMITTEES CONCERNED

RESEARCH COMMITTEE ON STRUCTURES

Chairperson Hitoshi Shiohara
Secretaries Yoshiaki Hisada, Hiroshi Isoda, Satoshi Yamada
Members (Omitted)

STEERING COMMITTEE FOR LOADS ON BUILDINGS

Chairperson Toru Takahashi (Chairperson for English version)
Co-Chair Tsuyoshi Takada (Chairperson for Japanese version (2015))
Secretaries Akie Hisagi, Hitomitsu Kikitsu, Masayuki Kohiyama
Members Hiroshi Arai, Takuji Hamamoto, Kyoko Hirata, Shuei Ikeda, Toru Ishii,
Takashige Ishikawa, Tatsuya Itoi, Jun Kanda, Akira Kokubo, Akihiro Kusaka,
Masahiro Matsui, Yasuhiro Mori, Hajime Nakajima, Naohiro Nakamura,
Hideo Nakashima, Yasuo Okuda, Naoki Satake, Tsuyoshi Takada,
Tetsuro Tamura, Tetsuro Taniguchi, Takuya Tsutsumi

WORKING GROUP MEMBERS FOR EDITING ENGLISH VERSION (2018)

Chapter 1 GENERAL CONCEPTS and Chapter 2 LOADS AND LOAD COMBINATIONS

WG on English Version
Chairperson Yasuhiro Mori
Secretary Masayuki Kohiyama (Chairperson for Japanese version (2015))
Members Kyoko Hirata, Takashige Ishikawa, Tatsuya Itoi, Akihiro Kusaka,
Tsuyoshi Takada, Toru Takahashi

Chapter 3 DEAD LOADS and Chapter 4 LIVE LOADS

Chairperson Akie Hisagi
Members Hideki Idota, Takashige Ishikawa, Takeshi Kikuchi, Youjirou Kikuoka,
Akira Kokubo, Hiroaki Ohta, Kazushige Yamamura, Yutaka Yokoyama

Chapter 5 SNOW LOADS

Chairperson Hajime Nakajima (Chairperson for English version)
Co-Chair Toru Takahashi (Chairperson for Japanese version (2015))
Secretaries Takahiro Chiba, Takayuki Sagawa
Members Koichiro Ishikawa, Hirotoshi Kikuchi, Tatsuya Kotake, Hiroki Matsushita,
Hirozo Mihashi, Kiyotoshi Otsuka, Yoshihide Tominaga, Takuya Tsutsumi
AIJ Recommendations for Loads on Buildings

Chapter 6 WIND LOADS

WG on English Version of Wind Load Provisions
Chairperson Tetsuro Tamura
Members Tomoko Aihara, Shinji Ito, Junji Katagiri, Akira Katsumura, Hiromasa Kawai,
Hitomitsu Kikitsu, Hirotoshi Kikuchi, Koji Kondo, Yuka Masuyama,
Masahiro Matsui, Koichi Miyashita, Naohiro Nakagawa, Osamu Nakamura,
Kazuo Ohtake, Akira Sarukawa, Ryoji Sasaki, Daiki Sato, Daisuke Somekawa,
Koji Takamori, Fumiyoshi Takeda, Tetsuro Taniguchi, Hiroshi Terazaki,
Yasushi Uematsu, Manabu Yamamoto

Chapter 7 SEISMIC LOADS

Chairperson Toru Ishii
Secretaries Tatsuya Itoi, Shinichi Matsushima
Members Seiichiro Fukushima, Tomoo Hirakawa, Takeshi Morii, Hirokazu Nakamura,
Shigeki Sakai, Tsuyoshi Takada, Ryoichi Tamura, Kohei Tanaka,
Hiroshi Tsunekawa

Chapter 8 THERMAL LOADS

Chairperson Hideo Nakashima (Chairperson for English version)
Co-Chair Takashige Ishikawa (Chairman for original version (2015))
Secretaries Shuei Ikeda, Satoru Nagase, Akie Hisagi
Members Hidekatsu Asai, Yukio Hayashi, Hiroo Ito, Hideo Kobayashi, Tetsuo Mochida,
Akihiro Nagata, Lee Zaijun

Chapter 9 EARTH AND HYDRAULIC PRESSURE

Hiroshi Arai, Makoto Suzuki

Chapter 10 TSUNAMI LOADS

Chairperson Yasuo Okuda
Secretaries Masanobu Hasebe, Tadashi Ishihara
Members Taro Arikawa, Tatsuya Asai, Takuji Hamamoto, Tatsuya Itoi,
Toshikazu Kabeyasawa, Hideo Matsutomi, Yoshiaki Nakano,
Kazuyoshi Nishijima, Minoru Sakata, Manabu Shoji, Tomokazu Tateno,
Hiroshi Terasaki,

Chapter 11 IMPULSIVE LOADS

Chairperson Takuji Hamamoto
Secretary Toshiya Kano
AIJ Recommendations for Loads on Buildings (2015)

Main text

March 2018

Architectural Institute of Japan

Index
Chapter 1 General concepts 1
Chapter 2 Loads and load combinations 3
Chapter 3 Dead loads 6
Chapter 4 Live loads 7
Chapter 5 Snow loads 9
Chapter 6 Wind loads 14
Chapter 7 Seismic loads 76
Chapter 8 Thermal loads 86
Chapter 9 Earth and hydraulic pressures 87
Chapter 10 Tsunami loads 90
Chapter 11 Impulsive loads 93
Chapter 12 Other loads 97
-1-

CHAPTER 1 GENERAL CONCEPTS

1.1 Scope of Applications

These recommendations apply to the estimation of loads on ordinary buildings and similar structures or
the parts thereof. Estimated loads may be used in the structural design of the buildings and their parts,
and in an assessment of the structural performance.

1.2 Fundamental Concept

1.2.1 Structural performance
Loads and their intensities shall be determined appropriately for buildings to be designed to have a
proper structural performance with respect to the following factors.
(1) Safety
Buildings shall have an appropriate degree of safety against various loads. To satisfy this requirement,
the appropriate loads and their intensities shall be evaluated. The appropriate degree of safety shall be
determined based on social and economic circumstances as well.
(2) Serviceability
Buildings shall satisfy an appropriate degree of serviceability to avoid losing their functionality under
the expected loads during normal use. To satisfy this requirement, the appropriate load intensity shall be
determined for relatively frequent loads. The appropriate degree of serviceability of buildings shall be
determined based on the particular importance, as well as the social and economic circumstances.
(3) Reparability
If necessary, buildings should be designed to maintain their reparability when they are damaged. The
appropriate degree of reparability of a building should be determined based on the importance and social
and economic circumstances.
(4) Robustness
Buildings shall be designed to be robust when suffering from unexpected type of loads or excess
intensity over their design loads.
(5) Accountability of the structural engineers to the owners regarding the designed performance level
When reflecting the requests of a building owner, the structural engineer should configure the
performance level of the building by considering the load states. The structural engineer should provide
the owner information on the results of the design.

1.2.2 Load modeling and structural analysis

Load effects in terms of stress or deformation of the structural members and/or joints are obtained from
a structural analysis based on the estimated loads. Structural analysis methods and procedures are not
specified in the recommendations; however, loads are estimated, in principle, for a static analysis.
Dynamic loads caused by strong wind, earthquake ground motions, live loads, and tsunamis, among
other factors, are evaluated for both the design as equivalent static loads.
-2- Recommendations for Loads on Buildings

1.2.3 Proper design and construction

The intensities of the design load should be evaluated by assuming that the buildings are not only
properly designed and constructed such that a reduction in safety and serviceability from human error
is minimal, but also appropriately maintained and managed.

1.3 Terms and Definitions

The terms used in the recommendations are defined as follows:
- Limit state: State beyond which either a structure or its parts no longer satisfy the prescribed
requirements related to safety or serviceability.
- Limit state design: Design for targeted safety as well as serviceability limit states.
- Load effect: Stress, deformation, displacement, and other factors induced within the structure by
single or combined loads.
- Basic value of load: Representative intensity of a load as a reference for estimation of the load
effects, essentially based on 100-year return period value of the load.
- Load factor: Partial safety coefficients by which the basic values of the load effects are multiplied
to obtain the design load effects.
- Return period: The expected time interval between which events greater than a certain magnitude
occur.
-3-

CHAPTER 2 LOADS AND LOAD COMBINATIONS

2.1 Loads
The following loads are to be considered in a structural design.
1. Dead load (𝐺𝐺)
2. Live load (𝑄𝑄)
3. Snow load (𝑆𝑆)
4. Wind load (𝑊𝑊)
5. Seismic load (𝐸𝐸)
6. Thermal load (𝑇𝑇)
7. Earth and hydraulic pressure (𝐻𝐻)
8. Tsunami load (𝑇𝑇𝑢𝑢 )
9. Impact load (𝐼𝐼)
10. Other loads

2.2 Basic Value of Load

The basic values of a load are to be determined based on the following policies:
- The basic value of a dead load (𝐺𝐺) is to be determined according to the actual conditions.
- The basic value of a live load (𝑄𝑄) is to be determined on the basis of a value giving a non-exceedance
probability of 99% under normal conditions, or by a value corresponding to this condition if the
former value is difficult to statistically evaluate.
- The basic value of a snow load (𝑆𝑆) is to be determined on the basis of a 100-year recurrence value
of the ground snow depth.
- The basic value of a wind load (𝑊𝑊) is to be determined on the basis of a 100-year recurrence value
of the 10-min mean wind speed.
- The basic value of a seismic load (𝐸𝐸) is to be determined on the basis of a 100-year recurrence value
of the horizontal acceleration response spectrum on the engineering bedrock.
- The basic value of a thermal load (𝑇𝑇) is to be determined on the basis of a 100-year recurrence value
of temperature under normal conditions, or by a value corresponding to this condition if the former
value is difficult to evaluate statistically.
- The basic value of earth and hydraulic pressure (𝐻𝐻) is to be determined on the basis of a value giving
a non-exceedance probability of 99% under normal conditions, or using a value corresponding to
this condition.
- The basic value of a tsunami load (𝑇𝑇𝑢𝑢 ) is to be determined on the basis of a 100-year recurrence
value of only the inundation depth, or the inundation depth and flow velocity, at the construction
site brought about by a (seafloor) earthquake or by a value corresponding to this condition if it is
difficult to statistically evaluate.
- The basic value of an impact load (𝐼𝐼) is to be determined on the basis of a value giving an exceedance
- 4 - Recommendations for Loads on Buildings

probability of 1% for accident occurrence, or by a value corresponding to this condition if it is

difficult to statistically evaluate.
- The basic values of the other loads are evaluated as described above.

2.3 Load Combination and Required Performance

(1) The load combination used for designing and evaluating an entire building or its parts is determined
according to the required performance level of the target building or its parts.
(2) The required performance level is determined by the structural engineer in consensus with the owner,
and in consideration of the use of the building or its parts, its importance, sociality, service life, economic
efficiency, degree of influence at the loss of the target performance, and relationship with the design
methods.
(3) In addition to a persistent load, the following loads should be considered depending on the actual
conditions. The state in which some of the loads act simultaneously or with a time lag when the influence
of the first load remains even after its unloading should be considered, if necessary.
- Extra-ordinary live load
- Snow load
- Wind load
- Seismic load
- Thermal load
- Earth and hydraulic pressure
- Tsunami load
- Impact load
- Other accidental load
(4) When other accidental loads are taken into account, they are combined with persistent loads .

2.4 Load Factors

2.4.1 Load factors for use in limit state design format
(1) When designing a building and/or structural member, the following limit states shall be considered
appropriately for each load combination.
- Ultimate limit state
- Serviceability limit state
(2) A target performance level shall be defined using a target reliability index considering the
predetermined reference period.
(3) Load combinations shall be considered as the sum of the products of the load effect corresponding
to the basic value of each load and the load factor for the load effect, namely,

𝛾𝛾p 𝑆𝑆pn + ∑𝑘𝑘 𝛾𝛾𝑘𝑘 𝑆𝑆𝑘𝑘n, (2.1)

in which 𝑆𝑆pn and 𝑆𝑆𝑘𝑘n are load effects from the basic value of a principal load and a 𝑘𝑘-th secondary
CHAPTER 2 LOADS AND LOAD COMBINATIONS - 5 -

load, respectively, and 𝛾𝛾p and 𝛾𝛾𝑘𝑘 are the corresponding load factors.
The load factors shall be determined appropriately by considering a target reliability index for a
considered limit state, the variability in the load effect of each load and resistance, the probability of
coinciding loads, and other elements.

2.4.2 Load factors for use in allowable stress and ultimate strength designs
(1) The return period of a principal load shall be determined appropriately by considering the service
life; importance; failure consequences of the buildings; properties of each load, such as the variability,
occurrence frequency, and duration time; and the bases of the allowable stress and load bearing capacity
used in the design.
(2) Load combinations shall be considered as the sum of the products of the load effect corresponding
to the basic value of each load and the load factor for the load effect, namely,

𝑘𝑘Rp 𝑆𝑆pn + ∑𝑘𝑘 𝛾𝛾𝑘𝑘 𝑆𝑆𝑘𝑘n, (2.2)

in which 𝑆𝑆pn and 𝑆𝑆𝑘𝑘n are the load effects owing to the basic value of a principal load and a 𝑘𝑘 -th
secondary load, respectively; 𝑘𝑘Rp is the return period conversion factor for the principal load; and 𝛾𝛾𝑘𝑘
is the load factors for a 𝑘𝑘 -th secondary load. When the load effect is not proportional to the basic
physical quantity such as a wind load, the load effect calculated using the basic value of the load factored
by the return period conversion factor corresponds to 𝑘𝑘Rp 𝑆𝑆pn.
(3) The return period conversion factors for a snow load presented in Clause A5.3, for a wind load
presented in Clause A6.1.4, and for a seismic load presented in Clause 7.2.4 are to be used for a design
return period 𝑡𝑡R differing from 100 years. The return period conversion factors for the other loads are
to be calculated according to the actual conditions.

2.4.3 Load factor of secondary thermal load

Regardless of the limit state design method, the allowable stress design method, or the ultimate strength
design method, the following values are used as the load factor of a secondary thermal load.
- Load combination with a snow load of 0.8
- Load combination with a wind load of 0.7
- Load combination with a seismic load of 0.4
-6-

CHAPTER 3 DEAD LOADS

3.1 Setting and calculation of dead load

Building dead load should be estimated based on the actual conditions of the building.
-7-

CHAPTER 4 LIVE LOADS

4.1 General

4.1.1 Definition
Live loads are vertical loads from furniture and people etc., which act randomly and vary with time
and space during the lifetime of the building. They are determined for each element of a structure when
considering the design limit states, the particular use of the building, the temporary concentration of
people and furniture, and their dynamic effects.

4.2 Estimation of Live Load

4.2.1 Basic value of live load

The basic value Q (N/m2) of a live load is estimated from the following equation:
Q = ke ka kn Q0, (4.1)
where
ke : conversion factor to equivalent uniformly distributed load (see 4.2.3),
ka : area reduction factor (see 4.2.4),
kn : multi-story reduction factor (see 4.2.5),
Q0 : basic live load intensity (N/m2) (see 4.2.2).
However, kn can be used for only calculating the axial force of the column and the vertical loads of
the foundations.

4.2.2 Basic live load intensity

The basic live load intensity Q0 is the 99th percentile of the sustained live load intensity for the
particular use, as shown in Table 4.1.

Table 4.1 Basic live load intensity Q0

Categories* ① ② ③ ④ ⑤ ⑥ ⑦ ⑧
Q0 (N/m2) 1,000 500 1,600 2,100 3,500 2,200 4,700 1,800
*Category means one of the following uses:
① residential flats, dwellings
② hotel rooms (not including unit baths)
③ offices, laboratories
④ supermarkets, department stores
⑤ computer rooms (not including cables)
⑥ lane loads and parking garage spaces
⑦ library rooms
⑧ buildings where loads are produced mainly by people, for example, theaters, cinemas,
halls, assembly rooms, conference rooms, and classrooms
- 8 - Recommendations for Loads on Buildings

4.2.3 Conversion factor for equivalent uniformly distributed load

Conversion factors for equivalent uniformly distributed load ke are listed in Table 4.2.

Table 4.2 Conversion factor to equivalent uniformly distributed load

Categories* ① ② ③ ④ ⑤ ⑥ ⑦ ⑧
Slab 1.8 2.0 1.6 1.8 1.6 1.8 1.6 1.6
Girder, Column, Foundation 1.2
*Categories are given in Table 4.1

4.2.4 Area reduction factor

Area reduction factor ka can be calculated from the following equations.
(1) For categories ① through ⑦ in Table 4.1
0.4
𝑘𝑘a = 0.6 + ≦ 1.0 (4.2)
�𝐴𝐴f /𝐴𝐴ref

(2) For category ⑧ in Table 4.1

0.3
𝑘𝑘a = 0.7 + ≦ 1.0 (4.3)
�𝐴𝐴f /𝐴𝐴ref

where Af is the influence area (m2) and Aref is the basic area (=18 m2).

4.2.5 Multi-story reduction factor

Multi-story reduction factor kn can be calculated from the following equation.
0.4
𝑘𝑘n = 0.6 + (4.4)
√𝑛𝑛
where n is the number of supported floors.
No reduction shall be permitted for category ⑧ in Table 4.1, and ka kn > 0.4 for all categories.

4.3 Live Loads Considering Concentration, Deflections, or Cracks

When the effects of furniture concentration, deflections, or cracks must be considered, it is necessary
to determine the appropriate intensity of the design live loads considering such effects.

4.4 Dynamic Effects of Live Loads

With regard to the dynamic effects of a live load, the effects of movement of people and objects
must be considered when it is necessary to evaluate the serviceability performance of the building in
relation to vibrations, such as habitability for the occupants and counter-vibration measures for precision
equipment. It is also desirable to consider the influence of the ambient environment and the source (or
sources) of vibrations located on other floors inside the building.
-9-

CHAPTER 5 SNOW LOADS

5.1 Scope and Procedure

A snow load can be categorized as (1) a snow load on a roof, (2) a partial snow load on a roof, (3) and
other snow loads. A snow load on a roof shall be defined by the product of the snow load on the ground
for the region considered, and a shape coefficient. It can be reduced if the snow load is controlled with
certainty.

5.1.1 Basic snow load on a roof

The basic snow load on a roof 𝑆𝑆(kN/m2) is given by the following equation:

𝑆𝑆 = 𝜇𝜇0 𝑆𝑆0, (5.1)

where
𝜇𝜇0 : shape coefficient, as defined in 5.3
𝑆𝑆0 (kN/m2): snow load on the ground, as defined in 5.2

5.2 Snow Load on the Ground

5.2.1 Equations for snow load on the ground
Snow load on the ground 𝑆𝑆0 is determined from the following equation:

𝑆𝑆0 = 𝑘𝑘env 𝑑𝑑0 𝜌𝜌0 , (5.2)

where
𝑘𝑘env: environmental coefficient, as defined in 5.2.4,
𝑑𝑑0 (m): basic snow depth on the ground, as defined in 5.2.2,
𝜌𝜌0 (kN/m3): equivalent unit weight for ground snow, as defined in 5.2.3.
The values of 𝑑𝑑0 and 𝜌𝜌0 can be defined through an appropriate method using data on the precipitation
and temperature.

5.2.2 Basic snow depth on the ground

Basic snow depth on the ground d0 is defined as the annual maximum value for the entire season with
a 100-year recurrence interval, and is estimated from meteorological data of the ground snow depth
observed for a certain period.

5.2.3 Equivalent unit weight for ground snow

Equivalent unit weight for ground snow 𝜌𝜌0 is given by the following equation:

𝜌𝜌0 = 0.72�𝑑𝑑0 ⁄𝑑𝑑ref + 2.32, (5.3)

where
- 10 - Recommendations for Loads on Buildings

𝑑𝑑0 (m): basic snow depth on the ground,

𝑑𝑑ref (m): reference snow depth (= 1 m).

5.2.4 Environmental coefficient

The environmental coefficient 𝑘𝑘env is generally defined as unity. When the snow depth on the ground
is estimated to locally increase because of geographical features, man-made and natural features, 𝑘𝑘env,
should be correspondingly larger than unity.

5.3 Shape coefficient

Shape coefficient 𝜇𝜇0 is defined by the following equation:

𝜇𝜇0 = 𝜇𝜇b + 𝜇𝜇d + 𝜇𝜇s , (5.4)

where
𝜇𝜇b : basic shape coefficient defined in 5.3.1,
𝜇𝜇d : shape coefficient for irregular distribution caused by snow drifting defined in 5.3.2,
𝜇𝜇s : shape coefficient for irregular distribution caused by sliding defined in 5.3.3.
The shape coefficient for large or specially shaped buildings should be determined from the special
field researches or experiments.

5.3.1 Basic shape coefficient

Basic shape coefficient 𝜇𝜇b is given in Fig. 5.1. In the figure, wind speed 𝑉𝑉(m/s) indicates the
average wind speed in January through February. For an intermediate value of 𝑉𝑉 , 𝜇𝜇b should be
determined through a linear interpolation from the lines with the specific velocities.

Figure 5.1 Relation between 𝜇𝜇b and roof slope.

5.3.2 Shape coefficient for irregular distribution caused by snow drift

(1) The shape coefficients for an irregular distribution caused by snow drift 𝜇𝜇d in the troughs of an
Chapter 5 SNOW LOADS - 11 -

M-shaped, multiple pitched, and multi-span roofs are given in Table 5.1. As shown in Figs. 5.2(a)
and 5.2(b), at a ridge, 𝜇𝜇d should be zero. At the halfway point, 𝜇𝜇d is calculated through linear
interpolation. For a fractional value of 𝑉𝑉, 𝜇𝜇d should also be determined using linear interpolation.
(2) For a multilevel roof, the distribution of the shape coefficient for a lower level roof should be
determined, as shown in Fig. 5.3. In the figure, 𝜇𝜇d at point O is given in Table 5.2. For an
intermediate value of 𝑉𝑉, 𝜇𝜇d should be determined through linear interpolation.

Table 5.1 The value of 𝜇𝜇d for the troughs of an M-shaped, multiple pitched, and multi-span roofs
Roof M-shaped and multiple pitched roofs Multi-span roof
slope average wind speed in Jan. through Feb. average wind speed in Jan. through Feb.
≤2 m/s 3 m/s 4 m/s 4.5 m/s≤ ≤2 m/s 3 m/s 4 m/s 4.5 m/s≤
≤ 10° 0 0 0 0 0 0 0 0
25° 0 0 0.15 0.20 0.10 0.20 0.35 0.55
40° 0 0.20 0.35 0.45 0.10 0.30 0.45 0.70
50°≤ 0 0.30 0.55 0.70 0.10 0.40 0.65 0.80

0 0 0 0

μd μd
μd
（a） M-shaped and multiple pitched roofs （b） multispan roof

Figure 5.2 Distributions of roof snow on M-shaped, multiple pitched, and multi-span roofs

Table 5.2 𝜇𝜇d for multilevel roofs

average wind speed in Jan. ≤2 m/s 3 m/s 4 m/s 4.5 m/s≤
through Feb.
𝜇𝜇d 0.01 0.30 0.50 0.60

hs
μd

O
2hs

Figure 5.3 𝜇𝜇d for multilevel roofs

5.3.3 Shape coefficient for irregular distribution caused by sliding

The shape coefficients for an irregular distribution caused by sliding 𝜇𝜇s on M-shaped, multiple
pitched, and multi-span roofs are determined from either Eq. (5.5) or Eq. (5.6) according to the roof
slope. The value of 𝜇𝜇s is positive in the troughs of these roofs and negative at the ridge. At the halfway
point, 𝜇𝜇s is determined through linear interpolation. When the slope of the roof is between those defined
in Eq. (5.5) and Eq. (5.6), 𝜇𝜇s is determined from the sliding performance of the roofing materials used.
- 12 - Recommendations for Loads on Buildings

(1) When the roof slope is smaller than 10°,

𝜇𝜇s = 0. (5.5)

(2) When the roof slope is larger than 25°,

𝜇𝜇s = 𝜇𝜇b . (5.6)

5.4 Partial Snow Load on a Roof

A partial snow load should be considered for the following cases:
(1) The snow load on the roof partially increases because of snowdrift caused by projecting structures.
(2) The snow load on the roof partially increases because of snow overhang, cornices, and icicles at the
eaves or verge.
(3) The snow slides from the upper roofs to the eaves or lower levels of multilevel roofs. Their distance
and impact should also be considered.

5.5 Snow Load on a Roof with Snow Control

5.5.1 Equation for snow load on a roof with snow control
Snow load on a roof with snow control is given through the following equation:

𝑆𝑆 = 𝑆𝑆n 𝜇𝜇n − 𝑆𝑆c , (5.7)

where
𝑆𝑆(kN/m2): design snow load on a roof with snow control,
𝑆𝑆n (kN/m2): characteristic snow load on the ground used for the design when the snow load on the
roof is controlled, as defined in section 5.5.2,
𝜇𝜇n : shape coefficient with snow control corresponding to 𝜇𝜇0 defined in section 5.3,
𝑆𝑆c (kN/m2): controlled snow load, as described in section 5.5.3

5.5.2 Ground snow load with accumulation for 𝒏𝒏 days

(1) Ground snow load with accumulation for 𝑛𝑛 days
When the roof snow is certainly controlled, snow load on the ground 𝑆𝑆n (kN/m2) is determined from
the following equation:

𝑆𝑆n = 𝑘𝑘env 𝑑𝑑n 𝜌𝜌n , (5.8)

where
𝑘𝑘env: environmental coefficient, as defined in section 5.2.4,
𝑑𝑑n (m): basic snow depth on the ground when the snow load on the roof is controlled, as defined in
(2) of section 5.5.2
𝜌𝜌n (kN/m3): equivalent unit weight for ground snow with roof snow control, as defined in (3) of
section 5.5.2.
Chapter 5 SNOW LOADS - 13 -

The value of 𝑆𝑆n in Eq. (5.8) can be estimated through the appropriate method using the precipitation
and temperature.

(2) Basic snow depth with accumulation for n days

The basic snow depth on the ground 𝑑𝑑n is defined as the annual maximum value of snow
accumulation for 𝑛𝑛 days with a 100-year recurrence interval, and is estimated from meteorological data
of the ground snow depth observed for a certain period.

(3) Equivalent unit weight for ground snow with roof snow control
Equivalent unit weight for ground snow with roof snow control 𝜌𝜌n is equal to 𝜌𝜌0 , as indicated in Eq.
(5.3).

5.5.3 Snow load with snow control

Controlled snow load, 𝑆𝑆c , is generally determined after a field research and experiments investigating
the capacity of the sliding or melting devices, where 𝑆𝑆c is the differential between the initial snow load
expected when heavy snow fall starts, and the snow load removed by a device whose performance is
guaranteed even during a heavy snow fall.

5.6 Other Snow Loads

The following should be considered as additional snow loads.
(1) When the lateral pressure from snow drifts on the outside wall of the building might be strong, it
should be considered.
(2) When roof snow might be integrated with snow cover on the ground, the snow load caused by
settlement should be considered.
(3) When snow adheres to the building or snow covers the member, falling adhered snow or the snow
load should be considered.
(4) When snow accumulates on external members such as eaves, the snow load and falling snow should
be considered.
(5) When snow blowing into balconies or the outside corridors might be excessive, it should be
considered.
(6) When snow drifts caused by neighboring buildings or trees might accumulate, they should be
considered.
(7) When a high-rise building and/or large roof building might cause a snowdrift, it should be
considered.
(8) For a membrane structure, a ponding phenomenon induced by the snow load should be considered.
-14-

CHAPTER 6 WIND LOADS

6.1 General

6.1.1 Scope of application

(1) This chapter describes wind loads for the design of buildings that respond elastically in strong winds.
(2) Two different wind loads are described. The first is for the design of structural frames, i.e., horizontal
wind loads on structural frames and roof wind loads on structural frames. The second is for the design
of components/cladding of buildings.

6.1.2 Estimation principle

(1) The basic value of wind load W is determined from the basic wind speed U0 given in A6.1.3, which
corresponds to the 10-min mean wind speed at 10 m above the ground over a flat open terrain with
a return period of 100 years.
(2) The design wind speed for a given return period shall be determined from the return period
conversion factor specified in A6.1.4.
(3) Wind loads for the design of buildings are individually specified for a horizontal wind loads on
structural frames, roof wind loads on structural frames, and wind loads on components/cladding. The
horizontal wind loads for the design of structural frames shall be individually determined in the
along-wind, across-wind, and torsional directions.
(4) For relatively flexible buildings with large aspect ratios, the horizontal wind loads on structural
frames in the across-wind and torsional directions given in A6.5 and A6.6 shall be considered. The
criteria for this are described in 6.1.3(1).
(5) For flexible buildings with very large aspect ratios, the structural safety against vortex-induced
vibration and aeroelastic instability shall be checked. The criteria for this are described in 6.1.3(2).
The wind loads on the structural frames and members with a round sectional shape caused by a
vortex-induced vibration shall be determined through A6.9.
(6) The horizontal wind loads on the structural frames are given by the product of the velocity pressure
given in A6.1, the wind force coefficient given in A6.2, the gust effect factor given in A6.3, and the
projected area or subject area, as shown in 6.2.
(7) The roof wind loads on the structural frames are given by the sum of the time-averaged and
fluctuating components, as shown in 6.3. The time-averaged component is given by the product of
the velocity pressure given in A6.1, the wind force coefficient given in A6.2 and the subject area.
The fluctuating component shall be determined through A6.4.
(8) The wind load on the components/cladding is given by the product of the velocity pressure given in
A6.1, the peak wind force coefficient given in A6.2, and the subject area.
(9) In principle, the wind load should be calculated based on the design wind speed for each wind
direction specified in A6.1.2.
(10) Wind loads on buildings that are constructed during only a specific season and removed during the
CHAPTER 6 WIND LOADS - 15 -

other seasons can be calculated by using the seasonal wind factors specified in A6.1.6.
(11) The reference height is generally the mean roof height of the building. The wind loads are calculated
from the velocity pressure at this reference height. However, wind loads on lattice type structures
shall be calculated from the velocity pressure at each height, as shown in A6.7.
(12) The wind loads on the structural frames of free roofs are given by the product of the velocity
pressure given in A6.1, the wind force coefficient given in A6.2, the gust effect factor given in A6.8.3,
and the subject area, as shown in A6.8.
(13) For wind loads on structural frames, the combination of horizontal wind loads in the along-wind,
across-wind, and torsional directions, and the combination of horizontal wind load and roof wind
load shall be considered according to A6.10. For the components of the cladding and structural frame,
or particular joints of the cladding and structural frames, the combination of horizontal wind load on
the structural frames and local wind load on the cladding shall be considered.
(14) When the computational fluid dynamics (CFD) technique is applied to the evaluation of the wind
loads, it is necessary to obtain sufficient accuracy for the numerical results based on A6.11.
(15) The increase in wind-induced vibration caused by neighboring buildings shall be considered from
A6.12.
(16) When the wind load shielding effects by surrounding topographies or buildings are considered, the
future changes shall be confirmed, and the shielding effect shall be investigated through the
appropriate wind tunnel test, CFD, or other suitable verification methods.
(17) The response acceleration for checking the habitability of a building against wind-induced vibration
shall be evaluated from A6.13. For this evaluation, the 1-year recurrence wind speed can be obtained
from A6.14.
(18) Cumulative loading effects such as fatigue damage should be evaluated in consideration of the
duration of the wind load, as shown in Section A6.15.

6.1.3 Buildings for which particular wind loads or wind-induced vibrations are taken into account
(1) Buildings for which horizontal wind loads on structural frames in the across-wind and torsional
directions are taken into account
For buildings that satisfy the following criteria, the wind load in the across-wind direction, as defined
in A6.5, and the wind load in the torsional direction, as defined in A6.6, shall be checked.
H
≥3, (6.1)
BD
where
H (m): reference height, as defined in 6.1.2(11)
B (m): building width,
D (m): building depth.
(2) Vortex-induced vibration and aeroelastic instability
For buildings that satisfy the following criteria, vortex-induced vibration and aeroelastic instability
- 16 - Recommendations for Loads on Buildings

shall be checked through appropriate wind tunnel tests, CFD, and so on. For buildings with a circular
section, the wind load is prescribed in A6.9.
1) Buildings with a rectangular section

H  U U 
≥ 4 and  ≥ 0.83U Lcr ≥ 0.83U Tcr ,
H * H *
or  (6.2)
BD  L
f BD f T BD 

where
U H (m/s): design wind speed as defined in A6.1.2 (wind directionality factor K D = 1 and
wind seasonal factor K S = 1),
*
U Lcr : non-dimensional critical wind speed for aeroelastic instability in the across-wind
direction calculated from Table 6.1,
*
U Tcr : non-dimensional critical wind speed for aeroelastic instability in the torsional direction
calculated from Table 6.2,
f L , f T (Hz): natural frequency for the first mode in the across-wind and torsional
directions.
2) Buildings with circular cross-section
H UH
≥ 7 and ≥ 3.5 , (6.3)
Dm f L Dm

where
D m (m): building diameter at height a of 2H / 3 .

Table 6.1 Non-dimensional critical wind speed for aeroelastic instability in the across-wind direction
*
U Lcr
*
Flat terrain Mass damping Critical wind speed U Lcr
Side ratio D / B
categories parameter δ L Note)
δ L ≤ 0.7 16 δ L
D / B ≤ 0.8
0.7 < δ L 11
0.8< D / B ≤ 1.5 - 1.2 δ L +7.3
δ L ≤ 0.2 2.3
I, II
1.5< D / B ≤ 2.5 0.2< δ L ≤ 0.8 12
0.8 < δ L 15 δ L
δ L ≤ 0.4 3.7
D / B >2.5
0.4 < δ L not necessary to evaluate
D / B ≤ 0.8 - 4.5 δ L +6.7
0.8< D / B ≤ 1.2 - 0.7 δ L +8.8
III, IV, V
1.2< D / B ≤2 - 11
D / B >2 - 15
Note that δ L is the mass damping parameter defined as δ L = ζ L M L /( ρBDH ) , where ζ L is the
damping factor for the first mode in the across-wind direction, M L (kg) is the generalized mass of the
CHAPTER 6 WIND LOADS - 17 -

building for the first mode in the across-wind direction in which the top of the mode shape is the unit
value, and ρ (1.22kg/m3) is the air density.

*
Table 6.2 Non-dimensional critical wind speed for aeroelastic instability in torsional direction U Tcr

Side ratio D / B Mass damping parameter δ T Note) *

Critical wind speed U Tcr

all δ T ≤ 0.05 1.7

0.05< δ T ≤ 0.1 9.2
D / B ≤ 1.5
0.1 < δ T not necessary to evaluate
0.05< δ T ≤ 0.15 8 δ T +3.3
1.5< D / B ≤ 2.5
0.15 < δ T 7.4 δ T +7.2
2.5< D / B ≤ 5 0.05 < δ T 10.5 δ T +4.2
Note that δ T is the mass damping parameter defined as δ T = ζ T I T /( ρB 2 D 2 H ) , where ζ T is the
damping factor for the first mode in the torsional direction, and I T (kg m2) is the generalized torsional
moment of the building for the first mode in the torsional direction in which the top of mode shape is
the unit value.

6.2 Horizontal Wind Loads on Structural Frames

6.2.1 Scope of application

This section defines the procedures for estimating the horizontal wind loads on the structural frames
in the along-wind direction.

6.2.2 Procedure for estimating wind loads

Wind loads on structural frames in the along-wind direction are calculated from Eq. (6.4).
WD = qHCDGDA, (6.4)
where
WD (N): wind loads in the along-wind direction at height Z (m),
qH (N/m2): velocity pressure, as defined in A6.1.1,
CD : wind force coefficient, as defined in A6.2,
GD : gust effect factor, as defined in A6.3,
A(m2): projected area at height Z (m).

6.3 Roof Wind Loads on Structural Frames

6.3.1 Scope of application

This section defines the procedures used for estimating the roof wind loads on the structural frames
of buildings.
- 18 - Recommendations for Loads on Buildings

6.3.2 Procedure for estimating wind loads

Roof wind loads on the structural frames are calculated from Eq. (6.5)

WR = qH CR AR ± WR' , (6.5)

where
WR (N): roof wind load,
qH (N/m2): design velocity pressure, as defined in A6.1.1,

CR : wind force coefficient, as defined in A6.2,

AR (m2): roof area supported by roof beam,

W ' R (N) : fluctuating roof wind load, as defined in A6.4.

6.4 Wind Loads on Components/Cladding

6.4.1 Scope of application

This section defines the procedures for estimating the wind loads on the building
components/cladding.

6.4.2 Procedure for estimating wind loads

Wind loads on the components/cladding of a building are calculated from Eq. (6.6).
W = q Cˆ A ,
C H C C (6.6)
where
WC (N): wind load on the components/cladding,
qH (N/m2): design velocity pressure, as defined in A6.1.1,
ĈC : peak wind force coefficient, as defined in A6.2,
AC (m2): subject area of components/cladding.
CHAPTER 6 WIND LOADS - 19 -

A6.1 Wind Speed and Velocity Pressure

A6.1.1 Velocity pressure

The design velocity pressure, qH(N/m2), shall be calculated from Eq. (A6.1).
1
qH = ρU H2 (A6.1)
2

where
ρ(kg/m3): air density, which is assumed to be 1.22,
UH (m/s): design wind speed, as defined in A6.1.2.

A6.1.2 Design wind speed

The design wind speed, UH (m/s), shall be calculated from Eq. (A6.2).
U H = U 0 k RW K D K S E H (A6.2)
where
U0: basic wind speed (m/s) depending on the geographic location of the construction site, as
defined in A6.1.3,
kRW: return period conversion factor, as defined in A6.1.4,
KD: wind directionality factor, as defined in A6.1.5,
KS: wind seasonal factor, as defined in A6.1.6,
EH: wind speed profile factor at reference height H defined in A6.1.7.
However, the one-year recurrence wind speed is defined in A6.1.14 for an evaluation of the
habitability.

A6.1.3 Basic wind speed

The basic wind speed U0 (m/s) corresponds to the 100-year recurrence of a 10-min mean wind speed
at 10 m above the ground for flat open terrain. The basic wind speed at various locations in Japan is
shown in Fig. A6.1.

A6.1.4 Return period conversion factor

The return period conversion factor kRW is calculated from Eq. (A6.3).
k RW = 0.63(λU − 1) ln t R − 2.9λU + 3.9 (A6.3)
U 500
λU =
U0
where
U500(m/s): 500-year recurrence of 10-min mean wind speed at 10 m above the ground for flat
open terrain, as defined in Fig. A6.2,
U0 (m/s): basic wind speed, as defined in A6.1.3.
- 20 - Recommendations for Loads on Buildings

Izu islands, not shown in the map 46

Ogasawara Islands, Satsunann Islands,
Okinawa Islands, Daitou Islands, 50
Sakishima Islands, not shown in the map

Figure A6.1 Basic wind speed U 0 (m/s)

CHAPTER 6 WIND LOADS - 21 -

Izu islands, not shown in the map 52

Ogasawara Islands, Satsunann Islands,
Okinawa Islands, Daitou Islands, 58
Sakishima Islands, not shown in the map

Figure A6.2 500-year recurrence 10-minute mean wind speed U500 (m/s) at 10 m above the ground
for flat open terrain
- 22 - Recommendations for Loads on Buildings

A6.1.5 Wind directionality factor

The wind directionality factor K D reflects the directional characteristics of an extreme wind, which
are influenced by the geographical location and topographic features of the construction site. It shall be
determined with reference to the wind directionality factors for the eight cardinal directions shown in
Table A6.1. When the wind seasonal factor KS is considered, KD = 1.
When the aerodynamic shape factors for each wind direction are known from an appropriate wind
tunnel experiment or CFD the following hold.
(1) When the aerodynamic shape factors for each wind direction are known
The wind directionality factor K D , which is used to evaluate the wind loads on structural frames and
components/cladding for a particular wind direction, shall take the same value as that for the cardinal
direction whose 45° sector includes this same wind direction.
(2) When the aerodynamic shape factors in A6.2 are used
1) When assessing wind loads on structural frames, the following occur.
a) When the aerodynamic shape factors are dependent on the wind direction, four wind directions
that coincide with the principal coordinate axis of the structure should be considered. If the wind
direction is within a 22.5° sector centered at one of the eight cardinal directions, the wind
directionality factor K D for this direction shall be adopted. If the wind direction is outside the
22.5° sector, the larger of the values for the two nearest cardinal directions shall be adopted.
b) When the aerodynamic shape factors are independent of the wind direction, the wind directionality
factor K D shall take the same value as that for the cardinal direction whose 45° sector includes
that wind direction.
2) When assessing wind loads on cladding,
KD = 1.
Table A6.1 Wind directionality factor K D
Wakkanai Kitamiesashi Haboro Omu Rumoi Asahikawa Abashiri Otaru Sapporo Iwamizawa
NE 0.95 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.9
E 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85
SE 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85 1 1
S 1 0.85 1 0.85 0.85 1 0.85 0.85 1 1
SW 1 1 1 1 0.95 1 0.85 1 0.85 0.85
W 0.85 1 1 1 0.95 1 0.95 1 0.95 1
NW 0.85 0.85 0.95 0.85 1 0.85 1 1 1 0.95
N 0.95 0.85 0.85 0.85 1 0.85 0.9 0.85 0.85 0.9

Obihiro Kushiro Nemuro Suttsu Muroran Tomakomai Urakawa Esashi Hakodate Kutchan
NE 0.85 0.85 0.9 0.85 0.85 0.85 0.85 0.85 0.95 0.85
E 0.85 0.9 0.9 0.85 0.85 0.85 0.85 0.85 0.95 0.95
SE 0.85 0.9 0.85 1 0.85 0.85 0.85 0.85 0.85 0.95
S 0.85 0.85 0.85 1 0.85 0.85 0.85 1 1 0.95
SW 0.85 0.85 0.85 0.85 0.85 0.85 0.85 1 1 0.95
W 1 1 0.95 1 1 0.85 1 1 0.95 1
NW 1 0.9 1 1 1 0.9 1 1 0.9 1
N 0.85 0.85 0.95 0.9 0.85 1 0.85 0.85 0.85 0.85
CHAPTER 6 WIND LOADS - 23 -

Table A6.1 (continued) Wind directionality factor K D

Mombetsu Hiroo Ofunato Shinjo Wakamatsu Fukaura Aomori Mutsu Hachinohe Akita
NE 0.9 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85
E 0.9 0.85 0.85 0.85 0.9 0.85 0.85 0.85 0.85 0.85
SE 0.85 0.85 0.85 0.85 0.9 0.85 0.85 0.85 0.85 0.85
S 0.85 0.85 0.85 0.85 0.85 1 0.85 0.85 0.85 0.95
SW 1 1 0.85 0.85 0.85 1 1 0.95 1 0.95
W 1 1 0.85 1 1 0.95 1 1 1 1
NW 1 0.85 1 1 1 0.95 0.9 0.85 0.95 1
N 0.9 0.85 0.9 0.85 0.85 0.9 0.85 0.85 0.85 0.85

Morioka Miyako Sakata Yamagata Sendai Ishinomaki Fukushima Shirakawa Onahama Wajima
NE 0.85 0.85 0.85 0.85 0.85 1 0.85 0.85 1 0.9
E 0.85 0.85 0.85 0.85 0.85 0.95 0.85 0.85 0.85 0.85
SE 0.85 0.9 0.85 0.85 0.85 0.9 0.85 0.85 0.9 0.85
S 0.85 0.9 0.85 0.9 0.85 0.85 0.85 0.85 0.85 0.9
SW 0.95 0.95 0.9 1 0.85 0.85 0.85 0.85 0.85 1
W 1 1 1 1 1 1 1 1 0.95 1
NW 1 0.95 1 0.95 1 1 1 1 1 0.95
N 0.95 0.95 0.85 0.85 0.85 1 0.85 0.95 1 0.95

Aikawa Niigata Kanazawa Fushiki Toyama nagano Takada Utsunomiya Fukui Takayama
NE 0.85 0.85 0.85 0.9 0.85 0.85 0.85 0.9 0.85 0.85
E 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85
SE 0.85 0.85 0.85 0.85 0.9 0.85 0.85 0.85 1 0.9
S 0.85 0.85 0.9 0.85 1 0.85 0.85 0.85 1 1
SW 0.85 1 1 1 0.9 1 0.85 0.85 0.85 0.85
W 0.85 1 0.9 0.95 0.9 1 1 0.85 0.85 0.85
NW 1 1 0.85 0.85 0.85 0.9 0.95 0.85 1 0.85
N 1 0.9 0.85 0.9 0.85 0.85 0.85 1 1 0.85

Matsumoto Suwa Kumagaya Mito Tsuruga Gifu Nagoya Iida Kofu Chichibu
NE 0.85 0.85 0.85 1 0.85 0.85 0.85 0.95 0.85 0.85
E 0.85 1 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85
SE 1 1 0.85 0.85 0.85 1 1 0.85 0.85 0.85
S 1 0.95 0.85 0.85 0.85 0.95 1 0.85 0.85 0.85
SW 0.9 0.85 0.85 0.85 0.85 0.85 0.85 1 0.85 0.85
W 0.85 1 0.95 0.9 0.85 0.85 0.85 1 0.85 0.9
NW 0.85 1 1 1 1 0.9 0.85 1 1 1
N 0.85 0.85 0.95 1 0.85 0.85 0.85 0.95 1 0.85

Choshi Ueno Tsu Irako Hamamatsu Omaezaki Shizuoka Mishima Tokyo Owase
NE 0.9 0.85 0.85 0.9 0.95 1 0.85 1 0.85 0.85
E 0.85 1 1 0.95 0.95 1 0.85 0.85 0.85 0.95
SE 0.85 0.9 1 0.95 0.85 0.85 0.85 0.85 0.85 0.85
S 0.85 0.85 0.85 0.9 0.85 0.95 0.95 0.85 0.85 0.85
SW 0.85 0.9 0.85 0.85 0.85 0.95 1 0.85 0.85 0.95
W 0.85 0.85 0.95 1 1 1 1 0.85 0.85 1
NW 0.95 0.85 0.9 1 1 0.95 0.85 0.85 1 0.95
N 1 0.85 0.85 0.85 0.85 0.85 0.85 1 1 0.85
- 24 - Recommendations for Loads on Buildings

Table A6.1 (continued) Wind directionality factor K D

Irozaki Ajiro Yokohama Tateyama Katsuura Oshima Miyakejima Hachijojima Chiba Yokkaichi
NE 0.85 0.95 0.85 0.85 0.85 1 0.85 0.85 0.85 0.85
E 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.9 1
SE 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.9 1
S 0.85 0.85 0.85 0.85 1 0.85 0.85 1 0.95 0.85
SW 0.9 1 0.85 0.85 1 0.95 0.85 1 0.95 1
W 1 1 0.85 0.85 0.85 0.9 0.95 0.9 0.85 1
NW 0.85 0.85 1 1 1 0.85 1 0.9 1 0.85
N 0.85 0.95 1 1 0.9 1 1 0.9 1 0.85

Saigo Matsue Sakai Yonago Tottori Toyooka Maizuru Hagi Hamada Tsuyama
NE 0.9 0.9 0.95 0.85 0.85 0.85 0.85 0.85 0.85 0.85
E 0.85 0.85 0.95 0.85 0.85 0.85 0.85 0.85 0.85 0.95
SE 0.85 0.85 0.85 0.85 0.85 0.85 1 0.9 0.85 0.95
S 0.85 0.85 0.85 0.85 1 0.85 1 0.9 0.85 0.85
SW 0.85 0.9 0.95 0.95 0.85 0.85 0.85 0.85 0.9 0.85
W 0.85 1 1 1 0.85 0.85 0.85 0.9 1 0.95
NW 0.85 0.95 0.85 1 0.9 1 1 1 0.85 0.95
N 1 0.95 0.85 1 0.9 1 1 1 1 1

Kyoto Hikone Shimonoseki Hiroshima Kure Fukuyama Okayama Himeji Kobe Osaka
NE 1 0.85 0.85 0.85 1 0.95 0.85 0.85 1 0.9
E 0.95 0.85 0.85 0.85 0.85 0.85 0.85 0.85 1 0.85
SE 0.85 0.85 0.85 0.95 0.85 0.85 0.85 1 0.85 0.85
S 0.85 0.85 0.85 1 0.9 0.85 0.85 1 0.85 1
SW 0.85 0.85 0.85 1 0.9 0.85 1 0.85 0.85 1
W 0.85 0.95 0.85 0.85 0.9 0.85 0.85 0.9 0.85 1
NW 0.95 1 1 0.85 1 0.85 0.85 0.9 1 1
N 0.95 0.85 0.95 0.85 1 1 0.85 0.85 1 1

Sumoto Wakayama Shionomisaki Nara Yamaguchi Izuhara Hirado Fukuoka Iizuka Sasebo
NE 0.85 0.85 0.95 1 0.85 0.85 0.9 0.85 0.85 1
E 0.85 0.85 0.95 1 0.9 0.85 0.85 0.85 0.85 0.85
SE 0.9 1 0.85 0.85 1 0.85 0.85 1 0.85 0.85
S 1 1 0.9 0.85 0.85 0.95 0.85 1 0.9 0.85
SW 0.85 0.95 0.95 0.85 0.85 1 0.85 0.85 0.9 0.85
W 0.85 1 1 0.85 0.85 0.9 0.85 1 0.85 0.9
NW 0.85 1 1 0.85 0.85 0.95 0.95 1 1 0.9
N 0.85 0.85 0.85 0.85 0.85 1 1 1 1 1

Saga Hita Oita Nagasaki Kumamoto Nobeoka Akune Kagoshima Miyakonojo Miyazaki
NE 0.85 0.85 0.85 0.9 0.85 0.85 0.85 0.85 0.85 0.9
E 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.9 0.85
SE 0.85 0.85 1 0.85 0.85 0.95 0.85 0.95 1 1
S 1 1 0.85 0.9 1 1 1 0.95 1 0.85
SW 0.85 1 0.85 1 1 0.85 0.9 1 0.9 0.85
W 0.85 0.95 0.9 1 0.95 0.85 0.85 0.85 0.85 0.85
NW 0.95 0.85 0.9 0.95 1 0.9 0.85 0.85 0.85 0.85
N 0.95 0.85 0.85 0.9 1 0.9 0.85 0.85 0.85 0.9
CHAPTER 6 WIND LOADS - 25 -

Table A6.1 (continued) Wind directionality factor K D

Makurazaki Aburatsu Yakushima Tanegashima Ushibuka Fukue Matsuyama Tadotsu Takamatsu Uwajima
NE 0.85 0.95 0.85 0.85 0.85 0.85 0.9 0.85 0.9 0.95
E 1 0.9 0.85 0.85 0.85 0.85 0.85 0.85 0.85 0.95
SE 1 0.9 0.95 0.95 0.85 0.9 1 0.85 0.85 1
S 1 1 1 1 0.85 0.95 1 0.85 0.85 1
SW 1 1 1 0.9 0.9 0.95 0.95 1 0.95 0.85
W 0.95 0.85 0.85 0.9 1 0.95 0.95 1 1 0.85
NW 0.95 0.85 0.85 0.85 0.85 1 0.9 0.85 0.9 0.85
N 0.95 0.95 0.85 0.85 0.85 1 0.9 0.85 0.9 0.85

Kochi Tokushima Sukumo Shimizu Murotomisaki Naze Miyakojima Kumejima Naha Nago
NE 0.85 0.85 0.85 0.85 1 0.85 0.95 0.85 0.85 0.85
E 1 0.85 0.95 0.95 1 0.85 0.85 0.85 0.85 0.85
SE 1 1 0.95 0.9 0.9 0.9 0.85 0.95 0.95 0.9
S 0.85 1 0.9 0.95 0.85 0.85 0.95 1 1 1
SW 0.85 0.85 1 0.95 0.95 0.85 0.85 1 1 1
W 0.85 0.85 1 1 1 0.85 0.95 0.85 1 0.85
NW 0.85 0.85 1 0.85 0.9 1 1 1 1 0.9
N 0.85 0.85 0.85 0.85 0.85 1 1 1 1 0.85

Okinoerabu Minamidaitojima
NE 0.85 0.9
E 0.85 0.85
SE 1 0.95
S 1 0.95
SW 0.85 0.85
W 0.85 0.85
NW 0.9 1
N 0.95 1

A6.1.6 Wind seasonal factor

The wind seasonal factor KS shall be adopted to estimate the wind loads in a specific season. The
values of KS for the winter season (4 months from December to March) are provided in Fig. A6.3.
For the other seasons (spring, summer, and autumn) the value is 1.
When the wind seasonal factor is adopted, the wind directionally factor shall not be considered (i.e.,
KD = 1).
- 26 - Recommendations for Loads on Buildings

0.85
N
0.8

0.75

0.8

0.9
0.7

0.65 0.8

0.7 0.7

0.6

0.8
0.6
0.7
0.55
0.55
0.6

0.7 0.75
0.7
0.55
0.7 0.6
0.6
0.55

0.6
0.6

0.5

Izu islands, Satsunann Islands, Okinawa

0.55 Islands, Daitou Islands, Sakishima Islands,
0.6
and Ogasawara Islands, not shown in the
map

Figure A6.3 Wind seasonal factor (winter season: 4 months from December to March)
CHAPTER 6 WIND LOADS - 27 -

A6.1.7 Wind speed profile factor

(1) Wind speed profile factor
The wind speed profile factor E is calculated from Eq. (A6.4).

E = E r Eg (A6.4)

where
E r : exposure factor for terrain categories as defined in (2),

Eg : topography factor as defined in A6.1.9.

(2) Exposure factor based on the terrain categories

The exposure factor for each terrain category is defined in 2) according to the terrain categories
defined in 1).
1) The terrain categories of the construction site are defined in Table A6.2. However, if the terrain
category changes from smooth to rough within a upwind region of the construction site, i.e., within the
smaller distance of 40H (where H is the reference height) and 3km, the terrain category of the
construction site is assumed to be the same as that of the smooth upwind terrain.

Table A6.2 Terrain categories

Category Condition at construction site and upwind region
Smooth I Open sea, lakes, flat surface without any significant obstacles
↑ II Open country, grassland, agricultural field with low vegetation or
few obstacles (e.g. isolated low buildings, or trees)
III Suburban, wooded terrain, area with low-rise buildings or with
scattered tall buildings (4 to 9 floors)
↓ IV Urban, built-up area with tall buildings (4 to 9 floors)
Rough V City center, area with heavy concentration of tall buildings (higher
than 10 floors) or high-rise buildings

2） The exposure factor based on the terrain categories is defined in Eq. (A6.5), according to the terrain
categories defined in 1)

  Z α
1.7  Zb < Z ≤ ZG
  Z G 
Er =  α
(A6.5)
  Zb 
1 .7
 Z   Z ≤ Zb
  G

where
Z (m): height above ground,
Z b , Z G ,α : parameters of the exposure factor E r , defined in Table A6.3.
- 28 - Recommendations for Loads on Buildings

Table A6.3 Parameters of the exposure factor E r

Category I II III IV V
Z b (m) 3 5 10 20 30
Z G (m) 250 350 450 550 650
α 0.1 0.15 0.2 0.27 0.35

A6.1.8 Turbulence intensity and turbulence scale

The turbulence intensity and turbulence scale in A6.2 and A6.3 are defined as follows:
(1) Turbulence intensity
1) The turbulence intensity I Z is defined according to the conditions of the construction site as

I Z = I rZ EgI , (A6.6)

where
I rZ : turbulence intensity at height Z for each terrain category, defined in 2),

EgI : topography factor defined in A6.1.9.

2) Turbulence intensity for each terrain category

The turbulence intensity I rZ for each terrain category is defined in Eq. (A6.7) according to the
terrain categories defined in Table A6.2.

  Z  −α −0.05
0.1  Zb < Z ≤ ZG
  Z G 
I rZ = −α −0.05
(A6.7)
  Zb 
 0.1 Z  Z ≤ Zb
  G

where
Z (m): height above ground,
Z b , Z G ,α : parameters of the exposure factor as defined in Table A6.3.
(2) Turbulence scale

The turbulence scale LZ (m) is defined according to the terrain category of the construction site as

  Z  0.5
100  Zb < Z ≤ ZG
  30 
LZ =  0.5 (A6.8)
100 Z b  Z ≤ Zb
  30 
CHAPTER 6 WIND LOADS - 29 -

where
Z (m): height above ground,

Z b , Z G : parameter of the exposure factor, defined in Table A6.3.

A6.1.9 Effects of topography

The effects of topography on the wind speed and the turbulence intensity shall be evaluated through
an appropriate wind tunnel experiment or CFD. Otherwise, when wind passes at right angles over an
escarpment or a ridge-shaped topography, it can be evaluated based on the topography factor for the
mean wind speed and turbulence intensity (Eg and EgI), shown in (1) and (2) below.
(1) Topography factor
The topography factor, which reflects the change in mean wind speed that occurs as wind passes at a
right angle over an escarpment or ridge-shaped topography, as shown in Figs. A6.4 and A6.5, is defined
in Eq. (A6.9). However, when the inclination θ s calculated from Eq. (A6.10) is less than 7.5°, or when
X s / H s is beyond the range shown in Tables A6.4 and A6.5, it is not necessary to consider the

topography factor; i.e., Eg = 1 .

 Z   Z 
Eg = (C1 − 1)C 2 ( − C3 ) + 1 exp− C 2 ( − C3 ) + 1 and Eg ≥ 1 , (A6.9)
 Hs   Hs 
Hs
θ s = tan −1 , (A6.10)
2 Ls
where the C1, C2, and C3, parameters of the topography factor are given in Tables A6.4 and A 6.5, and
depend on the topography shape, inclination θ s , and distance X s (m) from the top of the topographic
feature to the construction site. When the inclination θ s is greater than 60°, the topography factor is
assumed to be the same as that for 60°.
Here,
Z (m): height above ground, assumed to be the same value as Z b when smaller than Z b ,
H s (m): height of the topography,
Ls (m): horizontal distance from the top of topographic feature to the point where the height is
half the topography height, as shown in Figs. A6.4 and A6.5.
- 30 - Recommendations for Loads on Buildings

Figure A6.4 Escarpments

Figure A6.5 Ridge-shaped topography

Table A6.4 Parameters of the topography factor Eg (escarpments)

Xs /Hs
θs
−4 −2 −1 −0.5 0 0.5 1 2 4 8
C1 1.15 1.3 1.5 1.5 1.6 1.45 1.3 1.3 1.2 1.15
7.5° C2 0.8 0.8 0.8 0.8 0.8 0.7 0.6 0.6 0.5 0.4
C3 −2 −2 −2 −2 −2 −2 −2 −2 −2 −2
C1 0.4 1 1.2 1.55 2.1 1.65 1.5 1.3 1.2 1.15
15° C2 0.9 0 0.65 0.85 1 0.8 0.7 0.55 0.45 0.35
C3 −2 −2 −2 −2 −2 −2 −2 −2 −2 −2
C1 0.7 −0.5 1.05 1.1 1.3 1.3 1.25 1.2 1.15 1.1
30° C2 0.65 1.2 1.65 1.5 1.45 1.3 0.9 0.9 0.85 0.6
C3 −2 −2 1 0.8 0.3 0.3 0.5 0.7 1.2 1.4
C1 0.8 0 −3.5 1.1 1.2 1.35 1.3 1.2 1.15 1.1
45° C2 0.5 1 1.6 2 1.1 1.3 1.3 1.3 0.9 0.55
C3 −2 −2 −2 0.8 0.3 0.2 0.75 1.05 1.4 2
C1 0.6 0.1 −1.8 −2.4 1.2 1.4 1.35 1.25 1.15 1.1
60° C2 0.65 0.9 1.3 2.6 2 1.8 1.7 1.5 0.85 0.45
C3 −2 −2 −2 −1 0.5 0.5 0.8 1.2 1.9 3.1
CHAPTER 6 WIND LOADS - 31 -

Table A6.5 Parameters of the topography factor Eg (ridge-shaped topography)

Xs /Hs
θs
−4 −2 −1 −0.5 0 0.5 1 2 4 8
C1 1.1 1.2 1.35 1.35 1.4 1.3 1.3 1.2 1.1 1
7.5° C2 1 1 1 1 1.5 1.2 1.1 2 1.6 0
C3 0 0 0 0 0.2 0.2 0.2 0.5 0.9 0
C1 1 1.05 1.2 1.25 1.3 1.4 1.3 1.25 0.35 0.65
15° C2 0 0 1 1 1 1.5 1.5 2 3 2
C3 0 0 0 0 0 0.5 0.6 1.1 0.2 0.3
C1 0.75 0.55 0.85 1 1.2 1.3 1.25 1.2 1.1 1.02
30° C2 1.5 2 2 0 1 2 2 1.6 1.7 1.7
C3 0 0 0 0 0 1.1 1.3 2.1 2.2 2.8
C1 0.75 0.55 0.2 0.75 1.15 1.2 1.15 1.12 1.1 1.02
45° C2 1.5 2 2 3 1 2.5 2.5 2 1.6 1.3
C3 0 0 0 0 0 1.2 1.9 2.2 2.5 3.2
C1 0.75 0.55 0.2 0.2 1.15 1.12 1.15 1.12 1.1 1.02
60° C2 1.5 1.5 1.8 3 1 2.2 2.5 2 1.6 1.3
C3 0 0 0 0 0 1.8 2 2.3 2.6 3.4
Note) For a particular inclination θ s and a horizontal location X s / H s , the topography factor is
calculated by interpolating the linearly from the values at the nearest inclinations and horizontal
locations.

(2) Topography factor for turbulence intensity

The topography factor for the turbulence intensity when the wind passes at a right angle to an
escarpment or ridge-shaped topography, as shown in Figs. A6.4 and A6.5, is defined as follows:
EI
EgI = , (A6.11)
Eg

where
 Z   Z 
EI = (C1 − 1)C 2 ( − C3 ) + 1 exp− C 2 ( − C3 ) + 1 and EI ≥ 1 , (A6.12)
 Hs   Hs 

in which EI is the topography factor for the standard deviation of the fluctuating wind speed.
When the inclination θ s calculated from Eq. (A6.10) is less than 7.5°, or the distance from the top
of the topographic feature X s (m) is beyond the range of X s / H s in Tables A6.4 and A6.5, it is not
necessary to consider the topography factor, i.e., EI = 1 .
Here

E g : topography factor for the mean wind speed, defined in Eq. (A6.9),

C1 , C 2 , C 3 : parameters determining the topography factor EI , given in Tables A6.6 and A6.7,
depending on the topography shape, inclination θ s , and distance X s (m) from the
top of the topographic feature to the construction site (when the inclination θ s is
greater than 60°, the topography factor is assumed to be the same as that for 60
- 32 - Recommendations for Loads on Buildings

degrees),
Z (m): height above ground, and assumed to be the greater of Z b and Z c when smaller than
Z b , as shown in Table A6.3, or Z c , shown in Tables A6.6 and A6.7,
H s (m): height of the topography,
Ls (m): horizontal distance from the top of the topographic feature to the point where the height
is half the topography height.

Table A6.6 Parameters of the topography factor for turbulence intensity EI (escarpments)
Xs /Hs
θs
−4 −2 −1 −0.5 0 0.5 1 2 4 8
Zc/Hs 0 0 0 0 0 0 0 0 0 0
C1 1 1 1 1 1 1 1 1 1 1
7.5°
C2 0 0 0 0 0 0 0 0 0 0
C3 0 0 0 0 0 0 0 0 0 0
Zc/Hs 0 0 0 0 0 0 0 0 0 0
C1 1 1.05 1.1 1.1 1.1 1.1 1.1 1.1 1.1 1.1
15°
C2 0 0 0.2 0.2 0.3 0.4 0.5 0.5 0.3 0.3
C3 0 0 0 0 0 0 0 0 0 0
Zc/Hs 0 0 0 0 0 0 0 0 0 0
C1 1.05 1.05 1.1 1.15 1.2 1.3 2.5 1.8 1.4 1.25
30°
C2 0 0 0 0.7 2 2.5 10 8 4 1.5
C3 0 0.5 1 1 0.5 0 0 0.1 0.2 0.6
Zc/Hs 0 0 0 0 0 0 0 0.3 0.3 0.2
C1 1.05 1.1 1.15 1.3 1.8 1.5 3 3 1.8 1.5
45°
C2 0 0 0 1.4 1.7 1.6 6 5 3.5 2
C3 0 0 0.5 0 −1 −0.8 0 0 0.3 0.5
Zc/Hs 0 0 0 0 0 0 0 0.5 0.7 0.9
C1 1.1 1.15 1.2 1.3 6 8 4 3.5 2.2 1.7
60°
C2 0 0 0 0.7 2.5 5 8 5 3 1.5
C3 0 0 0.3 0.5 −1.3 −0.4 0.2 0.2 0.4 0.5
CHAPTER 6 WIND LOADS - 33 -

Table A6.7 Parameters of the topography factor for turbulence intensity EI (ridge-shaped
topography)

Xs /Hs
θs
−4 −2 −1 −0.5 0 0.5 1 2 4 8
Zc/Hs 0 0 0 0 0 0 0 0 0 0
C1 1 1 1 1 1 1 1 1 1 1
7.5°
C2 0 0 0 0 0 0 0 0 0 0
C3 0 0 0 0 0 0 0 0 0 0
Zc/Hs 0 0 0 0 0 0 0 0.6 1.5 2
C1 1 1 1 1 1 3.4 4.2 4 2.8 2
15°
C2 0 0 0 0 0 19 11 4.6 2 1.6
C3 0 0 0 0 0 0.1 0.2 0.3 0.6 0.7
Zc/Hs 0 0 0 0 0 0.6 0.8 1.5 1.6 2.2
C1 1 1 1 1 1 1.6 1.9 2.2 3.2 2.7
30°
C2 0 0 0 0 0 5 4 2 1.7 1.3
C3 0 0 0 0 0 0.4 0.6 1 0.7 0.5
Zc/Hs 0 0 0 0 0 0.8 1.5 1.7 1.9 2.5
C1 1 1 1 1 1 1.4 1.7 2.1 2.8 2.8
45°
C2 0 0 0 0 0 2.6 2.2 1.7 1.4 1.3
C3 0 0 0 0 0 0.8 1.1 1.2 0.9 0.5
Zc/Hs 0 0 0 0 0 1.35 1.6 1.8 2 2.6
C1 1 1 1 1 1 1.6 1.8 2.1 2.9 2.9
60°
C2 0 0 0 0 0 2 1.6 1.3 1.3 1.2
C3 0 0 0 0 0 1 1.2 1.2 0.8 0.6

Note) For a particular inclination θ s and a horizontal location X s / H s , the topography factor for a
fluctuating wind speed is calculated by interpolating the linearly from the values at the nearest
inclinations and horizontal locations.

A6.2 Wind Force and Pressure Coefficients

The wind force and pressure coefficients fall into two categories corresponding to the design of the
structural frames and components/cladding. The coefficients shall be estimated from appropriate wind
tunnel tests or CFD. For buildings that satisfy the criteria specified in A6.2.2 through A6.2.7, the
coefficients can be estimated from the following procedure using the wind pressure (external and
internal pressure coefficients) and wind force coefficients provided in these clauses.

A6.2.1 Procedure for estimating the wind force coefficients

(1) Wind force coefficients for the design of structural frames
1) Wind force coefficients CD for estimating the horizontal wind loads on structural frames
The wind force coefficients are given in A6.2.4(1) and A6.2.4(4), or calculated from Eq. (A6.13),
using the external pressure coefficients provided in A6.2.2.

CD = Cpe1 − Cpe2 (A6.13)

- 34 - Recommendations for Loads on Buildings

where

Cpe1 : external pressure coefficient on windward face,

Cpe2 : external pressure coefficient on leeward face.

2) Wind force coefficients CR for estimating the roof wind loads on structural frames
The wind force coefficients are given in A6.2.5(2), or calculated from Eq.(A6.14), using the external
pressure coefficients provided in A6.2.2 and the internal pressure coefficients provided in A6.2.3.

CR = Cpe − Cpi (A6.14)

where

Cpe : external pressure coefficient on a roof,

C pi : internal pressure coefficient.

3) Wind force coefficients CD for estimating the horizontal wind loads on lattice structures
The wind force coefficients are given in A6.2.4(3), or calculated from the wind force coefficients, for
individual members provided in A6.2.4(5).
(2) Peak wind force coefficients ĈC for design of the components/cladding
The peak wind force coefficients ĈC are given in A6.2.7, or calculated from Eq.(A6.15), using the
peak external pressure coefficients provided in A6.2.5 and the factor for effect of fluctuating internal
pressures provided in A6.2.6.
Cˆ C = Cˆ pe − C * pi (A6.15)
where
Ĉpe : peak external pressure coefficient,
C * pi : factor for effect of fluctuating internal pressures.
CHAPTER 6 WIND LOADS - 35 -

A6.2.2 External wind pressure coefficients for structural frames

(1) External wind pressure coefficients Cpe for buildings with rectangular plans and aspect ratios
H / BD of greater than 2
For buildings with rectangular plans and aspect ratios H / BD of greater than 2, the external wind
pressure coefficients on the windward and leeward walls and on the roof are given in Table A6.8. The
values in Table A6.8 are applicable to buildings whose aspect ratios H / BD are less than or equal to
8.

Table A6.8 External wind pressure coefficients Cpe for buildings with rectangular plans and aspect
ratios of H / BD > 2
i) Wall
External pressure coefficient Cpe
B≥D B<D
Windward wall Cpe1 0.8kZ
Leeward wall Cpe2 −0.5 −0.35

Factor for vertical profile kZ

Z ≤ Zb Z b < Z < 0.8H Z ≥ 0.8H

( Z b / H ) 2α ( Z / H ) 2α 0.82α
B (m) : building width
2α
Note: When 0.8H < Zb, put kZ = 0.8 . D (m) : building depth
H (m) : reference height
Z (m): height from ground
Zb (m): height defined in Table A6.3
α : parameter defined in Table A6.3
ii) Roof (flat)
External pressure coefficient Cpe
Cpe Ra Rb Rc B
Zone Ra −1.2
Zone Rb −0.6
Zone Rc −0.2
l1 : 0.5B
l1 l2 l2 : B
Roof

(2) External wind pressure coefficients Cpe for buildings with rectangular plans and aspect ratios
H / BD of less than or equal to 2
1) Buildings with flat, gable, mono-sloped, and hip roofs
The external pressure coefficients Cpe for buildings with rectangular plans and flat, gable, mono-
sloped, and hip roofs whose aspect ratios H / BD are less than or equal to 2 are listed in Table A6.9(1).
- 36 - Recommendations for Loads on Buildings

Table A6.9(1) External wind pressure coefficients Cpe for buildings with rectangular plans and aspect
ratios of H / BD ≤ 2
i) Wall
Zone WU (windward wall) Zone Sa, Sb, Sc (side wall)
B≤H B>H Sa Sb Sc
0.8kZ 0.6 −0.7 −0.4 −0.2
kZ : factor for vertical profile provided in Table A6.8
When 0.8H < Z b , put k Z = 0.82α.
Zone La, Lb (leeward wall)
La Lb
Wind
Roof angel θ B < 6H
direction D≤H D>H B ≥ 6H
D≤H D>H

W1 θ ≤ 45° −0.6 −0.4 −0.6 −0.4 −0.8

θ < 20° −0.6 −0.4 −0.6 −0.4 −0.8
W2
20° ≤ θ < 30° −0.6 −0.5 −0.6 −0.5 −0.8
W3
30° ≤ θ ≤ 45° −0.6 −0.6 −0.6 −0.6 −1.0
ii) Roof (flat)
Zone Ra, Rb, Rc (roof)
Ra
Rb Rc
D≤H D>H
Positive not necessary to evaluate
B ≤ 2H −1.0 −0.8 −0.4 −0.2
Negative B = 4H −1.1 −0.9 −0.5 −0.3
B ≥ 6H −1.2 −1.0 −0.6 −0.4
Note : Linear interpolation is permitted for values of B＝2H～(4H)～6H (m)
iii) gable and mono-sloped roofs
Zone RU (windward roof)
Negative
Roof angle θ Positive D≤H D>H
B ≤ 2H B = 4H B ≥ 6H B ≤ 2H B = 4H B ≥ 6H

10° 0 −1.0 −1.1 −1.2 −0.8 −0.8 −0.9

15° 0 −0.7 −0.8 −0.9 −0.6 −0.6 −0.6
20° 0.1 −0.5 −0.6 −0.7 −0.4 −0.4 −0.4
25° 0.1 −0.3 −0.4 −0.5 −0.2 −0.3 −0.3
30° 0.2 −0.2 −0.3 −0.4 0 −0.1 −0.2
35° 0.3 0 −0.1 −0.2 0 0 −0.1
40° 0.4 0 −0.1 −0.1 0 0 0
45° 0.4 0 0 0 0 0 0
CHAPTER 6 WIND LOADS - 37 -

Note : Linear interpolation is permitted for values of θ (°) and B＝2H～(4H)～6H (m).
When θ < 10°, the values for flat roofs can be used.
Zone RLa，RLb (leeward roof)
RLb
Roof angle θ RLa
B < 6H B ≥ 6H

10° ≤ θ < 15° −0.6 −0.6 −1.1

15° ≤ θ ≤ 45° −0.6 −0.6 −1.4
Note: When θ < 10°, the values for flat roofs can be used.
iv) Hip roof
Zone RU (windward roof), Ra, Rb (roof)
Positive Negative
Roof angle θ Ra
RU, Ra, Rb RU Rb
B≤D B>D
10° 0 −1.2 −0.9 −1.1 −0.9
15° 0 −1.1 −1.2 −1.1 −0.9
20° 0 −0.9 −1.4 −1.1 −0.9
25° 0 −0.8 −1.4 −1.1 −0.9
30° 0 −0.6 −1.4 −1.1 −0.9
35° 0.1 −0.5 −1.4 −1.1 −0.9
40° 0.2 −0.3 −1.4 −1.1 −0.9
45° 0.3 −0.2 −1.4 −1.1 −0.9
Note : Linear interpolation is permitted for values of θ (°) other than those shown..
Note: When θ < 10°, the values for flat roofs can be used.

Zone RL (leeward roof)

D≥B D<B
−0.9 −1.0

WU (windward wall)，RU(windward
Wind La,Lb (leeward wall),
roof)，Ra,Rb,Rc (roof), Sa,Sb,Sc (side
direction RLa,RLb,RL (leeward roof)
wall)
Building with flat roof

W1
- 38 - Recommendations for Loads on Buildings

Building with gable roof

Building with mono-slope roof

Building with hip roof

W1
CHAPTER 6 WIND LOADS - 39 -

B (m) : building width

D (m) : building depth
H (m) : reference height
l (m) : smaller of 4H and B

2) Buildings with vaulted roofs

External pressure coefficients for buildings with rectangular plans and vaulted roofs are given in
Table A6.9(2).

Table A6.9(2) External pressure coefficients Cpe for buildings with rectangular plans and vaulted roofs
i) Wall
External wind pressure coefficients are defined in Table A6.9(1), where the gradient angle of the
straight line connecting the eaves and top of the vaulted roof is used for θ (°) when evaluating the
external pressure coefficients on the leeward wall (La, Lb).

ii) Roof
Wind Zone Ra Zone Rb Zone Rc
f/B
direction h/B=0 h/B=0.3 h/B=0.7 h/B=0 h/B=0.3 h/B=0.7 h/B=0 h/B=0.3 h/B=0.7
0 −0.4 −0.9 −0.8 −0.4 −0.5 −0.4 −0.4 −0.3 −0.2
0.1 −1.2 −1.1 −1.1 −0.7 −0.5 −0.5 −0.4 −0.4 −0.4
W1
0.3 −1.1 −1.1 −1.1 −0.6 −0.5 −0.5 −0.4 −0.4 −0.4
0.4 −1.1 −1.1 −1.1 −0.5 −0.5 −0.5 −0.4 −0.4 −0.4
Note : Linear interpolation is permitted for values of f/B and h/B other than those shown.

Wind Zone Ra Zone Rb Zone Rc

f/D
direction h/D=0 h/D=0.3 h/D=0.7 h/D=0 h/D=0.3 h/D=0.7 h/D=0 h/D=0.3 h/D=0.7
0 −0.4 −1.0 −0.9 −0.4 −1.0 −0.9 −0.4 −0.6 −0.9
0.1 −0.5 −1.2 −1.5 −0.9 −1.0 −1.0 −0.5 −0.5 −0.5
W2
0.3 −0.1 −0.4 −0.9 −1.2 −1.4 −1.5 −0.5 −0.5 −0.5
0.4 0.2 0 −0.5 −1.2 −1.3 −1.4 −0.5 −0.5 −0.5
Note : Linear interpolation is permitted for values of f/D and h/D other than those shown.
- 40 - Recommendations for Loads on Buildings

B (m) : building width

D (m) : building depth
H (m) : reference height
f (m) : rise
h (m) : eave height
l (m) : the smaller of 4H and B
θ (°) : roof angle used for
evaluating the external
pressure coefficients on
the leeward wall
θ= tan-1(2f /B)
or
θ = tan-1(2f /D)
CHAPTER 6 WIND LOADS - 41 -

(3) External pressure coefficients Cpe for spherical domes

The external pressure coefficients for spherical domes are given in Table A6.9(3).

Table A6.9(3) External pressure coefficient Cpe for spherical domes

Zone Ra (Positive) Ra (Negative)

f/D h/D=0 h/D=0.25 h/D=1 h/D=0 h/D=0.25 h/D=1
0 Not necessary to evaluate −0.6 −1.4 −1.2
0.05 0.3 0 0 0 −1.0 −1.6
0.1 0.4 0 0 0 −0.6 −1.2
0.2 0.5 0 0 0 0 −0.4
0.5 0.7 0.6 0.6 Not necessary to evaluate

Zone Rb Rc Rd
f/D h/D=0 h/D=0.25 h/D=1 h/D=0 h/D=0.25 h/D=1 h/D=0 h/D=0.25 h/D=1
0 0 −0.8 −1.2 0 −0.1 −0.4 0 −0.1 −0.3
0.05 0 −0.4 −0.8 −0.2 −0.4 −0.4 −0.1 −0.3 −0.3
0.1 0 −0.4 −0.6 −0.4 −0.6 −0.6 −0.2 −0.4 −0.4
0.2 0 −0.4 −0.6 −0.6 −0.8 −1.0 −0.2 −0.4 −0.4
0.5 0 −0.3 −0.4 −1.1 −1.2 −1.3 −0.2 −0.4 −0.4
Note : Linear interpolation is permitted for values of f/D and h/D other than those shown.

D (m) : building diameter

H (m) : reference height
h (m) : eaves height
f (m) : rise
- 42 - Recommendations for Loads on Buildings

A6.2.3 Internal pressure coefficients

Internal pressure coefficients for structural frames shall be estimated appropriately by considering the
location and size of distributed gaps and openings on the walls. For buildings without dominant openings,
values of 0 or −0.4 shall be used.

A6.2.4 Wind force coefficients for design of structural frames

(1) Wind force coefficients CD for buildings with circular and elliptic plans
Wind force coefficients for buildings with circular and elliptic plans are given in Table A6.10. The
values in Table A6.10 are applicable to cases in which D2U H ≥ 6 (m2/s), H / D2 ≤ 8 , and D1 / D2 ≤ 3 .

Table A6.10 Wind force coefficients CD for buildings with circular and elliptic sections

C D = 1.2k1k 2 k 3 k Z
where
k1 : factor for aspect ratio
k2 : factor for surface roughness

k3 : factor for D1 D2

kZ : factor for vertical profile defined in Table A6.8 and k Z = 0.82α

when 0.8H < Z b
k1
H D2 < 1 1 ≤ H D2 ≤ 8

0.6(H D2 )
0.14
0.6

k2
smooth surface (metal, concrete, D1 (m) : major axis of plan
0.75
flat curtain walls, etc.) D2 (m) : minor axis of plan
rough surface (1% relative H (m) : reference height
0.9
roughness, rough curtain walls, etc.) Zb (m) : height defined in Table A6.3
very rough surface (5% relative α : parameter defined in Table A6.3
1
roughness)

k3
D1 D2 1 1.5 2 2.5 3
1 1.4 1.6 1.8 1.9
k3

Note: Linear interpolation is permitted for values of k3 other than those shown.
CHAPTER 6 WIND LOADS - 43 -

(2) Wind force coefficients CR for free roofs with rectangular plan
1) Pitched or troughed free roofs
Wind force coefficients for pitched or troughed free roofs with a rectangular plan are listed in Table
A6.11(1). The values in Table A6.11(1) are applicable to cases in which H ≤ 10 m, H 2 ≤ B ≤ 30
m,, and B ≥ D .

Table A6.11(1) Wind force coefficients CR for pitched or troughed free roofs with rectangular plan

Combination 1 for wind force Combination 2 for wind force

Roof angle θ
Zone Ra Zone Rb Zone Ra Zone Rb
−20° −0.8 0.6 0.0 0.6
−10° −0.6 0.4 0.1 0.6
0° −0.4 0.1 0.2 0.1
5° −0.4 −0.2 0.4 −0.1
10° −0.2 −0.4 0.5 −0.3
15° 0.0 −0.6 0.7 −0.5
20° 0.2 −0.5 0.8 −0.4
Note: Linear interpolation is permitted for values of θ(°) other than those shown.

B (m) : building width

D (m) : building depth
H (m) : reference height
θ( ο ) : roof angle
- 44 - Recommendations for Loads on Buildings

2) Mono-sloped free roofs

The wind force coefficients for mono-sloped free roofs with a rectangular base are listed in Table
A6.11(2). The values in Table A6.11(2) are applicable to cases in which H ≤ 10m , H 2 ≤ B ≤ 30m , and

B≥D.

Table A6.11(2) Wind force coefficients CR for mono-sloped free roofs with rectangular base

Combination 1 for wind force Combination 2 for wind force

Wind direction Roof angle θ
Zone Ra Zone Rb Zone Ra Zone Rb
0° −0.4 0.1 0.2 0.1
W1 15° −1.2 0.0 0.0 0.0
20° −1.4 −0.2 0.0 0.0
0° 0.1 −0.4 0.1 0.2
W2 15° 0.0 0.0 0.2 1.0
20° 0.0 0.0 0.3 1.3
Note: Linear interpolation is permitted for values of θ(°) other than those shown.

B (m) : building width

D (m) : building depth
H (m) : reference height
θ( ο ) : roof angle
CHAPTER 6 WIND LOADS - 45 -

(3) Wind force coefficients CD for lattice structures

Wind force coefficients for lattice structures are given in Table A6.12.

Table A6.12 Wind force coefficients CD for lattice structures

i) Square arrangement

Solidity ϕ Angle Circular pipe Angle Circular pipe

0 3.8 2.3 4.4 2.5
0.5 1.9 1.4 2.3 1.7
0.6 1.9 1.4 2.3 1.7

ii) Triangular arrangement

Solidity ϕ Angle Circular pipe The solidity ϕ is defined by

0 3.8 2.3 ϕ = AF / A0
0.5 1.9 1.4 AF(m ) : projected area per panel (
2 )
0.6 1.9 1.4 ( area) (=Bh)
A0(m2) : whole plane

Note: Linear interpolation is permitted for values of ϕ other than those shown.
- 46 - Recommendations for Loads on Buildings

(4) Wind force coefficients CD for fences on the ground

Wind force coefficients for fences on the ground are given in Table A6.13.

Table A6.13 Wind force coefficients CD for fences on the ground

Solidity ϕ CD
0 1.2
0.2 1.5
0.6 1.7
≥ 0.9
1.2
(Solid fences included)
Note : The area for calculating the wind loads is the overall area multiplied by the solidity ϕ .
The definition of ϕ is the same as that in Table A6.12.
Linear interpolation is permitted for values of ϕ other than those shown.
The height of a fence H is used for calculating the wind load.

(5) Wind force coefficients CX and C Y for members

Wind force coefficients for members are given in Table A6.14.

Table A6.14 Wind force coefficients CX and CY for members

CX θ CX CY θ CX CY θ CX CY
1.2 0° 2.1 0 0° 2.4 0 0° 2.1 0
45° 1.6 1.6 45° 1.6 0.7 30° 2.1 −0.2
90° 0 0.8 60° 0.7 1.1

θ CX CY θ CX CY θ CX CY θ CX CY
0° 1.2 0 0° 1.1 0 0° 2.0 0 0° 1.9 2.2
45° 0.8 0.8 45° 0.8 0.7 45° 1.8 0.1 45° 2.3 2.3
90° 0.6 0.5 90° 0.9 0.5 90° 0 0.1 90° 2.2 1.9
135° −1.7 0.6 135° −2.3 0.6 135° −1.9 −0.6
180° −2.3 0 180° −2.5 0 180° −2.0 0.3
225° −1.4 −1.4
CHAPTER 6 WIND LOADS - 47 -

θ CX CY θ CX CY θ CX CY θ CX CY
0° 2.0 1.1 225° −1.5 −0.6 0° 2.1 0 0° 2.6 0
45° 2.3 1.1 270° 0.6 −0.8 45° 2.1 0.6 45° 2.0 0.8
90° 1.8 0.8 315° 1.2 −0.2 90° ±0.6 0.7 90° ±0.6 0.8
135° −1.7 0 135° −1.6 0.6
180° −2.0 0.1 180° −2.0 0
2
Note: The area for calculating the wind loads is bl(m ) (b, member width; l, member length)
irrespective of the wind direction.

Net

Solidity ϕ CX
0 2.0
0.2 2.0
0.6 2.7
≥ 0.9
2.0
(Solid plate included)
Note : The area for calculating the wind loads is bl ϕ (m2) (l , net length)
The definition of ϕ is the same as that in Table A6.12.
Linear interpolation is permitted for values of ϕ other than those shown.
- 48 - Recommendations for Loads on Buildings

A6.2.5 Peak external pressure coefficients for components/cladding

(1) Peak external pressure coefficients Ĉpe for buildings with rectangular plans and aspect ratios
H/ B1 B2 of greater than 2
For buildings with rectangular plans and aspect ratios H / B1 B2 of greater than 2, the peak external
pressure coefficients are given in Table A6.15. The values in Table A6.15 are applicable to buildings
with H / B1 ≤ 8 .

Table A6.15 Peak external pressure coefficients Ĉpe for buildings with rectangular plans and aspect
ratios H / B1 B2 of greater than 2
i) Wall

a) Positive peak external pressure coefficients

Cˆ pe = k Z (1 + 7 I Z )
where kZ : factor for vertical profile defined in Table A6.8,
IZ : turbulence intensity at height Z defined by Eq. (A6.6).
When the effect of local topography is considered, the values of kZ and IZ at reference height H
can be used in the above equation.

b) Negative peak external pressure coefficients

Zone Case Ĉ pe B1(m) : smaller side length of plan
Wa all −3.0 B2(m) : larger side length of plan
Wb all −2.4 H(m) : reference height
b / B ≤ 0.2 −3.0 la1(m) : the smaller of H and B1
Wc
b / B > 0.2 −2.4 la2(m) : the smaller of H and B2
b' / B ≤ 0.2 −3.0 b/B : the smaller of b1/B1 and b2/B2
Wd
b' / B > 0.2 −2.4 b’/B : the larger of b1/B1 and b2/B2

Building with rectangular plan Building with recessed corners Building with chamfered
corners
CHAPTER 6 WIND LOADS - 49 -

ii) Roof

a) Positive peak external pressure coefficients

Not necessary to evaluate

b) Negative peak external pressure coefficients

Zone Ĉ pe
Ra −2.5
Rb −3.2
Rc −5.0kC
Reduction factor for area subject to local suction kC
Subject area AC (m2) kC
AC < 1 1
1 ≤ AC ≤ 5
0.18
1/ AC
5 < AC 0.75

(2) Peak external pressure coefficients Ĉpe for buildings with rectangular plans and aspect ratios
H / B1 B2 of less than or equal to 2
1) Buildings with flat, gable, mono-sloped, and hip roofs
Peak external pressure coefficients for buildings with rectangular plans and flat, gable, mono-sloped,
and hip roofs whose aspect ratios H / B1 B2 are less than or equal to 2 are listed in Table A6.16(1).

Table A6.16(1) Peak external pressure coefficients Ĉpe for buildings with rectangular plans and flat,
gable, mono-sloped, and hip roofs whose aspect ratios H / B1 B2 are less than or equal to 2.
i) Wall

a) Positive peak external pressure coefficients

Scope of application: all parts of the walls
Cˆ pe = k Z (1 + 7 I Z )
where kZ : factor for vertical profile defined in Table A6.8,
IZ : turbulence intensity at height Z defined by Eq. (A6.6).
When the effect of the local topography is considered, the values of kZ and IZ at the reference
height H (Z=H) can be used in the above equation.

b) Negative peak external pressure coefficients

Zone Ĉ pe
Wa −4.2
Wb −2.4
- 50 - Recommendations for Loads on Buildings

ii) Roof

a) Positive peak external pressure coefficients

Scope of application: all parts of the roofs except for flat roofs
Cˆ pe = C pe (1 + 7 I H )
where Cpe : positive external pressure coefficients for zone RU defined in Table A6.9(1),
IH : turbulence intensity at height H (m), obtained by substituting H (m) for Z (m)
in Eq. (6.6).

b) Negative peak external pressure coefficients

Zone Ĉ pe
Ra −4.3
Flat roof Rb −7.3kC
Rc −4.3
Rf −2.5
Roof angle θ
Zone
10° 20° 30° ≤ θ ≤ 45°
Ra −4.3 −3.4 −3.4
Rb −7.3kC −4.3 −4.3
Gable roof Rc −4.3 −4.3 −4.3
Rd −4.3 −7.3kC −4.3
Re −3.4 −4.3 −4.3
Rf −2.5 −2.5 −2.5
Rg −3.4 −7.3kC −4.3
Roof angle θ
Zone
10° 20° 30° ≤ θ ≤ 45°
Ra −4.3 −3.4 −3.4
Mono-sloped Rb −7.3kC −4.3 −4.3
roof Rc −4.3 −4.3 −4.3
Rd −7.3kC −8.8kC −7.3kC
Re −5.0 −5.0 −5.0
Rf −2.5 −2.5 −2.5
Roof angle θ
Zone
10° 20° 30° 45°
Ra −4.3 −3.4 −3.4 −3.4
Rb −6.8kC −5.0kC −4.1 −4.0
Hip roof Rc −4.3 −3.4 −3.4 −3.4
Rd −2.5 −3.5 −3.6 −4.6
Re −2.7 −6.0kC −3.3 −3.3
Rf −3.4 −2.5 −2.5 −3.3
Rg −3.4 −4.1 −3.6 −2.7
Note: kC represents the reduction factor for an area subjected to local suction, as defined in Table
A6.15. Linear interpolation is permitted for values of θ (°) other than those shown.
When θ < 10°, the values for flat roofs can be used.
CHAPTER 6 WIND LOADS - 51 -

Building with flat roof Building with gable roof

Building with mono-sloped roof Building with hip roof

l (m) : the smallest among 4H, B1, and B2

lb1 (m) : the smaller of 4H and B1
lb2(m) : the smaller of 4H and B2
θ (°) : roof angle
- 52 - Recommendations for Loads on Buildings

2) Buildings with vaulted roofs

The peak external pressure coefficients for buildings with rectangular plans and vaulted roofs are
given in Table A6.16(2).

Table A6.16(2) Peak External pressure coefficients Ĉpe for buildings with rectangular plans and
vaulted roofs
i) Wall

a) Positive peak external pressure coefficients

Positive peak external pressure coefficients are defined in Table A6.16(1).

b) Negative peak external pressure coefficients

Negative peak external pressure coefficients are defined in Table A6.16 (1).

ii) Roof
a) Positive peak external pressure coefficients
Zone Ra Zone Rb Zone Rc
f/B1
h/B1=0 h/B1=0.3 h/B1=0.7 h/B1=0 h/B1=0.3 h/B1=0.7 h/B1=0 h/B1=0.3 h/B1=0.7
0.1 0.8 0.8 0.5 0.5 0.4 0.3 0.2 0.1 0.0
0.3 2.0 2.3 1.8 1.6 1.4 1.2 0.6 0.4 0.4
0.4 2.2 2.4 2.4 1.9 1.8 1.8 0.8 0.6 0.5
Note: Linear interpolation is permitted for values of h/B1 and f/B1 other than those shown.

b) Negative peak external pressure coefficients

Zone Ra Zone Rb Zone Rc Zone Rd
f/B1 h/B1 h/B1 h/B1 h/B1 h/B1 h/B1 h/B1 h/B1 h/B1 h/B1 h/B1 h/B1
=0 =0.3 =0.7 =0 =0.3 =0.7 =0 =0.3 =0.7 =0 =0.3 =0.7
0 −2.5 −3.2 −3.2 −2.5 −2.5 −2.5 −2.5 −2.5 −2.5 −2.5 −5.4kc −5.4kc
0.1 −1.4 −4.2 −4.8 −1.8 −2.2 −3.2 −2.5 −2.5 −2.5 −3.4 −4.8 −4.4
0.3 −1.4 −2.4 −2.6 −2.0 −3.2 −3.2 −3.8 −4.4 −4.5 −4.0 −4.4 −4.5
0.4 −1.8 −2.4 −2.6 −2.4 −3.2 −3.2 −4.3 −4.4 -4.6 −4.0 −4.4 −4.8
Note: kC represents the reduction factor for an area subjected to local suction defined in Table
A6.15. Liner interpolation is permitted for values of h/B1 and f/B1 other than those shown.
CHAPTER 6 WIND LOADS - 53 -

Zones for positive peak pressure coefficients Zones for negative peak pressure coefficients
h (m) : eave height
f (m) : rise
B1 (m) : building length in span direction
B2 (m) : building length in ridge direction
H (m) : reference height
l (m) : the smallest among 4H, B1, and B2

(3) Peak external pressure coefficients Ĉpe for buildings with circular plans and spherical domes
The peak external pressure coefficients Ĉpe for buildings with circular plans and spherical domes
are given in Table A6.16(3). The values in Table A6.16(3) are applicable to buildings with h D ≤ 1 .

Table A6.16(3) Peak external pressure coefficients Ĉpe for buildings with circular plans and domes

i) Wall

a) Positive peak external pressure coefficients

Positive peak external pressure coefficients are defined in Table A6.17.

b) Negative peak external pressure coefficients

Negative peak external pressure coefficients are defined in Table A6.17.
- 54 - Recommendations for Loads on Buildings

ii) Roof

a) Positive peak external pressure coefficients

Zone Ra Zone Rb Zone Rc
f/D h/D= h/D= h/D= h/D= h/D= h/D=
h/D = 0 h/D = 0.25 h/D= 1
0 0.25 1 0 0.25 1
0 0.6 0.4 0.4 1.1 0.5 0.5 0.6 0.6 0.6
0.05 1.3 0.5 0.5 1.0 0.4 0.4 0.5 0.1 0.1
0.1 1.7 0.7 0.7 0.9 0.3 0.3 0.4 0 0
0.2 0.9(1+7IH) 0.6(1+7IH) 0.4(1+7IH) 1.2 0.6 0.6 0.2 0 0
0.5 1+7IH 1+7IH 1+7IH 1.9 1.3 0.7 0.3 0 0
IH : turbulence intensity at reference height H (m), obtained by substituting H (m) for Z (m) in Eq.
(A6.6).
Note: Linear interpolation is permitted for values of f/D and h/D other than those shown.

b) Negative peak external pressure coefficients

Zone Ra Zone Rb Zone Rc
f/D h/D= h/D= h/D= h/D= h/D= h/D=
h/D = 0 h/D = 0.25 h/D= 1
0 0.25 1 0 0.25 1
0 −4.4 −5.1 −3.3 −1.5 −3.7 −3.0 −0.4 −2.3 −2.3
0.05 −3.0 −4.8 −3.3 −1.5 −2.7 −2.7 −1.3 −1.3 −1.3
0.1 −2.0 −4.2 −3.0 −1.5 −2.2 −2.2 −1.4 −1.4 −1.4
0.2 −2.0 −2.0 −2.0 −1.9 −1.9 −1.9 −2.1 −2.1 −2.1
0.5 −2.6 −2.6 −2.6 −2.8 −2.8 −2.8 −3.0 −3.0 −3.0
Note: Linear interpolation is permitted for values of f/D and h/D other than those shown.

D (m) : building diameter

f (m) : rise
H (m) : reference height
h (m) : eave height

(4) Peak external pressure coefficients Ĉpe for buildings with circular and elliptic plans
Peak external pressure coefficients for buildings with circular and elliptic sections are given in Table
A6.17. The values in Table A6.17 are applicable to cases in which H D2 ≤ 8 and D1 D2 ≤ 3 .
CHAPTER 6 WIND LOADS - 55 -

Table A6.17 Peak external pressure coefficients Ĉpe for buildings with circular and elliptic plans

a) Positive peak external pressure coefficients

Cˆ pe = k Z (1 + 7 I Z )
where
kZ : factor for vertical profile defined in Table A6.8
When 0.8H < Zb, apply kZ = 0.82α. upper part
IZ : turbulence intensity at height Z (m) defined by Eq.
(A6.6)
When the effect of local topography is considered, the
values of kZ and IZ at the reference height H (m) (Z = lower part
H) can be used in the above equation.

b) Negative peak external pressure coefficients

Cˆ pe = −{(k1 − 1)k 2 + k 4 + 1.4}(1 + 7 I H )k 5 D1(m) : major axis of plan
where D2(m) : minor axis of plan
k1: factor for aspect ratio defined in Table A6.10 H(m) : reference height
k2: factor for surface roughness defined in Table A6.10
k4: factor for edge effects defined in the following table
k5: factor for D1/D2 defined in the following table
k6: factor for local negative pressure areas defined in the following table
IH: turbulence intensity at height of H (m), obtained by substituting H (m) for Z (m) in Eq. (6.6)

k4
Upper part
Lower part
H D2 ≤ 2 2 < H D2 ≤ 7 7 < H D2 ≤ 8

0.2 0.2 0.1(H/D2) 0.7

D1/D2 1.0 1.5 2 2.5 3

Zone Wa 1.0 1.3 1.5 1.6 1.75
Zone Wb 1.0 0.95 0.9 0.85 0.8
Note: Linear interpolation is permitted for values of k5 other than those shown.

k6
D1/D2 1.0 1.5 2 2.5 3
k6 0.5 0.35 0.25 0.23 0.22
Note: Linear interpolation is permitted for values of k6 other than those shown.
- 56 - Recommendations for Loads on Buildings

A6.2.6 Factor for effect of fluctuating internal pressures

The factors for effect of fluctuating internal pressures when designing the components/cladding shall
be estimated appropriately by considering the location and size of the distributed gaps and openings on
the walls. For buildings without dominant openings, the values of 0 or −0.5 shall be used.

A6.2.7 Peak wind force coefficients for components/cladding

(1) Peak wind force coefficients ĈC for pitched free roofs with rectangular plan
1) Pitched free roofs
Peak wind force coefficients for pitched free roofs with a rectangular plan are given in Table A6.18(1).
The values in Table A6.18(1) are applicable to cases in which H ≤ 10 m, H / 2 ≤ B ≤ 30 m, and B ≥ D .

Table A6.18(1) Peak wind force coefficients ĈC for pitched free roofs with a rectangular plan

Positive peak wind force coefficients Negative peak wind force coefficients
Zone Roof angle Roof angle
0° 5° 10° 15° 20° 0° 5° 10° 15° 20°
Ra 2.0 2.2 2.4 2.6 2.9 −1.6 −1.6 −3.2 −4.2 −4.0
Rb 2.0 2.2 2.4 2.6 2.9 −1.6 −1.6 −2.4 −3.0 −2.9
Rc 2.0 2.2 2.4 2.6 2.9 −1.6 −1.6 −1.8 −2.5 −2.4
Rd 2.2 2.6 2.9 3.3 3.4 −2.8 −2.8 −1.6 −1.4 −1.7
Re 2.2 2.6 2.9 3.3 3.4 −2.8 −2.8 −2.8 −2.8 −2.7
Rf 2.2 2.2 2.2 2.2 2.4 −2.8 −2.8 −3.2 −4.2 −4.0
Note: Linear interpolation is permitted for values of θ (°) other than those shown.

B(m) : building width

D(m) : building depth
H(m) : reference height
θ (°) : roof angle
CHAPTER 6 WIND LOADS - 57 -

2) Troughed free roofs

The peak wind force coefficients for troughed free roofs with a rectangular plan are given in Table
A6.18(2). The values in Table A6.18(2) are applicable to cases in which H ≤ 10 m, H / 2 ≤ B ≤ 30 m,

and B ≥ D .

Table A6.18(2) Peak wind force coefficients ĈC for troughed free roofs with a rectangular plan

Positive peak wind force coefficients Negative peak wind force coefficients
Zone Roof angle Roof angle
0° 5° 10° 15° 20° 0° 5° 10° 15° 20°
Ra 2.0 1.9 2.4 3.2 3.2 −1.6 −1.9 −2.6 −3.0 −3.2
Rb 2.0 1.9 1.9 2.4 2.6 −1.6 −1.9 −2.6 −3.0 −3.2
Rc 2.2 2.0 1.9 2.4 2.6 −2.8 −3.6 −3.6 −3.6 −3.8
Rd 2.2 2.0 1.9 2.4 2.6 −2.8 −3.6 −4.6 −4.8 −4.9
Re 2.2 2.0 2.4 3.2 3.2 −2.8 −2.8 −2.8 −2.8 −3.0
Note: Linear interpolation is permitted for values of θ (°) other than those shown.

B(m) : building width

D(m) : building depth
H(m) : reference height
θ (°) : roof angle
- 58 - Recommendations for Loads on Buildings

3) Mono-slope free roofs

The peak wind force coefficients for mono-slope free roofs with a rectangular plan are given in Table
A6.18(3). The values in Table A6.18(3) are applicable to cases in which H ≤ 10 m, H / 2 ≤ B ≤ 30 m,

and B ≥ D .

Table A6.18(3) Peak wind force coefficients ĈC for mono-slope free roofs with rectangular plan.

Positive peak wind force coefficients Negative peak wind force coefficients
Zone Roof angle Roof angle
0° 5° 10° 15° 20° 0° 5° 10° 15° 20°
Ra 2.0 2.1 2.2 2.3 2.7 −1.6 −2.0 −2.5 −3.3 −3.5
Rb 2.0 2.1 2.2 2.5 3.1 −1.6 −4.2 −5.1 −6.3 −6.7
Rc 2.0 2.7 3.3 3.5 4.6 −1.6 −2.8 −2.8 −2.8 −3.6
Rd 2.2 2.1 2.2 2.5 3.1 −2.8 −4.2 −5.1 −6.3 −6.7
Re 2.2 2.7 3.3 3.5 4.6 −2.8 −2.8 −2.8 −2.8 −3.6
Rf 2.2 2.7 3.3 3.5 4.6 −2.8 −3.0 −5.1 −6.3 −6.7
Note: Linear interpolation is permitted for values of θ (°) other than those shown.

B(m) : building width

D(m) : building depth
H(m) : reference height
θ (°) : roof angle

A6.3 Gust Effect Factor on Along-Wind Loads on Structural Frames

The gust effect factor GD for along-wind loads on structural frames is estimated from Eq. (A6.16).

C 'g
GD = 1 + g D 1 + φ D2 RD (A6.16)
Cg
where
g D = 2 ln(600ν D ) + 1.2
CHAPTER 6 WIND LOADS - 59 -

0.49 − 0.14α k = 0.07 ( H / B ≥ 1)

C 'g = 2I H
 0.63 BH L
1 +
( H )
0.56



k = 0.15 ( H / B < 1)
k
 ( H / B) 
1 1
Cg = +
3 + 3α 6
3 M D1
φD = λ
2+ β MD
πF
RD = D
4ζ D
RD
ν D = fD
1 + RD

FD =
(
I H2 FS D 0.57 − 0.35α + 2 R 0.053 − 0.042α )
C ' g2
1
R=
fDB
1 + 20
UH
f L
4 D H
UH
F= 5/6
  f D LH   2

1 + 71  
  UH  
0.9
SD = 0.5
  f D H  
2
 f B
1 + 6   1 + 3 D 
  U H    UH 
λ = 1 − 0.4 ln β
M D = ∫0 m(Z )µ 2 (Z )dZ
H

β
Z
µ (Z ) =  
H
where
g D : peak factor for along-wind vibration,
Cg : overturning moment coefficient in along-wind direction,
Cg' : rms overturning moment coefficient in along-wind direction,
φ D : mode shape correction factor for along-wind load,
RD : resonance factor for along-wind vibration,
ν D (Hz): level crossing rate,
α : exponent of power law for wind speed profile as defined in Table A6.3,
I H : turbulence intensity at reference height given by Eq. (A6.6), in which H (m) is
substituted for Z (m),
H (m): reference height as defined in 6.1.2(11),
- 60 - Recommendations for Loads on Buildings

B (m): building width,

LH (m): turbulence scale at reference height given by Eq. (A6.8), in which H (m) is
substituted for Z (m),
FD : along-wind force spectral factor,
ζ D : damping factor for the first translational mode in along-wind direction,
f D (Hz): natural frequency for the first translational mode in along-wind direction,
F : wind force spectrum factor,
S D : size effect factor,
R : correlation coefficient between wind pressures on the windward and leeward faces,
U H (m/s): design wind speed as defined in A6.1.2,
β : exponent of power law for the first translational vibration mode in along-wind direction,
M D (kg): generalized mass of building for along-wind vibration,
M D1 (kg): generalized mass of building for along-wind vibration calculated based on β = 1,
λ : mode correction factor of general wind force,
m(Z ) (kg/m): mass per unit height at Z(m),
µ : first mode shape of building in each direction.

A6.4 Fluctuating Roof Wind Loads on Structural Frames

A6.4.1 Scope of application

This section defines the procedures for estimating fluctuating roof wind loads on structural frames of
buildings. These procedures can be applied to buildings that satisfy the following conditions when the
wind is normal to the front face.
i) Rectangular plan without a change in the vertical direction
ii) Roof angle of less than 10°
H
iii) ≤1
BD
D
iv) 0.4 ≤ ≤ 2.5
B
where
H (m): reference height as defined in 6.1.2(11),
B (m): building width,
D (m): building depth.

A6.4.2 Procedure for estimating wind loads

Fluctuating roof wind loads on the structural frames are calculated from Eq. (A6.17).
πx
WR' = 2q H CR' AR sin g R 1 + RR , (A6.17)
L
CHAPTER 6 WIND LOADS - 61 -

where
g R = 2 ln (600 f R ) + 1.2
πFR
RR =
4ς R
FR = FB + FF
r1 f R*
FB =
(1 + r f ) 2
* r3
R
r4

exp{− 0.5(ln ( f ) )}
r 2
F = 5 *
r6 + 0.5
2π
F R

fR B
f R* =
UH

Parameters C R and r1 − r6 in Eq. (A6.17) are defined as follows for each roof beam direction.
'

(1) Roof beam parallel to the wind direction

H H
C R' = 0.14 − 0.08 + 0.17
D B
r1 = 0.2 , r2 = 1.0 , r3 = 2.0 , r4 = 4.0 , r5 = 1.0 , r6 = 0.1
(2) Roof beam perpendicular to the wind direction
'
CRC − CRE
'
y 5CRE'
− CRC
'
CR' = +
0.4 D 4
H H H H
'
C RE = 0.04 + 0.02 + 0.29 , '
C RC = 0.18 − 0.08 + 0.16
D B D B
2
D D
r1 = 0.3 , r2 = 0.76  − 2.84  + 3.20 , r3 = 3.0 , r4 = 3.0 , r5 = 0.8
 
B B
2
D
r6 = 0.003  + 0.068
B
where
W ' R (N): fluctuating roof wind load,
qH (N/m2): design velocity pressure as defined in A6.1.1,

CR' : rms generalized wind force coefficient of roof beam,

AR (m2): roof area supported by a roof beam,

x (m) : distance from the end of a roof beam,
L (m): span of a roof beam,
f R (Hz): natural frequency of the first mode of a roof beam,
ζ R : damping factor of the first mode of a roof beam,
U H (m/s): design wind speed as defined in A6.1.2,
y (m): distance from the leading edge of the roof beam normal to the wind direction.
- 62 - Recommendations for Loads on Buildings

A6.5 Across-Wind Loads on Structural Frames

A6.5.1 Scope of application

This section defines the procedures for estimating the horizontal across-wind loads on structural frames.
These procedures can be applied to buildings that satisfy the following conditions when the wind is
normal to the front face.
i) Rectangular plan without a change in the vertical direction
H
ii) ≤6
BD
D
iii) 0.2 ≤ ≤5
B
UH
iv) ≤ 10
f L BD

where
H (m): reference height as defined in 6.1.2(11),
B (m): building width,
D (m): building depth,
U H (m/s): design wind speed as defined in A6.1.2,
f L (Hz): natural frequency for the first mode in across-wind direction.

A6.5.2 Procedure for estimating wind loads

Across-wind loads on structural frames are calculated from Eq. (A6.18).
Z
WL = 3q H C 'L A g L 1 + φL2 RL (A6.18)
H
where
C 'L = 0.0082( D / B) 3 − 0.071( D / B) 2 + 0.22( D / B)
g L = 2 ln(600 f L ) + 1.2
πFL
RL =
4ς L
m
4κ j (1 + 0.6β j ) β j ( f L / f sj ) 2
FL = ∑
j =1 π {1 − ( f L / f sj )2 }2 + 4β j2 ( f L / f sj )2
1 D/B <3
m=
2 D/B ≥3
κ 1 = 0.85
κ 2 = 0.02
( D / B ) 4 + 2.3( D / B ) 2 0.12
β1 = +
2.4( D / B ) − 9.2( D / B ) + 18( D / B ) + 9.5( D / B ) − 0.15 ( D / B )
4 3 2

0.28
β2 =
( D / B ) 0.34
CHAPTER 6 WIND LOADS - 63 -

0.12 UH
f s1 =
{1 + 0.38( D / B) } 2 0.89 B
0.56 U H
f s2 =
( D / B ) 0.85 B
β −1
M L1  Z 
φL =   λ
ML  H 
λ = 1 − 0.4 ln β
H
ML = ∫0
m( Z ) µ 2 ( Z )dZ
β
Z
µ (Z ) =  
H
where
WL (N): across-wind load at height Z (m),
q H (N/m2): velocity pressure as defined in A6.1.1,
C ' L : rms overturning moment coefficient in across-wind direction,
A (m2): projected area at height Z (m),
Z (m): height above ground,
H (m): reference height as defined in 6.1.2(11),
g L : peak factor for across-wind vibration,
φ L : mode shape correction factor for across-wind load,
RL : resonance factor for across-wind vibration,
D (m): building depth,
B (m): building width,
f L (Hz): natural frequency for the first mode in across-wind direction,
FL : spectral coefficient of overturning moment in across-wind direction,
ς L : damping factor for the first mode in across-wind direction,
U H (m/s): design wind speed as defined in A6.1.2,
M L (kg): generalized mass of building for across-wind vibration,
M L1 (kg): generalized mass of building for across-wind vibration with β = 1,
β : exponent of power law for the first transitional vibration mode in across-wind direction,
λ : mode correction factor of generalized wind force,
m(Z ) (kg/m): mass per unit height at height Z (m),
µ : first mode shape of building in across-wind direction.

A6.6 Torsional Wind Loads on Structural Frames

A6.6.1 Scope of application

This section defines the procedures for estimating the torsional wind loads on structural frames. These
procedures can be applied to buildings that satisfy the following conditions when wind is normal to the
- 64 - Recommendations for Loads on Buildings

front face.
i) Rectangular plan without a change in the vertical direction
H
ii) ≤6
BD
D
iii) 0.2 ≤ ≤5
B
UH
iv) ≤ 10
f L BD
where
H (m): reference height as defined in 6.1.2(11),
B (m): building width,
D (m): building depth,
U H (m/s): design wind speed as defined in A6.1.2,
f T (Hz): natural frequency of the first mode in torsional direction.

A6.6.2 Procedure for estimating wind loads

Torsional wind loads on structural frames are calculated from Eq. (A6.19).
Z
WT = 1.8q H C 'T AB g T 1 + φT2 RT (A6.19)
H
where
C 'T = 0.04( D / B) 2 + 0.02
g T = 2 ln(600 f T ) + 1.2
πFT
RT =
4ς T
FT = 0.8 FB + v1 FV + w1 FW
18 f m*
FB =
{1 + 0.46(18 f ) } * 1.8
m
2.3

FV =
1 

exp − 0.5
( )
 ln f s* / v3 + 0.5v 22  
2
 
2π v 
v2   2  

FW =
1 
 
exp − 0.5
(
 ln f m* / w3 + 0.5w22  

2
)
2π w2  
w2    

f s* =
{
8.3 f T B 1 + 0.38(D / B ) }
1.5 0.89

UH
f B
f m* = T
UH
β −1
I T1  Z 
φT =   λ
IT  H 
H
IT = ∫
0
i ( Z ) µ 2 ( Z )dZ
CHAPTER 6 WIND LOADS - 65 -

i ( Z ) = m( Z )( B 2 + D 2 ) / 12
β
Z
µ (Z ) =  
H
λ = 1 − 0.4 ln β
The parameters of the spectral coefficient for torsional moment FT are given in Table A6.19.

Table A6.19 Parameters of spectral coefficient for torsional moment FT

D/B ≦ 0.9 D/B > 0.9
v1 －2{(D/B)2－(D/B)} 0.1(D/B)-1
v2 0.56IH(D/B)-1 2IH(D/B)0.5
w1 0.04(D/B)-0.5 0.06
w2 1.75 I H 0.7 I H
-0.4
w3 0.24(D/B)

D/B ≦ 1 D/B > 1

-0.2
v3 0.8(D/B) 0.18(D/B)
where
WT (Nm): torsional wind load at height Z (m),
q H (N/m2): velocity pressure as defined in A6.1.1,
C ' T : rms torsional moment coefficient,
A (m2): projected area of building at height Z (m),
B (m): building width,
Z (m): height above ground,
H (m): reference height as defined in 6.1.2(11),
g T : peak factor for torsional vibration,
φT : mode shape correction factor for torsional wind load,
RT : resonance factor for torsional vibration,
D (m): building depth,
f T (Hz): natural frequency of the first mode in torsional direction,
FT : spectral coefficient of torsional moment,
ς T : damping factor of the first mode in torsional direction,
U H (m): design wind speed as defined in A6.1.2,
β : exponent of power law for the first torsional vibration mode,
I T (kgm): generalized mass of building for torsional vibration,
I T1 : generalized mass of building for torsional vibration with β = 1,
λ : mode correction factor of generalized wind force,
i (Z ) (kgm2/m): inertial moment per unit height at height Z (m),
m(Z ) (kg/m): mass per unit height at height Z (m),
- 66 - Recommendations for Loads on Buildings

I H : turbulence intensity at reference height given by Eq. (A6.6), in which H (m) is

substituted for Z (m),
µ : first mode shape of building for torsional vibration.

A6.7 Horizontal Wind Loads on Lattice Structural Frames

A6.7.1 Scope of application

This section defines the procedures for estimating horizontal wind loads owing to gusts on the lattice
structures built directly on the ground.

A6.7.2 Procedure for estimating wind loads

Horizontal wind loads on the lattice structures are calculated from Eq. (A6.20).
WD = q Z C D GD AF (A6.20)
where
WD (N): wind load,
qZ (N/m2): velocity pressure at height Z (m) obtained by substituting H (m) to Z (m) in
Eq. (A6.1),
C D : wind force coefficient as defined in A6.2.4(3),
GD : gust effect factor calculated using the method described in A6.7.3,
2
AF (m ): projected area of one face of a lattice structure at height Z (m).

A6.7.3 Gust effect factor

Gust effect factor is calculated from Eq.(A6.21).
C g′
GD = 1 + g D φ D 1 + RD , (A6.21)
Cg
where
g D = 2 ln(600vD ) + 1.2
2I H
Cg′ = BD
α +3
1 λB
Cg = −
2α + 3 2α + 4
  
− 0.3 (β − 2 ) + 1.4 (1 − 0.4 ln β )
M D2 BH
φD =  0.5
MD  B0  
π SD F
RD =
4ζ D BD
RD
vD = f D
1 + RD
2
 3  1
BD = 1 − λB 
 4  1 + 2 HB / LH
CHAPTER 6 WIND LOADS - 67 -

BH
λB = 1 −
B0
2
 3  1
S D = 1 − λ B 
 4  {1 + 3 . 5 f D B / U H }{1 + 2 f D H / U H }

4 f D LH / U H
F=
{1 + 71( f D LH / U H ) }
2 5/6

B = (BH + B0 ) / 2
H
M D = ∫ m( Z ) µ 2 ( Z )dZ
0
β
Z
µ (Z ) =  
H
where
φ D : mode shape correction factor of along-wind load,
I H : turbulence intensity at a reference height given by Eq. (A6.6), in which H (m) is
substituted for Z (m),
LH (m): turbulence scale at reference height given by Eq. (A6.8), in which H (m) is
substituted for Z (m),
α : exponent of power law in wind speed profile defined in A6.1.7,
f D (Hz): natural frequency of the first translational mode in along-wind direction,
ζ D : damping factor of the first translational mode in along-wind direction,
H (m): reference height, i.e., height of the lattice structure,
B0 (m): structure width at ground level,
BH (m): structure width at height H (m),
BD : background excitation factor,
U H (m/s): design wind speed as defined in A6.1.2,
M D (kg): generalized mass for along-wind vibration,
M D2 (kg): generalized mass for along-wind vibration calculated based on β = 2,
m(Z ) (kg/m): mass of unit height at Z (m),
β : exponent of power law for vibration mode,
µ : first mode shape in along-wind direction.

A6.8 Wind Loads on Free Roofs

A6.8.1 Scope of application

This section defines the procedures for estimating wind loads on structural frames of free roofs. This
procedure can be applied to free roofs with H ≤ 10 m, H / 2 ≤ B ≤ 30 m, and B ≥ D .
where
- 68 - Recommendations for Loads on Buildings

H(m): reference height defined in 6.1.2(11),

B(m): building width,
D(m): building depth.

A6.8.2 Procedure for estimating wind loads

Wind loads on structural frames of free roofs are calculated from Eq. (A6.22).
WF = qH CR GR AF (A6.22)
where
WF (N): wind load,
qH (N/m2): design velocity pressure as defined in A6.1.1,
CR : wind force coefficient as defined in A6.2.4(2),
GR : gust effect factor for roof wind load as defined in A6.8.3,
AF (m2): subject area.

A6.8.3 Gust effect factor

The gust effect for estimating the wind loads on structural frames of free roofs is calculated from Eq.
(A6.23).
GR = (1 + 3I u ) 2 (A6.23)
where
IH: turbulence intensity at reference height given by Eq. (A6.6), in which H (m) is substituted
for Z (m).

A6.9 Vortex-Induced Vibration

A6.9.1 Scope of application

This section describes the wind loads on buildings and members with circular sections caused by
vortex-induced vibration.
A6.9.2 Vortex-induced vibration and resulting wind loads on buildings with circular plans
Wind loads on buildings with circular plans caused by vortex-induced vibration are calculated from
Eq. (A6.24).
Z
Wr = 0.8 ρU r2 C r A (A6.24)
H
where U r (m/s) is the resonance wind speed calculated from Eq. (A6.25).
U r = 5 f L Dm (A6.25)
where
Wr (N): wind load at height Z (m),
ρ (kg/m3): air density (= 1.22),
C r : wind force coefficient at resonance as defined in Table A6.20,
CHAPTER 6 WIND LOADS - 69 -

Z (m): height above ground,

H (m): reference height as defined in 6.1.2(11),
A (m2): projected area at height Z (m),
f L (Hz): natural frequency of the first mode in the across-wind direction,
Dm (m): building diameter at height 2H / 3 (m).

Table A6.20 Wind force coefficient at resonance C r

U r Dm ρs ζ L < 5 ρs ζ L ≥ 5
1.3 1.5 ρ 1.7
U r Dm < 3 +
ζL ζ L ρS ζL
3 ≤ U r Dm < 6 Linear interpolation Linear interpolation
0.53 0.16 ρ 0.57
6 ≤ U r Dm +
ζL ζ L ρs ζL
where
ζ L : damping factor of the first mode in across-wind direction,
ρ s (kg/m3): building density as given by ρ s = M /( HDm DB ) ,
M (kg): total building mass,
DB (m): building diameter at the base.

A6.9.3 Vortex-induced vibration and resulting wind load on building components with circular
sections
Wind loads on building components with circular sections caused by vortex-induced vibration are
calculated from Eq. (A6.27) when the conditions of Eq. (A6.26) are satisfied.
L UH
≥ 15 and ≥ 3.5 (A6.26)
D fLD

where
L (m): length of the member
D (m): diameter of the member
U H (m/s): design wind speed at height H (m), which is the mean height of the component,
as defined in A6.1.2
f L (Hz): natural frequency of the first bending mode

 πx  M 0.26U r*
Wr = (2πf L ) 2 sin   A (A6.27)
 L  L 0.75δ + 0.36U r
1.1 *

where U r* is the non-dimensional resonant wind speed, and δ is a mass-damping parameter,

calculated from Eqs. (A6.28) and (A6.29), respectively.

- 70 - Recommendations for Loads on Buildings

3
U r* = 5 + (A6.28)
δ
4πζ L M
δ= (A6.29)
ρD 2 L
where
Wr (N): wind load at position x (m),
x (m): distance from the end of the component,
M (kg): total mass of the component,
L (m): span of the component,
A (m2): projected area at x (m)
ζ L : damping factor of the first bending mode of the component,
ρ (kg/m3): air density (= 1.22).

A6.10 Combination of Wind Loads

A6.10.1 Scope of application

This section defines the procedures used for estimating the combination of horizontal and roof wind
loads.
For buildings not satisfying the conditions of Eq. (6.1), a combination of the along-wind load and
across-wind load should be considered with reference to (1) in A6.10.2. For buildings satisfying the
condition of Eq. (6.1), a combination of horizontal wind loads should be considered with reference to
(2) in A6.10.2.
For horizontal and roof wind loads, the combination load shall be considered with reference to
A6.10.3.

A6.10.2 Combination of horizontal wind loads

(1) Buildings with aspect ratio H / BD < 3
The along-wind load calculated using 6.2 and the across-wind load calculated from Eq. (A6.30) shall
be considered together.
WLC = γWD (A6.30)
where
D
γ = 0.35 and γ ≥ 0.2
B
WLC (N): combined across-wind load,
WD (N): along-wind load defined in 6.2,
B (m): building width,
D (m): building depth.
(2) Buildings with aspect ratio of H / BD ≥ 3
CHAPTER 6 WIND LOADS - 71 -

The three cases of load combinations described in Table A6.21 shall be considered.

Table A6.21 Horizontal wind load combinations

Combination along-wind force across-wind force torsional moment
1 WD 0.4WL 0.4WT
 0.6 
2 WD  0.4 + 
GD 
WL ( )
2 + 2 ρ LT − 1 WT

 0.6 
3 WD  0.4 + 
GD 
( )
2 + 2 ρ LT − 1 WL WT

Note) WD, WL, and WT are the load effects from the along-wind load, across-wind load, and torsional
load, as defined in 6.2, A6.5, and A6.6, respectively. Here, GD is the gust effect factor for the
along-wind loads defined in A6.3, and ρ LT is the correlation coefficient between the across-
wind vibration and torsional vibration defined in Table A.6.22.

Table A.6.22 Correlation coefficient ρ LT between across-wind and torsional vibrations

ρLT
D/B f1B/UH
ξ=1.0 ξ=1.1 ξ≧1.4
≤0.1 0.9 0.7 0.5
0.2 0.3 0.6 0.5
≤0.5 0.3 0.4 0.6 0.6
0.6 0.6 0.6 0.6
≥1 0.7 0.7 0.7
≤0.1 0.8 0.5 0.2
1 0.2 0.6 0.5 0.5
≥0.3 0.5 0.5 0.5
≤0.05 0.6 0.4 0.3
≥2 0.1 0.6 0.2 0.2
≥0.2 0.2 0.2 0.2
Note) For intermediate values of ξ , f1 B /U H , and D / B , linear interpolation may be applied.
 fL / fT fL ≥ fT
ξ =
 fT / fL fL < fT
f L (Hz): natural frequency of the first translational mode in across-wind direction,
f T (Hz): natural frequency of the first torsional mode,
f1 (Hz): the smaller between f L and f T .
- 72 - Recommendations for Loads on Buildings

A6.10.3 Combination of horizontal and roof wind loads

A combination of horizontal wind loads defined in A6.10.2 and roof wind load calculated from 6.3
shall be considered together.

A6.11 Wind Loads Estimation using CFD

For CFD applied to the wind load estimation, the required physical quantities are the wind velocity,
velocity fluctuation, and time histories of the wind pressure or wind force. Then, the governing equations
for computation should be constructed for an incompressible viscous fluid, and can express the solution
as unsteady flows and wind pressures based on the engineering scale of the buildings and their
surroundings. Because the wind is a turbulent flow, it cannot be obtained through an analytical solution.
To obtain a numerical solution with sufficient accuracy, a discrete approximate solver and turbulent
modeling should be selected appropriately.

A6.12 Effects of Neighboring Tall Buildings

Effects of interference by neighboring buildings and structures shall be considered for an estimation
of the design wind loads on the structural frames and components/cladding of the buildings, when such
effects may increase the wind loads.

A6.13 Response Acceleration

A6.13.1 Scope of application

This section defines the maximum along-wind response acceleration for ordinary buildings, the
maximum across-wind response acceleration for buildings with rectangular plans satisfying the
conditions of A6.5.1, and the maximum torsional response acceleration for buildings with rectangular
plans satisfying the conditions of A6.6.1.

A6.13.2 The maximum response acceleration in along-wind direction

The maximum response acceleration in the along-wind direction at the top of a building is calculated
from Eq. (A6.31).
qH g aD BHCH Cg′ λ RD
aD max = (A6.31)
MD
where
g aD = 2 ln (600 f D ) + 1.2
where
aD max (m/s2) : maximum response acceleration in along-wind direction at the building top,
CHAPTER 6 WIND LOADS - 73 -

qH (N/m2) : velocity pressure as defined in A6.1.1,

B (m) : building width,
H (m) : reference height as defined in (11) of 6.1.2,
C H : wind force coefficient CD at reference height as defined in A6.2,

Cg′ : rms overturning moment coefficient as defined in A6.3,

λ : mode correction factor of generalized wind force as defined in A6.3,

RD : resonance factor as defined in A6.3,
M D (kg): generalized mass of building for along-wind vibration as defined in A6.3,
f D (Hz) : natural frequency of the first mode in along-wind direction,
g aD : peak factor for along-wind vibration,

A6.13.3 Maximum response acceleration in across-wind direction

The maximum response acceleration in the across-wind direction at the top of the building is
calculated from Eq. (A6.32).
q H g L BHC 'L λ RL
aLmax = (A6.32)
ML
where
a Lmax (m/s2): maximum response acceleration in across-wind direction at the building top,
q H (N/m2): velocity pressure as defined in A6.1.1,
g L : peak factor for the across-wind vibration as defined in A6.5.2,
B (m): building width
H (m): reference height as defined in 6.1.2(11),
C ' L : rms overturning moment coefficient as defined in A6.5.2,
λ : mode correction factor of generalized wind force as defined in A6.5.2,
RL : resonance factor as defined in A6.5.2,
M L (kg): generalized mass of building for across-wind vibration as defined in A6.5.2.

A6.13.4 Maximum torsional response acceleration

The maximum torsional response acceleration at the top of a building is calculated from Eq. (A6.33).
0.6q H g T B 2 HC 'T λ RT
aTmax = (A6.33)
IT
where
a Tmax (m/s2): maximum torsional response angular acceleration at the top of a building,
q H (N/m2): velocity pressure as defined in A6.1.1,
g T : peak factor for torsional vibration defined in A6.6.2,
B (m): building width,
H (m): reference height as defined in (11) of 6.1.2,
- 74 - Recommendations for Loads on Buildings

C 'T : rms torsional moment coefficient as defined in A6.6.2,

λ : mode correction factor of generalized wind force as defined in A6.6.2,
RT : resonance factor as defined in A6.6.2,
I T (kgm2): generalized inertia moment of building for torsional vibration as defined in A6.6.2.

A6.14 One-Year Recurrence Wind Speed

One-year recurrence wind speed U1H is calculated from Eq. (A6.34).

U 1H = U 1 E H (A6.34)
where
U1 (m/s): one-year recurrence of 10-min mean wind speed at 10 m above the ground
corresponding to the flat terrain category of II, defined in Fig. A6.6,
EH: wind speed profile factor at reference height H, defined in A6.1.7, according to the surface
roughness condition at the construction site.

A6.15 Duration Time of Wind Loads

In a fatigue evaluation of the structural members, the duration time of the wind load should be
evaluated by taking into account the service life of the member and/or its maintenance cycle.
To examine the wind-induced fatigue on members (such as structural members, components /cladding,
base-isolation devices, and vibration control devices), it is necessary to define the distribution of the
wind speed and its duration time at the construction site.
The safety of the members against fatigue damage accumulated during the evaluated duration time of
the wind load shall then be confirmed.
CHAPTER 6 WIND LOADS - 75 -

22
20
20

20 18
20

16
22
18
16
20 20 24
22
18
20 22

18
22
20

Izu islands, not shown in the map and

Ogasawara Islands, Satsunan Islands,
27
Okinawa Islands, Daitou Islands,
Sakishima Islands, not shown in the map

Figure A6.6 One-year recurrence 10-minute mean wind speed U1 (m/s) at 10 m above the ground for
flat open terrain
-76-

CHAPTER 7 SEISMIC LOADS

7.1 Principles of Defining Seismic Loads

7.1.1 Component factors of seismic loads

The seismic load on buildings is set by taking into consideration the various spectral characteristics
relating to the seismic source, the seismic wave propagation, the ground motion at the engineering
bedrock, the soil amplification in the surface layer and the soil-structure interaction, and the dynamic
property of the buildings.
(1) For ordinary buildings, the seismic load is evaluated using the acceleration response spectrum,
and the equivalent static force procedure is applied (see Sec.7.2).
(2) For buildings that are irregular in plan or elevation, the design earthquake motion is
represented by the acceleration time history and the dynamic response analysis is applied. In
this case, it is recommended that the equivalent static force procedure also be applied (see
Sec.7.3).

7.1.2 Idealization of building and soil

Prior to the estimation of the equivalent static seismic loads or the response analyses using design
earthquake motions, the building and its foundation (including basement) should be appropriately
idealized. The equivalent static seismic loads should then be estimated or response analyses should be
conducted using the design earthquake motion for the idealized model.

(1) Analytical model

In the case of a regular shaped building with rigid floors, the building is idealized as a multi- degree-
of-freedom system with the mass lumped at the floor levels, considering the effects of sway and rocking
of the foundation. In this case, the input ground motion is assumed to be transmitted at the bottom of
the foundation.
When a lesser effect of the interaction between the building and soil can be estimated, the building
can be idealized as being fixed at the foundation. In this case, the seismic input motion is assumed to be
transmitted at the bottom of the foundation or at the ground level. It is also recommended to consider
the effects of soil-structure interaction.
For irregular shaped buildings or those where torsional vibration is expected to be dominant, the
building is idealized as a three-dimensional frame. The seismic input motion is assumed to be
transmitted at the bottom of the foundation or at the ground level according to the idealized model.
When the soil layers above the engineering bedrock are idealized, the assumed location of the seismic
input motion is at the surface of the engineering bedrock.
CHAPTER 7 SEISMIC LOADS - 77 -

(2) Analysis method

When response analyses of the buildings and soil are conducted, the analysis method should be
selected depending on the degree of required nonlinearity.
Based on the plastic deformation capabilities of the structural members, the building hysteresis
characteristics of the linear and non-linear domain should be appropriately reflected for analysis.

(3) Weight for seismic load

The weight for seismic load estimation is calculated as the sum of the dead load and the applicable
portion of the live load (snow load should be considered in heavy snow districts). The live load QE for
calculating the seismic load is evaluated from the actual conditions of the building, including fixed
objects, or from the values given in Table 7.1 in the case of the room types listed in Table 4.1.
・
・Table 7.1 Live load for seismic load QE
Room type† 1 2 3 4 5 6 7 8
QE (N/m2) 300 200 700 900 1,100 900 2,200 300
† Room types are given in Table 4.1.

7.2 Estimation of Equivalent Static Seismic Loads

7.2.1 Methods for estimating seismic load

The seismic story shear force at the i-th story 𝑉𝑉E𝑖𝑖 is estimated as follows:
𝑆𝑆a �𝑇𝑇1 , ζ1 �
𝑉𝑉E𝑖𝑖 = 𝑘𝑘D𝑖𝑖 𝑘𝑘F𝑖𝑖 𝑊𝑊𝑖𝑖 (7.1)
𝑔𝑔
where 𝑘𝑘D𝑖𝑖 : factor to reduce the story shear force according to the ductility of the building (see 7.2.5),
𝑘𝑘F𝑖𝑖 : factor to increase the story shear force according to the irregularity of the building (see 7.2.6),
𝑆𝑆a �𝑇𝑇1 , ζ1 �: spectral acceleration for the fundamental mode (natural period 𝑇𝑇1 and damping
factor ζ1 ) (see 7.2.4),
𝑔𝑔: gravitational acceleration,
𝑊𝑊𝑖𝑖 : equivalent weight of the i-th story calculated considering the vibration characteristics of a
multi-story building (see 7.2.3).

7.2.2 Natural period and damping factor considering the soil-structure interaction

When it is inappropriate to assume that the ground is rigid, e.g., for soft soil, an eigenvalue analysis
is conducted considering the sway and rocking motions.
Even if an eigenvalue analysis is not performed, the fundamental natural period 𝑇𝑇1 and the
- 78 - Recommendations for Loads on Buildings

corresponding damping factor 𝜁𝜁1 are calculated for a specified building, considering the soil-structure
interaction, as follows:

T1 = T f2 + Ts2 + Tr2 , (7.2)

where 𝑇𝑇𝑓𝑓 : fundamental natural period of the building fixed at the base,
𝑇𝑇𝑠𝑠 : natural period of sway assuming the building is rigid,
𝑇𝑇𝑟𝑟 : natural period of rocking assuming the building is rigid.
3 3 3
Tf  T  T 
ζ1 = ζ f   +ζs s
T  + ζ r  r  , (7.3)
T 
 1   1   T1 
where ζ𝑓𝑓 : damping factor for the fundamental mode of the building fixed at the base,
ζ𝑠𝑠 : damping factor of sway assuming the building is rigid,
ζ𝑟𝑟 : damping factor of rocking assuming the building is rigid.

7.2.3 Methods for estimating the equivalent building weight

(1) Procedure using an eigenvalue analysis

Equivalent weight of the i-th story of a building is estimated as follows:

𝑗𝑗
𝑊𝑊𝑖𝑖 = �∑𝑗𝑗=1
𝑐𝑐
𝑊𝑊𝑖𝑖𝑖𝑖 2 , (7.4)

where 𝑗𝑗𝑐𝑐 : the maximum number of natural modes considered,

𝑊𝑊𝑖𝑖𝑖𝑖 : the equivalent weight of the i-th story for the j-th natural mode.
The equivalent weight of the i-th story for the j-th natural mode 𝑊𝑊𝑖𝑖𝑖𝑖 is estimated as follows:

𝑊𝑊𝑖𝑖𝑖𝑖 = 𝑘𝑘Saj ∑𝑛𝑛𝑘𝑘=𝑖𝑖�𝑤𝑤𝑘𝑘 𝛽𝛽𝑗𝑗 ∅𝑘𝑘𝑘𝑘 �, (7.5)

where 𝑘𝑘Saj : ratio of 𝑆𝑆a �𝑇𝑇𝑗𝑗 , ζ𝑗𝑗 � to 𝑆𝑆a �𝑇𝑇1 , ζ1 � �= 𝑆𝑆a �𝑇𝑇𝑗𝑗 , ζ𝑗𝑗 ��𝑆𝑆a �𝑇𝑇1 , ζ1 ��,
𝑛𝑛: number of stories,
𝑤𝑤𝑘𝑘 : weight of the k-th story, including the dead and live loads,
𝛽𝛽𝑗𝑗 : participation factor for the j-th natural mode,
∅𝑘𝑘𝑘𝑘 : the j-th natural mode shape of the k-th story.
The j-th natural mode of the k-th story ∅𝑘𝑘𝑘𝑘 , the j-th natural period 𝑇𝑇𝑗𝑗 , and the participation factor 𝛽𝛽𝑗𝑗
are obtained through an eigenvalue analysis of a model for a specified building. The weight of the k-th
story 𝑤𝑤𝑘𝑘 is obtained from the dead and live loads specified in 7.1.2 for combination with the seismic
load.

When the natural periods for different modes are expected to be close to each other, instead of Eq.
(7.5), 𝑊𝑊𝑖𝑖 is calculated as follows:

𝑗𝑗 𝑗𝑗
𝑊𝑊𝑖𝑖 = �∑𝑗𝑗=1
𝑐𝑐
∑𝑘𝑘=1
𝑐𝑐
𝑊𝑊𝑖𝑖𝑖𝑖 𝜌𝜌𝑗𝑗𝑗𝑗 𝑊𝑊𝑖𝑖𝑖𝑖 , (7.6)

where 𝜌𝜌𝑗𝑗𝑗𝑗 is the correlation factor for the j-th and k-th natural modes, and is given as follows:
CHAPTER 7 SEISMIC LOADS - 79 -

8�ζ𝑗𝑗 ζ𝑘𝑘 �ζ𝑗𝑗 −𝑟𝑟𝑗𝑗𝑗𝑗 ζ𝑘𝑘 �𝑟𝑟𝑗𝑗𝑗𝑗 3⁄2

𝜌𝜌𝑗𝑗𝑗𝑗 = 2 , (7.7)
�1−𝑟𝑟𝑖𝑖𝑖𝑖 � +4ζ𝑗𝑗 ζ𝑘𝑘 𝑟𝑟𝑖𝑖𝑖𝑖 �1−𝑟𝑟𝑗𝑗𝑗𝑗 2 �+4�ζ𝑗𝑗 2 −ζ𝑘𝑘 2 �𝑟𝑟𝑗𝑗𝑗𝑗 2

where 𝑟𝑟𝑗𝑗𝑗𝑗 is the natural frequency ratio of the j-th and k-th modes (𝑟𝑟𝑗𝑗𝑗𝑗 = 𝜔𝜔𝑗𝑗 ⁄𝜔𝜔𝑘𝑘 = 𝑇𝑇𝑘𝑘 ⁄𝑇𝑇𝑗𝑗 ), ζ𝑗𝑗 and ζ𝑘𝑘
are damping factors for these modes, respectively.

(2) Procedure without eigenvalue analysis

For ordinary buildings that are regular in plan and elevation, the equivalent weight 𝑊𝑊𝑖𝑖 can be
estimated without an eigenvalue analysis as follows:
𝑊𝑊𝑖𝑖 = 𝑘𝑘V𝑖𝑖 𝜇𝜇m ∑𝑛𝑛𝑘𝑘=𝑖𝑖 𝑤𝑤𝑘𝑘 , (7.8)
where 𝑘𝑘V𝑖𝑖 is the seismic shear distribution factor for the i-th story,
𝜇𝜇m : adjustment factor for a single-degree-of-freedom assumption in the case of a multi-degree-
of-freedom system.
Here, 𝑘𝑘V𝑖𝑖 can be calculated as follows:
1
𝑘𝑘V𝑖𝑖 = 1 + 𝑘𝑘1 (1 − 𝛼𝛼𝑖𝑖 ) + 𝑘𝑘2 � − 1�, (7.9)
�𝛼𝛼𝑖𝑖

where 𝛼𝛼𝑖𝑖 is the normalized weight, and is given as follows:

∑𝑛𝑛 𝑤𝑤
𝛼𝛼𝑖𝑖 = ∑𝑛𝑛𝑘𝑘=𝑖𝑖 𝑤𝑤𝑘𝑘 . (7.10)
𝑘𝑘=1 𝑘𝑘

The factors 𝑘𝑘1 and 𝑘𝑘2 are determined depending on the number of stories (or height or fundamental
natural period 𝑇𝑇1 ):
• when the natural period of the first mode 𝑇𝑇1 is within the range where the acceleration response
spectrum is constant in that period range, e.g., low-rise buildings, 𝑘𝑘1 ≈ 1, 𝑘𝑘2 ≈ 0,
• when the natural period of the first mode 𝑇𝑇1 is within the range where the velocity response
spectrum is constant in that period range, e.g., high-rise buildings, 𝑘𝑘1 ≈ 0, 𝑘𝑘2 ≈ 1,
• and for other buildings, e.g., medium-rise buildings, k1 and k2 have intermediate values.
The factor 𝜇𝜇m is 0.82 when 𝑇𝑇1 is small, 0.90 when 𝑇𝑇1 is large, and 1.0 for a single-degree-of-freedom
system.

7.2.4 Acceleration response spectrum

(1) The acceleration response spectrum at the base of the foundation (or at the ground level)
The acceleration response spectrum at the base of the foundation (or at the ground level) to evaluate
the seismic load is given as follows:
𝑆𝑆a (𝑇𝑇, ζ) = 𝑘𝑘p (𝑇𝑇, ζ)𝜎𝜎a (𝑇𝑇, ζ), (7.11)
where 𝑘𝑘p (𝑇𝑇, ζ): peak factor of the acceleration response of a single-degree-of-freedom
system,
𝜎𝜎a (𝑇𝑇, ζ)：root mean square of the acceleration response of a single-degree-of-
freedom system,
- 80 - Recommendations for Loads on Buildings

T: period (s),
ζ: damping factor.

Here, 𝑘𝑘p (𝑇𝑇, ζ) is usually assumed to be 3.0, and may also be calculated through appropriate methods.
In addition, 𝜎𝜎a (𝑇𝑇, ζ) is calculated as follows:

∞
𝜎𝜎a (𝑇𝑇, ζ)2 = ∫0 |𝐻𝐻(𝜔𝜔)|2 𝐺𝐺a (𝜔𝜔)𝑑𝑑𝑑𝑑, (7.12)
Here, 𝐻𝐻(𝜔𝜔) is the transfer function of acceleration, and is calculated as follows:
(𝜔𝜔0 2 )2 + (2𝜁𝜁𝜔𝜔0 𝜔𝜔)2
|𝐻𝐻(𝜔𝜔)|2 = (7.13)
(𝜔𝜔0 2 − 𝜔𝜔 2 )2 + (2𝜁𝜁𝜔𝜔0 𝜔𝜔)2
where 𝜔𝜔 = 2𝜋𝜋⁄𝑇𝑇 and 𝜔𝜔0 is the natural circular frequency for a single-degree-of-freedom system.
In addition, 𝐺𝐺a (𝜔𝜔) is the acceleration power spectrum at the ground level or at the base of the
foundation, and is calculated as follows:
𝐺𝐺a (𝜔𝜔) = |𝐻𝐻GS (𝜔𝜔)|2 |𝐻𝐻SSI (𝜔𝜔)|2 𝐺𝐺aE (𝜔𝜔), (7.14)
where 𝐻𝐻GS (𝜔𝜔): soil amplification function representing the amplification characteristics of the surface
soil above the engineering bedrock,
𝐻𝐻SSI (𝜔𝜔): adjustment function to represent the soil-structure interaction (equal to 1 when the
soil-structure interaction is not considered),
𝐺𝐺aE (𝜔𝜔): acceleration power spectrum at the engineering bedrock.

The acceleration power spectrum of earthquake motion on the engineering bedrock 𝐺𝐺aE (𝜔𝜔) is obtained
by converting the acceleration response spectrum of earthquake motion on the engineering bedrock
𝑆𝑆aE (𝑇𝑇, 0.05). The duration of ground motion 𝑇𝑇d for spectral conversion is assumed to be 20 s.

(2) Acceleration response spectrum of earthquake motion on the engineering bedrock

The acceleration response spectrum of earthquake motion on the engineering bedrock for a return
period of r years, SaE(T, 0.05), can be evaluated by either of the following two ways.

1) Uniform hazard spectrum

A uniform hazard spectrum (UHS) for a return period of r years is used for the acceleration response
spectrum in the engineering bedrock. The spectrum is evaluated by multiplying the basic acceleration
response spectrum given in Sec.7.2.4 (5) by the return period conversion factor krE. When calculating
the basic value of a seismic load, the return period conversion factor krE is set to 1.0.

2) Spectrum evaluated from normalized acceleration response spectrum

The spectrum is evaluated from the normalized acceleration response spectrum S0(T,0.05) given in
Eq. (7.15), considering the basic peak acceleration on the engineering bedrock a0 and the return period
conversion factor krE. The shape of the normalized acceleration response spectrum is shown in Fig. 7.2.6.
CHAPTER 7 SEISMIC LOADS - 81 -

When calculating the basic value of a seismic load, the return period conversion factor krE is set to 1.0.

𝑆𝑆𝑎𝑎E (𝑇𝑇, 0.05) = 𝑘𝑘𝑟𝑟𝑟𝑟 𝑎𝑎0 𝑆𝑆0 (𝑇𝑇, 0.05)

𝑘𝑘𝑟𝑟𝑟𝑟 𝑎𝑎0 (1 + (𝑘𝑘𝑅𝑅0 − 1)𝑇𝑇⁄𝑇𝑇𝐶𝐶 ′) : (𝑇𝑇 < 𝑇𝑇𝐶𝐶 ′) (7.15)

= � 𝑘𝑘𝑟𝑟𝑟𝑟 𝑎𝑎0 𝑘𝑘𝑅𝑅0 : (𝑇𝑇𝐶𝐶 ′ ≤ 𝑇𝑇 < 𝑇𝑇𝑐𝑐 )
𝑘𝑘𝑟𝑟𝑟𝑟 𝑎𝑎0 𝑘𝑘𝑅𝑅0 𝑇𝑇𝑐𝑐 ⁄𝑇𝑇 : (𝑇𝑇𝑐𝑐 ≤ 𝑇𝑇)

where krE： return period conversion factor given in Sec. 7.2.4(6),

a0： basic peak acceleration of earthquake motion on the engineering
bedrock in Sec. 7.2.4(5),
S0(T,0.05)： normalized acceleration response spectrum of earthquake motion on the
engineering bedrock,
Tc, Tc’： corner periods related to the predominant period of the engineering
bedrock, Tc = 0.3-0.5 s and Tc’/Tc = 0.2-0.5 s,
kR0： acceleration response ratio of approximately 2 to 3, where the
acceleration response becomes approximately constant.

kR0

2
S0(T, 0.05)

0
Tc' Tc
period(s)
Figure 7.2.6 Normalized acceleration response spectrum of earthquake motion on the engineering
bedrock

(3) Soil amplification function

The amplification of seismic motion from the engineering bedrock to the ground level or to the base
of the foundation is evaluated using the transfer function obtained from a one-dimensional shear wave
propagation analysis (equivalent linear analysis). However, if this analysis is difficult, it can be
simplified through the formula of soil amplification function 𝐻𝐻GS (𝜔𝜔) given by

1 2
|𝐻𝐻GS (𝜔𝜔)|2 = � � , (7.16)
cos 𝐴𝐴+i𝛼𝛼 G sin 𝐴𝐴
- 82 - Recommendations for Loads on Buildings

𝜔𝜔𝑇𝑇G
𝐴𝐴 = , (7.17)
4�1+2𝑖𝑖𝜁𝜁G

where 𝛼𝛼G ：impedance ratio from the engineering bedrock to the surface soil,
𝑇𝑇G ：predominant period (s) of soil above the engineering bedrock,
𝜁𝜁G ：damping factor of soil above the engineering bedrock,
𝑖𝑖 ：imaginary unit (𝑖𝑖 = √−1).

(4) Adjustment function of soil-structure interaction

The adjustment function used to represent a kinematic soil-structure interaction is evaluated by
considering the embedment effects of a basement, and is calculated as follows:

 1 
 : δ d ≤ 1
2
1 + 2ηδ d ,
H SSI (ω)
2 (7.18)
= 
 1 : δd > 1
1 + 2η 

where η : ratio between embedment depth of foundation d and the equivalent width l (= �𝐴𝐴𝑓𝑓 , 𝐴𝐴𝑓𝑓 , area
of foundation) ( = d/l ),
δ d : normalized depth of foundation embedment (2𝜔𝜔𝜔𝜔⁄(𝜋𝜋𝑉𝑉𝑠𝑠 )),
𝑉𝑉S: shear wave velocity of the soil adjacent to the basement.

(5) Basic acceleration response spectrum

The basic acceleration response spectrum of earthquake motion on the surface of the engineering
bedrock Sa0(T, 0.05) is given by the uniform hazard spectrum for a return period of 100 years. Instead
of using a uniform hazard spectrum, it is also given by the normalized acceleration response spectrum
S0(T, 0.05) multiplied by the basic peak acceleration at the surface of the engineering bedrock a0 shown
in a standard hazard map. It may also be evaluated through an appropriate seismic hazard analysis.

(6) Return period conversion factor

The return period conversion factor k rE (T ) of the acceleration response spectrum for a return period
of r years is determined from a hazard analysis at the construction site.

7.2.5 Factor to reduce story shear force according to plastic deformation capacity and response
displacement

The factor kDi related to the ductility of the i-th story is calculated according to the inelastic
characteristics and limit deformation of the building. The seismic story shear force of each story is
reduced using the factor related to ductility. This is verified during the estimation of the seismic story
shear force by confirming that the global safety is assured and damage to the building members are
within the allowable limit.
CHAPTER 7 SEISMIC LOADS - 83 -

7.2.6 Factor to increase the story shear force according to building irregularity

The factor kFi of the i-th story of a building with irregularities in plan or elevation is calculated by
considering the possibility of damage concentrated on a certain story, torsional vibration, or other
element.

7.3 Design Earthquake Motions and Building Response Evaluation

7.3.1 Fundamental concept

For a building whose dynamic behavior should be precisely captured, such as a very important
building, the design earthquake motions are generated according to this section, and a dynamic response
analysis is carried out for the appropriate model taking into account the soil-structure interaction and
dynamic properties of the building.
The procedures used to generate the design earthquake motions are classified into two types as
follows:
• generation of design earthquake motions that are compatible with the design response
spectrum,
• generation of design earthquake motions based on scenario earthquakes considering local
site conditions and design requirements of the building.

7.3.2 Design earthquake motions fit to the design response spectrum

The design earthquake motions are generated as simulated earthquake motions (time history) that are
compatible with the design response spectrum, described in Sec.7.2, as a target response spectrum.
The target response spectrum is defined either
• at the free surface of the engineering bedrock, or
• at the location of input motion to the idealized model of the building.
In the latter case, the target response spectrum is evaluated by considering the soil conditions between
the engineering bedrock and the definition location of the target response spectrum.

7.3.3 Design earthquake motions based on scenario earthquakes

First, the service life of the building and the engineering bedrock are decided according to the
conditions of the construction site and the requirements of the building. Second, the scenario earthquakes
that need to be considered are assumed from the numbers of earthquake sources that needs be taken into
account for the design. Third, the earthquake motions at the construction site are evaluated considering
- 84 - Recommendations for Loads on Buildings

the path characteristics from the seismic source to the site, the amplification characteristics due to the
local soil conditions, and so on, in addition to the characteristics of the assumed scenario earthquakes.
Finally, the design earthquake motions are set from the evaluated earthquake motions at the construction
site.

(1) Assumption of scenario earthquakes that need to be considered

The engineering bedrock is selected after considering the conditions of the construction site (e.g.,
seismic potential, geography, geology of the surrounding area) and the service life of the building is
determined based on its properties (e.g., building use, structural characteristics, or function). The
scenario earthquakes are assumed after considering those earthquakes that need to be taken into account
for the building design, such as inter-plate, intra-plate (crustal and oceanic), and shallow-crustal
earthquakes, based on the conditions mentioned above. The parameters of the scenario earthquakes are
determined according to the characteristics of the local seismic activity and/or the characteristics of the
previously known individual earthquakes.

(2) Evaluation of earthquake motions

Earthquake motions at the construction site are evaluated by reflecting the source characteristics of
the assumed scenario earthquakes, the path characteristics from the seismic source to the site, the
amplification characteristics owing to the local soil conditions at the construction site, and so on. Among
the various methods available, appropriate methods are adopted to evaluate the earthquake motions
according to the quality and quantity of information on the source, path, and site characteristics,
depending on the conditions of the construction site and building.

(3) Setting the design earthquake motions

The design earthquake motions at the engineering bedrock of a construction site are set from the
evaluated earthquake motions at the construction site, determining from the response characteristics and
the design requirements of the building, among other factors. It is suggested that not only the spectral
amplitude but also the characteristics of the Fourier phase-dependence, or time-dependence, and the
duration of the earthquake motions are taken into account.

7.3.4 Earthquake response analyses

The seismic load acting on the structural members and components can be estimated based on
earthquake response analyses.
Measures for quality assurance shall be taken when the seismic load is estimated using earthquake
response analyses. High-level measures shall be taken, in particular, when the seismic load is estimated
based only on earthquake response analyses, i.e., without employing a method of equivalent static
seismic load.
CHAPTER 7 SEISMIC LOADS - 85 -

(1) Model for earthquake response analyses

Earthquake response analysis models for the building and foundation (including basement) are
constructed in accordance with Sec. 7.1.2. The natural period, vibration modes, and damping
characteristics are modelled to be consistent with reality. The yield point, strength, dynamic hysteresis
characteristics, and ductility are modelled as compatible with the actual characteristics when nonlinear
response analyses are conducted. Necessary measures including experiments are taken when required
to obtain the characteristics needed to construct the analysis models.

(2) Estimation of seismic load based on earthquake response analyses

The seismic load is estimated in accordance with Sec. 7.3.2 and Sec. 7.3.3, and is determined by
considering all results of the multiple response analyses conducted. Appropriate parameters such as the
maximum stress or deformation are specified for each structural member and component to confirm the
required performance.

(3) Seismic load owing to flow and deformation of the soil

The effects of flow and deformation of the soil, caused by liquefaction and other factors, are taken
into consideration when estimating the seismic load.
-86-

CHAPTER 8 THERMAL LOADS

8.1 Scope of Application

(1) This chapter describes the thermal load for a building design.
(2) When a building is subject to changes in temperature owing to its site layout, time, size, use, or
service conditions, the thermal load needs to be considered.
(3) When thermal stress or deflection in the structural members is reduced through expansion joints or
some other effective measures, the thermal load may be disregarded.

8.2 Calculation of Thermal Load

The basic value of the thermal load shall be the 100-year recurrence of a temperature change such as
the outside air temperature, solar radiation, underground temperature, or equivalent value.
-87-

CHAPTER 9 EARTH AND HYDRAULIC PRESSURES

9.1 Overview
(1) The earth and hydraulic pressures acting against structures such as exterior basement walls and retaining
walls, which touch the soil directly, are considered. The buoyancy is also considered below the groundwater
level.
(2) The earth pressure that occurs continuously on the exterior basement walls and other areas is generally
assumed to be the earth pressure at rest, and its influence is appropriately considered when there is a surcharge
on the ground surface.
(3) The earth pressure that occurs continuously on the retaining walls is generally assumed to be the active
earth pressure, and its influence is appropriately considered when there is a surcharge on the ground surface.
(4) When the earth and hydraulic pressures increase remarkably from an earthquake, such an increase is
appropriately taken into account.
(5) When calculating the continuously occurring earth pressure, the value of the 99% non-exceedance
probability considering the accuracy of the calculation method and the variation of the geotechnical
parameters is assumed to be the characteristic value of the load. The characteristic value of the hydraulic
pressure is assumed to be the highest free water level with a return period of 100 years, and the influence of
time variance is also considered.

9.2 Earth and hydraulic pressures occurring on exterior basement walls

The earth and hydraulic pressures that occur continuously on exterior basement walls are calculated
according to Eq. (9.1) for the above groundwater level, and are calculated according to Eq. (9.2) for below
the groundwater level:
p = K0γt z + ∆p0 (9.1)
p = K0γt h + K0γ' (z−h) + γw (z−h) + ∆p0 (9.2)
where p is the earth and hydraulic pressures for each unit area at depth z (kN/m 2),
γt is the total unit weight
of soil (kN/m3), γ’ is the submerged unit weight of the soil (kN/m3), γw is the unit weight of water (= 9.8
kN/m3), K0 is the earth pressure coefficient at rest, h is the groundwater depth (m), and ∆p0 is the earth
pressure increment for a surcharge on the ground surface (kN/m2).

9.3 Soil pressure that occurs continuously on the retaining walls

The earth pressure occurring continuously on the retaining walls is generally assumed to be the active earth
pressure. When there is no surcharge on the ground surface, the active earth pressure is calculated as follows:
p A = K Aγ t z − 2c K A (9.3)
where pA is the active earth pressure for each unit area at depth z (kN/m2), KA is the active earth pressure
coefficient, γt is the total unit weight of the soil (kN/m3) (submerged unit weight of soil below the water
table), and c is the cohesion of the soil (kN/m2). However, when pA becomes negative, it is assumed to be
zero. The active earth pressure coefficient, KA, can be calculated using the following equation:
- 88 - Recommendations for Loads on Buildings

cos 2 (φ − θ ) , (9.4)
KA = 2
 sin(φ + δ ) sin(φ − α ) 
cos 2 θ cos(θ + δ )1 + 
 cos(φ + δ ) cos(θ − α ) 

where φ is the angle of internal friction of the backfill soil (deg), θ is the angle between the backfill soil and
the vertical plane, δ is the angle of the wall friction, and α is the angle between the ground surface and the
horizontal plane. Here, it is assumed that sin(φ−α) = 0 at φ−α < 0. When it is necessary to consider the
hydraulic pressure, such influence should be taken into account, and is considered when there is a surcharge
on the ground surface.

9.4 Earth pressure during earthquake acting on the retaining walls

The active earth pressure acting on the retaining walls shall be calculated based on reliable ground
investigations and proper analyses, or using either of the following methods.
(1) Active earth pressure of the earthquake as proposed by Monobe and Okabe
pEA = KEA γt z (9.5)
where pEA is the active earth pressure of an earthquake for each unit area at depth z (kN/m2), KEA is the earth
pressure coefficient of an earthquake based on the Monobe and Okabe equation, γt is the total unit weight of
the soil (kN/m3) (submerged unit weight of soil below the water table), and z is the depth (m) where the active
earth pressure pEA acts on the wall.
cos 2 (φ − θ − θ k ) (9.6)
K EA = 2
 sin(φ + δ ) sin(φ − α − θ k ) 
cosθ k cos 2 θ cos(θ k + θ + δ )1 + 
 cos(φ + δ + θ k ) cos(θ − α ) 

where φ is the angle of internal friction of the backfill soil (deg), θ is the angle between the backfill soil and
the vertical plane, δ is the angle of wall friction, α is the angle between the ground surface and the horizontal
plane, θk is the earthquake synthesis angle θk = tan-1 kh (deg), and kh is the design horizontal seismic intensity.
Here, it is assumed that sin(φ−α−θk) = 0 at φ−α−θk < 0. When it is necessary to consider hydraulic pressure,
such influence should be taken into account, and is considered when there is a surcharge on the ground
surface.

(2) Trial wedge method

a) When the cohesion in the Monobe and Okabe expression is not considered,
sin(ψ − ϕ + θ k ) ⋅ W
PEA = . (9.7)
cos(ψ − ϕ − δ − θ ) ⋅ cos θ k

b) When the cohesion in the Monobe and Okabe expression is considered,

W ⋅ secθ k sin(ψ − φ + θ k ) − c ⋅ l cos φ
PEA = , (9.8)
cos(ψ − φ − δ − θ )

where PEA is the resultant force of the active earth pressure for an earthquake (kN/m), W is the weight of the
soil wedge (kN/m), ψ is the active slip angle of the earthquake (deg), δ is the angle of wall friction, c is the
cohesion in the Monobe and Okabe expression (kN/m2), and l is the length of the active slip surface (m).
Chapter 9 EARTH AND HYDRAULIC PRESSURES - 89 -

9.5 Groundwater level for design

(1) The groundwater level for the design is determined as the highest water level with a return period of 100
years when considering the free water level based on the continuity of the soil strata and other factors.
(2) Buoyancy vPw occurring on the bottom of foundation shall be calculated through the following equation:
vPw = γw (z−h), (9.9)
where γw is the unit weight of water (kN/m ), z is the depth (m) of the position at which the buoyancy is
3

required from the ground surface level, and h is the depth (m) from the ground surface level to the water
table.

9.6 Uncertainties of earth pressure and the geotechnical parameters for design
(1) The earth pressure (active earth pressure, earth pressure at rest, and earth pressure during an earthquake)
acting on the exterior basement and retaining walls shall be determined by taking into consideration the
influence of the uncertainties of the parameters used in the calculations.
(2) A 99% non-exceedance probability shall be adopted for the characteristic value of the geotechnical
parameter used for the earth pressure calculations.
-90-

CHAPTER 10 TSUNAMI LOADS

10.1 General

10.1.1 Scope of application

The scope of application is as follows:
(1) All buildings and structures on the ground.
(2) Incoming and receding tsunamis caused by submarine earthquakes.
(3) Design inundation depths and flow velocities are evaluated through tsunami hazard maps,
computational fluid dynamics, or hydrodynamic tests, and the design inundation depths should not
exceed the heights of the designed buildings and structures.
(4) Seawater is an incompressible fluid, and the density of the water depends on the surrounding
conditions.
(5) The Froude number of an applicable tsunami is between zero and 2.

10.1.2 Estimation principle

Tsunami loads include loads at the surge front, flow behind the surge front, in still water, and of
waterborne debris. The following principals (1) through (3) are supposed in an evaluation of these loads.
(1) Design tsunami loads
This article defines the evaluation methods of the design tsunami loads using the design inundation
depths and design flow velocities. The design tsunami loads are given by the appropriate combination
of loads at the surge front, flow behind the surge front, in still water, and of waterborne debris.
(2) Tsunami forces on buildings
Tsunami forces on buildings can be derived from the difference in wave pressures between the front
and back of the building surface.
(3) Combination of tsunami and other loads
The combination of tsunami and other loads should be appropriately taken into account.

10.1.3 Three methods of estimation

The design tsunami loads can be calculated based on any of following three methods according to the
available input parameters. They have larger values between the loads at the surge front and the flow
behind the surge front for incoming tsunamis, and are the loads of flow behind the surge front for a
receding tsunami.
A．Time history of inundation depths and flow velocities
B．Maximum inundation depth and maximum flow velocity
C．Maximum inundation depth

10.1.4 Design inundation depth and design flow velocity

CHAPTER 10 TSUNAMI LOADS - 91 -

The design inundation depths and design flow velocities are defined based on tsunami hazard maps
formulated by local governments or through appropriate computational fluid dynamics (CFD).

10.1.5 Estimation of tsunami wave forces using CFD

The time history data or the maximum value of the wave forces and wave pressures on buildings can
be evaluated through sophisticated fluid simulation models that appropriately evaluate three-
dimensional surrounding conditions including the target building, as well as the time history and spatial
distribution of the inundation depths and flow velocities.

10.2 Tsunami Loads at Surge Front

10.2.1 Wave force of tsunami bore

Tsunami bore forces should be appropriately calculated according to the calculation methods of 10.1.3
depending on information such as the inundation depth and flow velocity.

10.2.2 Impulsive bore force

When buildings are located close to the embankment (sea wall), the impulsive bore force should be
taken into consideration as a force acting on the parts of the building.

10.2.3 Vertical force

Vertical forces should be appropriately evaluated depending on the conditions of the incident tsunami,
buildings, ground, and other elements.

10.3 Tsunami Loads of Flow behind Surge Front

10.3.1 Drag
The horizontal flow directional forces on buildings induced by the flow behind a surge front can be
estimated as drag depending on the density of seawater, flow velocities, and horizontally projected area
of the submerged portion of the buildings.

10.3.2 Vertical force

The calculation procedure is the same as that indicated in Section 10.2.3, with the exception of the
buoyancy.

10.4 Hydrostatic Loads in Still Water

10.4.1 Horizontal hydrostatic force

Horizontal hydrostatic forces, which are the resultant forces of hydrostatic pressures acting on vertical
- 92 - Recommendations for Loads on Buildings

members in the horizontal direction, must be considered. The horizontal hydrostatic forces acting on
entire building structures may not be considered.

10.4.2 Buoyancy
Buoyancy, which is the upward resultant force of hydrostatic pressure acting on horizontal members
in the vertical direction, should be considered.

10.5 Loads on Members with Openings and/or Open Structures

The tsunami loads on members with openings and/or open structures should be evaluated based on
their area exposed to the wave pressure.

10.6 Loads from Waterborne Debris

10.6.1 Debris impact loads

Waterborne debris impact loads should be evaluated considering the expected conditions during a
tsunami event around the building site of concern.

10.6.2 Damming effect caused by waterborne debris

The increase in tsunami loads caused by a damming effect of waterborne debris should be
appropriately considered.

10.7 Consideration of Resistant Tsunami Design of Buildings

10.7.1 Damage from earthquake ground motion

The load bearing capacity of buildings should be evaluated appropriately accounting for damage from
earthquake ground motions.

10.7.2 Liquefaction of ground

Liquefaction of the ground from earthquakes should be taken into consideration because it may
degrade the capacities of buildings, such as the tensile capacities of piles for the overturning resistance.

10.7.3 Scouring of ground

Local scour holes should be taken into consideration because they may be caused mainly along the
corner of the front resulting from flows around the buildings.
-93-

CHAPTER 11 IMPULSIVE LOADS

11.1 Scope and evaluation policy

11.1.1 Scope

The scope of the impulsive loads addressed in these recommendations is as follows.

(1) Target buildings
The target buildings cover offices, apartments, hotels, hospitals, schools, and public and commercial
buildings. The structural type is limited to multi-story frame structures, including reinforced concrete
and steel buildings.
(2) Target events
The target events are limited to artificial phenomena. Any events resulting from natural phenomena
are out of the scope of the recommendations. Artificial phenomena exclusively indicate accidental
events resulting in an impulsive load, as described below:
1. Impact load: Road vehicles, derailed trains, small aircraft, helicopters, forklifts
2. Blast load: Internal or external explosion
(3) Target limit states
For parts of a building subject to direct load action, only the safety limit state is considered as the
target limit state.

11.1.2 Evaluation policy

The evaluation policy of an impulsive load is described below.

(1) Evaluation policy of impact load
If any experiment data on an impact load are available, the maximum value of its time history is given
as a static design load. If no experiment data on an impact load are available, its time history is obtained
through crash simulations or by using a simplified energy method. The resulting time history is
approximated as a triangular wave with an equivalent impulse, and the maximum value of the triangular
wave is given as a static design load.
(2) Evaluation policy of blast load
Existing empirical formulas that are derived from experimental study are considered to obtain the
time history of the blast pressure for internal and external explosions. The maximum value of the
resulting time history is given as a static design load.
(3) Combination of impulsive and other loads
An impulsive load is combined exclusively with dead and live loads.
- 94 - Recommendations for Loads on Buildings

11.2 Impact load

11.2.1 Road vehicles

For buildings that are located adjacent to vehicle roads, buildings enclosing vehicle roads, or buildings
entirely or partially used as parking lots, it is recommended to consider road vehicle impacts. The impact
force of a road vehicle depends on the mass of the vehicle, impact velocity, incident angle, geometry,
and stiffness, and is determined according to the following methods.
- The maximum force in the time history of the impact load observed during the impact tests is used.
- The time history of the impact load is obtained through crash simulations and approximating the
resulting time history as a triangular wave with the equivalent impulse. The maximum force of the
triangular wave is then used.
- The design load obtained using a detailed model is applied individually.

11.2.2 Derailed train

For buildings that are adjacent to train tracks or buildings enclosing them, it is recommended to
consider a train impact. The impact force of a derailed train depends on the mass of the train, impact
velocity, incident angle, geometry, and stiffness, and is determined according to the following methods.
- The time history of the impact load is obtained through crash simulations, and the resulting time
history is approximated as a triangular wave with an equivalent impulse. The maximum force of the
triangular wave is used.
- The design load obtained using a detailed model is applied individually.

11.2.3 Small aircraft

Buildings with a height of 60 m or more above the ground or water surface are ordered by law to be
equipped with airplane warning lights. For buildings taller than 60 m, it is recommended to consider a
small aircraft impact event. The term “small aircraft” does not have any strict definition, although it
generally indicates a small two to six seat aircraft with a single reciprocating engine. The impact force
of a small aircraft depends on its mass, impact velocity, incident angle, geometry, and stiffness, and is
determined according to the following methods.
- The time history of the impact load is obtained using a simplified energy method, and the resulting
time history is approximated as a triangular wave with an equivalent impulse. The maximum force of
the triangular wave is used.
- The design load obtained using a detailed model is applied individually.
CHAPTER 11 IMPULSIVE LOADS - 95 -

11.2.4 Helicopter

For buildings that have heliports on their rooftop, it is recommended to consider a helicopter crash
event. The impact force of a helicopter crash depends on the mass of the helicopter, impact velocity,
incident angle, geometry, and stiffness, and is determined according to the following methods.
- The time history of impact load is obtained using a simplified energy method, and the resulting time
history is approximated as a triangular wave with an equivalent impulse. The maximum force of the
triangular wave is used.
- The design load obtained using a detailed model is applied individually.

11.2.5 Forklift

For buildings that permit the entrance of a forklift, it is recommended to consider a forklift impact.
The impact force of a forklift depends on its mass, impact velocity, incident angle, geometry, and
stiffness, and is determined according to the following methods.
- The time history of an impact load is obtained using a simplified energy method, and the resulting time
history is approximated as a triangular wave with an equivalent impulse. The maximum force of the
triangular wave is used.
- The design load obtained using a detailed model is applied individually.
- 96 - Recommendations for Loads on Buildings

11.3 Blast load

11.3.1 Internal explosion

For buildings that have any enclosed or semi-enclosed spaces where natural or city gas can be trapped,
it is recommended to consider the effect of an explosion on the building and its structural and non-
structural members, along with consideration of the volume of the space (room) and the open areas
(glass windows). Existing empirical formulas derived from experimental studies conducted to obtain
the time history of the blast pressure for an internal explosion are considered. The maximum value of
the time history is given as a static design load.

11.3.2 External explosion

For buildings near sites where an explosion may possibly occur, it is recommended to consider the
effects of such an explosion on the building and its structural and non-structural members, along with
the size of the explosion source and the offset distance. Existing empirical formulas derived from
experimental studies conducted to obtain the time history of the blast pressure for an external
explosion are considered. The maximum value of the time history is given as a static design load.
-97-

CHAPTER 12 OTHER LOADS

12.1 Other Loads

Other loads are all loads other than dead loads, live loads, snow loads, wind loads, seismic loads,
thermal loads, soil pressure, hydraulic pressure, tsunami loads, and impact loads.
AIJ Recommendations for Loads on Buildings (2015)

Commentary

March 2019

Architectural Institute of Japan

Index
Chapter 1 General concepts C1-1
Chapter 2 Loads and load combinations C2-1
Chapter 3 Dead loads C3-1
Chapter 4 Live loads C4-1
Chapter 5 Snow loads C5-1
Chapter 6 Wind loads C6-1
Chapter 7 Seismic loads C7-1
Chapter 8 Thermal loads C8-1
Chapter 9 Earth and hydraulic pressures C9-1
Chapter 10 Tsunami loads C10-1
Chapter 11 Impulsive loads C11-1
- C1-1 -

CHAPTER 1 GENERAL CONCEPTS

1.1 Scope of Application

This chapter describes the fundamental concepts in a structural design of buildings and the definitions
used by these recommendations to estimate the design load.
In this revision, the following are newly described by taking into consideration the domestic opinions
regarding safety after the Great East Japan Earthquake Disaster and international trends of the basic
concept of structural design. Namely, basic approaches are straightened to determine the design loads
caused by events with a large uncertainty. Robustness and the accountability of structural engineers are
added in the clause of the fundamental concepts. Tsunamis and impact loads, which were dealt with as
“other loads” in the 2004 and former versions, are commented on in an independent chapter.
The recommendations have been revised to reflect the most recent findings as far as possible. During
the period of past revisions in 1993 and 2004, the Hanshin-Awaji Earthquake Disaster and the Great
East Japan Earthquake Disaster occurred in 1995 and 2011, respectively, and have drawn significant
public attention on structural safety. In AIJ, in addition to the allowable stress design, and the ultimate
strength design, various design methods have been developed and compiled into the design guidelines
during the last 20 years. Considering the situation stated above, this version provides a common
methodology for determining the design load while avoiding a dependence on any particular design
methods, as the former versions did, having their basis on the limit state design method. Effort has been
made to include the outcomes from state-of-the-art research as much as possible; however, those
considered as incomplete or insufficient for use in a practical design are cited in either commentaries or
the appendices.
The scope of the recommendations includes every part of a building that resists a load, that is, not
only the entire buildings and structural components but also the non-structural components. However,
direct application shall not be considered for buildings that are constructed through a totally new
construction system or with novel materials, extremely large buildings, or buildings with an
unprecedented unique use. Nonetheless, when consulting the fundamental concepts explained in Clause
1.2, each prescription in these recommendations is applicable even to such cases, and the application is
encouraged.
The outcomes of various researches are cited with an unbiased perspective as much as possible. They
shall be replaced when new surveys or research suggest any parts to need modification or improvement.
In particular, although the basic values of load are estimated on the basis of objectively examining most
of the presently available data, they should be replaced with more reliable values whenever extensive
studies yield better data.
These recommendations are based on a stochastic model of the load intensity. In other words, to
ensure the target level of safety and serviceability of structures, they have certain standpoints in that the
design loads are to be properly determined when considering the degree of physical variation of the
loads themselves, and the accuracy of their estimation. This is because adequate statistical properties
- C1-2 - Commentary on Recommendations for Loads on Buildings

are to be available to determine the maximum loads for examining the safety and serviceability required
for ordinary buildings. For instance, it would be reliable to determine loads caused by 50- or 100-year
recurrence values of snowfall, wind, or earthquakes, or by the maximum of one hundred live load
samples. However, for very rare events such as a weather phenomenon with a return period of thousands
of years, or only a single phenomenon in tens of thousands of samples, it would not be adequate to
determine the values by extrapolating data currently available for engineering purposes.
For such rare events, it is desirable, if possible, to determine the values as the maximum loads that
are expected with a certain probability within a specified period. The probability may be evaluated, for
instance, by taking the following steps: modeling the mechanisms of events that produce the load,
supposing event scenarios including the site and magnitude producing the load, and assuming the
probabilities of occurrence of event scenarios and uncertainty of the model parameters. On the other
hand, for particularly important and/or specialized buildings, it might be necessary to determine the
design load in a deterministic way as the maximum considerable load based on a rational judgement
with state-of-the-art knowledge and technology, such as numerical modeling of earthquake sources for
a seismic load, numerical weather prediction for wind and snow loads, and future demand forecasts for
live loads.
It should be noted, however, that these two attitudes do not necessarily contradict each other. For
example, the concept of probability and statistics may be useful for setting the particular values of the
parameters used in a numerical simulation and/or an interpretation of the meaning of the outcomes. It
would rather be necessary to take into account the degree of “uncertainty” in each process even in a
deterministic approach. In the first place, the determination of the design load is a decision making
process toward “uncertainty.” The decision should be supported by the judgement of structural engineers
as experts, and agreed upon not only by the owners but also by all of the stakeholders of the project.
In general, the loads can be estimated within the context of a particular design procedure. The
recommendations have their basis on the limit state design; however, they are also applicable to the
allowable stress design and the ultimate strength design. Namely, the “basic values of loads” are
provided as the reference for any kinds of design procedure, and they should be multiplied with an
appropriate load factor. The recommendations do not present any particular return periods or load factors
corresponding to certain levels of safety and serviceability, but they describe a framework to determine
the values with some illustrative examples in the commentaries. The values should be specified as parts
of individual design procedures.
Structural engineers, in principle, are in charge of determining the design loads to satisfy the required
levels of safety and serviceability. For the purpose of satisfying the required levels, the provisions of
laws are regarded as one of the references.
It should be noted that the recommendations do not discuss the design load already prescribed in the
Building Standards Law of Japan. The provisions and technical standards in the law have been
historically prescribed on the basis of the results of research and surveys by the AIJ and relevant societies,
and do not always reflect the latest findings. The standards and recommendations published by the AIJ
CHAPTER 1 GENERAL CONCEPTS - C1-3 -

have not only complemented them but have also provided a scientific basis of their amendments by
organizing the discussion of new technologies. It should be noted again, however, that design loads
adequately determined according to the recommendations may not be legal.
These recommendations, when taking the circumstances of the structural design described above into
consideration, provide the present knowledge of the design loads based on the accumulated research
outcomes.

1.2 Fundamental Concept

1.2.1 Structural performance
(1) Safety
In the structural design of a building, structural engineers should make use of their knowledge and
technology to satisfy the requirements that are imposed on the structural performance. There are several
performances required for a building. “Safety” and “serviceability” are closely related to the structural
design. The goal of a structural design is to determine the appropriate structural dimensions and
materials so as to allow the building to maintain a level of safety and serviceability during its service
life against any type of anticipated load and disturbance such as a dead, live, snow, wind, or seismic
load, etc.
Buildings of all kinds and sizes, for both private and public use, shall be structurally safe, not only
against natural hazards but also against dead and live loads. This is required by the Building Standard
Law of Japan, and this requirement is regarded as the minimum level to ensure public safety and
protection of property. In these recommendations, a state in which the safety limit is exceeded is
considered as a state of total or partial collapse of the building, which would provoke the possibility of
injury or death within and/or around the building.
Society does not expect buildings to be safe regardless of the extremity of the load applied. Because
building construction is an economic activity, safety requirements must take into account an economic
rationale. An appropriate balance between safety and economic rationality is essentially required.
However, if the economic aspect is excessively emphasized in the design, construction, and use of a
building, the safety level may be less regarded. A low safety standard could lead to a considerable loss
of economic activity in the event of an earthquake, typhoon, or heavy snowfall.
Thus, when buildings are owned by the private sector, their safety requirement should not depend
only on the building owners’ individual values. The minimum safety requirements should instead be
legislated with balance among various factors, such as social and economic circumstances. In fact, this
is the purpose of the establishment of the Building Standard Law.
However, the minimum requirement of legislation does not exist at an absolute level, and rather
varies according to the level of technology, economic situation, social demand, accumulation of disaster
experience, and so on. Therefore, design regulations can change along with these environmental changes.
Structural engineers may be asked by their clients for the buildings to satisfy a higher safety criterion
than that required by law. The load determined in a structural design for safety directly controls the
- C1-4 - Commentary on Recommendations for Loads on Buildings

safety level for the whole of the building. It is essential for structural engineers to take care of various
requests revealed in the planning stage when they determine the design load.
(2) Serviceability
During the structural design procedure, in addition to the safety requirements, the serviceability of the
building should be examined under sustained loads or relatively frequent loads. In the limit state design,
the serviceability limit state is generally defined for the serviceability criteria. Examples of the
exceedance of serviceability limit are as follows:
1. Excessive vibrations that cause discomfort to the occupants or influence non-structural components
or equipment (in particular, resonance).
2. Deflections of structural or non-structural components that influence their designed appearance and
functions.
3. Minor damage that influences the durability of the structural components, the efficiency of non-
structural components, or the exterior appearance (including cracks).
Examples classified in terms of serviceability include tall building vibrations excited by the wind,
deflections of opening components by snowfall, deflections and vibrations of long-span floors, minor
cracking, creeping, and drying shrinkage of concrete components, etc.
Serviceability may be examined not only against moderate loads but also against much severer loads.
For example, hospitals, fire stations, and buildings used for disaster management are expected to be
functional even during a disaster. After experiencing recent disasters, not a few private companies and
the public sector have established a business continuity planning (BCP), and required their buildings to
maintain functionality even against higher loads.
For long-term load, creep may need to be considered. Some particular materials show creep
characteristics if they are loaded for a long period of time. Thus, components made from such types of
materials may be examined to avoid losing their functionality during a long period.
In the allowable stress design for long-term loads, serviceability is examined for sustained loads,
although a procedure is also usually executed for safety. Additional design criteria may be needed for
serviceability, particularly for buildings with long-span components, broad floors, large aspect ratios,
and disturbances inside as well as outside of the buildings.
(3) Reparability
In a structural design, the reparable limit may also be a concern along with the safety and serviceability
limits. For example, a serious problem appeared right after the 1995 Kobe earthquake; there were many
buildings that were structurally safe but not functioning at all and required proper repair for their
expected use. The reparable limit is not as clear as the safety and serviceability limits because it is related
not only to the degree of importance of the building, and social and economic circumstances, but also
to the modes of thinking of the building owners and engineers. Reparability is occasionally explained
as not being intended in the design process but resulting from the building itself.
Because of current social trends, environmental consciousness, and an effective use of resources,
which lead to the promotion of long-life buildings these recommendations include a section of
CHAPTER 1 GENERAL CONCEPTS - C1-5 -

“reparability”. However, the “load combination for limit state design” in section 2.4.1 does not include
reparability as a load to be considered because it is not yet a design target in the current design practice.
(4) Robustness
We have learned from past disaster experience that buildings can be subjected to loads that have never
been considered in the design process, or of which intensities and/or probability of occurrence are too
uncertain to be quantitatively estimated even if they could be supposed to occur. On the other hand,
certain types of loads are not necessarily considered as design loads from social and economic
viewpoints, because their occurrence probability can be estimated as sufficiently low with the best
knowledge of the present science and technology. However, it is desirable to take such “accidental loads”
into consideration such that they will not cause catastrophic events, even if they are not explicitly
adopted in the design requirements. “Robustness” is a concept of such structural performance of
buildings. An improvement of robustness does not always improve safety against the loads that are
explicitly included in the design requirements, but decreases the probability of being in danger caused
by an unexpected load. Although an extra safety margin against design loads and additional redundancy
can improve the robustness, there are some other possible means, including the following:
- a structural design with which the safety margin does not depend much on how its design loads are
determined,
- measures through which damage and/or a loss of some of the members or parts of the structure will
not lead to a total collapse, and
- a structural system that exhibits phenomena indicating the possibility of its total collapse well before
it happens.
For buildings with the potential of severe consequences if they collapse, robustness should be more
seriously considered against a large intensity loads and/or rare load combinations. For instance, Annex
F of ISO 2394 (2015)1) classifies the severity of consequences into five classes: from Class 1,
“Predominantly insignificant material damages,” to Class 5, “Disastrous events causing severe losses of
societal services and disruptions and delays at the national scale over periods in the order of years,”
based on their influencing period and range, and the number of casualties. It also provides examples of
structures classified into each of these classes.
Clause 2.3 4), “Treatment of accidental load” in these recommendations, describes a basic concept
of treatment of the following types of loads, such as those caused by human errors:
- the loads whose probability of occurrence is challenging to estimate quantitatively, although they
can be at least supposed to occur,
- the loads whose influence on society is not negligible because of uncertainty in the estimation of
their probability of occurrence despite such estimated probability being low.
(5) Accountability of the structural engineer(s) to the owner(s) regarding the designed
performance level
Reflecting the request of building owners, structural engineers should configure the performance level
of the building when considering the load states. Structural engineers should not only listen to the owners’
- C1-6 - Commentary on Recommendations for Loads on Buildings

requests, but also communicate well with them through the processes of explaining the context of the
structural design in plain language, listing to the owners’ own opinions in line with the engineers’
considerations, and reaching a consensus. The engineers should determine the target design performance
level under a consensus with the owners. If the users of the building are not owners, the engineers should
sufficiently explain to them the designed structural performance.
The fact of the matter is, most owners have much less knowledge regarding the structural
performance than structural engineers. In addition, they may not have explicit opinions or requests
regarding the performance level. Nevertheless, structural engineers should go ahead and present their
considerations regarding the balance between safety and economy, and to explain the process of setting
the service life and design loads, verification of the performance criteria, and other aspects. Achieving
this procedure provides the engineer’s adequate accountability for the design. It is also necessary for the
owners and users to understand their own responsibility through the use of the building, the performance
level of which they understand, and their participation in making decisions regarding the performance
level. Through the collaborative effort of owners and engineers, a reasonable decision on the
performance level and the desirable quality of the buildings should be realized.

1.2.2. Load modeling and structural analysis

Buildings are subjected to various types of loads and external forces. Such loads can be characterized
by their intensity, direction, and location to which they act. Some loads are always present and invariant
to time, and some loads vary slowly with time. In addition, some loads act suddenly and their intensity
and/or direction vary momentarily. Wind and seismic loads are typical loads categorized into the last
group. A structural analysis is a procedure used to evaluate their effect, and can be categorized into two
groups, static and dynamic analyses, which should be compatible with the characteristics of the loads.
It is considered at the present and at least in the near future that a dynamic analysis is used only for
buildings with a special structural system or those for special purposes, and that a static analysis is used
for ordinary buildings. Thus, in the recommendations, in principle, design loads are provided as static
loads. When they are inherently dynamic, their static design loads are determined by considering their
dynamic characteristics. In other words, a design load is a model of a load.
To conduct a structural analysis, a building is modelled for a practical analysis. Then, some analytical
conditions are defined and assumptions are made if necessary. Therefore, the structural engineers should
keep in mind that there are indispensable discrepancies between the analytical results and the responses
of real buildings.

1.2.3 Proper design and construction

To satisfy the safety and serviceability requirements of a building, proper consideration should be
provided, not only during the design stage but also during the construction and use. An error during any
of these stages may cause the safety requirement to be unsatisfied. However, there are many technical
problems when evaluating the design loads and taking such errors into account. Furthermore, a design
CHAPTER 1 GENERAL CONCEPTS - C1-7 -

load determined by considering the effect of such errors will have additional characteristics other than
the inherent characteristic of the load itself. Therefore, the design loads in the recommendations are
determined by assuming that buildings are properly designed, constructed, and used.
Therefore, devices and systems, such as snow removing mechanisms, and vibration control devices
against seismic and wind loads, should be properly maintained, managed, and applied if control of the
load intensity is considered in the structural design.

1.3 Terms and Definitions

Among the terms related to a load evaluation, the definitions of the terms associated with the limit state
design are given below:
Limit state: In a design, it is important to clarify the performances required for a structure or part
thereof, and to know the limits the performances do not achieve. These recommendations deal with
safety, which is the capability of reducing the risks of human casualties, asset losses, or damage, and
serviceability, which is the capability of offering functions for normal use, as the performances required
for the building. The limit states of safety and serviceability are defined as the “safety limit state” and
“serviceability limit state,” respectively.
The ultimate limit state, which relate to the maximum load bearing capacity of a building, has
almost the same meaning as the safety limit state. In the “Recommendations for Limit State Design of
Buildings,” the AIJ2) describes the ultimate and serviceability limit states of different structural types for
reference.
Limit state design: A design used to achieve the performance level with adequate reliability of not
exceeding the limit of the performance clarified by the limit state. These recommendations call the
designs for safety and serviceability as the “safety limit state design” and “serviceability limit state
design,” respectively.
Load intensity: A physical quantity expressing the magnitude of a load. It has, in principle, the
dimensions of “force” or “force per unit area.” In these recommendations, the load intensities are defined
as follows:
Dead and live loads: Average weight per unit area.
Snow load: Weight per unit area calculated from the ground snow depth or ground snow weight.
Wind load: Wind pressure calculated from 10-min mean wind speed.
Seismic load: Story shear force calculated from the peak acceleration on the engineering bedrock or
peak acceleration response.
Load effect: Reaction force, member force, deformation, and acceleration, etc. of a structure
induced by loads acting on the building. The concept of the load effect, which is transformed into force,
deformation, and acceleration, enables directly comparing the load and resistance, and is useful in the
design.
Basic value of load: Representative load intensity in calculation of the load effect. In principle, the
value is calculated based on the magnitude of a physical quantity corresponding to a return period of
- C1-8 - Commentary on Recommendations for Loads on Buildings

100 years. In cases in which it is difficult to determine a 100-year recurrence value, the basic value of
the load is determined based on a value with a non-exceedance probability of 99%. This method provides
consistency because a value with a non-exceedance probability of 99% has the same non-exceedance
probability with that of a 100-year recurrence value during a 1-year period.
When using a return period other than 100 years in the design, the load intensity is estimated using
the return period conversion factor described in each chapter.
Load factor: A partial safety factor taking into account the uncertainty of the load. Its value is
determined in consideration of the target performance level and variability in the load. The variability
in the load includes both variability in the load itself and uncertainty associated with load evaluation
methods. Originally, it represents a factor to be multiplied with the load effect calculated based on the
basic value of the load, i.e., the basic value of the load effect, in a limit state design. In the
recommendations, however, the design load is basically calculated by multiplying the basic value of the
load effect with a load factor regardless of the design methods.
Resistance factor: A partial safety factor taking into account the uncertainty of the resistance, the
value of which is determined in consideration of the target performance level and variability in resistance.
In a limit state design, the design resistance is calculated on the basis of a nominal value of resistance
multiplied by the resistance factor. The factor is determined according to the target resistance, materials,
and structural type, etc.
Reliability index: An index expressing a margin to the specified limit state, and one of quantitative
expressions of reliability. A reliability index 𝛽𝛽 is related with a probability of failure 𝑃𝑃𝑓𝑓 by 𝑃𝑃𝑓𝑓 = 1 −
Φ(𝛽𝛽), where Φ(⋅) denotes the cumulative distribution function of a standard normal random variable.
Probability of failure (limit state probability): The probability that the state of a building exceeds
the limit state associated with safety or serviceability within the specified reference period. These
recommendations refer to 𝑃𝑃𝑠𝑠 = 1 − 𝑃𝑃𝑓𝑓 as the reliability, where 𝑃𝑃𝑓𝑓 is the probability of failure.
Return period: An expected time interval 𝑡𝑡R of the events in which the load intensity exceeds a
specified value. If the events are statistically independent of each other, 𝑡𝑡R can be obtained with the
occurrence probability of the event within a particular time period (typically 1 year), 𝑝𝑝, as follows:
∞
1
𝑡𝑡R = � 𝑖𝑖 (1 − 𝑝𝑝)𝑖𝑖−1 𝑝𝑝 = (1.3.1)
𝑝𝑝
𝑖𝑖=1
Note that the return period determined according to the building use and other factors applied in
calculating the design loads is called the design return period.
The probability that a certain load will exceed the intensity with a return period of 𝑡𝑡R at least once
during a period of 𝑡𝑡R is about 0.63 (≈ 1 − 𝑒𝑒 −1) when 𝑡𝑡R is large, and the occurrence of the events is
independent and follows a Poison process3).
Reference period: A load acting on a building structure is more likely to have a large value when
the time period considered is longer, and thus a specific period should be determined for an evaluation
of the reliability index. These recommendations call the time period set to evaluate the reliability index
CHAPTER 1 GENERAL CONCEPTS - C1-9 -

or a probability of failure the “reference period.” A period of 1 year is used for the serviceability limit
state, and 50 years is used for a safety limit state.
Design service life: An assumed time period considered at the design stage, and during which a
building can be maintained as scheduled, free from major repair, and fulfilling the original usage. This
term is sometimes referred to as a “period of endurance.” However, these recommendations refer to the
time period considered in the design stage, during which a building provides its intended functionality,
as the “design service life,” while referring to the time period during which a building can provide
functionality as a building and endure its use physically, economically, and socially after its design or
construction as the “period of endurance.”

References
1) ISO: General Principles on Reliability for Structures. ISO 2394, Switzerland, 2015
2) AIJ: Recommendation for Limit State Design of Buildings, 2003 (in Japanese)
3) Ang, A. H-S. and Tang, W. H.: Probability Concepts in Engineering Planning and Design — Decision,
Risk, and Reliability, John Willey & Sons, Inc., New York, NY, USA, 1984
- C2-1 -

CHAPTER 2 LOADS AND LOAD COMBINATIONS

2.1 Loads
Conventionally, dead, live, snow, wind, and seismic loads, and earth and hydraulic pressure, have been
considered in the design of ordinary buildings. However, thermal load cannot be ignored because certain
types of buildings with a significant scale and/or for special usage have become common. In addition to
these loads, tsunamis, explosions, and impacts from collisions or fallings with a man-made object should
be considered in the design of a building at a specific construction site or with special features, and such
cases have recently garnered attention. Therefore, descriptions about these loads are newly included in
this edition of the recommendations. The symbols for each load are adopted in the main text to
harmonize with international standards. In addition, there are cases in which the weight of the
mechanical equipment, which is not covered as a live load, vibration, or impact load, and accidental
load need to be considered. Such loads are treated as “other loads” in these recommendations. Vibrations
during the use of a building, and temporary load during the construction, should also be taken into
consideration depending on the construction site, size, use, and construction method.

2.2 Basic Values of Loads

As clearly shown in the load formulae in each chapter, a load denoted by 𝑆𝑆 can generally be expressed
as the product of coefficients 𝐶𝐶1 , 𝐶𝐶2 , ⋯ , 𝐶𝐶𝑚𝑚 and a physical quantity 𝑋𝑋 as follows:
𝑆𝑆 = 𝐶𝐶1 ⋅ 𝐶𝐶2 ⋯ 𝐶𝐶𝑚𝑚 ⋅ 𝑋𝑋, (2.2.1)
in which 𝑋𝑋 is a random variable with inherent (aleatory) uncertainty related to natural phenomenon,
and 𝐶𝐶𝑗𝑗 (𝑗𝑗 = 1,2, ⋯ , 𝑚𝑚) are also random variables whose uncertainty could be reduced through the
accumulation of information, research findings, and so on, but cannot be reduced strictly to zero.
Examples of 𝑋𝑋 are the ground snow depth for a snow load, the squared mean wind speed for a wind
load, and the peak ground acceleration on the engineering bedrock for a seismic load. The values of 𝐶𝐶𝑗𝑗
include parameters such as the roof shape factor for a snow load, and the wind pressure coefficient for
a wind load. Furthermore, analytical errors, errors associated with modeling, and errors from insufficient
statistical data (epistemic uncertainty) can be taken into account.
When considering multiple loads expressed through this manner in a design, the design value of
each load on a common basis can be evaluated using the common rule for determining 𝑋𝑋. The basic
value of load 𝑆𝑆n is now expressed as
𝑆𝑆n = 𝐶𝐶1 ⋅ 𝐶𝐶2 ⋯ 𝐶𝐶𝑚𝑚 ⋅ 𝑋𝑋n . (2.2.2)
Here, 𝑋𝑋n in Eq. (2.2.2) can be determined as the 𝑡𝑡R -year recurrence value of a natural phenomenon if
the return period of the load can be specified. It can be determined as the value with a predetermined
non-exceedance probability under normal conditions (e.g., a non-exceedance probability of 1 − 1/𝑡𝑡R ),
if it is difficult to evaluate the return period of the load. As for 𝐶𝐶𝑗𝑗 in Eq. (2.2.2), either the most
appropriate value for 𝐶𝐶𝑗𝑗 from an engineering viewpoint, or the mean value of 𝐶𝐶𝑗𝑗 if the statistical data
are available, should be used.
- C2-2 - Commentary on Recommendations for Loads on Buildings

The basic value of a load can be determined by the “maximum” value considered. The maximum
event in the past might be easily accepted by the general public as a design load against natural
phenomena such as snow, wind, and earthquakes. As maximum events, for example, the 1963 or 1981
great snow falls, the 1934 Muroto typhoon, and the 1923 great Kanto or 1995 Kobe earthquakes can be
considered. However, it is more rational to objectively evaluate the design load intensity on the basis of
multiple past events or an ample statistical database rather than a single past event. Although load events
can be modeled using a Poisson process, a Markov process, or one of various extreme-value distribution
models, it is preferable for the basic value of a load to be evaluated independent of such models.
Therefore, it would be rational to determine the basic value of a load as a value corresponding to a
specified return period 𝑡𝑡R (equal to the annual exceedance probability of 1/𝑡𝑡R ). For future
diversification of structural design methods, the same return period should be adopted for snow, wind,
and seismic loads.
The design load in the recommendations is basically evaluated as the product of the basic value of
a load determined as above, and a load factor for use in not only a limit state design but also allowable
stress and ultimate strength designs. For practical application to the allowable stress and ultimate
strength designs, a return period conversion factor is introduced as a load factor in order to modify the
𝑡𝑡R -year recurrence value to the value corresponding to the return period considered in the design.
Because a design load is calculated as the product of a basic value of a load and an adequate load
factor, as described above, any value of the return period used to specify a basic value of the load can
give the same design load. Although some design codes in European and American countries adopt a
basic value of a load with a return period of 50 years, we should consider that the life span of human
beings is 80 years or so. Because events with a return period of 100 years roughly correspond to events
with a probability of 50% to occur at least once during the life span of a human being, since the 1993
edition, these recommendations have adopted a return period of 100 years to determine the basic value
of a load. However, the accumulation of data is expected to deal with tsunami and impact loads in a
statistic manner, except for limited regions.
On the contrary, a live load and earth and hydraulic pressure are considered as a stationary load, and
accordingly, the above discussion is not directly applied. However, conceptual compatibility is
maintained by adopting the same annual exceedance probability. In other words, the load value with a
non-exceedance probability of 99% is adopted as the basic value of a live load and earth and hydraulic
pressure. This probability coincides with the probability that the annual maximum load does not exceed
the 100-year recurrence value. It should be noted that a dead load is to be determined on the basis of an
average value according to the actual conditions owing to its small variability.

2.3 Load Combination and Required Performance

1) Concept of load combination
Buildings are subjected to some of the loads listed in Section 2.1 at the same time. Such a state is called
a “load combination.” However, in these recommendations, the “combination of load effects” such as
CHAPTER 2 LOADS AND LOAD COMBINATIONS - C2-3 -

the combined stress and deformation under simultaneous multiple loads is discussed.
As long as a building exists, it is affected by gravity. Because a building usually contains certain objects,
it is always subjected to a vertical force downward owing to gravity acting on its own weight and the
weight of the objects. On the basis of the description “persistently” in Sec. 2.3(3) of these
recommendations, it is assumed that a building is subjected to these two types of loads under normal
conditions. It is reasonable to add the snow load of a normal winter in heavy snowy areas. Depending
on the shape and conditions of the site and buildings, the influence of a thermal load owing to changes
in the outside air temperature and solar radiation, earth pressure, and hydraulic pressure should be
considered.
On the other hand, an “extra-ordinary live load” refers to a case in which the limit state can be
reached mainly owing to moving, remodeling, and concentration of people. The same applies to the
items below a “snow load.” Naturally, one cannot deny the possibility that a building will be subjected
to simultaneous multiple “principal loads” and will reach the limit state. Conventionally, including the
Building Standard Law of Japan, it is considered that only a snow load is applied for a relatively long
period of time, and is reduced to a certain extent; the reduced snow load is then combined with the load
from a strong wind or strong ground motion.
It should be noted that when multiple loads occur simultaneously, the combined effect is not always
a simple sum of the stress in each loaded state. For example, the simultaneous action of heavy snow and
strong wind can occur relatively frequently in the Japan Sea side region during winter. A negative
pressure on the roof surface owing to the wind load does not directly act on the roofing material but
often only raises the snow of the roof, which is a worse state for the structure than when all load effects
are simply summed.
The decision regarding whether a load combination is considered should be based on the “possibility”
that the loads considered in the combination act simultaneously. For example, during the service life of
buildings on the order of several decades, the possibility that the largest ground motion with a duration
of a few minutes will simultaneously occur with the strongest wind of the maximum class with a duration
of about half a day is very rare, and is considered as a stochastic process. However, in view of the
approaching typhoon after the 1923 Great Kanto Earthquake, it is reasonable to consider the
combination of typhoons of an annual average value and the largest ground motion. The degree of
“possibility of simultaneous action” considered in the design shall be dependent on the use and
importance of the building and the required performance level. The more important a building is, the
rarer the combinations that should be considered.
In the design of general buildings, load combinations where the possibility of simultaneous
occurrence is extremely low can be neglected; however, the following combinations should be
considered.
(1) When the live load 𝑄𝑄 is the “principal load,” namely, 𝐺𝐺 + 𝑄𝑄
(2) When the snow load 𝑆𝑆 is the “principal load,” namely, 𝐺𝐺 + 𝑄𝑄 + 𝑆𝑆
(3) When the wind load 𝑊𝑊 is the “principal load,” namely, 𝐺𝐺 + 𝑄𝑄 + 𝑊𝑊 (general area) and 𝐺𝐺 + 𝑄𝑄 +
- C2-4 - Commentary on Recommendations for Loads on Buildings

𝑆𝑆 + 𝑊𝑊 (heavy snow area)

(4) When the seismic load 𝐸𝐸 is the “principal load,” namely, 𝐺𝐺 + 𝑄𝑄 + 𝐸𝐸 (general area) and 𝐺𝐺 + 𝑄𝑄 +
𝑆𝑆 + 𝐸𝐸 (heavy snow area)
However, the combinations are not limited to the above. Depending on the situation, a load combined
with a tsunami, or an impact load as the principal load, should be considered. In addition, depending on
the degree of influence of the length of the building or the temperature environment (such as the heat
insulating condition), a load combination with the thermal load as the principal load should be
considered. As aforementioned, it should be noted that the thermal load can be a persistent load.
Events such as heavy snow after the 2004 Niigata Prefecture Chuetsu Earthquake, and the tsunami
and repeated aftershocks after the 2011 Tohoku Region Pacific Coast Earthquake, revealed that it may
be necessary to evaluate the performance of a building against not only the load combinations in which
loads act “simultaneously,” but also those in which several “principal loads” act with certain time lags.
For example, it may be possible that a volcano will erupt following an earthquake and volcanic ash will
fall and accumulate on a roof. In such a case, it should be noted that the analyzed problem becomes non-
linear with several loading directions, as if a building damaged by an earthquake is subjected to a snow
load.

2) Required performance level

2-1) Required performance level and design load
The recommendations clearly state that the intensity of a design load should be determined according
to the performance level required for a target building. The design loads determined herein are not the
minimum “design loads” prescribed in conventional standards or codes, but should reﬂect the use,
importance, and design service life of the building. The higher the required performance level is, the
larger the design load should be. By doing so, the performance level of the building can be clariﬁed
based on the concept of a performance-based design. Although it is widely and practically used to adjust
the allowable limit of a building capacity under a constant design load, the relation of the reduction to
the required performance level should be clarified.
The function of a warehouse or similar structure is retained as long as there is no damage to its
contents such as shelves and goods, even if it suffers some damage at the time of a disaster. A simple
reinforcement to the structure may be possible before aftershocks occur. On the contrary, facilities in
disaster areas that should exert their functions during the time of a disaster should have an appropriate
performance level. If the influence remains even after the removal of the preceding load, they should
also deal with the possibility of exceeding multiple limit states occurring with time lags. In such a way,
it is reasonable for the design load to be differentiated according to the importance of the building, social
characteristics, and economic efficiency, and differentiation can be carried out through adjusting the
reliability index using the limit state design method.
A design load can be determined by selecting the return period, as can be seen in Vision20001) and
a draft of the performance-based design in the AIJ2, 3). Table 2.3.1 shows the required performance matrix
CHAPTER 2 LOADS AND LOAD COMBINATIONS - C2-5 -

by SEAOC, in which the seismic loads are determined on the basis of return periods of 43 to 970 years,
or equivalently, exceedance probabilities within 30, 50, or 100 years. The idea of determining the
intensity of a load on the basis of return period is associated to a certain extent with the design service
life of the building from the viewpoint of the loads, and is widely used. In contrast, the allowable limit
of each part of the building should correspond directly to the limit state required in the design. As for
the relation between the limit state and the allowable limit (design criteria), studies4, 5)
have been
conducted recently, and cases of performance-based designs referring to these studies have also been
reported6). In essence, however, not only the load intensities but also the allowable limit is a random
variable, and the limit state design method treats both loads and resistance as random variables.

Table 2.3.1 Required performance matrix for seismic design in Vision20001)

Ground motion intensity Damage level
Frequency Exceedance Negligible Light Moderate Severe
(return period) probability (fully operational) (operational) (life safe) (near collapse)
Frequent (43 years) 50% in 30 years ● × × ×
Occasional (72years) 50% in 50 years ■ ● × ×
Rare (475 years) 10% in 50 years ★ ■ ● ×
Very rare (970years) 10% in 100 years ★ ■ ●
● ordinary building ■ important building ★ very important building

2-2) How to determine the performance level

Ultimately, the performance level of the building should be determined responsibly by the structural
engineer; nevertheless, the decision should be based on an agreement among the engineer, owners, users,
and so on, after a careful and thorough exchange of their opinions. When exchanging opinions, the
required state of the building under each load state should be clarified by considering the use of the
building and by assuming the situation in which the building is subjected to each load. In addition, the
length of the design service life is also an important factor in determining the performance level. There
is a trade-off relation between the performance level and economic efficiency, and so-called lifecycle
cost management (LCCM), which is an economic evaluation including the building construction, service
life, and demolition, will be more important in the future. In other words, the balance between economic
efficiency and performance level shall be thoroughly studied.
As a method for setting the performance level, it is also common to investigate the performance
level of buildings designed according to existing standards, and consider the level as a standard
performance level. Such procedure is called a code calibration based on existing design methods. Some
trial calculations of the standard performance level are conducted and presented in “Guidebook of
Recommendations for Loads on Buildings 126).”

3) Load state
- C2-6 - Commentary on Recommendations for Loads on Buildings

The load states to be considered in a building design shall be selected after appropriately understanding
the use of the building, the environment to which it is exposed, and the site conditions. In these
recommendations, the term “load state” is used to indicate a state to which a building and the structural
members are exposed, and the structural design is then carried out for each load state.
The load state can be defined for each structural member as well as for the entire structure. This is
closely related to how the required performance is considered and how to confirm the performance.
When confirming the performance of the structural member, the load state must be selected for the
member according to the actual situation. To satisfy their performances at an adequate level, a reasonable
cost is required. Because it is impossible to set the performance level by ignoring the economic aspect,
the performance level required for the building is in a trade-off relation with that of the economy.
Therefore, the relation should be clarified as much as possible, and efforts should be made to explore
the performance levels to find a reasonable level from the viewpoint of economics and building
performance.

4) Treatment of accidental load

While the other loads treated in the recommendations represent physical entities such as the weight of
objects, snow, earthquake, and so on, it should be noted that the expressions of accidental loads represent
the statistical properties of the load itself, i.e., the frequency of occurrence. Although classified as
accidental loads (𝐴𝐴), various cases may occur. For example, a massive tsunami or tornado is a family of
tsunami or strong wind, respectively, that occurs at a certain frequency, and the occurrence of such
events will be estimated as extremely rare based on past statistics and probabilistic models. However, if
a certain amount of uncertainty exists in the estimation of the probability of occurrence, and if the
magnitude of social influence is large, a massive tsunami or tornado may be treated as an “accidental
load” separately from the ordinary design loads.
In the treatment of an accidental load, the decision of whether to consider the influence of the load
is the first significant issue. In addition, because of the difficulty in obtaining statistical data for
estimating the probability of occurrence, the design load will be determined based on the scenario. The
ground of the scenario should then be clearly presented such that the accountability can be ensured.
As for the load combinations, because accidental loads act infrequently, only “𝐺𝐺 + 𝑄𝑄 + 𝐴𝐴” in a
general area or “𝐺𝐺 + 𝑄𝑄 + 𝑆𝑆 + 𝐴𝐴” in a heavy snow area can be considered. However, a load arising from
events occurring in a chain from a single event should be considered together as a series. It should be
noted that when a structure is subjected to a load such as an accidental load, its response may reach into
a plastic range, and the superposition of the load effects by a linear sum will not be established.

2.4 Load Factors

2.4.1 Load factors for use in limit state design format
1) Limit state and reliability index
This section discusses load combinations and load factors to be considered in a limit state design (LSD).
CHAPTER 2 LOADS AND LOAD COMBINATIONS - C2-7 -

Although the term “limit state design” has a broader meaning, it is considered as a reliability-based LSD
in these recommendations.
Except for a dead load, which can be predicted fairly accurately based on a building’s weight, a great
deal of uncertainty and variability exist in the intensity of loads such as live, snow, wind, and seismic
loads. The intensity of these loads depends, for instance, on the location of the construction site, the size
of the building, and when it is to be designed or constructed, among other factors.
LSD is a fairly flexible structural design method. In LSD, a limit state, which is a boundary between
favorable and unfavorable states of a building or a structural member, is explicitly determined. The
performance of a building is measured quantitatively by accounting for the uncertainty in the loads, the
configuration of a structure, and the strength of the materials, etc., and thus, it can be controlled
intentionally. The performance level is often measured by a limit state probability, 𝑃𝑃𝑓𝑓 , which is the
probability that the state of a building will exceed a limit state during a reference period, or by a
reliability index, which is equivalent to a limit state probability
Limit states can be classified into two categories, an ultimate limit state and a serviceability limit
state. An ultimate limit state is the boundary beyond which human lives are in danger, such as the loss
of stability in an entire structure, or a portion of a structure, and a floor or total collapse of a building
owing to a lack of load bearing capacity and/or ductility. A serviceability limit state is a boundary beyond
which some difficulty initiates in the daily use of a building. It includes damage that decreases the
inhabitability and durability, such as cracking affecting the appearance of a building, and excessive
deformation or vibrations that lead to discomfort of the occupants and/or a harmful influence on non-
structural members and equipment. It also includes the initiation of non-reversible behavior such as the
yielding of a building member.
Recently, longer building service lives, such as long-life quality housing promoted by the Ministry
of Land, Infrastructure, Transport, and Tourism, have been sought out owing to concerns regarding
environmental issues, whereas demands of temporal commercial buildings have also increased for
economic reasons. The longer the service life of a building is, the higher the possibility of the building
being subjected to a large loading, and accordingly, the larger the design load that shall be assumed.
However, the concept of a service life is not explicitly considered in a conventional design method, and
the design load intensity for ordinary buildings is also employed for such buildings without considering
the difference in the design service life. Although there are certain cases in which the design load is
determined based on the return period, it is not clearly stated how the return period is determined
considering the service life.
On the contrary, the limit state probability for a reference period, typically 50 years for the ultimate
limit state and 1 year for the serviceability limit states, is used as a quantitative measure of the structural
performance, and accordingly, a building can be designed by considering the difference in the length of
the service life by altering the reference period with the same target limit state probability.
Note here that the reference period is used to evaluate the limit state probability, which is the
probability of exceeding a limit state during a reference period. There are other types of “periods” or
- C2-8 - Commentary on Recommendations for Loads on Buildings

“lives” in a structural design, such as the “service life,” “design service life,” and “(expected) return
period.” These are periods during which a building is physically in use, the service life assumed in the
design stage, and the mean time interval during which a loading larger than a certain intensity occurs,
respectively. These “periods” should be distinguished from one another. In addition, it should be noted
that either the same reference periods or the same limit state probabilities should be considered when
comparing the structural performances because the limit state probability varies depending on the
reference period.
A reliability index, 𝛽𝛽, is a measure of the reliability for a considered limit state, and can be defined
using the limit state probability, 𝑃𝑃𝑓𝑓 , as
𝛽𝛽 = Φ−1 (1 − 𝑃𝑃𝑓𝑓 ), (2.4.1)
in which Φ(•) is the cumulative distribution function (CDF) of the standard normal random variable,
and Φ−1 is its inverse function. Figure 2.4.1 illustrates the relation between a reliability index, 𝛽𝛽, and
a limit state probability, 𝑃𝑃𝑓𝑓 .

Figure 2.4.1 Relation between a reliability index, 𝛽𝛽, and a limit state probability, 𝑃𝑃𝑓𝑓

An estimation of a limit state probability requires knowledge of probability and statistics. However,
reliability-based LSD can also be carried out in a similar manner as conventional deterministic design
procedures using the load and resistance factors determined through probability and statistics by
considering the uncertainty in the strength of the materials and the load intensity, as well as the target
performance level. Such design procedures have already been used in practice in North America and
Europe7–9), and the following recommendations have already been published from the AIJ in Japan:
“Standards for Limit State Design of Steel Structures (draft) (1990)10),” “Recommendations for Limit
State Design of Steel Structures (1998, revised in 2010)11),” and “Recommendations for Limit State
Design of Buildings (2002)12).”
It has been recognized among researchers and engineers that the load intensity should be treated as
a random variable. In these recommendations, wind and snow loads have been treated as random
variables since the revision in 1981, and live and seismic loads have been treated as random variables
since the revision in 1993. A thermal load has been treated as a random variable since the revision in
CHAPTER 2 LOADS AND LOAD COMBINATIONS - C2-9 -

2004; the revision in 2004 also includes a database of the statistical characteristics for each load. In this
revision, tsunami and impact loads are treated as random variables.

2) Design format and load combinations based on Turkstra’s rule

As the design requirement in LSD, the limit state probability of a building and/or structural member
during a reference period shall be equal to or smaller than the target (or acceptable) probability. The
design format can be generally expressed as
𝑃𝑃𝑓𝑓 = Prob{𝑔𝑔(𝐗𝐗) ≤ 0 ; 0 < 𝑡𝑡 ≤ 𝑡𝑡ref } ≤ 𝑃𝑃𝑓𝑓𝑎𝑎 = 1 − Φ�𝛽𝛽T �, (2.4.2)
in which 𝐗𝐗 is a vector of basic random variables, 𝑡𝑡ref is the reference period, 𝑃𝑃𝑓𝑓𝑎𝑎 is the maximum
acceptable limit state probability, 𝛽𝛽T is the target reliability index, and 𝑔𝑔(𝐗𝐗) is the limit state function.
A limit state function indicates the state of a building and/or structural member, and 𝑔𝑔(𝐗𝐗) > 0, 𝑔𝑔(𝐗𝐗) <
0, and 𝑔𝑔(𝐗𝐗) = 0 represent a building in a favorable state, an unfavorable state, and at the boundary
(limit) between favorable and unfavorable states, respectively.
A building is subjected to multiple load processes that vary over time. Let us consider a combination
of the effect of all loadings (load combination) as follows:
𝐺𝐺(𝑡𝑡) + 𝑄𝑄(𝑡𝑡) + 𝑆𝑆(𝑡𝑡) + 𝑊𝑊(𝑡𝑡) + 𝐸𝐸(𝑡𝑡), (2.4.3)
in which 𝐺𝐺(𝑡𝑡), 𝑄𝑄(𝑡𝑡), 𝑆𝑆(𝑡𝑡), 𝑊𝑊(𝑡𝑡), and 𝐸𝐸(𝑡𝑡) are the load effect owing to a dead load 𝐺𝐺 , live load 𝑄𝑄,
snow load 𝑆𝑆, wind load 𝑊𝑊, and seismic load 𝐸𝐸 at time 𝑡𝑡. Other load effects owing to such as a thermal
load shall be combined according to the actual conditions.
A limit state function can be expressed as
𝑔𝑔�𝑅𝑅, 𝐒𝐒(𝑡𝑡)� = 𝑅𝑅 − {𝐺𝐺(𝑡𝑡) + 𝑄𝑄(𝑡𝑡) + 𝑆𝑆(𝑡𝑡) + 𝑊𝑊(𝑡𝑡) + 𝐸𝐸(𝑡𝑡)}, (2.4.4)
in which 𝑅𝑅 is the resistance (either in strength or deformation) and 𝐒𝐒(𝑡𝑡) is a vector consisting of the
load effect of each loading at time 𝑡𝑡 under a load combination. When the coincidence possibility of
two loads is small, one of them can be excluded from the load combination.
It seems cumbersome to evaluate Eq. (2.4.2) using Eq. (2.4.4). However, when the change in
resistance over time is negligible, the limit state probability can be simply expressed as

𝑃𝑃𝑓𝑓 = Prob �𝑅𝑅 − max {𝐺𝐺(𝑡𝑡) + 𝑄𝑄(𝑡𝑡) + 𝑆𝑆(𝑡𝑡) + 𝑊𝑊(𝑡𝑡) + 𝐸𝐸(𝑡𝑡)} ≤ 0� ≤ 1 − Φ�𝛽𝛽T �. (2.4.5)
0<𝑡𝑡≤𝑡𝑡ref

The probability distribution of the maximum load combination during a reference period can be
estimated directly14–16) or approximated using Turkstra’s rule13). Turkstra’s rule presumes that this
maximum value can be approximately estimated as the sum of the maximum load effect owing to a
principal load during the reference period, and the arbitrary-point-in-time intensity of the other
(secondary) load effects. Applying Turkstra’s rule, the combination of time-variant loads can be treated
as a combination of time-invariant loads as

max {∑𝑛𝑛𝑖𝑖=1 𝑆𝑆𝒊𝒊 (𝑡𝑡)} ≈ 𝑆𝑆p + ∑𝑘𝑘 𝑆𝑆𝑘𝑘 , (2.4.6)

0<𝑡𝑡≤𝑡𝑡ref

in which 𝑆𝑆𝒊𝒊 (𝑡𝑡) is the load effect of an 𝑖𝑖-th load (including both the principal and secondary loads) at
time 𝑡𝑡, 𝑆𝑆p is the maximum load effect in the reference period owing to a principal load, and 𝑆𝑆𝑘𝑘 is the
- C2-10 - Commentary on Recommendations for Loads on Buildings

annual maximum load effect owing to a 𝑘𝑘-th secondary load.

Note that it is not appropriate to apply Turkstra’s rule to cases in which more than two load processes
are intermittent or impulsive. In addition, the rule can provide erroneous estimates when a principal load
is not dominant (roughly when the standard deviation of the principal load < the standard deviation of
the sum of the secondary loads).

3) Load factors
In the implementation of LSD, limit state probabilities can be evaluated directly; however, this is
cumbersome in practice. Alternatively, the load and resistance factor format has been proposed. The
load and resistance factors can be determined based on the target performance levels determined by the
structural engineers themselves. This design format has the advantage of an easy and straightforward
adjustment. The design format employing load and resistance factors is expressed as

𝜙𝜙 𝑅𝑅n ≥ 𝛾𝛾p 𝑆𝑆pn + ∑𝑘𝑘 𝛾𝛾𝑘𝑘 𝑆𝑆𝑘𝑘n , (2.4.7)

in which 𝑆𝑆pn and 𝑆𝑆𝑘𝑘n are the load effect owing to the basic value of a principal load and a 𝑘𝑘 -th
secondary load, respectively, and 𝛾𝛾p and 𝛾𝛾𝑘𝑘 are the corresponding load factors. In addition, 𝑅𝑅n and
𝜙𝜙 are the nominal value of resistance and the resistance factor, respectively.
It is difficult to evaluate directly the load factors for time-variant loads16). However, applying the
aforementioned Turkstra’s rule, the combination of time-variant loads can be approximated as a
combination of time-invariant loads. A flowchart of the load and resistance factors is shown in Fig. 2.4.2.

START

Determine Target Performance Level

Determine Target Reliability Index

Determine Type of CDF and Statistics Data from Each Chapter

for Each Load of the Recommendations

Evaluation of Load and Resistance Factors

Figure 2.4.2 Flowchart of load and resistance factors

Figure 2.4.3 schematically illustrates the concept of load and resistance factors. In a structural design,
the nominal value of resistance 𝑅𝑅n is determined such that the target reliability index is satisfied. If the
ratio between the nominal value and the mean value is constant, the location of the joint probability
density function (PDF) of resistance 𝑅𝑅 and load effect 𝑆𝑆 , 𝑓𝑓𝑅𝑅𝑅𝑅 (𝑟𝑟, 𝑠𝑠) , is determined by adjusting
𝑓𝑓𝑅𝑅𝑅𝑅 (𝑟𝑟, 𝑠𝑠) in the horizontal direction so that the limit state probability, that is, the volume of 𝑓𝑓𝑅𝑅𝑅𝑅 (𝑟𝑟, 𝑠𝑠)
CHAPTER 2 LOADS AND LOAD COMBINATIONS - C2-11 -

in the domain of (𝑟𝑟 − 𝑠𝑠 < 0), shall be smaller than the maximum acceptable limit state probability. Once
𝑓𝑓𝑅𝑅𝑅𝑅 (𝑟𝑟, 𝑠𝑠) is fixed, the design point (𝑟𝑟∗ , 𝑠𝑠∗ ) is determined as the most likelihood point of 𝑓𝑓𝑅𝑅𝑅𝑅 (𝑟𝑟, 𝑠𝑠) on
the limit state surface 𝑟𝑟 − 𝑠𝑠 = 0, and the load and resistance factors are defined as the ratio of 𝑠𝑠∗ to
the basic value of load 𝑆𝑆n , and the ratio of 𝑟𝑟∗ to the nominal value of the resistance 𝑅𝑅n, respectively.

Design Point
fR(r)
fS(s)

Design Point

(a) Load and resistance factors and (b) Projection of PDFs and
design point in 2-D space design point on

Figure 2.4.3 Load and resistance factors and design point

The load and resistance factors can be evaluated using AFOSM17); however, this requires a
probability analysis and is cumbersome in practice. In the “Recommendations of Limit State Design for
Buildings12),” tables of the load and resistance factors are presented by considering representative
statistical characteristics of loads. When the conditions are different from the assumptions used to
evaluate these factors, the factors should be evaluated using an approximation method such as the third
moment method18~20) or the simplified method12) presented in the “Recommendations of Limit State
Design for Buildings.” The revised version of the simplified method is described briefly in the following.

4) Simplified method for load factors

For a considered limit state, the load factors for the load effect owing to the basic value of a load
(basic load effect) and the resistance factor for the nominal resistance are evaluated through the
following three equations (refer to the “Guidebook of Recommendations for Loads on Buildings 126)”
for further details).
The load factor for a principal load, 𝛾𝛾p , is

1 𝜇𝜇𝑆𝑆p
𝛾𝛾p = exp ( 𝛼𝛼𝑆𝑆̃p 𝛽𝛽T 𝜎𝜎ln 𝑆𝑆̃p ) . (2.4.8)
2 𝑆𝑆pn
�1+𝑉𝑉𝑆𝑆�p
- C2-12 - Commentary on Recommendations for Loads on Buildings

The load factor for a 𝑘𝑘-th secondary load, 𝛾𝛾𝑘𝑘 , is

1 𝜇𝜇𝑆𝑆𝑘𝑘
𝛾𝛾𝑘𝑘 = exp ( 𝛼𝛼𝑆𝑆̃𝑘𝑘 𝛽𝛽T 𝜎𝜎ln 𝑆𝑆̃𝑘𝑘 ) . (2.4.9)
2 𝑆𝑆𝑘𝑘n
�1+𝑉𝑉𝑆𝑆�𝑘𝑘

The resistance factor, 𝜙𝜙, is

1 𝜇𝜇𝑅𝑅
𝜙𝜙 = exp ( − 𝛼𝛼𝑅𝑅 𝛽𝛽T 𝜎𝜎ln 𝑅𝑅 ) 𝑅𝑅n
, (2.4.10)
�1+𝑉𝑉𝑅𝑅 2

in which
𝛽𝛽T : target reliability index during the reference period (that is, 50 years for an ultimate limit state
and 1 year for a serviceability limit state),
𝜇𝜇𝑆𝑆p : mean value of the maximum load effect of a principal load during the reference period, 𝑆𝑆p
(note that 𝑆𝑆p is the 50-year maximum value for an ultimate limit state and the annual
maximum value for a serviceability limit state),
𝜇𝜇𝑆𝑆𝑘𝑘 : mean value of the load effect owing to the basic value of a 𝑘𝑘-th secondary load,
𝜇𝜇𝑅𝑅 : mean value of the resistance of a structural component, 𝑅𝑅,
𝑆𝑆pn : load effect owing to the basic value of a principal load,
𝑆𝑆𝑘𝑘n : load effect owing to the basic value of a 𝑘𝑘-th secondary load,
𝑅𝑅𝑛𝑛 : nominal value of the resistance of a structural component, 𝑅𝑅,
𝑆𝑆̃𝑖𝑖 : log-normally approximated random variable of 𝑆𝑆𝑖𝑖 , where 𝑆𝑆𝑖𝑖 is either 𝑆𝑆p or 𝑆𝑆𝑘𝑘 ,
𝑉𝑉𝑋𝑋 : coefficient of variation (COV) of a random variable 𝑋𝑋, which equals the standard deviation
of 𝑋𝑋 divided by the mean value of 𝑋𝑋,
𝜎𝜎ln 𝑋𝑋 : logarithmic standard deviation of 𝑋𝑋, i.e., the standard deviation of ln 𝑋𝑋.
In addition, the values of 𝛼𝛼𝑋𝑋 are the sensitivity factors showing the relative importance of each random
variable, and are evaluated as

𝛼𝛼𝑅𝑅 = 𝛼𝛼𝑅𝑅∗ ∙ 𝑢𝑢, (2.4.11)

𝛼𝛼𝑆𝑆𝑖𝑖 = 𝛼𝛼𝑆𝑆∗𝑖𝑖 ∙ 𝑢𝑢, (2.4.12)

in which 𝛼𝛼𝑅𝑅∗ , 𝛼𝛼𝑆𝑆∗𝑖𝑖 , and 𝑢𝑢 are determined by

𝜎𝜎ln𝑅𝑅
𝛼𝛼𝑅𝑅∗ = ,
2
�𝜎𝜎ln𝑅𝑅 2 +∑𝑗𝑗�𝑐𝑐𝑗𝑗 𝜎𝜎ln𝑆𝑆� �
(2.4.13)
𝑗𝑗

𝑐𝑐𝑖𝑖 𝜎𝜎ln𝑆𝑆�
𝛼𝛼𝑆𝑆∗𝑖𝑖 = 𝑖𝑖
,
2 (2.4.14)
�𝜎𝜎ln𝑅𝑅 2 +∑𝑗𝑗�𝑐𝑐𝑗𝑗 𝜎𝜎ln𝑆𝑆� �
𝑗𝑗

1.05
𝑢𝑢 = 2 max 𝑉𝑉𝑆𝑆
� −0.6
,
∗ �2 +�max 𝛼𝛼 ∗ � �∙Φ�
1−�1−��𝛼𝛼𝑅𝑅 𝑖𝑖 �
(2.4.15)
𝑆𝑆𝑖𝑖 0.4

in which 𝑐𝑐𝑖𝑖 is the ratio of the mean of 𝑆𝑆̃𝑖𝑖 to the total sum average of 𝑆𝑆̃𝑖𝑖 values, and is expressed as
CHAPTER 2 LOADS AND LOAD COMBINATIONS - C2-13 -

1 𝑆𝑆n 𝜇𝜇𝑆𝑆
exp�𝜆𝜆∗𝑆𝑆� + 𝜎𝜎ln𝑆𝑆� 2 � 𝑖𝑖 𝑖𝑖
𝑖𝑖 2 𝐺𝐺n 𝑆𝑆n
,
𝑖𝑖
𝑐𝑐𝑖𝑖 = 𝑖𝑖
𝑆𝑆n 𝜇𝜇𝑆𝑆 (2.4.16)
1 𝑗𝑗 𝑗𝑗
∑𝑗𝑗 exp�𝜆𝜆∗𝑆𝑆� + 𝜎𝜎ln𝑆𝑆� 2 �
𝑗𝑗 2 𝑗𝑗 𝐺𝐺n 𝑆𝑆n
𝑗𝑗

in which 𝜆𝜆𝑆𝑆∗̃𝑖𝑖 is the mean value of ln(𝑆𝑆̃𝑖𝑖 �𝜇𝜇𝑆𝑆𝑖𝑖 ) (hereafter referred to as the normalized logarithmic
mean), 𝐺𝐺n is the load effect owing to the basic value of the dead load 𝐺𝐺, and the values or formulae in
Table 2.4.1 are used for the ratio between the mean value of a load effect to its basic load effect, 𝜇𝜇𝑆𝑆𝑖𝑖 ⁄𝑆𝑆n𝑖𝑖 .
The statistics of 𝑆𝑆̃𝑖𝑖 , i.e., the normalized logarithmic mean, 𝜆𝜆𝑆𝑆∗̃𝑖𝑖 , and the logarithmic standard
deviation, 𝜎𝜎ln 𝑆𝑆̃𝑖𝑖 , shall be calculated depending on its type of probability distribution as follows:

- Gumbel distribution
𝜆𝜆𝑆𝑆∗̃𝑖𝑖 = ln�1 − 0.164 ∙ 𝑉𝑉𝑆𝑆𝑖𝑖 �, (2.4.17)
1+3.14∙𝑉𝑉𝑆𝑆
𝜎𝜎ln 𝑆𝑆̃𝑖𝑖 = 0.430 ∙ ln �1−0.164∙𝑉𝑉 𝑖𝑖 �, (2.4.18)
𝑆𝑆𝑖𝑖

in which 𝑉𝑉𝑆𝑆𝑖𝑖 is the COV of the annual maximum or 50-year maximum load effect, 𝑆𝑆𝑖𝑖 , before it is log-
normally approximated.

- Fréchet distribution
If 𝛽𝛽T ≤ 2.5,
0.367 1
𝜆𝜆𝑆𝑆∗̃𝑖𝑖 = − ln �Γ �1 − ��. (2.4.19)
𝑘𝑘 𝑘𝑘

Table 2.4.1Ratio between the mean value of a load effect to its basic load effect
Principal load for serviceability limit Principal load for ultimate limit state
state or secondary load
Dead load, 𝐺𝐺 1.0 －
Live load, 𝑄𝑄 Directly calculated from Table 4.2.5
Snow load, 𝑆𝑆 0.19 𝑆𝑆0 0.46 0.977 𝑆𝑆0 0.004
22) If 𝜇𝜇𝑆𝑆𝑎𝑎 ⁄𝑆𝑆0 can be directly calculated If 𝑆𝑆0 is listed in Table A5.1.1 or A5.1.2,
from Table A5.1.1 or A5.1.2, the the ratio can be calculated using its 1⁄𝑎𝑎
directly calculated value should be and 𝑏𝑏 values as,
used. 4.49 + 𝑎𝑎𝑎𝑎
𝜇𝜇𝑆𝑆P ⁄𝑆𝑆0 =
4.60 + 𝑎𝑎𝑎𝑎

Wind load, 𝑊𝑊 {1 − 2.49(𝜆𝜆𝑈𝑈 − 1)}2 {1 − 0.0688(𝜆𝜆𝑈𝑈 − 1)}2

Seismic load, 𝐸𝐸 1 1
0.011⁄𝑘𝑘 ∙ Γ �1 − � 0.51⁄𝑘𝑘 ∙ Γ �1 − �
𝑘𝑘 𝑘𝑘
𝑆𝑆0 (kN/m2): snow weight on the ground calculated using the basic physical quantity (note that the
coefficients are dimensionless, see Chapter 5),
𝜆𝜆𝑈𝑈 : the ratio of a 500-year recurrence wind speed to the basic physical quantity, which is a 100-year
recurrence wind speed (see Chapter 6),
- C2-14 - Commentary on Recommendations for Loads on Buildings

𝑘𝑘: a coefficient representing the statistical characteristics of seismic hazard at a site,

Γ(∙): the Gamma function.
1.82
𝜎𝜎ln 𝑆𝑆̃𝑖𝑖 = (2.4.20)
𝑘𝑘
If 𝛽𝛽T > 2.5,
0.170 1
𝜆𝜆𝑆𝑆∗̃𝑖𝑖 = − − ln �Γ �1 − �� (2.4.21)
𝑘𝑘 𝑘𝑘
2.29
𝜎𝜎ln 𝑆𝑆̃𝑖𝑖 = (2.4.22)
𝑘𝑘

Table 2.4.2 Physical quantities as the basis of a load and their COV
Physical quantity as Principal load for service- Principal load for ultimate
the basis of load ability limit state or limit state (50-year
secondary load (annual maximum)
maximum)
CDF* COV CDF* COV
Dead load,
Own weight Normal 0.1 ―
𝐺𝐺
Live load, Weights of people, Based Use the values in
furniture, supplies, on Table Table 4.2.5 ⟵
𝑄𝑄
machines, stores, etc. A4.2.1
Snow load, Weight of snow on Gumbel 1.1 𝑆𝑆0 −0.50 Gumbel If 𝑆𝑆0 is listed in
the ground If 𝜇𝜇𝑆𝑆𝑎𝑎 ⁄𝜎𝜎𝑆𝑆𝑎𝑎 can be Table A5.1.1 or
𝑆𝑆 22)
directly calculated A5.1.2, the COV
from Table A5.1.1 can be calculated
or A5.1.2, the using its 1⁄𝑎𝑎 and
directly calculated 𝑏𝑏 as
value should be 1.28
used. 𝑉𝑉𝑆𝑆P =
4.49 + 𝑎𝑎𝑎𝑎
Wind load, 10-min mean wind Gumbel 1.28(𝜆𝜆𝑈𝑈 − 1) Gumbel 1.28(𝜆𝜆𝑈𝑈 − 1)
speed at 10-m height 5.64 − 4.02𝜆𝜆𝑈𝑈 1.72 − 0.111𝜆𝜆𝑈𝑈
𝑊𝑊
above the ground
Seismic Spectral acceleration Fréchet � Fréchet �
⃓ 2 ⃓ 2
at the first natural ⃓ Γ �1 − � ⃓ Γ �1 − �
load, 𝐸𝐸 ⃓ 𝑘𝑘 ⃓ 𝑘𝑘
⃓
⃓ 2−1 ⃓
⃓ 2−1
period or peak ⃓ 1 ⃓ 1
⃓
⃓�Γ �1 − �� ⃓
⃓�Γ �1 − ��
acceleration on the ⃓ 𝑘𝑘 ⃓ 𝑘𝑘
engineering bedrock ⎷ ⎷

* The CDF of the dead load is time-independent because its variation in time can be negligible. The
CDF of the live load is one at an arbitrary point in time.
- Other distributions
𝜆𝜆𝑆𝑆∗̃𝑖𝑖 = 𝑒𝑒0 + 𝑒𝑒1 𝑉𝑉𝑆𝑆𝑖𝑖 + 𝑒𝑒2 𝑉𝑉𝑆𝑆𝑖𝑖 2 + 𝑒𝑒3 𝑉𝑉𝑆𝑆𝑖𝑖 3 (2.4.23)
2 3
𝜎𝜎ln 𝑆𝑆̃𝑖𝑖 = 𝑠𝑠0 + 𝑠𝑠1 𝑉𝑉𝑆𝑆𝑖𝑖 + 𝑠𝑠2 𝑉𝑉𝑆𝑆𝑖𝑖 + 𝑠𝑠3 𝑉𝑉𝑆𝑆𝑖𝑖 (2.4.24)
CHAPTER 2 LOADS AND LOAD COMBINATIONS - C2-15 -

The coefficients 𝑒𝑒𝑗𝑗 and 𝑠𝑠𝑗𝑗 used in Eqs. (2.4.23) and (2.4.24) are presented in Table 2.4.3 for a
lognormal approximation of the annual maximum or 50-year maximum, and in Table 2.4.4 for an
approximation of the 50-year maximum using the CDF of the annual maximum.
The COV of 𝑆𝑆̃𝑖𝑖 , 𝑉𝑉𝑆𝑆� , can be calculated as
𝑖𝑖

2
𝑉𝑉𝑆𝑆�𝑖𝑖 = �exp ��𝜎𝜎ln 𝑆𝑆�𝑖𝑖 � � − 1, (2.4.25)

Table 2.4.3 Lognormal approximation of annual and 50-year maximums

（0.05 ≤ 𝑉𝑉𝑆𝑆𝑖𝑖 ≤ 2.0）
CDF of annual max. Coefficients in Eq. (2.4.23) Coefficients in Eq. (2.4.24)
e0 e1 e2 e3 s0 s1 s2 s3
Normal 0.000 0.000 0.000 0.000 0.012 0.830 −0.437 0.111
Gamma 0.002 −0.022 −0.267 −0.079 0.009 0.925 −0.205 0.087

Table 2.4.4 Lognormal approximation of 50-year maximum using CDF of annual maximum
（0.05 ≤ 𝑉𝑉𝑆𝑆𝑖𝑖 ≤ 2.0）
CDF of annual max. Coefficients in Eq. (2.4.23) Coefficients in Eq. (2.4.24)
e0 e1 e2 e3 s0 s1 s2 s3
Normal 0.055 1.683 −0.704 0.137 0.029 0.244 −0.158 0.036
Lognormal 0.094 1.915 −0.614 0.069 0.002 0.609 −0.153 0.015
Gamma 0.021 2.027 −0.704 0.044 0.015 0.447 −0.209 0.044

For a wind load, careful treatment is required. The statistical characteristics of the basic wind speed
are generally given, but a wind load is proportional to the square of the wind speed. For such a case, the
maximum wind speed is approximated as a lognormal random variable using the COV calculated based
on Table 2.4.2. The statistics of the annual maximum or those of the 50-year maximum of the wind load
effect are then estimated based on the logarithmic standard deviation of the corresponding (annual or
50-year maximum) basic wind speed. If the variability in a wind load depends only on the wind speed,
the statistics of the wind load can be expressed using the logarithmic standard deviation of the
corresponding wind speed, 𝜎𝜎ln 𝑈𝑈� , as follows.

𝜆𝜆∗𝑊𝑊 ∗
� = −2 𝜆𝜆𝑈𝑈
� (2.4.26)
𝜎𝜎ln 𝑊𝑊
� = 2 𝜎𝜎ln 𝑈𝑈
� (2.4.27)
2
� = �exp (4 (𝜎𝜎ln 𝑈𝑈
𝑉𝑉𝑊𝑊 � ) ) − 1 ≈ 2 𝑉𝑉𝑈𝑈
� (2.4.28)
The approximation in Eq. (2.4.28) can be sufficient if 𝑉𝑉𝑈𝑈� is smaller than around 0.2.
In general, uncertainty exists not only in the most basic physical quantity of a load, 𝑋𝑋, such as the
ground snow depth and basic wind speed, but also in the coefficients that are multiplied with 𝑋𝑋 to
estimate the load intensity (Eq. (2.2.1)) and in the model used for estimating the load effect from the
- C2-16 - Commentary on Recommendations for Loads on Buildings

load intensity. If such variability needs to be considered, the normalized logarithmic mean and the COV
of the load effect can be estimated as

𝜆𝜆𝑆𝑆∗̃𝑖𝑖 ≈ 𝜆𝜆𝑋𝑋∗� + ∑𝑗𝑗 𝜆𝜆∗𝐶𝐶𝑗𝑗 , (2.4.29)

𝑉𝑉𝑆𝑆𝑖𝑖 ≈ �𝑉𝑉𝑋𝑋 2 + ∑𝑗𝑗 𝑉𝑉𝐶𝐶𝑗𝑗 2. (2.4.30)

in which 𝜆𝜆𝑆𝑆∗̃𝑖𝑖 and 𝑉𝑉𝑆𝑆𝑖𝑖 are the normalized logarithmic mean and COV of the load effect 𝑆𝑆𝑖𝑖 , respectively;
𝜆𝜆𝑋𝑋∗� and 𝑉𝑉𝑋𝑋 are the normalized logarithmic mean and COV of the most basic physical parameter, or the
basic value of a load using the most basic physical quantity, 𝑋𝑋, such as the square of the wind speed,
respectively; and 𝜆𝜆∗𝐶𝐶𝑗𝑗 and 𝑉𝑉𝐶𝐶𝑗𝑗 are the normalized logarithmic means and COVs of the coefficients, 𝐶𝐶𝑗𝑗 ,
that are multiplied with 𝑋𝑋 to estimate the load intensity, respectively.

5) Target performance level

The performance level (target reliability index) required for a building or component should be
determined by considering the safety of human life, importance of the building, usage, length of the
service life, structural type of the building, type of loading, social recognition of safety, consequence
and social inconvenience brought about by failure, the minimum requirements prescribed in the
regulations, the economic efficiency, and the amount of expense and effort required to reduce the
probability of failure. In some cases, a building as a whole has a required reliability level, whereas in
other cases the required levels differ from part to part in a building.
The reliability indices of a building designed based on the current Building Standard Law and
enforcement orders in Japan vary depending on the construction site, the structural type of the building,
the components, the load combination, and the limit state considered, among other factors. Examples of
reliability indices for RC columns subjected to a load combination of 𝐺𝐺 + 𝑄𝑄 + 𝐸𝐸 are 𝛽𝛽U = 0.6–1.8
for an ultimate limit state with a reference period of 50 years, and 𝛽𝛽S = 0.2–4.0 for a serviceability
limit state with a reference period of 1 year; for steel columns, 𝛽𝛽U =1.9–2.4 and 𝛽𝛽S = 1.3–2.4 (see the
“Guidebook of Recommendations for Loads on Buildings 126)”). Considering these examples, the target
reliability index can be set higher or lower than that achieved by following the current Building Standard
Law if necessary.

2.4.2 Load factors for use in allowable stress and ultimate strength designs
1) Design return period
Loading states in an allowable stress design are determined according to the conditions where the
building is located, and in general, are classified into long- and short-term loading states based on the
duration of the loading. A short-term loading state rarely occurs during the service life of a building, and
includes large earthquakes, typhoons, extraordinary snow falls, and so on, of which the loads are treated
as a principal load. A long-term loading state implies the ordinary states under the normal state of a
building, and the load is treated as a secondary load when considered with a short-term load. When only
CHAPTER 2 LOADS AND LOAD COMBINATIONS - C2-17 -

long-term loads are considered, one of them is treated as a principal load. When dead and live loads are
considered, the latter is treated as a principal load.
The values of a principal load used in the allowable stress design and ultimate strength design are
determined in terms of the design return period (or exceedance probability for a live load), whereas the
values of the secondary loads are determined using the ratio of the mean value to the basic value of each
load as a respective load factor.
Let us consider the service life of a building. The possibility of experiencing extremely rare events,
such as large earthquakes, increases with the service life of a building. This implies that, on the basis of
probability, the longer the service life of a building is, the larger the design load required. Given the
probability that a certain load 𝑆𝑆 exceeds the magnitude 𝑠𝑠 within an arbitrary 1 year period is 𝑃𝑃1 , the
probability that the load will exceed the magnitude 𝑠𝑠 at least once within its service life 𝑡𝑡L , i.e., 𝑃𝑃T ,
can be expressed as

𝑃𝑃T = 1 − (1 − 𝑃𝑃1 )𝑡𝑡L ≈ 𝑃𝑃1 𝑡𝑡L, (2.4.31)

in which each exceedance event is assumed to occur independently during each year. If the annual
probability 𝑃𝑃1 is small, 𝑃𝑃T is approximately equal to 𝑃𝑃1 𝑡𝑡L, as in Eq. (2.4.31). In the case of a longer
service life, 𝑃𝑃1 should be reduced, i.e., a larger value of the design loads should be selected to keep the
exceedance probability of a load during a selected service life constant. In other words, a longer design
return period should be selected for a longer service life.
Determination of the design loads based on the return periods has not been explicitly achieved in a
conventional allowable stress design or ultimate strength design in Japan. The service life, however, has
been regarded in the life cycle cost management and life of the buildings. Therefore, even if conventional
design methods are used, it is important to account for design loads based on the service life of a building.
The importance and failure consequences of a building are mentioned in the following. If the usage
itself of a building is important or a building itself is valuable, the building is considered as important.
If the direct and indirect influence on the surroundings owing to the loss of the expected functions and/or
failure of a building is serious, the building is considered as the one with significant influence on the
surroundings. When designing such buildings, larger design loads are generally adopted, and in a
conventional design, design loads are increased through the introduction of an overdesign factor. Such
a factor is often called an “importance factor” or a “factor of building use” and is multiplied with a
standard design load to increase the design load. This adjustment of the design loads by multiplying the
factors is equivalent to selecting a longer return period. In other words, the longer the return period is,
the larger the load intensity, and the greater the increase in the above factors. At the same time, it is
beneficial to consider the design loads based on the return period because doing so always increases the
awareness of the non-exceedance or exceedance probability of the loads themselves. More specifically,
by uniformly treating different kinds of loads based on the probability of occurrence, the relative
intensity of the loads can be compared with one another on the basis of the degree of occurrence or
exceedance.
- C2-18 - Commentary on Recommendations for Loads on Buildings

In 2000, the Housing Quality Assurance Act was enacted in Japan, and the factored design loads
have been introduced to easily achieve the performance grading, which is not based on the importance
or use of a house, but based on the requirement of the owner, in the current provision on securing the
housing quality. In Vision 2000 shown in Table 2.3.1, the design loads are specified based on the return
period. The return period concept has been playing an extremely important role in the performance-
based design.

2) 𝑡𝑡R -year recurrence value

Let us consider the annual maximum of load 𝑋𝑋 with the CDF, 𝐹𝐹𝑋𝑋 (𝑥𝑥). Then, the interval between the
events in which 𝑋𝑋 exceeds value 𝑥𝑥, i.e., 𝑇𝑇R , is also a random variable. Let us derive the expected value
of time interval 𝑇𝑇R . First, the probability that 𝑋𝑋 exceeds 𝑥𝑥 during a year is 𝑝𝑝 = 1 − 𝐹𝐹𝑋𝑋 (𝑥𝑥), and the
probability that 𝑋𝑋 exceeds 𝑥𝑥 for the first time during the second year is (1 − 𝑝𝑝)𝑝𝑝 . Similarly, the
probability that 𝑋𝑋 exceeds 𝑥𝑥 for the first time during an 𝑖𝑖 -th year is (1 − 𝑝𝑝)𝑖𝑖−1 𝑝𝑝. Therefore, the
expected value of the time interval, 𝑡𝑡R , can be estimated as

1 1
𝑡𝑡R = 𝐸𝐸[𝑇𝑇R ] = ∑∞
𝑖𝑖=1 𝑖𝑖 (1 − 𝑝𝑝)
𝑖𝑖−1
𝑝𝑝 = 𝑝𝑝 = 1−𝐹𝐹 . (2.4.32)
𝑋𝑋 (𝑥𝑥)

This 𝑡𝑡R is called the “mean return period” or simply the “return period,” and the value corresponding
to this period, 𝑥𝑥, is referred to as the 𝑡𝑡R -year recurrence value, 𝑥𝑥(𝑡𝑡R ). Here, 𝑥𝑥(𝑡𝑡R ) can be obtained
by solving Eq. (2.4.32) with respect to 𝑥𝑥, and can be expressed using the inverse function of 𝐹𝐹𝑋𝑋 (𝑥𝑥) as

1
𝑥𝑥(𝑡𝑡R ) = 𝐹𝐹𝑋𝑋 −1 �1 − 𝑡𝑡 �. (2.4.33)
R

Naturally, the longer 𝑡𝑡R is, the larger 𝑥𝑥(𝑡𝑡R ) is.

The relation between the importance factor, 𝐼𝐼F , discussed in Clause 1), and the return period can be
expressed as

𝑥𝑥(𝑡𝑡R )
𝐼𝐼F = 𝑥𝑥(𝑡𝑡0 )
, (2.4.34)

where 𝑡𝑡0 is the return period of the design load in a conventional design, and 𝑡𝑡R is the return period
of the load factored by 𝐼𝐼F .

3) Return period of design load and performance level23)

When a design load is determined in terms of the return period in the allowable stress design or
ultimate strength design, and the uncertainties in resistance can be considered as negligible compared
with those in the load effect, for example, a predetermined interstory drift ratio is considered as a limit
state, the exceedance probability of the load effect directly corresponds to the performance level of the
structure, i.e., the failure probability. However, there generally exist negligible uncertainties in the
resistance, and it is not an easy task to select the return period for design loads for use in the allowable
CHAPTER 2 LOADS AND LOAD COMBINATIONS - C2-19 -

stress design or ultimate strength design. Here, some examples of the relation between the return period
and the limit state probability are presented as references considering two load combinations: one is the
combination of a dead load and a live load (long-term loading), and the other is a combination of a dead
load, a live load, and a principal load (short-term loading). The following are assumed:

- The dead load is normally distributed with the COV, 𝑉𝑉𝐺𝐺 = 0.1,
- The live load is lognormally distributed with the COV, 𝑉𝑉𝑄𝑄 = 0.4,
- The ratios of the mean of the live load and a principal load to that of the dead load is 0.2 and 0.7,
respectively,
- The resistance is lognormally distributed with the COV, 𝑉𝑉𝑅𝑅 = 0.15, 0.3, or 0.45, and its 5th
percentile value is considered as the nominal value, i.e., the mean of the resistance is determined
such that it satisfies both of the following two design criteria:
- the 5th percentile value of the resistance
≥ the mean of the dead load + the basic value of the live load (99% non-exceedance value)
- the 5th percentile value of the resistance
≥ the mean of the dead load + the mean of the live load + the 𝑡𝑡R -year recurrence value of a
principal load.

It should be noted that the examples are shown only as a guide. In order to express the level of the
structural performance more accurately and achieve an efficient design, it is desirable to adopt the load
and resistance factor design based on the failure probability or the reliability index that is directly related
to the importance of a building (refer to “Guidebook of Recommendations for Loads on Buildings 126)”
for details).
Considering snow and wind loads, Fig. 2.4.4 illustrates the relation between the return period of the
design load and (a) serviceability limit state probability with the reference period equal to 1 year, and
(b) the ultimate limit state probability with the reference period equal to 50 years when the annual
maximum value of a principal load is described through a Gumbel distribution. It is assumed that the
COV of the annual maximum value, 𝑉𝑉𝑆𝑆 , is equal to 0.45 considering the annual maximum of the wind
load, and that the COV of the resistance, 𝑉𝑉𝑅𝑅 , is equal to 0.15, 0.3, or 0.45. The dotted line in the figure
shows the exceedance probability of a principal load with the return period equal to 𝑡𝑡R , which
corresponds to the limit state probability when 𝑉𝑉𝑅𝑅 is equal to zero. It can be seen in the figure that the
relation depends on the ratio among the loads. When the return period of a principal load is relatively
short, where the intensity of the load is small compared with the other loads, the limit state probability
is constant regardless of the return period. Within this range, the resistance is determined through the
load combination of the dead and live loads. In contrast, when the return period of a principal load is
relatively long, the limit state probability decreases more as the return period is longer. It should be
noted that the rate of decrease tends to be smaller when 𝑉𝑉𝑅𝑅 is larger.
Considering the snow and wind loads, as shown in Fig. 2.4.4, Fig. 2.4.5 illustrates the relation
between the return period of the design load and (a) the serviceability limit state probability with a
- C2-20 - Commentary on Recommendations for Loads on Buildings

reference period equal to 1 year, and (b) the ultimate limit state probability with a reference period equal
to 50 years. It is assumed that the COV of the resistance, 𝑉𝑉𝑅𝑅 , is equal to 0.3, and that the annual
maximum value of the principal load is described through a Gumbel distribution with the COV, 𝑉𝑉𝑆𝑆 ,
equal to not only 0.45 but also to 0.35 and 0.7 considering the annual maximum of the snow load in a
heavy snow area and an ordinary area, respectively. The limit state probability is larger within the range
of long return periods as the COV of a principal load, 𝑉𝑉𝑆𝑆 , is smaller. This trend becomes more obvious
as the COV of the resistance, 𝑉𝑉𝑅𝑅 , increases.

Figure 2.4.4 Limit state probability and return period of design loads of wind and snow (𝑉𝑉𝑆𝑆 = 0.45)

Figure 2.4.5 Limit state probability and return period of design loads of wind and snow (𝑉𝑉𝑅𝑅 = 0.3)

Considering a seismic load, Fig. 2.4.6 illustrates the relation between the return period of the design
load and (a) the serviceability limit state probability with the reference period equal to 1 year, and (b)
the ultimate limit state probability with the reference period equal to 50 years when the annual maximum
value of the principal load is described through a Fréchet distribution with the COV, 𝑉𝑉𝑆𝑆 , equal to 1.4.
It can be seen in the figure that the relations do not depend on the return period unlike those of the snow
load and wind load shown in Fig. 2.4.4. The relations do not change much when 𝑉𝑉𝑆𝑆 is larger than about
CHAPTER 2 LOADS AND LOAD COMBINATIONS - C2-21 -

0.8. On the other hand, when the return period of the principal load is relatively short, and accordingly,
when the intensity of the load is small compared with the other loads, the limit state probability is
constant regardless of the return period, as observed in Fig. 2.4.4. Within this range, the resistance is
determined through a load combination of the dead and live loads.

Figure 2.4.6 Limit state probability and return period of design earthquake (𝑉𝑉𝑆𝑆 = 1.4)

2.4.3 Load factor of secondary thermal load

Regardless of the limit state design method, the allowable stress design method, and the ultimate
strength design method, the following values are used for the load factor of a secondary thermal load.
- Load combination with a snow load: 0.8
- Load combination with a wind load: 0.7
- Load combination with a seismic load: 0.4
Even when considering the snow, wind, and seismic loads as the principal load, the thermal load,
which occurs on a daily basis, needs to be appropriately combined. The load factor of a secondary
thermal load was evaluated using Turkstra’s rule (Method 1) and the probability distribution of the
directly determined combination load (Method 2). Figures 2.4.7 through 2.4.9 show an example of the
load combination factor of the thermal load obtained using the values for a 100-year return period of the
combination load obtained using Methods 1 and 2. Although the factors shown in Figs. 2.4.7 through
2.4.9 have certain variations, they can be evaluated on the safety side using the values shown through
the broken lines in the figures.

Sapporo Asahikawa Takada

method1 method2 method1 method2 method1 method2

Average Summer Average Summer Average Summer

Reference temperature Reference temperature Reference temperature

Figure 2.4.7 Load factor of a secondary thermal load with a snow load as the principal load
- C2-22 - Commentary on Recommendations for Loads on Buildings

Sapporo Tokyo Naha

method1 method2 method1 method2 method1 method2

Average Summer Winter Average Summer Winter Average Summer Winter

Reference temperature Reference temperature Reference temperature

Figure 2.4.8 Load factor of a secondary thermal load with a wind load as the principal load

Sapporo Tokyo Naha

method1 method2 method1 method2 method1 method2

Average Summer Winter Average Summer Winter Average Summer Winter

Reference temperature Reference temperature Reference temperature

Figure 2.4.9 Load factor of a secondary thermal load with a seismic load as the principal load

References
1) SEAOC: VISION 2000 — Performance Based Seismic Engineering of Buildings, 1995
2) AIJ/Special Research Committee on Seismic Disaster Mitigation: Report of Sub-committee on
Overall Seismic Safety, 2001 (in Japanese)
3) AIJ/Special Research Committee on Seismic Disaster Mitigation/Sub-committee on Evaluation of
Hazard and Seismic Safety: Report of “Earthquake Resistance Menu 2004,” 2004 (in Japanese)
4) Kitamura, H., Miyauchi, Y., Fukushima, J., Fukada, Y., and Mori, N.: Study on standards for
judging structural performances in performance based design: Verification of JSCA seismic
performance gradation system with time history analysis, Journal of Structural and Construction
Engineering (Transactions of AIJ), No. 576, pp. 47-54, 2004 (in Japanese)
5) Iwata, Y., Sugimoto, H., and Kuwamura, H.: Reparability limit of steel structural buildings: Study
on performance based design of steel structural buildings – Part 2, Journal of Structural and
Construction Engineering (Transactions of AIJ), No. 588, pp. 165-172, 2005 (in Japanese)
6) AIJ: Seismic Loading — Toward Performance-based Design, pp. 247-259, 2008 (in Japanese)
7) ASCE: Minimum Design Loads for Buildings and Other Structures, ASCE 7-02, Reston, VA, USA,
2002
8) ISO: General Principles on Reliability for Structures, ISO 2394, Switzerland, 2015
9) British Standard: Eurocode － Basis of Structural Design, (EN1990:2002), 2002
10) AIJ: Standard for Limit State Design of Steel Structures (Draft), 1990 (in Japanese)
11) AIJ: Recommendation for Limit State Design of Steel Structures, 1998 (in Japanese)
CHAPTER 2 LOADS AND LOAD COMBINATIONS - C2-23 -

12) AIJ: Recommendation for Limit State Design of Buildings, 2003 (in Japanese)
13) Turkstra, C.J.: Theory of structural design decision, Study No. 2, Solid Mechanics Division,
University of Waterloo, Waterloo, Ontario, Canada, 1970
14) Wen, Y.K.: Statistical combination of extreme loads, Journal of the Structural Division, ASCE, Vol.
103, No. 5, pp. 1079-1093, 1977
15) Kohno, M., Sakamoto, J., and Aoki, K.: Exceedance probability in combinations of stochastic loads
and its application to probability-based limit state design, Journal of Structural and Construction
Engineering (Transaction of AIJ), No. 405, pp. 31-41, 1989 (in Japanese)
16) Mori, Y. and Murai, K.: Probabilistic modeling of combined load effect taking temporal variation
of load processes into account, Journal of Structural and Construction Engineering (Transaction of
AIJ), No. 525, pp. 33-39, 1999 (in Japanese)
17) Hoshiya, M. and Ishii, K.: Reliability-based Structural Design, Kajima Institute Publishing, Tokyo,
Japan, 1986 (in Japanese)
18) Ugata, T.: Reliability analysis considering skewness of distribution — Simple evaluation of load
and resistance factor, Journal of Structural and Construction Engineering (Transaction of AIJ), No.
529, pp. 43-50, 2000 (in Japanese)
19) Zhao, Y.G. and Lu, Z.H.: Determination of load and resistance factors by method of moments,
Journal of Structural and Construction Engineering (Transaction of AIJ), Vol. 73, No. 625, pp. 383-
389, 2008
20) Zhao, Y.G. and Lu, Z.H.: Estimation of load and resistance factors by the third-moment method
based on the 3P-lognormal distribution, Frontiers of Architecture and Civil Engineering in China,
Vol. 5, No. 3, pp. 315-322, 2011
21) Nakao, M. and Mori, Y.: Load and Resistance Factors for Use in Serviceability Limit State Design
Based on Maximum Seismic Response, Proc. AIJ Tokai Chapter Architectural Research Meeting,
Vol. 51, pp. 125-128, 2013 (in Japanese)
22) Furuki, Y., Chiba, T., Mori, Y., and Takahashi, T.: Statistics of snow load for limit state design, Proc.
JSSI & JSSE Joint Conference – 2014/Hachinohe, p. 104, 2014 (in Japanese)
23) Yamato, Y. and Mori, Y.: A study on performance-based design with design load determined based
on return period, Proc. AIJ Tokai Chapter Architectural Research Meeting, Vol. 52, pp. 89-92, 2014
(in Japanese)
24) Nakashima, H., Saito, T., and Ishikawa, T.: Study on combination of seismic load and thermal load,
Summaries of Technical Papers of Annual Meeting, AIJ, Vol. B-1, pp. 71-72, 2013 (in Japanese)
25) Nakashima, H., Ishikawa, T., and Itoh, H.: Study on combination of snow load, wind load and
thermal load, Summaries of Technical Papers of Annual Meeting, AIJ, Vol. B-1, pp. 109-110, 2014
(in Japanese)
26) AIJ: Guidebook of Recommendations for Loads on Buildings 1, 2016 (in Japanese)
- C3-1 -

CHAPTER 3 DEAD LOADS

Introduction

The dead load is the weight of a structure itself, such as the framing members in a building or the
weight of the finishing, materials, equipment, piping, or other objects continuously fixed to the structure.
Characteristically, a dead load is static, vertical, and unlikely to fluctuate during the service life of the
building.
In the calculation of the dead load, it is necessary to conduct an appropriate evaluation according to
the actual conditions of the building. This means that the dead load at the time of the design should be
set to evaluate the actual conditions of the building, taking into consideration the construction conditions,
and what objects are permanently installed in the building during use. In this case, it is difficult to strictly
distinguish a dead load from a live load, as described in Chapter 4, but it is necessary to appropriately
classify a dead load as much as possible and to include anything that actually exists in either load.

3.1 Setting and calculation of dead load

(1) Setting and calculation of dead load
The weight of the structure itself, such as the framing members in the building or the weight of objects
such as the finishing materials, equipment, or piping, which are continuously fixed to the structure, is
classified as a dead load. Setting the dead load must be done reasonably and accurately according to the
actual situation and conditions. For the dead load assumed at the time of design, it is necessary to pay
close attention and not underestimate the load. In addition, the following items should be considered:
deviation of weight estimate at the time of design, dimensional deviations at the time of construction,
attachments applied during use of the building, and the expected load during a scheduled expansion or
remodeling, among other factors. Although it is difficult to strictly distinguish a dead load from a live
load, referring to 4.1 “Definition of live load and applicable range,” it is necessary to appropriately
classify such a load as much as possible and to include anything that actually exists in either load. Here,
a flexible and appropriate engineering judgment by the engineer is required. However, using a limit state
design method, it is necessary to set the dead load and live load as probabilistic models, and a live load
considered in the present recommendations is set based on concrete survey data, and thus items not
included in the survey data should be set as a dead load. For reference, specific examples to be set as a
dead load are described below.
Structure: foundation, column, beam, wall, floor slab, stairs, additional cross section from finishing
details or construction reasons, reinforcement materials and attached fasteners of the
interior and exterior materials, etc.
Finishing: finishing materials for roofs, floors, walls, ceilings, stairs, and exteriors, unit baths (UBs),
joinery, etc.
- C3-2 - Recommendations for Loads on Buildings

Building equipment: equipment fixed on the foundation or frame, pipes, ducts, and equipment installed
in the ceiling, piping, trunk lines, and ducts installed in the PS, EPS, and DS, sanitary
appliances fixed to buildings, elevator facilities, escalator facilities, mechanical parking
facilities, cables in free access floors in computer rooms, etc. (equipment such as home
appliances not fixed to buildings, wiring installed under the OA floor of an office is included
in the live load).
Other: rooftop advertisement towers, screen walls on roofs, signs and objects fixed to buildings, etc.

As a general rule, for each part of the roof, floor, wall, ceiling, joinery, and so on, the dead load is
calculated by multiplying the volume (or area/length) by the density or unit volume (or unit area/unit
length) weight of the building material, and it is calculated by adding the weight of fixed facility
equipment/piping, etc., to the building.
These recommendations are based on the probability/statistical model of a load, and to ensure safety
and usability of a structure, after grasping the variations in the load and the accuracy of the evaluation
method, and conducting an examination, the appropriate value is used as the design load. However, at
present it is considered difficult to quantify the factors of deviation of a dead load, and research progress
in this regard remains insufficient; therefore, it is assumed that the deterministic method is applied at
this time in setting and calculating the dead load. For reference purposes, the deviating factors of a dead
load are described in the next section.

(2) Variable factor of dead load

For a performance design including safety, it is desirable for the dead load to be set using a
probabilistic/statistical method. For this purpose, the actual value of the dead load and data on the
variation are indispensable. However, it is difficult to grasp the data on the dead load owing to the nature
of the load because no practical method of investigation has been studied, and the fluctuation of the load
itself is small compared to the other loads. An investigation into the actual situation is not ongoing at
this time.
A discussion of the magnitude of the actual fluctuation value of a dead load remains a future task.
Alternatively, herein we consider the fluctuation factors of a dead load, which can be largely classified
into those at the design and construction times, as shown in Fig. 3.1.1.
These variations are influenced based on the structure type, building shape, and designer. In addition,
they are difficult to be determined uniformly. Moreover, the environmental conditions and the conditions
of use after the completion of the building construction are considered factors of change, and
furthermore, how to calculate the dead load with a degree of accuracy seems to differ based on the
engineer.
In this way, it is difficult to quantify the factor of fluctuation in a dead load, and the difficulty of
actually measuring the true building weight complicates this problem.
CHAPTER 3 DEAD LOADS - C3-3 -

According to a previous study1), when comparing the weight of the building with the measured and
calculated values, the average value of a dead load is almost equal, and the fluctuation is 5–10%. There
have also been reports indicating that the density and unit volume weight fluctuation of the structural,
finishing, and other materials are 1–10%, and the fluctuation of the dimensional accuracy is about 1–
4%.
On the other hand, according to another study2), it was reported that the dead load is considered to
have an error of about ± 10% for a normal case.
a. Accuracy of set values such as density
（Accuracy of the adopted numerical value）
b. Designer's way of thinking （Precision assumed by the designer）

Variation at design c. Accuracy of weight calculation

（Presence or absence of misunderstanding）
d. Performance of computerized program
（Difference of assumption condition）
Dead loads e. Due to uncertainty
（Fluctuation of assumed numerical value）
f. Error in building size
（Error in length, height, area）
g. Error in member dimension
（Error in member cross-sectional dimension）
Variation at construction h. Error in material constants
（Variation such as material density）
i. Variation due to detail
（Finishing details, member joint, rebar anchor, etc.）
j. Change in load on construction
（Unexpected appending of cross section, etc.）

Figure 3.1.1 Variable factor of dead load

(3) Material density

The basis of a dead load calculation is the material density of the building materials. Therefore, for
various materials in categories of cement, wood, metal, stone, glass and other materials,Table 3.1.1
shows the density of the material based on various documents such as Chronological Scientific Tables.
The densities of the materials shown in Table 3.1.1 are used to calculate a dead load, although the
density for wood stone, and other materials may greatly differ depending on the references. In the case
of stone, its value is clearly indicated in the remarks column when known, but in calculating a dead load,
it is necessary for the engineer to judge the load to avoid an underestimation.
- C3-4 - Recommendations for Loads on Buildings

Table 3.1.1 (1) Density of building materials

Density
Material name Source Remarks
（×103ｋｇ/ｍ3）
portland cement 3.1～3.2 Textbook for Building Materials3）
（high-early strength, ordinary, moderate-heat,
low-heat）
cement portland blast-furnace slag cement B type 3.0
・portland fly-ash cement B type
ordinary eco cement 3.2
Japanese sawara cypress 0.34 Textbook for Building Density is a numerical
Japanese cedar 0.38 Materials3） value in an air-dried state.
Japanese cypress 0.41～0.44
・hiba false arborvitae・todo fir Wood placed under normal
Japanese *mark
・Yezo spruce・momi fir atmospheric pressure for a
coniferous ：supervision by Forestry
Japanese hemlock・Japanese torreya 0.51～0.54 long period shows a certain
Research Institute
・Japanese red pine・Japanese larch moisture content
，Wood industry handbook
・Japanese yew tree （Equilibrium moisture
(Revised fourth edition)4）
Japanese black pine 0.57 content ... average 15% on
average in Japan). The
western red cedar ・formosan cypress 0.37～0.38 **mark state at this time is called
European spruce 0.43 ：Editing by Ministry of an air-dried state.
Western hemlock・Sitka spruce 0.46～0.51 Agriculture and Forestry
・ponderosa pine・Lawson cypress Forestry Experiment Station
Imported ・white fir・Scots pine Wood Department
coniferous ・Taiwan yellow cypress・agathis ，300 kinds of useful wood
・Radiata pine・Korean pine in the world,1975
・Siberian larch・Norfork island pine ,Japan Wood Processing
Oregon pine 0.55
Technology Association5）
paulownia tree 0.29
Japanese linden・Japanese judas tree 0.48～0.50
・castor-aralia
timber camphor tree・Japanese chestnut 0.52～0.55
・Fraxinus platypoda oliv
persimmon tree・Cerasus jamasakura** 0.60～0.63
Japanese ・Japanese zelkova tree
hardwoods ・Japanese beech
Japanese ash・Japanese oak 0.65～0.69
**
・painted maple ・Japanese red birch
distylium racemosum 0.89～0.92
**
・Japanese white oak
・Japanese evergreen oak

aspen 0.39
jongkong 0.48
white lauan・red lauan 0.53
・yellow meranti
Imported red meranti 0.58
hardwoods black walnut・ramin・mahogany 0.63～0.66
teak・taun・yellow birch 0.69～0.71
white oak・apiton・Bombay black wood 0.77～0.82
ebony 0.98
rosewood 1.09

others moso bamboo・Japanese timber bamboo 0.76～0.80*

CHAPTER 3 DEAD LOADS - C3-5 -

Table 3.1.1 (2) Density of building materials

Density
Material name Source Remarks
（×103ｋｇ/ｍ3）
metal magnesium 1.74 Textbook for Building Materials3）
aluminum 2.69
*mark：Chronological Scientific Tables6）
titanium 4.51
**mark：ISO 91947）
zinc 7.12
cast iron 7.25**
tin 7.28
iron 7.87
brass 8.50
copper 8.93
nickel 8.90
silver 10.5
lead 11.3
mercury 13.6*
uranium 18.95*
gold 19.3
platinum 21.45*

Table 3.1.1 (3) Density of building materials

Density
Material name Source Remarks
（×103ｋｇ/ｍ3）
stone material tuff 1.4～2.6* Textbook for Building Materials3）
sandstone 2.0～2.4 *mark：Chronological Scientific
granite（Mikageishi） 2.6～2.9 Tables6）
marble 2.5～2.9
slate 2.7～2.9*
glass material sheet glass for buildings About 2.5 Textbook for Building Materials3）

others asphalt 1.04 Textbook for Building Materials3）

*
gasoline 0.66～0.75
methyl alcohol 0.79* *mark：Chronological Scientific
petroleum (kerosene) 0.80～0.83* Tables6）
heavy oil 0.85～0.90*
- C3-6 - Recommendations for Loads on Buildings

(4) Unit volume (or unit area/unit length) weight of a material

To obtain a dead load, each member and dimension constituting the building are set, the weight per
unit volume (or unit area/unit length) of each material and its volume (or area/length) are accurately
obtained, and all items are multiplied and summed.
The density of a material is the value of the material itself, and in the case of mixed materials or
materials with empty volumes, the density differs from the raw material itself. For this reason, the unit
weight per volume (or area/length) of typical building materials is shown in Tables 3.1.2 and 3.1.3.
Table 3.1.2 shows the unit volume weights of mixed materials, namely, aggregate, filler, concrete,
glass, brick, tile, concrete product, treated wood, and others. In the table, the unit volume weight or unit
area weight are abbreviated as the unit weight.
In addition, Table 3.1.3 shows the unit area (or length) weight of the material handled in the unit area
or unit length when calculating the weight.
CHAPTER 3 DEAD LOADS - C3-7 -

Table 3.1.2 (1) Unit volume weight of building materials

Unit Weight
Material name Source Remarks
（kN/m3）
aggregate・ artificial lightweight concrete fine aggregate 9～12 Textbook for Building
filler aggregate coarse aggregate 5～8 Materials3)
Recommendations for Textbook for Building Materials3)
dry state 17 Loads on Buildings9) f.m.3.3：17.5kN/m3
sand f.m.2.8：17.0kN/m3
wet state 20 f.m.2.2：16.5kN/m3 here f.m.：
fineness modulus

dry state 17 Textbook for Building Materials3)

gravel Maximum size 25mm：17kN/m3
wet state 21 Maximum size 20mm：16.5kN/m3

Textbook for Building Materials3)

dry state 15 Maximum size 20mm：
crushed stone 14.5～15.5kN/m3

wet state 19 (14.0～15.5kN/m3)

( ) Blast furnace slag crushed stone

dry state 20
sand mixed gravel
wet state 23
JIS A 5007 Depending on the unit volume mass,
perlite dry state 0.2～5 there are divisions of F, S, L
concrete concrete Fc ≦36 23 AIJ Standard for The unit weight depends on the
Structural Calculation of specific gravity of the aggregate to be
normal concrete 36＜Fc ≦48 23.5 Reinforced Concrete used and the formulation, so it depends
Structures10） on the actual situation, but it is better to
48＜Fc ≦60 24
indicate on the left if not particularly
Fc ≦27 19 investigated.
light-weight
*mark：When sand, crushed sand or
concrete type1 27＜Fc ≦36 21 slag sand is added to the lightweight
light-weight aggregate, add 1 to 2 kN/m3 to the left
Fc ≦27 17＊ weight.
concrete type2
reinforced Fc ≦36 24 Weight increment by rebar is
concrete calculated by actual condition. If it is
normal concrete 36＜Fc ≦48 24.5
not specially dense reinforcement,
48＜Fc ≦60 25 especially if you do not investigate it, it

light-weight Fc ≦27 20 can be weight increment（１kN/m3)

concrete type1 by non-muscle concrete + rebar.
27＜Fc ≦36 22
light-weight
Fc ≦27 18
concrete type2
steel Fc ≦36 25 AIJ Standards for Weight increment by steel bars and
reinforced Structural Calculation of steel is calculated based on the actual
concrete normal concrete 36＜Fc ≦48 25.5 Steel Reinforced [Link] it is not specially dense
Concrete Structures11) reinforcement, and if the amount of
48＜Fc ≦60 26 steel used per unit concrete amount is
not particularly large, especially if you
light-weight Fc ≦27 21
do not investigate it, weight increment
concrete type1 by muscle concrete + rebar（１
27＜Fc ≦36 23
kN/m3) + weight increment by steel
light-weight
Fc ≦27 19 bone（１kN/m3) may be used.
concrete type2
JASS 512) Unit volume weight is specified by
radiation shielding concrete 22～60
special mention
mortar 20 Recommendations for
Loads on Buildings9)
Cement 1: In the case of preparation
perlite mortar 10
of perlite 3
- C3-8 - Recommendations for Loads on Buildings

Table 3.1.2 (2) Unit volume weight of building materials

Unit Weight
Material name Source Remarks
（kN/m3）

glass sheet glass 25 Textbook for Building Materials 3) Wired sheet glass(26kN/m3 )

brick lightweight brick 11 Recommendations for Loads on ・Clay refractory brick

9)
Buildings Type1 more than 20kN/m3 (JIS R 2304)
・ hollow brick 13 ・High alumina refractory brick
tile ordinary brick 19 Type1 more than 28kN/m3 (JIS R 2305)

refractory brick
20
（chamotte brick for flue）
slug brick 21
Technical data of JAPAN Pottery tile for interior：Less than 20kN/m3
tile 22～24 CERAMIC TILE ASSOCIATION

Table 3.1.2 (3) Unit volume weight of building materials

Unit Weight
Material name Source Remarks
（kN/m3）
JIS A 5416 ALC panel structural design guidelines13)
autoclaved light-weight
concrete
concrete panel (ALC Weight for structural calculation 6.5kN/m3
4.5～5.5
products
panel) （Including Panel internal reinforcement,
panel mounting hardware, Joint mortar）
Large section wooden It depends on the type of trees to be used, so
building design and it depends on the actual condition, but it is the
treated wood glued laminated timber 5 construction manual14) value on the safe side as the approximate
calculation value of the design weight of
ordinary coniferous laminated wood.
dry state 13 Recommendations for Included in the types of clay and roam

others soil Loads on Buildings9)

ordinary state 16
wet state 18
CHAPTER 3 DEAD LOADS - C3-9 -

Table 3.1.3 (1) Unit area (or length) weight of building materials
Unit Weight
Material name Source Remarks
（N/m2）
roof tile clay roof tile 470 Recommendations ・Including overlay margin
for Loads on
Buildings9)
cement roof tile flat shape 370 JIS A 5402 ・Roof tile thickness 11～
flat S shape 370 12mm
・In the heavy snow fall region,
Japanese style shape 410
increases thickness of the roof
S shape 470 tile
block hollow t ＝100mm lightweight typeA 1,080 Recommendations ・Including joint mortar, filled
concrete for Loads on concrete, rebar
t ＝150mm typeA 1,930
block lightweight Buildings9) ・When including
typeB 2,250 circumferential girder
lightweight：Increase more
heavyweight typeC 2,560
than 200N/m2
t ＝190mm lightweight typeB 3,010
heavyweight：Increase
heavyweight typeC 3,400
more than 150N/m2
hollow Block size 115×115×80 1,100 Design and Including joint mortar（width
glass block (mm) 145×145×95 1,100 construction 10mm） and reinforcing
manual of glass bars(SUS5.5φ)
190×190×80 880
block wall16）
190×190×95 940
300×300×95 930
asphalt asphalt waterproof t ＝9mm 150 Recommendations ・Contain no protective
products layer for Loads on concrete layer
t ＝12mm 180
Buildings9)

Table 3.1.3 (2) Unit area (or length) weight of building materials
Unit Weight
Material name JIS name prestressed concrete wire(N/m) Source Remarks
welded wire mesh（N/m2）
steel prestressed SWPR2 stranded wire by 1.0 JIS G 3536
material concrete wire two pieces 2.9mm

SWPR7 stranded wire by 9.3mm 4.0

seven pieces 10.8mm 5.4
12.4mm 7.1
15.2mm 10.8
SWPR7 stranded wire by 9.5mm 4.2
seven pieces 11.1mm 5.7
12.7mm 7.6
15.2mm 10.8
SWPR1 stranded wire by 17.8mm 16.2
nineteen pieces 19.3mm 18.9
20.3mm 21.1
21.8mm 24.3
welded wire 6φ×100×100 40 Calculated using
mesh 6φ×150×150 30 cross section and
6φ×200×200 20 placement
9φ×150×150 70
9φ×200×200 50
9φ×250×250 40
- C3-10 - Recommendations for Loads on Buildings

Table 3.1.3 (3) Unit area (or length) weight of building materials
Unit
Material name JIS name Weight Source Remarks
（N/m2）
deck plate keystone plate 25×90×50×40×0.8 90 JIS G 3352 Weight as it is rolled
25×90×50×40×1.0 110 In the case of zinc plating
add the following numerical
25×90×50×40×1.2 130
values
deck plate 60×200×100×80×1.2 140 ・In case of Z12：3N/m2
60×200×100×80×1.6 190 ・In case of Z27：6N/m2
75×200×88×58×2.3 290
deck plate for plate thickness t ＝1.2mm 140 Each height：h＝75mm
composite slab plate thickness t ＝1.6mm 190 manufacturer Intermediate value of each
technical product
steel deck plate plate thickness t ＝1.2mm 190
document
for floor form
(flat deck) plate thickness t ＝1.6mm 250

Table 3.1.3 (4) Unit area (or length) weight of building materials
Column Beam
Unit Weight
Material name (thickness: Unit mm) (thickness: Unit mm) Remarks
（kN/m3 ）
1 hour 2 hour 3 hour 1 hour 2 hour 3 hour
sprayed rock wool Rock Wool Association
(dry type / semi- more than 2.8 25 45 65 25 45 60 Japan
dry type)
fire calcium Fibre-reinforced Cement
resistive Type 1 3.5～7.0 20 35 55 20 35 50
silicate Sheets Association
covering board Type 1：For finishing
Type 2 1.5～3.5 25 45 60 25 40 55 Type 2：For appearance
ceramic fiber 6 20 30 40 20 30 40

Table 3.1.3 (5) Unit area (or length) weight of building materials

outer Unit Weight（N/m）

nominal
Material name diameter Empty Piping filled Source Remarks
diameter
（mm） piping with water
piping gas pipe 50 61 50 80 Building Piping filled with
100 114 120 200 facilities water shall be
Seismic design weighing safely and
200 216 290 620
Construction not expecting a decay
300 319 520 1,240 margin.
guidelines15)
carbon steel 50 61 50 80
pipe for 100 114 160 240
pressure
200 216 410 730
piping
300 319 770 1,450
CHAPTER 3 DEAD LOADS - C3-11 -

Table 3.1.3 (6) Unit area (or length) weight of building materials
Unit Weight（N/m）
Plate
Size
Material name thickness No heat With heat Source Remarks
(mm×mm)
（mm） insulating insulating
material material
duct rectangular 450× 300 0.5 90 100 Building equipment Heat insulating
duct 750× 500 0.6 180 200 technology material：
1,500× 800 0.8 370 420 conference session glass wool
2,200× 1,000 1.2 880 940 material
spiral duct 150φ 0.5 20 30 Duct material：
200φ 0.5 30 40 zinc plating steel sheet
250φ 0.5 40 50
300φ 0.5 50 60
- C3-12 - Recommendations for Loads on Buildings

References
1) Kanda, J.: PROBABILISTIC LOAD MODELLING AND DETERMINATION OF DESIGN
LOADS FOR BUILDINGS, Research report, Dept. of Architecture, Faculty of Engineering,
University of Tokyo, 1990
2) Nishizawa, T., Ohno, T., Tobita, J., and Fukuwa, N.: Evaluation of the building weight of high
rise building with irregular shape for building health monitoring, Summaries of Technical Papers
of Annual Meeting, Architectural Institute of Japan, Structures II, pp. 717-718, 2009 (in Japanese)
3) Architectural Institute of Japan: Textbook for Building Materials, 2013 (in Japanese)
4) Supervision by Forestry Research Institute, Wood industry handbook (Revised fourth edition),
Maruzen Publishing Co. Ltd., 2004 (in Japanese)
5) Editing by Ministry of Agriculture and Forestry Experiment Station Wood Department, 300 kinds
of useful wood in the world, Japan Wood Processing Technology Association, 1975 (in Japanese)
6) Editing by National Astronomical Observatory of Japan: Chronological Scientific Tables, Maruzen
Publishing Co. Ltd., 2013 (in Japanese)
7) ISO: ISO 9194- Bases for design of structures - Actions due to the self-weight of structures, non-
structural elements and sorted materials - Density, 1987
8) Hagura, H., Miyahara, H., Sugino, T., and Adachi, I.: Handbook on load of civil engineering and
architecture, Ohmsha, Ltd., 1976 (in Japanese)
9) Architectural Institute of Japan: Recommendations for Loads on Buildings, 1981 (in Japanese)
10) Architectural Institute of Japan: AIJ Standard for Structural Calculation of Reinforced Concrete
Structures, 2010 (in Japanese)
11) Architectural Institute of Japan: AIJ Standard for Structural Calculation of Steel Reinforced
Concrete Structures, 2001 (in Japanese)
12) Architectural Institute of Japan: Japanese Architectural Standard Specification JASS5 Reinforced
Concrete Construction, 2009 (in Japanese)
13) ALC Association: ALC panel Structural Design Guidelines, 2004 (in Japanese)
14) The Building Center of Japan: Large Section Wooden Building Design and Construction Manual,
1988 (in Japanese)
15) The Building Center of Japan: Building Facilities Seismic Design and Construction Guidelines,
1982 (in Japanese)
16) Electric Glass Building Materials Co., Ltd.: GLASS BLOCK, Design and Construction Manual of
Glass Block Wall, 2013 (in Japanese)
- C4-1 -

CHAPTER 4 LIVE LOADS

4.1 General

4.1.1 Definition
Live loads are the weights of people, furniture, supplies, machines, stores, and so on, borne by a
building during its use and occupancy.
Live loads are distinguished from dead loads, which are the weights of the building itself, the
secondary members, and the finishing materials. Live loads are movable and variable during the use and
occupancy of a building, and occasionally cause dynamic effects. Therefore, they are easily affected by
social transitions, such as the rapid advances in building services equipment and mechanization. The
loads of small or movable pieces of equipment are considered live loads, but equipment belonging to
the building that is fixed and heavy is regarded as a dead load.
Live loads are specified as the weight per unit area corresponding to the use of the floor. In terms of
their concentration, they are estimated differently, depending on the type of structural component.
Live loads are produced by the gravity of people and equipment during use and movement, and do
not include environmental loads such as snow, wind, or earthquake loads. In the design of a building,
the design live load of must be calculated by considering the maximum load effect for the particular use
caused by the specific disposition of people and equipment.
The recommendations are based on data from recent surveys of live loads conducted in Japan. There
are two problems with using these data for the recommendations.
1) Not all possible floor uses are surveyed.
2) Spatial scatter may be comprehended using sufficient data, but temporal scatter, particularly that
resulting from the concentration of people and furniture occurring only once during a period of several
years, or even once in more than ten years, cannot be determined using few or no data.
For problem 1), it is impossible to survey all possible uses of a floor because future human activities
cannot be predicted. Therefore, design live loads for unspecified uses should be estimated from loads
caused by similar uses. The classification of uses in the recommendations is based on available data for
present typical uses. The recommendations apply to the normal building use. For special uses, the design
live load should be reconsidered with reference to the estimation method of the recommendations. The
disposition of furniture and people depends on the building uses, which causes the relationship between
the stochastic and design values for the maximum load effect to vary.
Problem 2) is related to the decision regarding the level of the building’s serviceability and safety in
its structural design. Because there have been few claims against using the live loads in a conventional
structural design of existing buildings, it is considered that the current sustained live loads in practice
can be referred to without serious danger or loss of serviceability. Therefore, in the recommendations,
the basic value of a live load is estimated based on the sustained load data surveyed. In accordance with
engineering judgment, the scenario of a rarely occurring concentration of people and furniture is
- C4-2 - Recommendations for Loads on Buildings

considered, and safety is verified through a probabilistic simulation model. This calculation assumes
that the estimation of the basic value is adequate. In the future, if sufficient stochastic data from temporal
variations are stored, it is thought that it will be possible to reconsider the design live load directly from
the probabilistic model of the maximum value during the design lifetime, as is currently done for snow,
wind, and earthquake loads.
The design load for safety and serviceability is based on the basic value referred to above. Therefore,
the basic value of a live load in the recommendations may be used as the design load in the allowable
stress design for sustained loading.
If the levels of safety and serviceability are modified, the percentile determining the basic value may
be varied from 99%, for example, to 95 or 99.9%, from the stochastic value of the surveyed data.
When a design load lower than the basic value is used, it should be carefully applied based on the
examination of the maximum value during its design lifetime such that the level of safety does not
become too low.
When the design load in the limit state design is estimated using the basic value in the
recommendations, it is necessary to determine the appropriate load factor. At the present time, there
may not be a sufficient amount of stochastic data, but for designs in which a determination of the
serviceability in a normal state and that of safety during the design lifetime are considered, the
relationship between performance and quality, which are ambiguous in a conventional allowable stress
design, are specified. Therefore, the recommendations are expected to be applicable to a limit state
design.

4.2 Estimation of Live Load

4.2.1 Equation for live load
The basic value of a live load is estimated as a sustained load, and calculated as a product of the
basic live load intensity, which is obtained statistically, a conversion factor for an equivalent uniformly
distributed load, an area reduction factor, and a multi-story reduction factor.
The basic live load intensity is in the 99th percentile based on the statistic data of the average weight
of people and furniture in an area of 18 m2 for the particular use of a floor. Considering the temporal
concentration, people and furniture should be estimated separately because of their different dispositions.
However, owing to an insufficient amount of data, they are estimated together in the recommendations.
The conversion factor for an equivalent uniformly distributed load is estimated differently for
members such as slabs, beams, girders, columns, and foundations because the influence of their
disposition state on the load effect is different. In general, the equivalent uniformly distributed load Qe
is defined as

∫𝐴𝐴 𝐼𝐼𝑖𝑖 𝑤𝑤(𝑥𝑥, 𝑦𝑦)𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑

𝑄𝑄𝑒𝑒 = max𝑖𝑖 � � (4.2.1)
∫𝐴𝐴 𝐼𝐼𝑖𝑖 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑
CHAPTER 4 LIVE LOADS - C4-3 -

where A is the influence area of the specific member, which is regarded as the floor area influencing the
load on the member, Ii is the influencing function defining the load effect on section i of the member,
and w(x, y) is the load of people and furniture at coordinates (x, y).
When considering Eq. (4.2.1), the basic equation for a load estimation expresses Q0 as a
representative value of the essentially ambiguous random variate X, and ke, ka, and kn are factors given
as mean values if there is a stochastic basis for doing so. However, in the recommendations, ke is defined
for convenience as the ratio of the 99th percentile of Q to Q0, where ka is given as described in the
following section 4.2.4, kn is 1, and Q is estimated from the mean influence area for each member.
Although ke is generally different for beams, girders, and columns, herein the difference is insignificant,
and thus the same value is used.

4.2.2 Basic live load intensity

The basic live load intensity Q is estimated based on surveys of several normal uses. The scatter of
the averaged load, that is, the live loads divided by the area upon which they act, becomes smaller as
the assumed area becomes larger because live loads are averaged over the whole area. Therefore, the
basic intensity of a live load should be determined by considering the influence of area.
In calculating statistic values, the surveyed data are divided into square unit areas, such as 1 m2
(1 m × 1 m), 4 m2 (2 m × 2 m), 9 m2 (3 m × 3 m), etc., and the averaged loads are calculated for
each case. This analysis is called an analysis of averaged live load intensities for square unit areas. The
calculated values are regarded as the statistic values of load intensity, and are estimated using the method
of moments to derive parameters of the probability distributions.
Four main probability distributions are applied: Normal, Log-normal, Gumbel (Type I extreme
distribution), and Gamma. Occasionally, other distributions are applied, but have no significant
influence on the results.
After an estimation of the parameters, the goodness of fit is examined based on the normalized error,
and probabilistic models for the respective areas should be selected as a distribution with the smallest
normalized error. One probability distribution is selected to specify the influence of the area on the loads.
A Gamma distribution is selected because it generally shows a good fit for various uses, and the
percentiles are calculated.
The basic live load intensity is estimated as the 99th percentile of the load models. To investigate
the influence of the area on the load intensities, the relationship between the percentiles and area is
expressed as a regressive equation.
For a detailed calculation method, see section 4.2.4. When considering the actual area of a building,
areas smaller than 16 m2 (4 m × 4 m) are not used. Fig. 4.2.1 shows examples of averaged weights and
regression curves. This figure shows that the load intensities are influenced by the area.
The basic live load intensities must be specified as the values normalized to a specific area. In the
recommendations, they are normalized to an area of 18 m2, based on the area of a single slab, the
arrangement of frames, and the mean surveyed area.
- C4-4 - Recommendations for Loads on Buildings

For particular uses not specified in the recommendations, the basic live load intensity is estimated
from surveyed data based on the principles of the recommendations. These principles should be adopted
in all cases. Where the use may change, the basic live load intensity should be re-estimated by the
designer, considering the probability of change, the use to which the building can be applied, and any
extra load.

Figure 4.2.1 The influence of the area on the load intensities (furniture and people)

4.2.3 Conversion factor for equivalent uniformly distributed load

The members are analyzed elastically to investigate the influence on the structure in normal use
based on a furniture disposition obtained from a survey. The equivalent uniformly distributed loads are
calculated through the following analysis.
The finite difference method is applied to the analysis.1) A reinforced concrete slab with four fixed
sides and a Poisson ratio of 0.167 is assumed. The boundary conditions of the girders are fixed. When
considering the effect of actual loads on a slab, the loads are assumed to be distributed uniformly within
CHAPTER 4 LIVE LOADS - C4-5 -

25 cm2 square areas.

The load effects on the bending moments and shear forces are analyzed for the slabs (short or long
direction, and support or mid-span), the load effects on the bending moments (support or mid-span) and
shear forces are analyzed for girders, and the load effects on the axial forces are analyzed for columns.
The equivalent uniformly distributed loads are examined in terms of the fixed end moments for the short
directions of the slabs, fixed end moments for the girders, and axial forces for the columns.
All analyses are based on the influence area, which is defined as the floor area whose load has an
influence on the assumed member. For a slab, it is equal to the tributary and panel areas, and for a girder
and column, is defined as shown in Fig. 4.2.2. In this analysis, a girder indicates a member that supports
the beams, and a beam indicates a member without a support. If there is no available information
regarding the location of the beams, such information is assumed for the respective use.
A stochastic analysis is conducted on an equivalent uniformly distributed load for each case of stress
and the averaged weight on each influence area. The load is estimated according to the probabilistic
model, which shows the best fit for the respective stresses. The conversion factor for a uniformly
distributed load is the ratio of the 99th percentile of the equivalent uniformly distributed load to the
averaged load on the influence area, and is calculated for each member.

Figure 4.2.2 Definition of influence area

Table 4.2.1 Statistics of equivalent uniformly distributed load
- C4-6 - Recommendations for Loads on Buildings
CHAPTER 4 LIVE LOADS - C4-7 -

Table 4.2.1 shows the results of these calculations. The conversion factor for a uniformly distributed
load is estimated based on Table 4.2.1. The conversion factor for a uniformly distributed load of slabs
is rounded off to 1.6, 1.8, or 2.0, whereas that of the frames is rounded to about 1.0 to 1.3, and thus 1.2
is adopted.
In a conventional design, the value of the beams used to be the same as that of the slabs or girders,
or the median value between them. In the recommendations, the designer may adopt a value according
to judgment. The value of a foundation is thought to be the same as that of a column, and is estimated
considering the effect of a reduction for changing the influence area, as indicated in sections 4.2.4 and
4.2.5. A reduction may also be applied to a multiple-story column.
In the equation for estimating the basic value, the basic live load intensity Q is multiplied by the
conversion factor for a uniformly distributed load. Thus, it is impossible to estimate the equivalent
uniformly distributed load of a standardized area of 18 m2 because of the difficulty in adjusting the area
for an equivalent uniformly distributed load analysis, which is not equal to 18 m2, to the area for an
analysis of the averaged live load intensities for the square unit areas. Therefore, it is assumed that the
relationship between the equivalent uniformly distributed load and its area is the same as that between
the averaged live load intensities for the square units and area. The product of the basic live load intensity
and the conversion factor for a uniformly distributed load can be regarded as an equivalent uniformly
distributed load.
The conversion factor for an equivalent uniformly distributed load is a ratio of the 99th percentile
of the equivalent uniformly distributed load to that of the averaged weight over the influence area of
each member. Fig. 4.2.3 compares the analyses of the equivalent uniformly distributed load2),3) and the
averaged live load intensities for the square unit areas. The broken lines and dash-dotted lines in the
figure connect the estimated values of the analyses for each member. The broken lines indicate the 99th
percentile of the equivalent uniformly distributed load, and the dash-dotted lines indicate that of the
averaged weight over the influence area. The solid line shows the results of the averaged live load
intensities for the square unit areas. Their gradients are regarded as the same.
Where the area is small, the equivalent uniformly distributed loads have a large scatter. According
to the above results, the estimated values for each application should be used, considering the
characteristics of the analysis of the equivalent uniformly distributed load and the averaged load
intensities for the square unit areas.
The International Organization for Standardization (ISO) published some live loads in ISO 2103,
“Loads due to use and occupancy in residential and public buildings”4). These are the lowest nominal
values of uniformly distributed loads and are defined as “the most unfavourable values for certain (or
expected) conditions of normal use of a building”.
Table 4.2.2 shows the load values of the ISO and AIJ. The load values of the ISO are the lowest
nominal values, whereas those of the AIJ are the rounded values of a product of the basic live load
intensity Q0 and the conversion factor for equivalent uniformly distributed load ka. In the use and
occupancy in which the loads are produced mainly by equipment (Nos. 1 7), the load values of the ISO
- C4-8 - Recommendations for Loads on Buildings

are mostly smaller than those of the AIJ, whereas in the use and occupancy in which the loads are
produced mainly by people (Nos. 8 13), the load values of the ISO are mostly greater than those of the
AIJ.

Figure 4.2.3 Comparison of analyses between the equivalent uniformly distributed load
and averaging live load intensities for square unit areas for office

Table 4.2.2 Comparison of loads (unit: kN/m2 = kPa)

CHAPTER 4 LIVE LOADS - C4-9 -

4.2.4 Area reduction factor

As the area increases, the variation in live loads becomes smaller because the live loads are averaged
over the area.
According to the recommendations, the design live load differs depending on the type of member.
One reason for this is that the relationship between a nominal live load intensity and an equivalent
uniformly distributed load differs for each member. Another reason is that the influence areas are
different with the slabs, beams, and columns. In 4.2.3, ke is defined as a conversion factor for converting
the nominal live load intensity into the equivalent uniformly distributed load based on an area of 18 m2.
This section presents a reduction factor for reducing the equivalent uniformly distributed load for areas
greater than 18 m2. Fig. 4.2.2 shows the influence area for evaluating the stress in the structural
components.
The area reduction factor is defined based on the following procedure. First, the type of probability
distribution for the averaged load intensity on the square units is examined for four types of probability
distribution: normal, log-normal, Gumbel (Extreme type I), and Gamma. The 99th percentile load,
which is calculated based on each selected probability distribution type, is formulated as a function of
the unit area.
The function of the area reduction factor is defined as follows:
𝑏𝑏
𝑄𝑄1 = 𝑎𝑎 + (4.2.2)
�𝐴𝐴f ⁄𝐴𝐴ref

where Q1 indicates a reduced live load intensity (N/m2), Af is the influence area (m2), and Aref indicates
the reference area (m2).
Parameters a and b in Eq. (4.2.2) are estimated using the method of least squares in the relation of
the 99th percentile to the unit area. The statistical data of the square units with an area of 4 × 4 (= 16)
m2 or more are used for the parameter estimation.
Next, parameters a and b in Eq. (4.2.2) are normalized by dividing Eq. (4.2.2) by the basic live load
intensity, namely, Q1, when Aref = 18 m2. The normalized formula for the reduction factor 𝑘𝑘a is
presented as follows:
𝑏𝑏�
𝑘𝑘a = 𝑎𝑎� + (4.2.3)
�𝐴𝐴f ⁄𝐴𝐴ref

The results of this analysis show that the Gamma distribution fits well for a probabilistic model of
the square unit loads for every type of use. Table 4.2.3 shows the values of parameters a and b, and
normalized parameters a� and 𝑏𝑏� of the reduction factor, for every type of use.
The area reduction factor should actually be derived using statistical data of the uniformly distributed
load, and thus the equivalent uniformly distributed load must be statistically analyzed to formulate the
load reduction factor. However, the reduction factor for changing the influence area is defined using the
statistical results of the square unit loads because data on an equivalent uniformly distributed load are
lacking.
- C4-10 - Recommendations for Loads on Buildings

The ISO recommends that uniformly distributed loads (except loads from stationary equipment and
stocked materials) be reduced.
For the use and occupancy in which loads are produced mainly by equipment, the loads can be
reduced by multiplying through the following factor:
3
𝛼𝛼1 = 0.3 + (if 𝐴𝐴 > 18 m2 ).
√𝐴𝐴

For the use and occupancy in which the loads are produced mainly by people, the loads can be
reduced by multiplying through the following factor:
3
𝛼𝛼2 = 0.5 + (if 𝐴𝐴 > 36 m2 ).
√𝐴𝐴

It should be noted that A indicates the load tributary area (not the influence area).
In the use and occupancy in which the loads are produced mainly by equipment, the reduction factor
of the ISO is smaller than that of the AIJ. In the use and occupancy in which the loads are produced
mainly by people, the reduction factor of the ISO is almost same as that of the AIJ.

Table 4.2.3 Parameters of reduction factor

4.2.5 Multi-story reduction factor

The axial compression stress in building columns caused by live loads is the cumulative stress of
the live loads on every floor that the column supports. Therefore, the variation in axial compression in
a multi-story column caused by live loads becomes smaller than the variation in axial compression in a
single-story column as the number of floors supported increases because the variation on every floor is
averaged. Thus, in calculating the axial compression caused by the live loads, the design live load can
be reduced according to the number of stories supported by the column.
However, this load reduction does not apply for loads that are produced mainly by people for two
reasons. One is that a temporary concentration of human load can easily occur, and the other is that the
load distribution over different floors cannot be clearly described. When the multi-story reduction factor
kn is used, the influence area of a single-story column is used to calculate the reduction factor kn.
The variation δi of an equivalent uniformly distributed load for a single column varies according to
the size of the tributary area of the column. The value of kn becomes smaller with an increase in δi.
CHAPTER 4 LIVE LOADS - C4-11 -

Although the tributary area of a column greatly varies with its position and the building use, δi is
assumed to be 0.4 in determining the reduction factor, considering the actual dimensions of the tributary
area of the column based on the statistical results of the square unit loads for office buildings, as shown
in Fig. 4.2.45).

Figure 4.2.4 Relationship between unit area and coefficient of variation based on a unit analysis

The correlation coefficient ρ of live loads between two different floors is determined to be 0.119,
based on the survey results for office buildings5). The reliability index, denoted by k, is 2.33 for a 99%
limit value based on the second moment method. Substituting these values into Eq. (4.2.4), removing
the square by the relation and rounding the coefficients, the multi-story
reduction factor kn is derived as shown in Eq. (4.2.5).

(4.2.4)

(4.2.5)

Table 4.2.4 shows the mean values, standard deviations, and coefficients of variation of equivalent
uniformly distributed loads of columns obtained through the survey results for office buildings5). Both
the mean values and standard deviations vary considerably for single-story columns on every floor;
however, for multi-story columns, the mean values converge to 540 N/m2, and the standard deviations
become smaller with an increase in the number of supported floors.
- C4-12 - Recommendations for Loads on Buildings

Table 4.2.4 Statistics of equivalent uniformly distributed load for columns

A temporary concentration of furniture often occurs with a relocation or change in occupancy. A

temporary concentration of live loads produces excessive axial compression on the columns. A live load
reduction factor for multi-story columns has been investigated6), considering the live load concentration
on multiple stories.
The thick solid line in Fig. 4.2.5 indicates the design live load intensity, and the thin solid lines show
the expected live load intensity when the number of simultaneous occurrences of concentrated load
changes from 2 to 6. In the figure, pex indicates the occurrence probability of k concentrated loads. The
99th percentile loads based on the survey data (indicated by in the figure) are calculated by
considering the effect of the difference in probability distributions.
Although there is a difference in the expected live load intensity for all numbers of supported floors,
it has been shown that the multi-story reduction factors in the recommendations are on the safe side and
are therefore reasonable.
The ISO recommends that uniformly distributed loads (except those loads from stationary equipment
and stocked materials) be reduced.
For the use and occupancy in which loads are produced mainly by equipment, the loads can be
reduced by multiplying through the following factor:
0.6
𝜂𝜂1 = 0.3 + (for 𝑛𝑛 ≥ 2).
√𝑛𝑛

For the use and occupancy in which the loads are produced mainly by people, the loads can be
reduced by multiplying through the following factor:
0.6
𝜂𝜂2 = 0.5 + (for 𝑛𝑛 ≥ 2),
√𝑛𝑛

where n is the number of completely loaded floors considered in the analysis (over the cross-section
CHAPTER 4 LIVE LOADS - C4-13 -

considered), for n = 1, 𝜂𝜂1 = 1, and 𝜂𝜂2 = 1.

In the use and occupancy in which loads are produced mainly by equipment, the reduction factor of
the ISO is smaller than that of the AIJ. In the use and occupancy in which the loads are produced mainly
by people, the AIJ recommendations do not apply a reduction.

Figure 4.2.5 Load intensity and occurrence probability of concentrated live loads

4.3 Live Loads Considering Uneven Distribution, Deflections, or Cracks

The analyses described in section 4.2 are based on the surveyed data for normal use. Because a
concentration or uneven distribution of loads may occur during normal use, they should be taken into
account in the estimation.
It is necessary to consider the influence of an unequal load in the following cases.
1) Cases that consider a temporary uneven distribution or concentration of furniture or human load
2) Cases that take into consideration deflections or cracks
3) Cases taking into account the dynamic effect of the loading of the load

In this section, research results on cases 1) and 2) are introduced, and those on case 3) are described
in section 4.4.

1) Cases that consider a temporary uneven distribution or concentration of furniture or human load
Concrete examples of cases in which loads of goods and human beings are temporarily unevenly
distributed or concentrated may include changes in tenants, remodeling, concentration conditions during
an emergency evacuation, and the like. However, it is difficult to precisely assume such a state
beforehand during the design stage. Because there are several studies that have examined the loading of
loads under ubiquitous circumstances, such as a temporary concentration of goods and human beings,
it is important to set the appropriate design load based on engineering judgment with reference to these
- C4-14 - Recommendations for Loads on Buildings

results as well.7)
The effect of unevenly distributed loads on the members is examined through a simulation analysis
when assuming a dynamic model. It is assumed that live loads are distributed uniformly on a slab. As
the loading area is made smaller, i.e., the ratio of distribution unevenness is made greater, the effect on
the stress of the member is examined. The effect on a square slab (ℓ × ℓ) is estimated as an example. It
is assumed that a square loading area moves on the slab. The fixed end moments are calculated.8) Figure
4.3.1 shows the results.
It is logical that, as the loading area decreases, the effect of the distribution unevenness on the stress
becomes greater. However, there is a limit to the actual concentration of furniture and the smallness of
the loading area, and thus the probability of greater unevenness of the distribution is small. In a structural
design, it is important to determine the design load in view of the particular use.

Figure 4.3.1 Effects of an uneven distribution of load on the fixed end moments of the slabs

When the loads are distributed almost uniformly during normal use, an analysis of an equivalent
uniformly distributed load gives nearly the same results as the averaged load intensities. Therefore, the
conversion factor for a uniformly distributed load needs to be determined such that the effect on each
member is apparent. In these recommendations, this is estimated in consideration of the stochastic data
on personnel loads based on the number of people on a floor during normal use, and in consideration of
an uneven load distribution.
Particularly when the loads consist mainly of personnel, the stochastic values of the density of people,
which is the number of people divided by the area in which they are located, are estimated from the data
on a survey of building users. The number of people during a single event is regarded as one sample. If
CHAPTER 4 LIVE LOADS - C4-15 -

the total number of people in an event occurring N times is obtained, that number divided by N is
regarded as 1. This assumption only applies to those events in which almost the same numbers of people
are gathered each time. Figure 4.3.2 shows the results.

Figure 4.3.2 Result of the analysis of personnel loads

During the design lifetime of a building, live loads vary with time. As explained above, live loads
for a building during normal use, i.e., sustained live loads, have been analyzed. Over the design lifetime,
the variation in live loads may be as shown in Fig. 4.3.3.
For example, in an office building, the occupancy may change several times during the design
lifetime, and the live loads will vary each time. During a single occupancy, transient loads may occur.
If the live load is determined synthetically based on this frequency, the maximum live load for the design
lifetime can be estimated. 9),10)

Figure 4.3.3 State of loading during the design lifetime of office buildings

Figure 4.3.4 shows the state of goods and humans, assuming extraordinary daily activities and the
equivalent distributed load at each time. Such concrete demonstration is also a reference at the time of
the design.
- C4-16 - Recommendations for Loads on Buildings

Figure 4.3.4 Live load for the types of uneven state of distribution11)

2) Cases that take into consideration deflections and cracks

The deflection and cracking of floorboards and beams are caused by complicated factors. Therefore,
it may not be appropriate to use the design load indicated in 4.2 as described. However, even if the same
design load value is used, the level of reliability is different for each design object. Some studies have
investigated the load of office buildings and analyzed the deflection of the floor plate from the result. It
has therefore been reported that a distributed load such as the equivalent deflection will be 0.24 to 1.75
times the loaded load strength.
In addition, analytical studies of wooden floors have been conducted. As a result, although the
rigidity of the joists and major draw has been mostly secured, and although most of the beams meet the
design criteria, the results indicate that about 6% of the members exceed the design criterion. When
designing a wooden structure, it is necessary to make an appropriate judgment by referring to these
results, such as a cross-sectional design, the materials to be used, and the placement interval of the pillars.
Cracks in concrete-type slabs and beam members are related to various cross section amounts, the
CHAPTER 4 LIVE LOADS - C4-17 -

member length, rebar strength, fog thickness, material tensile strength, and the like. It has been pointed
out that even if the stress of the tensile reinforcing bars is less than the allowable stress level, the
possibility of cracking is extremely high.
In research targeting cracks in the floorboards, a study on the design crack width owing to the design
load for an RC slab of an existing building was conduced. Cracks on the upper concrete surface at the
center of the short side of the fixed end for applications such as houses, offices, and classrooms are
obtained from the calculation formula of the RC standard.
Figure 4.3.5 shows the relationship between the crack width and loading load, using the thickness
of the cover fog and the shrinkage distortion as the parameters. If the shrinkage distortion reaches up to
about 300 μ, cracks exceeding 0.3 mm do not occur much even if the coat thickness is 3 cm.

Figure 4.3.5 Relationship between crack width and live loads 12)

Next, the loading value of a load that generates deflection and cracks when using an RC floor slab
design drawing is calculated and shown in Fig. 4.3.6.
- C4-18 - Recommendations for Loads on Buildings

Figure 4.3.6 Live load causing deflections, cracks, and surrender13)

The figure shows the results of a trial calculation of loading to surrender taken from design manuals
on houses and offices, in which the elastic deflection reaches 1/4000 of the span, and the maximum
crack width reaches 0.2 mm.
Although the loading value corresponding to the strength leading to a yield of the reinforcing bars
is large, it can be seen that there is a possibility of generating elastic deflection and cracking in the
building even with a relatively small load.
Therefore, when considering deformation in a design, it is also necessary to understand that in
certain cases it may be necessary to use a load value larger than that set up based on stress.
There have been a few systematic research reports on the influence of loading on the deflection and
cracking of concrete, and future progress is expected.
CHAPTER 4 LIVE LOADS - C4-19 -

4.4 Dynamic Effects of Live Loads

In recent times, light weight or low rigidity floor structures have been more commonly adopted in
office buildings and stores owing to the progress made in weight reduction, the strengthening of
materials, and the development of structural design technology, including long spanning floors. On these
floors, vibrations generated from the daily motions of humans such as walking and the operation of
various types of machinery and equipment may be a problem from the viewpoint of habitability. These
vibrations are rarely a problem for conventional rigid floors. However, for light weight or low rigidity
floors, the dynamic effects of human beings and equipment need to be considered depending on the use
of the building in order to achieve a preferred level of habitability.
Moreover, in concert halls, stadiums, and aerobic studios where many spectators gather, a very large
dynamic load is applied to the floor through such actions as many people jumping at the same time in
rhythm. Vibrations from this type of dynamic load may be a problem for the habitability of other floors
in the same building or floors of neighboring buildings owing to vibrations propagating through the
structure or ground. In addition, such vibration may, in some cases, cause a collapse of the floor where
the movement is being performed. It is preferable to consider these dynamic effects during the design
stage.
In buildings that are constructed on premises close to roads or railroad tracks with heavy traffic, the
vibrations generated by the running of the vehicle may be a problem from the viewpoint of habitability.
At present, there are many problematic cases with relatively lightweight buildings. However, in the
future, as the density of the city progresses, it is expected that many buildings with closer vibrating
sources and living spaces will appear, such as buildings with structures integrated with roads and tracks.
In such a building, it is necessary to study the dynamic effects from an operating vehicle.
In addition, plants/equipment, such as manufacturing or testing machines, sensitive to the effects of
vibrations, may be installed in facilities having ultra-precision environments, including semi-conductor
fabrication plants or research laboratories. Therefore, it is often necessary to control the floor responses
to the dynamic effects of humans, machinery, and equipment from the viewpoint of productivity and
workability. At such facilities, vibrations smaller than the human perception are generally a problem.
Depending on the site and plan, there are cases in which the design considerations, such as increasing
the mass and rigidity of the floor, are not possible, and there are many examples in which seismic
isolation devices and damping devices have been installed.
Based on specific research results, the dynamic effects of such live loads as those caused by human
- C4-20 - Recommendations for Loads on Buildings

movement, the operation of plants/equipment, and vehicular traffic are presented below.

4.4.1 Dynamic Effects of Human Movements

(1) Outline
Floor vibrations from various types of human movement cause many diverse types of problems.
Table 4.4.1 shows typical vibration-forcing activities and subsequent evaluation areas that consider the
actual problems caused by floor vibrations from human movement.
In this table, the term “baseline” is used for a movement under the column “one-step by one person.”
This implies that the level of vibration should first be based on understanding the vibration-forcing
power and floor vibrations from “one-step by one person.” Next, the number of persons or exciters
creating the vibration forcing, their positions, and the duration of their actions can be considered.
CHAPTER 4 LIVE LOADS - C4-21 -

Table 4.4.1 Floor vibrations caused by human movements

One step A few steps by A few semi- Random by Synchronized
by one several persons synchronized many persons, by many
person steps by several continuous persons,
persons continuous
Walking Baseline Habitability of Habitability of Habitability of
residence (for offices, etc. (for shopping malls,
those other than those other than pedestrian
floor exciters), floor exciters), decks, etc.
habitability of productivity (during
residence, and operability movement or
offices, etc. of facilities with standing still)
(including floor precision
exciters equipment
themselves) installed
Running Baseline Habitability of
offices, etc. (for
those other than
floor exciters)
Stepping Baseline Habitability of
up/down staircase (for
those other than
floor exciters
and for exciters
themselves)
Aerobics Baseline Habitability of
neighboring
rooms (for
those other than
floor exciters)
Dancing Baseline Habitability of
up and the said
down to building and
rhythm of neighboring
music or buildings (for
“tatenori” those other than
floor exciters),
structural safety
of the said
building
Jumping Baseline,
and ease of
landing use of
athletic
facilities
Semi-synchronized by several persons: Movement by two or three persons standing side by side and
unconsciously synchronizing their strides and walking pace
Random by many persons: Movement by various persons taking various positions and moving in various
directions at various speeds
Synchronized by many persons: Movement by many persons concurrently to the music, etc.
- C4-22 - Recommendations for Loads on Buildings

(2) Characteristics of Dynamic Loads from Human Movements

Figure 4.4.1 shows examples of a load-time curve for walking and running14-17). The peak p1 shown
in the figure is attributable to the impact created when one’s heel makes initial contact with a floor. The
peak p1 does not always appear, however; such incidence is within the range of 80-95% for walking and
70-85% for running. Walking loads other than peak p1 show a double-peak pattern. The first peak is due
to one’s heel making contact with the floor, and the second to one’s foot leaving the floor in preparation
for the next step. On the other hand, when running, both movements occur as a continuous movement,
and thus the running loads show a single-peak pattern.

Figure 4.4.1 Examples of load time curve for walking and running14-17)

Figure 4.4.2 shows the relationships between one’s stride/height, walking pace, foot-to-surface
contact time T0, (see Figure 4.4.1) and walking speed/height16). During normal walking, the speed/height
is approximately 0.7-0.9 /s, stride/height is around 0.4, the pace is approximately 0.5 s, and T0 is about
0.6 s. This indicates that the duration of time that both feet are in contact with the surface is
approximately 0.1 s. On the other hand, when running around the office, the upper limit of the
speed/height is approximately 1.5 /s. In this case, the stride/height is approximately 0.525, the pace is
about 0.35 s, and T0 is approximately 0.3 s.
Figure 4.4.3 shows the relationships between the magnitudes of the peaks p1 and p2, L1/W and L2/W
(where W refers to the exciter’s weight), the time for each load to reach its peak, T1 and T2, and the
speed/height16,17). When walking at a normal speed/height, the L1/W distribution centers at around 0.5,
and its upper limit is about 1.0. In addition, the T1, L2/W, and T2 distributions center at around 0.012 s,
1.2, and 0.15 s, respectively. When running at a speed/height of about 1.5 /s, the L1/W distribution centers
at around 1.5, and its upper limit is about 2.0. Finally, the T1, L2/W, and T2, distributions center at around
0.012 s, 2.4, and 0.11 s, respectively.
Figure 4.4.4 summarizes the information presented above by showing typical examples of loads
generated by walking and running.
CHAPTER 4 LIVE LOADS - C4-23 -

Figure 4.4.2 Relationships among stride/height, pace, T0, and speed/height16)

・
Figure 4.4.3 Relationships ampng L1/W, T1, L2/W, T2, and speed/height16,17)

Figure 4.4.4 Typical example of load time curve for walking and running
- C4-24 - Recommendations for Loads on Buildings

However, it should also be noted that all of the above data are applicable to a situation in which a
person (or persons) walks or runs around on a floor of normal hardness wearing socks. Peak p1 changes
when the floor is finished with a very soft flooring material such as felt-lined carpet, or when shoes are
worn (data other than peak p1 do not significantly change)14, 15, 18).

(3) Characteristics of Floor Vibrations Owing to Human Movements

Figure 4.4.5 shows examples of floor vibrations owing to one-step walking by one person (the
deformation time curve and acceleration time curve)14,15), together with the load time curve. Floor
vibrations from walking generally show complex and complicated characteristics of damped vibrations
at a natural frequency of a floor excited by peak p1, and so on (see the acceleration time curve), and
vibrations proportional to a double-peak patterned load (peaks p2, p3, etc.) (see the deformation time
curve).
1,000
Acceleration A (cm/s2) Deformation D (mm) Load L (N)

p2(L2, T2 ) p3(L3, T3 ) (0, T0 )

p1(L1, T1 )
0
1

0
100

-100
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Time T (s)
Weight W : 600N
(Calculated from actual measurements with a subject
weighing W : 754N)

Figure 4.4.5 Example of the load time curve and floor vibration for walking14,15)

An assessment of human sensitivity toward floor vibrations from walking is influenced by both
damped vibrations at the natural frequency of the floor and by vibrations proportional to the double-
peak patterned load19-22). Therefore, it is difficult to properly evaluate the dynamic effects of human
movements from the viewpoint of habitability without establishing a load model that enables us to
examine vibrations with two different frequency components.
CHAPTER 4 LIVE LOADS - C4-25 -

(4) Dynamic Load Model

(a) Time History Waveform
The time history waveform set on the basis of the load time curve for human movements serves as
the most basic dynamic load model. Figure 4.4.6 shows a typical time history waveform for walking.
The waveform shown in the figure is developed by setting the walker’s weight, W, as 600 N and
superposing sections supported by both legs (0.1 s each) in a typical load time curve, as shown in Figure
4.4.4.
(b) Fourier Series
The vibration-forcing power caused by continuous human movements generally involves many
components of the forcing frequency and its harmonics. The time history waveform consisting of such
components and the harmonics of continuous movements is generally expressed through the following
equation using a Fourier series.
𝑘𝑘

𝐹𝐹(𝑡𝑡) = 𝑊𝑊 �1 + � 𝛼𝛼𝑛𝑛 sin(2𝜋𝜋𝑛𝑛𝑛𝑛𝑛𝑛 + 𝛷𝛷𝑛𝑛 )�

𝑛𝑛=1

where: F(t) : time history waveform of the load (N)

W : exciter’s weight (N)
t : time (s)
αn : ratio of amplitude of n harmonic components to exciter’s weight
f : forcing frequency (Hz)
Φn : phase gap between n harmonic components and first harmonic components
n : harmonic number
k : upper limit of target n harmonic

Figure 4.4.6 Example of time history waveform for walking

Table 4.4.2 shows the ranges of nf. In addition, αn for various movements is derived from several
types of data. As for movements by one person, α1 is approximately 1/2 to 1/3 of the gap between the
maximum and minimum values of the actual loads. In the case of movements by many persons, α1 for
each person still tends to be smaller owing to the effects of phase differences among individual
movements.
The dynamic load model using a Fourier series is basically used for a floor with a relatively low
natural frequency in order to calculate and evaluate the amplitude of resonance that is induced between
the forcing frequency or its harmonics and its natural frequency by subtly changing the forcing
- C4-26 - Recommendations for Loads on Buildings

frequency, f, according to its natural value. This model may also be used for a floor whose natural
frequency is not low to predict vibrations from aerobics, “tatenori,” or other types of movements.
However, for walking and running, because this dynamic load model does not involve components of
the load equivalent to the peak p1, among others, it is necessary to separately examine the damped
vibrations at the natural frequency of the floor excited by the peak.

Table 4.4.2 Example of f , αn for walking and running

f (Hz) α1 α2 α3 α4
One person walking 1.6 - 2.3 0.38 - 0.5 0.086 - 0.2 0.057 about 0.05
One person running 2.0 - 3.3 1.2 - 1.4 0.33 - 0.4 0.1 - 0.15
One person jumping to landing 2.0 - 3.0 1.07 - 1.9 0.44 - 0.69 0.087 - 0.31
Dancing by many persons 1.5 - 3.0 0.5 0.2 0.05
Jumping and dancing by many persons 1.5 - 4.0 - 2.0 - 0.8 - 0.2
Aerobics by many persons 2.0 - 2.75 1.5 0.2 0.1
Concert by many persons 1.5 - 3.0 0.25 0.1 0.025
Jumping to landing by many persons 1.5 - 3.0 0.7 - 1.5 0.25 - 0.6 0.078 - 0.15

(c) Impulsive Load

The “Design Recommendations for Composite Constructions of the Architectural Institute of
Japan”23) indicates that an impulsive force created by one person walking is “almost equal to the impact
generated by a 3 kg object freefalling from a height of 5 cm”. On the other hand, the “AIJ Standards for
Structural Calculation of Reinforced Concrete Structures (1991 edition)”24) indicates that the effective
impulsive force from walking is about 3 N.s of the impulse (the half-sine wave with a maximum load
of 118 N and an action time of 0.04 s). This impulse nearly corresponds to the aforementioned 3 kg and
5 cm (3.03 Ns).
Any dynamic load model based on an impact with the above-mentioned impulse of about 3 N.s is
applicable to peak p1, and the momentum of the response to be calculated can be regarded as the
maximum amplitude in the early stage of damped vibrations at a natural frequency of a floor excited by
such a peak. In other words, this load model does not involve components of the double-peak patterned
load, and vibrations proportional to the double-peak patterned load must be separately studied. Moreover,
because vibrations at a natural frequency may be excited by peak p2 on a floor with a relatively low
natural frequency (within a range of approximately 7 to 10 Hz or lower), it is also difficult to predict
vibrations at the natural frequency with only the impact applicable to peak p1.
In this connection, when the impulse is calculated by transforming the load time curve up to peak p1
into a 1/4 sine wave with a maximum load of 300 N (0.5 times the average weight of 600 N) and an
action time of 0.012 s in accordance with the typical example of walking loads shown in Figure 4.4.4,
the value is 2.29 Ns. With the maximum load taken as 600 N (1.0 times the average weight of 600 N),
the value is 4.58 Ns.
CHAPTER 4 LIVE LOADS - C4-27 -

4.4.2 Dynamic Effects of Operations of Machinery and Equipment

There is a wide variety of machinery and equipment that can be vibration sources, including air-
conditioners for ordinary buildings, plants/equipment for industrial installations, and production
machines, and it is difficult to estimate the dynamic loads they cause in a unified way. In practice, the
vibration-forcing power is estimated from the information available from manufacturers by
understanding the vibration-inducing mechanisms of individual machinery and equipment. The
following shows an outline of the dynamic loads created by machinery and equipment25-27).
In general, vibration-inducing mechanisms can be largely classified into rotary motions,
reciprocating motions, and impulsive motions. Rotary machines such as electric fans and motors are
designed to eliminate unbalanced components caused by rotary motions. In reality, however, with a
complicated machine, it can be difficult to eliminate unbalanced components completely, resulting in a
vibration-inducing force. The vibration-inducing force caused by such a rotary machine generally
becomes a harmonic waveform and its frequency corresponds to the rotational speed of the machine.
An internal-combustion engine and a steam engine are mechanisms used to transform reciprocating
motions into rotary motions, whereas a compressor and others components are mechanisms used to
transform rotary motions into reciprocating motions. These transforming mechanisms are called crank
mechanisms, and their inertial force includes a frequency corresponding to the rotation speed of the
crank axis and its frequency components of two, four, six, or more times. However, the fourth or higher
unbalance force, which is less than the first and second forces, can be ignored for practical purposes.
For a multi-cylinder engine, although the force can be offset to some extent owing to the phase
relationship among the different cranks, the unbalanced inertial force and the unbalanced inertial
moment remain and cause vibrations. Therefore, when designing the fundamentals of an internal-
combustion engine, it is important to understand its mechanism and predict the occurrence of possible
vibrations.
If a heavy object falls into or collides with a machine such as a forger or a caster, the impulsive force
generated causes vibrations. Although it is difficult to accurately measure the magnitude of the actual
impulsive force, it is possible to express the time history waveform of the impulsive force with a half-
sine wave pulse or a rectangular pulse by assuming the impulse as the case of a free-fall or collision.
Becase it is possible to reduce the effects of the vibration-forcing power generated by the above-
mentioned machinery and equipment on buildings through appropriate vibration counter-measures, it is
important to fully understand the characteristics of the vibration sources and reflect them in the design28).
In addition, in ordinary buildings, a large number of various devices are installed, and the operating state
changes with time. Therefore, with unexpected vibrations such as from the coupling of multiple devices,
resonance in an unsteady state at the start of operation or at the time of stoppage may occur, and attention
is required when creating the design.
For reference, Table 4.4.3 shows a summary of the types of plants/equipment that are potential
vibration sources in ordinary buildings and the characteristics of such vibrations.
- C4-28 - Recommendations for Loads on Buildings

Table 4.4.3 Plants/equipment and induced vibrations26)

Type Vibration source Features of vibration
Heating Pump Rotary motion (axial rotation speed,
number of blades)

Heat pump Vibration with ultra-low frequency

Boiler

Burner
Air-conditioning/ventilation Fan Rotary motion (axial rotation speed,
number of blades)

Pump
Cooling tower
Refrigerator Reciprocating motion of pistons
Freezer Compressor (axial rotation speed,
higher level
Multi-purpose air-conditioner /
heating unit
Electric power Transformer Double the frequency at power
source
Generator Axial rotation speed
Transportation Elevator Rotary motion (axial rotation speed)
Escalator

4.4.3 Dynamic Effects of Vehicular Traffic

When a vehicle drives through an indoor parking space or along a road in front of a building,
disturbances from vibrations such as indoor floor vibrations may occur. In addition, when a train runs
above the ground or underground in close proximity to a building, the vibrations created propagate
throughout the building structure and radiate as sound in certain areas inside the building. This is a
problem with solid borne sound. Vibrations from a running of a train generally have a significant
component within the 63 to 125 Hz band, and thus there are many cases in which problems from sound
propagation through a solid occur.
The excitation force when the vehicle travels along a road or rail depends on the vibration
characteristics of the vehicle, including its mass, the spring in each part, the traveling speed, the structure
of the road or track (e.g., a plane or elevated), and the state of maintenance. Vibrations generated from
roads or tracks are mainly collisional vibrations that occur when a vehicle passes over a step, crack, or
rail joint of the road surface or track. Therefore, repairing road surfaces and rails will often significantly
suppress vibrations. In addition, depending on the structure of the ground, roads, and railroad tracks,
collision vibrations may resonate with the natural frequency of the vehicle, resulting in a vibration with
a relatively long duration.
When a car runs inside a building, its dynamic effects on the floor differ, depending on the type of
car itself, its operating speed, floor conditions, and the floor structure. Therefore, it is extremely difficult
to accurately assess the dynamic effects of a running vehicle and to reflect them within a building design.
CHAPTER 4 LIVE LOADS - C4-29 -

Instead, a simplified way to assess the dynamic effects of a running vehicle is to regard the dynamic
effects as a ratio of the dynamic deflection of the floor from the operating vehicle to the static deflections
from the vehicle’s own weight. In designing a floor, a formula has been developed to calculate the
additional static load of the car based on this ratio. In most cases, a ratio within the range of 1.2 to 1.3
is used.
Moreover, the dynamic effects are caused by a braking force generated when a running car comes
to a sudden stop. In this case, the braking force depends on the friction coefficient of the contact surface
of the car tires and load applied vertically to the contact surface of the tires. The maximum horizontal
load is generally approximately 0.6 to 0.7 G (where G is the gravitational acceleration). However, there
are cases in which an acceleration equivalent to 1 G or more from the fixed live load of a vehicle must
be considered owing to varying load sharing among the tires, depending on the type of vehicle or the
driving style.
- C4-30 - Recommendations for Loads on Buildings

4.5 Consideration of rooftop gardening loads

Rooftop gardening of residential houses has been increasing with the rise environmental awareness.
If we can determine the gardening plan and management during an in-service period, it will be easy to
evaluate the design live load of roof gardens. On the other hand, the structural engineer usually uses the
live load intensities for residential flats as the design loads for rooftop gardens because it is not easy to
grasp the live loads on the rooftop garden of a residential home. However, the adequacy of using the
design live loads for residential flats when designing a rooftop garden has not yet been confirmed. This
chapter introduces an example for evaluating the design live loads for the rooftop garden of a residential
home, and addresses the focal point in considering the rooftop gardening loads.

4.5.1 Survey of rooftop gardening loads for residential home

The samples investigated are 24 steel residential homes with ALC panels located in the Tokyo
metropolitan area29. A bathroom scale was employed for weighing movable objects such as potted plants
or gardening implements. The weights of unmovable objects, such as flowerbeds or turfed areas, were
determined using catalogues.
Planted areas on a rooftop can be divided into two areas. One is an area with turf or a flowerbed,
and the other is an area with tables, chairs, or potted plants on a normal surfaced rooftop. Each of these
areas has different characteristics in terms of the live loads on the layout and movement of objects or
humans. Therefore, the area surveyed is categorized into three use types: “planted area” (area with turf
or a flowerbed), “surfaced area” (area with tables, chairs, or potted plants on the normally surfaced
rooftops), and “others.” Typical appearances of each categorized area are shown in Fig. 4.5.1.
When evaluating the weight of objects on a rooftop, it is reasonable to categorize all rooftop objects
into movable and unmovable objects. Examples of movable objects include potted plants or gardening
implements, and examples of unmovable objects include soil or flowerbeds. The former type is named
“movable live loads,” and the latter is named “unmovable live loads.”.
CHAPTER 4 LIVE LOADS - C4-31 -

flowerbed (middle size

grass
tree)

flowerbed (flowers) surfaced rooftops

flowerbed (bushy tree) others

Figure 4.5.1 Typical appearance of each categorized area

4.5.2 Survey results

Table 4.5.1 shows an area ratio of each use area and the mean and maximum values of “movable live
loads.” The mean value of the ratio of a planted rooftop area is 26%, and maximum ratio is 45%. The
area ratios of “planted area,” “surface area,” and “others” are 25%, 25%, and 50%, respectively. The
“movable live load” also varies widely. The mean value is 59 N/m2, and maximum value is 197 N/m2.

Table 4.5.1 Area ratio of each use area and the mean and maximum values of “movable live loads”

Ratio of Area of “Movable live Load” “Movable live Load”

Rooftop
planted surfaced per residential home per residential home
area (m2)
area rooftop (N) unit area (N/ m2)
Mean value 68.4 0.264 0.257 3,263 59
Maximum
132.7 0.453 0.767 9,501 197
value
- C4-32 - Recommendations for Loads on Buildings

Table 4.5.2 shows the typical specifications of a “surfaced area” and “planted area.” The table also
shows the live load unit of each area type. The weight of the soil in a planted area is used as the value
of weight of artificial light-weight soil under wet conditions. The specific weight of light-weight soil
under wet conditions is 0.6. If the thickness of light-weight soil is over 150 mm, the live load is more
than 1,000 N/m2, which was determined in these recommendations.

Table 4.5.2 Typical specifications of “surfaced area” and “planted area”

specifications thickness weight
of soil (N/m2)
(mm)
Wood deck － 300
surface of

－
rooftop

tile 10mm 300

Precast concrete － 700
panel 30mm
grass 70 850
flowers 150 1,250
150 1,250
planter area

bushy tree
flower bed

(0.8–1.2m)
vegetable 200 1,600
garden
middle size 200 1,800
tree (1.2–2m)

For movable loads such as pods and planters, they are usually placed in an unfinished area, tables
and chairs are placed on a wooden deck or grass area. The occurrence frequencies of all movable loads
are listed in Table 4.5.3 in descending order. The average weight of a flower pot is 64 N, and the average
total weight of the flower pots in a residential house is 900 N. The variation in weight of the planters is
very large, with an average weight of 101 N, and maximum weight of up to 700 N.

Table 4.5.3 Survey results for movable live loads (N)

average load of maximum load
average standard
load quantity a residential of a residential
load deviation
home home
1. Flower pot 379 64 101 897 4,341
2. Planter 201 101 109 942 5,660
3. Chair 27 81 74 92 960
4. Outdoor unit of A/C 26 251 113 272 1,617
5. Storeroom 14 638 623 372 3,577
6. Table 7 171 74 50 314
7. Bucket 5 28 30 6 78
8. Concrete block 5 588 608 98 1,470
9. Hose 4 69 20 11 98
10. Toolbox 4 130 183 22 402
CHAPTER 4 LIVE LOADS - C4-33 -

4.5.3 Statistical characteristics of movable live load

To clarify the stochastic characteristics of a movable live load, the unit load of the movable live
load is analyzed based on the stochastic method employed in the AIJ recommendations for building
loads. Table 4.5.4 shows example unit load evaluations in areas where the movable load frequency
rate is high. Figure 4.5.2 shows a histogram of movable loads for each unit size, the mean values, the
standard deviations, the coefficient of variations, and the 99th percentile loads. Although the flower
pots and planters are closely spaced in images ①, ②, and ③, the mean unit load of unit size “a” is
300–500 N/m2, and the mean unit load of unit size “c” is half of unit size “a.” The heaviest unit load is
unit size “a,” shown in image ⑥. The mean value of a 624 N/m2 unit is larger than the 617 N/m2 of
the 99th percentile load shown in Fig. 4.5.3, and thus it can be said that such a case is very rare
because of the existence of a large flowerbed full of soil. The basic value of the design live loads for a
residential home determined in the AIJ recommendations for buildings loads is 1,000 N/m2, which
states that the 99th percentile for movable live loads for a rooftop gardens is less than the 99th
percentile of live loads for living rooms.

Table 4.5.4 Utilization of rooftop gardens and statistics of unit loads (N)
Sample size
Sample size
Sample size

a (1,830×610mm unit) b (1,830×1,830mm unit) c (3,660×3,660mm unit)

Figure 4.5.2 Distribution of unit loads for movable live loads

- C4-34 - Recommendations for Loads on Buildings

4.5.4 Examples of determining design live loads for rooftop gardens

This chapter introduces examples of determining the design live loads for rooftop gardens of
residential homes having steel frames and ALC panels or RC floors.
For the case of ALC panels, the most serious scenario is that the heaviest unmovable live load, such
as from a flowerbed with a vegetable garden and mid-sized tree, is throughout all areas of an ALC panel.
The live load of such a worst case will be the maximum equivalent uniformly distributed load. An
analytical model for calculating the design live loads for frames is a 3D static model. Applied loads on
the frame model are determined as the support reaction under the equivalent uniformly distributed loads,
as mentioned before. The assumed column section is an 80 × 80 × 3.2 box pin-connected at both ends.
The assumed beam section is an H-shaped beam of 250 × 100 × 4.5 × 6 in dimensions supported at both
pin ends. The selected stress when calculating the equivalent uniformly distributed load is the axial force
on the columns. For beams, the bending moment, shear force, and deflections are the selected stresses.
In the calculation of the RC-slab, some assumptions are employed to consider the effects of an
adjacent span and sectional shape of the members as follows. (1) The slab is modeled as a grid beam
with 305 mm intervals. (2) The boundary condition on four sides of the floor is a fixed support. (3) A
static pushover analysis is employed using uniformly distributed out-of-plane loads on the slab. (4) An
equivalent uniformly distributed load for a slab is determined based on the support reactions. (5) The
beams are modeled as the edge elements of the slab modeled. (6) An equivalent uniformly distributed
load for a beam is determined based on the end moments of the beam under the concentrated loads
applied at slab supports as a reaction. (7) An equivalent uniformly distributed load for a column is
calculated by using the sum of the shear forces at each end of the beam.
CHAPTER 4 LIVE LOADS - C4-35 -

Table 4.5.5 Probability distributions of equivalent uniformly distributed loads (N/m2)

The probability distributions, approved probability distribution types, and 99th percentile load
intensities are listed in Table 4.5.5. A Gumbel distribution type is adopted for a movable live load, and
most of the other loads have a Gamma distribution. The 99th percentile load has a tendency to decrease
with an increase the area of influence in the order of “floors,” “beams,” and “columns.” In the case of
containing unmovable live loads, the decreasing tendency is small because an unmovable live load is
often placed throughout all of the ALC panel surface. It can be concluded that the equivalent uniformly
distributed load for an ALC floor is almost the same as that for the RC floors when considering the
maximum data for a column as an extreme value. The effect of the human load reaches 100 N/m2.
Table 4.5.6 shows the calculated design live loads for each use considering the characteristics of the
rooftop live loads.
- C4-36 - Recommendations for Loads on Buildings

Table 4.5.6 Calculated design live loads for each use

floor frame seismic

movable
－－－
live loads

grass
unmovable
850 850 850
live loads
human loads 100 100 －
total 950 950 850
movable
－－－
live loads
bushy tree)
flowerbed
(flowers,
unmovable
1,250 1,250 1,250
live loads
human loads －－－
total 1,250 1,250 1,250
movable
(vegetable garden,
middle size tree )

－－－
live loads
flowerbed

unmovable
1,800 1,800 1,800
live loads
human loads －－－
total 1,800 1,800 1,800
movable
750 750 80
rooftop,others

live loads
Surface of

unmovable
300 300 300
live loads
human loads 100 100 －
total 1,150 1,150 380

Table 4.5.6 shows that the use of unmovable live loads is reflected in the design live loads. A shaded
cell indicates that live load intensity is smaller than the live loads regulated by the building standards of
Japan.

This chapter introduced the survey results for steel residential homes in an urban area. The examples
showed the determination of the design live loads for each type of use. Although the intended buildings
introduced herein are steel homes with ALC panels, the results for RC floors will be almost the same as
those of steel homes. If the soil used in a rooftop garden is ultra-lightweight-soil, the design live loads
for residential indoor areas are available for design live loads for a rooftop garden.

References
1) Tsuboi, Y.: Theory of Plates, Maruzen, 1960 (in Japanese).
2) Ishikawa, T., Hisagi, A.: A study on evaluation of live load, Journal of Structural Engineering,
Vol.38B, pp. 31-38 (in Japanese with English abstract), 1992.
3) Kinoshita, K., Kanda, J.: Equivalent uniformly distributed loads for office buildings, Summaries of
Technical Papers of Annual Meeting Architectural Institute of Japan, pp. 1023-1024 (in Japanese),
1984.
CHAPTER 4 LIVE LOADS - C4-37 -

4) ISO: ISO2103 --Loads due to use and occupancy in residential and public buildings, 1986
5) Idota, H. and Ono, T.: Study on live load of office buildings using measured data, summaries of
technical papers of annual meeting AIJ, pp. 163-164 (in Japanese), 1991.
6) Idota, H., Ono, T. and Hayakawa, Y.: Live load reduction of multiple-story column for office
buildings Part 2, Summaries of Technical Papers of Annual Meeting AIJ, pp. 215-216 (in Japanese),
1992.
7) Ishikawa, T., Hisagi, A.: A study on the effect of extraordinary live load on evaluation value,
Summaries of Technical Papers of Annual Meeting Architectural Institute of Japan, pp. 217-218 (in
Japanese), 1992.
8) Komori, S., Hayashi, M.: The effects of partial load on the fixed end slab being unevenly distributed:
The simple design of the fixed end slab that is supposed partial load, Proceeding of the 7th
Architectural Research Meetings (CHUGOKU and KYUSHU) Architectural Institute of Japan, pp.
129-136 (in Japanese), 1987.
9) Kanda, J., Kinoshita, K.: A probabilistic model for live load extremes in office buildings, Proc.4,
ICOSSAR, 1985.
10) Kanda, J., Yamamura, K.: Extraordinary live load model in retail premises, Proc. 5, ICOSSAR,
1989.
11) Hisagi, A., Ishikawa, T.: Evaluation of live load considering effects of unevenly distributed load
-Simulation assuming structural design for undaily situation, J. Struct. Constr. Eng., AIJ, No. 522,
August, 1999, pp. 21-27
12)Iwahara, S., Kakion, T.: Studies for evaluation of live load for crack on floor slabs of existing RC
buildings, AIJ Kyushu Chapter Architectural Research Meeting, No. 42, March 2003, pp.221-224.(in
Japanese).
13)Ishikawa, T., Hisagi, A.: Research on Strength in Some Limit States of the Floor Slabs in the Actual
Buildings, Journal of Structural Engineering, Vol. 41B, pp. 139-146 (in Japanese with English
abstract), 1995.
14) Yokoyama, Y.: Study on excitation apparatus and perception apparatus for evaluating floor vibration
caused by human walking, establishment of dynamic excitation apparatus and perception apparatus,
J. Struct. Constr. Eng., AIJ, No. 466, December, 1994, pp. 21-29
15) Yokoyama, Y. and Sato, M.: Study on excitation apparatus and perception apparatus for evaluating
floor vibration caused by human walking, development of impactive excitation apparatus and
verification of appropriateness of method to compute duration of vibration, J. Struct. Constr. Eng.,
AIJ, No. 476, October, 1995, pp. 21-30
16) Yokoyama, Y., Ito, K., Matsunaga, K. and Moritoki, H.: The relationship between moving velocity
and load caused by human walking and running, from a viewpoint of slipperiness and floor vibration,
Summaries of Technical Papers of Annual Meeting, Tokai branch, AIJ, No. 35, February, 1997, pp.
53-56
17) Yokoyama, Y. and Matsunaga, K.: Study on excitation apparatus for measurement of floor vibration
- C4-38 - Recommendations for Loads on Buildings

caused by human tripping and computation method of vibration damping time, J. Struct. Constr. Eng.,
AIJ, No. 519, May, 1999, pp. 13-20
18) Yokoyama, Y. and Sato, M.: Study on excitation apparatus and perception apparatus for evaluating
floor vibration caused by human walking, establishment of measurement method of vibration of floor
covered with finishing material, J. Struct. Constr. Eng., AIJ, No. 490, December, 1996, pp. 17-29
19) Ono, H. and Yokoyama, Y.: Study on vertical vibration of building floors occurred by human actions
and its indicating method from a viewpoint of human sense, in case of that the vibration source and
receiver are the same, J. Struct. Constr. Eng., AIJ, No. 381, November, 1987, pp. 1-9
20) Ono, H. and Yokoyama, Y.: Evaluation method for vertical vibration on building floors caused by
human activities from a viewpoint of comport, in case the same person causes and perceives the
vibration, J. Struct. Constr. Eng., AIJ, No. 394, December, 1988, pp. 8-16
21) Yokoyama, Y. and Ono, H.: Indicating methods of floor vibrations caused by human activities based
on human sensations, in the case of difference the vibration cause and the perceiver, J. Struct. Constr.
Eng., AIJ, No. 390, August, 1988, pp. 1-9
22) Yokoyama, Y. and Ono, H.: Presentation of the evaluation method for floor vibration when a
different person causes and perceives the vibration, study on method for evaluation vibrations of
building’s floors caused by human activities from a viewpoint of comport (part2), J. Struct. Constr.
Eng., AIJ, No. 418, December, 1990, pp. 1-8
23) Design Recommendations for composite constructions, Architectural Institute of Japan, 1985
24) AIJ Standard for structural calculation of reinforced concrete structures, Architectural Institute of
Japan, February, 1991
25) Yamahara, H.: Design of vibration isolation for preserving environment, Shokokusya, December,
1974
26) Mugikura, K.: Prediction of exciting force of equipment and floor impedance, Architectural
Acoustics and Noise Control, No. 101, March, 1998, pp. 37-43
27) Tano, M., Andou, K., Minemura, A. and Magikura, K.: Experimental study on the exciting force of
fans for building equipment, simplified methods for determining the exciting force of fans (Part 2),
J. Archit. Plann. Environ. Eng., AIJ, No. 427, September, 1991, pp. 49-55
28) Tano, M.: Estimation of vibration isolation, Architectural Acoustics and Noise Control, No. 101,
March, 1998, pp. 45-51
29) Koyama, T., Minagawa, T., Hanai, T., and Idota, H. “Design live loads 14 for rooftop gardens on
survey results of urban detached houses,” Transactions of AIJ Journal of 15 Structural and
Construction Engineering, Vol. 76, No. 660, pp. 255-261, February, 2011(in Japanese).
- C5-1 -

CHAPTER 5 SNOW LOADS

5.1 Scope and Procedure

Because the Japanese islands are surrounded by sea and are scattered over a rather wide range of
latitudes, snowfall mechanisms and the snow depth distributions vary from region to region. Hokkaido
and northwestern Honshu are covered with heavy snow during winter. These areas are subjected to the
snowfall mechanism shown in Fig. 5.1.1. However, areas on the Pacific Ocean side occasionally receive
heavy snow at the end of winter owing to approaching low pressures in the temperate zone.

Figure 5.1.1 Typical snowfall mechanism in Japan

In spite of the climate difference from region to region, and these two different snowfall mechanisms,
the present recommendation uses unified concepts as much as possible in determining snow loads on
roofs, as based on recent studies. A flowchart of the procedure used for determining the snow loads on
roofs is shown in Figs. 5.1.2 and 5.1.3. The main portion of the procedure is subdivided into three parts:
setting up the design concept on how to deal with roof snow loads, considering the snow conditions at
the construction site, and considering the snow conditions on a roof itself.
First, the designers should set up a design concept for a roof snow load, regardless of whether an
active control system is to be introduced to reduce the snow load accumulating on the roof. If an
appropriate system is introduced, the snow load design of the roof will be reduced. Because there have
been a number of accidents in which the control apparatuses of the buildings have failed, their reliability
needs to be critically assessed, while taking human error into account. Although recent developments of
control system technologies have been remarkable in Japan, this new recommendation is expected to
encourage their further development.
Snow loads on roofs vary as a function of the characteristic snow load on the ground, climate
(including temperature and wind speed during the winter), roof shape and roofing material, and the
differences in one winter to another. However, most available snow load data comprise the snow depth
on the ground. Therefore, this recommendation assumes that the snow load on the roof is proportional to
that on the ground. As a second step, the snow depth on the ground at the construction site is then
estimated through a statistical treatment and using an interpolation technique. The basic load value in
these recommendations is defined as that for a snow depth on the ground with a 100-year recurrence
interval. If a building is designed for another recurrence interval, a conversion factor is introduced to
- C5-2 - Commentary on Recommendations for Loads on Buildings

modify the value. The snow load on the ground is calculated from the snow depth multiplied by an
equivalent unit weight for the ground snow.
As the third step, the snow load on a roof is estimated from the snow load on the ground multiplied by
a ground-to-roof conversion factor such as a shape coefficient. Because the shape coefficient varies
according the roof shape, temperature, snow type (dry or wet), wind direction, and wind speed, it is
impossible to determine precisely. Therefore, it is recommended in principle for the shape coefficient
for the roof to be estimated through an experimental study using models. However, the tables in this
recommendation provide values for several typical shapes. For a large roof, even if its shape is simple,
an experimental study is also recommended because the shape coefficient is usually different from the
coefficient of small roofs.
CHAPTER 5 SNOW LOADS - C5-3 -

Figure 5.1.2 Flowchart of procedure used for determining snow load on roof
(without roof snow control)
- C5-4 - Commentary on Recommendations for Loads on Buildings

Figure 5.1.3 Flowchart of procedure for determining snow load on roof

(in the case of active roof snow control)
CHAPTER 5 SNOW LOADS - C5-5 -

5.1.1 Basic Snow Load on a Roof

In general, the snow load on a roof cannot be determined directly in the design. In these
recommendations, such load is estimated from the ground snow load. The snow load on a roof is usually
less than that on the ground because of wind, sunshine, and other factors. In addition, the roof shape and
slope influence the ratio of roof snow to ground snow. Therefore, the shape coefficients 𝜇𝜇0 and 𝜇𝜇n are
introduced to take into account the meteorological conditions. A partial snow load such as snowdrift
caused by projecting structures, snow eaves, and snow sliding from the upper roofs at the eaves or the
lower levels of multilevel roofs is considered as described in 5.4.
The basic equations of a snow load on a roof differ in the design depending on whether the snow on
the roof is controlled. If the roof snow is not controlled, the snow load on roof 𝑆𝑆 in the design is given
based on the characteristic snow load on the ground 𝑆𝑆0 and the shape coefficient 𝜇𝜇0 :

𝑆𝑆 = 𝜇𝜇0 𝑆𝑆0. (5.1.1)

5.2 Snow Load on the Ground

5.2.1 Equations for snow load on the ground
(1) Estimation of the ground snow weights based on the snow depth and equivalent unit weight
Because direct observations of snow loads on roofs are limited, characteristic snow loads on the
ground are used as characteristic values. For ordinary buildings without control systems, the annual
maximum load for the entire season is used as the characteristic snow load S0, which is given by Eq.
(5.2.1) as a product of the factors:

𝑆𝑆0 = 𝑘𝑘env 𝑑𝑑0 𝜌𝜌0 , (5.2.1)

where
𝑆𝑆0 (kN/m2) : characteristic snow load on the ground used for the design when the snow load on the
roof is not controlled,
𝑘𝑘env : environmental coefficient, as defined in 5.2.4,
𝑑𝑑0 (m): basic snow depth on the ground when the snow load on the roof is not controlled, as
defined in 5.2.2, and
𝜌𝜌0 (kN/m3): equivalent unit weight for ground snow, as defined in 5.2.3.

(2) Estimation of ground snow weights based on data of daily precipitation and daily mean air
temperature
To estimate the ground snow weights, in addition to a conventional method using the equivalent unit
weight for ground snow, as described in 5.2.3, this section presents the methods based on the
meteorological data such as the daily precipitations and daily mean air temperatures. Three types of
equations were proposed by Joh & Sakurai, Kamimura & Uemura, and Takahashi et al. as follows:
1) Joh & Sakurai’s equation1)
- C5-6 - Commentary on Recommendations for Loads on Buildings

They proposed a practical method to estimate the ground snow weights during a snow cover period,
as defined below:
𝑛𝑛−1

𝑃𝑃n = � 𝑃𝑃i �𝑇𝑇i < 2℃� (5.2.2)

𝑖𝑖=1

where 𝑃𝑃i (N/m2) is the daily snowfall weight converted from the daily precipitation (mm), and 𝑃𝑃n (N/m2)
is the accumulated value of 𝑃𝑃i from the first day (𝑖𝑖=1) of a period of continuous snow cover to the
(𝑛𝑛 − 1)th day, which is regarded as the ground snow weight of the nth day. Data on the Pi values are
summed only when the daily mean air temperature 𝑇𝑇i (℃) is less than the air temperature limit, namely,
+2℃, which is used for a distinction between rain and snow. Here, the daily melting snow weight is
neglected for practical application.
Figure. 5.2.1(a)–(d) show examples of the comparison of the estimated values through Eq. (5.2.2)
with the observed snow weights at cities located in heavy snow regions, Sapporo, Kamabuchi, Nagaoka,
and Tokamachi, respectively. In these four cities, the continuous snow cover periods are almost 3
months in Sapporo, and 5 months in the other three observed regions. It is clear from these examples that
the proposed equation can approximately reproduce the increasing process of the observed ground snow
weights from the first day to the day with the peak weight value.
2) Kamimura & Uemura’s equation2)
This equation is expressed as follows:

𝑆𝑆m = 𝑆𝑆m−1 + 𝐺𝐺b 𝑃𝑃m − 𝐶𝐶a 𝑇𝑇m (5.2.3)

⎧ 0 �𝑇𝑇 < 0℃�

⎪
𝐶𝐶a = 𝐶𝐶a1 �0℃ ≤ 𝑇𝑇 and 𝐶𝐶b 𝑃𝑃m > 0�
⎨
⎪𝐶𝐶 �0℃ ≤ 𝑇𝑇 and 𝐶𝐶b 𝑃𝑃m = 0�
⎩ a2

1 (𝑇𝑇 ≤ 𝑇𝑇l )
𝐶𝐶b = �(𝑇𝑇u − 𝑇𝑇)⁄(𝑇𝑇u − 𝑇𝑇l ) (𝑇𝑇l < 𝑇𝑇 ≤ 𝑇𝑇u )
0 (𝑇𝑇u < 𝑇𝑇)

where 𝑆𝑆m (N/m2) indicates the estimated ground snow weight on the mth day after the starting day of the
calculation; 𝑃𝑃m (N/m2) is the snowfall weight on the 𝑚𝑚th day, which is converted from the daily
precipitation (mm); 𝑇𝑇m (℃) is daily mean air temperature on the 𝑚𝑚th day; and 𝐶𝐶a (N/m2℃) is the
snowmelt coefficient. In addition, 𝐶𝐶b is the reference coefficient used to judge whether the data on 𝐶𝐶b
indicates the value of snowfall or rain, which depends on the upper limit temperature 𝑇𝑇u and lowest limit
temperature 𝑇𝑇l . In short, 𝐶𝐶b·𝑃𝑃m and 𝐶𝐶a ·𝑇𝑇m in Eq. (5.2.3) indicate the increasing snow weight and the
melting snow weight on the 𝑚𝑚th day, respectively
Figures 5.2.2(a) and (b) show the calculation examples in Nagaoka and Tokamachi using Eq. (5.2.3),
in which the estimated values are compared with the observed values for the ground snow weights. It is
clearly shown that this equation can approximately reproduce the ground snow weights over the
CHAPTER 5 SNOW LOADS - C5-7 -

continuous snow cover period.

The calculation was applied using the different and optimized values of snowmelt coefficient 𝐶𝐶a at
each observed region, which can be obtained from the observed data of the snow weights and other
factors. Therefore, further investigation is needed to determine how to optimize the suitable snowmelt
coefficient 𝐶𝐶a in other observed regions that have no observed data on the snow weight.

Figure 5.2.1 Estimation of ground snow weight using Eq. (5.2.2)1)

Figure 5.2.2 Estimation of ground snow weight using Eq. (5.2.3)2)

- C5-8 - Commentary on Recommendations for Loads on Buildings

3) Takahashi’s equation3)
This equation assumes that snowfall each day forms one layer of new snow on accumulated ground
snow. After a snowfall, the snow consolidates and metamorphoses in proportion to the number of days
and air temperature. Finally, the snow melts and percolates into the ground. In Eq. (5.2.4), numerator
indicates the estimated snow weight of the 𝑚𝑚th snow layer, and the denominator indicates the estimated
snow density of the 𝑚𝑚th snow layer.

m𝑃𝑃 − m𝑃𝑃c
m𝑑𝑑n = (5.2.4)
𝜌𝜌min �𝑛𝑛 + m𝑘𝑘

𝑃𝑃c = � m𝑃𝑃C = 𝐶𝐶(𝑇𝑇 − 𝑇𝑇0 ) (𝑇𝑇 > 𝑇𝑇0 )

⋯� 𝑚𝑚
𝑃𝑃c = 0 (𝑇𝑇 ≤ 𝑇𝑇0 )

m𝜌𝜌0 = 𝜌𝜌min �𝑇𝑇 ≤ 0℃�

⋯ � m𝜌𝜌0 = 𝜌𝜌min + m𝑇𝑇(𝜌𝜌max − 𝜌𝜌min )⁄3 0℃ < T ≤ 3℃
m𝜌𝜌0 = 𝜌𝜌max 3℃ < T

where 𝑚𝑚 is the number of days from the first day of continuous snow cover, 𝑛𝑛 is the number of days
from the snowfall of the layer, m𝑑𝑑n (m) is the estimated depth of the 𝑚𝑚th snow layer, m𝑃𝑃 is the snowfall
weight on the 𝑚𝑚th day converted from the daily precipitation (mm), 𝑃𝑃c is the weight of melting snow on
the 𝑚𝑚th day, 𝐶𝐶(N/m2℃) is the snowmelt coefficient, 𝑇𝑇 is the daily mean air temperature, 𝑇𝑇0 is the
reference temperature for snowmelt fixed as -2℃, m𝑘𝑘 is the correction coefficient for the initial
condition, 𝜌𝜌0 is the unit weight of snowfall on the 𝑚𝑚th day, 𝜌𝜌min is the minimum unit weight of a
snowfall, and 𝜌𝜌max is maximum unit weight of a snowfall fixed at 4.9 kN/m3. This equation needs
repeated calculations with a one-day step using the minimum unit weight of snow 𝜌𝜌min and the
snowmelt coefficient 𝐶𝐶 as unknown factors. That is, the error between the estimated daily snow depth
and the observed daily snow depth are minimized, leading to the optimal solution for the transition
process of the snow depth.
Figures 5.2.3(a)–(f) show the calculation examples at Sapporo, Shinjo, Takada, Toyama, and Fukui
using Eq. (5.2.4), in which the estimated values are compared with the observed values for the ground
snow weights. It is clearly shown that this equation can approximately reproduce the observed ground
snow weights over a continuous snow cover period.
CHAPTER 5 SNOW LOADS - C5-9 -

Figure 5.2.3 Estimation of weight and depth of ground snow using Eq. (5.2.4)3)

As a result, considering the difficulty in determining a suitable value regarding the regional
characteristics of the snowmelt coefficient in advance, methods using Eqs .(5.2.2) and (5.2.4) are
recommended at present to estimate the snow weight on the ground when only the data on the daily
precipitation and daily mean air temperature are available.

5.2.2 Basic snow depth on the ground

For a roof snow load without any control systems, the basic snow depth on the ground d0 is defined as
the annual maximum value for the entire season with a 100-year recurrence interval. The statistical
properties of snow depths at 423 meteorological observatories in Japan were studied to estimate the
values with a long recurrence interval4). Five types of probability distribution functions were applied
using two plotting techniques: Hazen and Thomas. The statistical properties in Japan were typically
subdivided into three groups, as shown in Fig. 5.2.4. It was concluded that the statistical properties of
the annual maximum snow depth are unable to be described using a single distribution or plotting
technique. As a result, a method is proposed in which a linear regression analysis is applied only to a
maximum of 1/3 of all data plotted on Gumbel probability paper.
- C5-10 - Commentary on Recommendations for Loads on Buildings

Figure 5.2.4 Type classification of statistical properties of snow depth on Gumbel probability paper4)

In this recommendation, statistical values with a 100-year recurrence interval were estimated through
the proposed methods using the annual maximum snow depth for the entire season, and for the
maximum annual accumulated snow depth for 7 days. Because the Gumbel distribution is given by Eq.
(5.2.5), the snow depth 𝑑𝑑0 on the ground with a recurrence interval of 𝑡𝑡R years (0 < 𝑡𝑡R < 200) can be
estimated using Eq. (5.2.6).

𝐹𝐹X (𝑥𝑥) = exp[−exp{−𝑎𝑎(𝑥𝑥 − 𝑏𝑏)}] −∞<x<∞ (5.2.5)

where
a: scale parameter,
b: position parameter.

𝑑𝑑 = 𝑏𝑏 + 1⁄𝑎𝑎 ln(𝑟𝑟) (5.2.6)

In general, a construction site differs from the weather station of the Meteorological Agency.
Therefore, this recommendation provides an empirical formula for estimating the snow depth at a
location without any observatories. According to previous studies5), the altitude and sea ratio, which is
defined as the ratio of sea area to total area around the site (see Fig. 5.2.5), are selected as the dominant
topographic factors influencing the snow depth on the ground. With this recommendation, the annual
maximum value of snow depth on the ground is estimated for a given point by
CHAPTER 5 SNOW LOADS - C5-11 -

𝑑𝑑 = �𝛼𝛼･altitude� + �𝛽𝛽･sea ratio� + 𝛾𝛾, (5.2.7)

where coefficients 𝛼𝛼 for altitude, 𝛽𝛽 for the sea ratio, and constant 𝛾𝛾 are given in tables in the original
Japanese version.

Nishikawa-cho (Yamagata Prefecture)

Figure 5.2.5 Definition of sea ratio5) Figure 5.2.6 Relation between snow depth and
altitude of observatories5)

Figure 5.2.7 Effectiveness of snow depth estimation based on two parameters5)

The snow depth is almost proportional to the altitude when the site is inland and in a narrowly
confined area (Fig. 5.2.6). In general, however, the altitude alone is insufficient as an interpolating factor
(Fig. 5.2.7 (a)). On the other hand, the values estimated using Eq. (5.2.7) nearly coincide with the
observed data, as shown in Fig. 5.2.7 (b).
The basic snow depth and characteristic ground snow weight for a 100-year recurrence interval at 142
meteorological offices are shown in Table 5.2.1.
- C5-12 - Commentary on Recommendations for Loads on Buildings

Table 5.2.1 (a) Annual maximum snow depth and snow weight (return period: 100 years)
Annual maximum snow depth on Annual maximum snow weight on Annual maximum snow depth on Annual maximum snow weight on
ground ground ground with accumulation for 7 days ground with accumulation for 7 days
Code Station n r=100 1/a b μda σda r=100 1/a b μda σda r=100 1/a b μda σda r=100 1/a b μda σda
(m) (m) (m) (kN/m2) 2 2
(kN/m ) (kN/m ) (m) (m) (m) (kN/m2) (kN/m2) (kN/m2)
401 Wakkanai 53 1.76 0.233 0.68 0.87 0.29 5.06 0.393 3.25 3.07 0.89 0.77 0.060 0.49 0.43 0.15 1.46 0.194 0.57 0.73 0.21
402 Kitamiesashi 53 1.90 0.183 1.06 1.13 0.26 6.96 0.764 3.44 3.80 1.06 0.73 0.075 0.39 0.43 0.10 1.39 0.138 0.76 0.78 0.23
404 Haboro 53 1.77 0.178 0.95 1.08 0.22 6.92 0.682 3.78 3.97 1.05 0.95 0.130 0.35 0.45 0.14 1.63 0.212 0.65 0.82 0.23
405 Oumu 53 1.31 0.111 0.79 0.76 0.23 3.84 0.420 1.90 1.99 0.67 0.60 0.058 0.33 0.33 0.10 1.20 0.129 0.61 0.58 0.24
406 Rumoi 53 1.82 0.174 1.02 1.00 0.32 5.88 0.488 3.64 3.42 1.04 0.85 0.099 0.40 0.45 0.13 1.35 0.127 0.77 0.77 0.22
407 Asahikawa 53 1.40 0.098 0.95 0.93 0.20 5.18 0.522 2.77 2.91 0.84 0.65 0.064 0.36 0.39 0.09 0.93 0.084 0.55 0.57 0.13
409 Abashiri 53 1.33 0.183 0.49 0.60 0.23 3.61 0.385 1.85 1.79 0.69 0.75 0.119 0.20 0.30 0.13 1.53 0.203 0.60 0.65 0.30
411 Otaru 53 1.72 0.103 1.24 1.17 0.24 7.25 0.647 4.27 4.33 1.12 0.73 0.074 0.39 0.42 0.10 1.51 0.158 0.78 0.88 0.20
412 Sapporo 53 1.50 0.103 1.02 1.00 0.20 5.45 0.519 3.06 3.14 0.84 0.87 0.094 0.43 0.47 0.13 1.40 0.152 0.70 0.75 0.22
413 Iwamizawa 53 2.01 0.170 1.23 1.24 0.30 7.58 0.819 3.82 4.05 1.26 1.05 0.114 0.53 0.59 0.15 1.72 0.215 0.73 0.90 0.25
417 Obihiro 53 1.62 0.250 0.47 0.65 0.30 4.20 0.731 0.84 1.41 0.81 1.14 0.180 0.31 0.43 0.21 2.16 0.404 0.30 0.65 0.42
418 Kushiro 53 0.86 0.122 0.30 0.37 0.15 2.25 0.349 0.64 0.86 0.43 0.71 0.106 0.22 0.30 0.13 1.29 0.191 0.42 0.53 0.23
420 Nemuro 53 0.95 0.163 0.19 0.32 0.18 2.35 0.401 0.50 0.77 0.47 0.65 0.102 0.19 0.24 0.13 1.22 0.190 0.34 0.46 0.23
421 Suttsu 53 1.25 0.095 0.82 0.76 0.22 4.99 0.523 2.58 2.57 0.92 0.79 0.098 0.34 0.37 0.14 1.40 0.164 0.64 0.70 0.23
423 Muroran 53 0.58 0.066 0.28 0.28 0.12 2.06 0.318 0.59 0.78 0.40 0.47 0.061 0.18 0.21 0.08 1.24 0.215 0.25 0.47 0.21
424 Tomakomai 53 0.78 0.123 0.22 0.30 0.14 1.85 0.278 0.57 0.69 0.38 0.58 0.097 0.14 0.23 0.10 1.12 0.182 0.29 0.41 0.21
426 Urakawa 53 0.52 0.076 0.16 0.21 0.10 1.13 0.185 0.28 0.39 0.22 0.45 0.074 0.11 0.17 0.08 0.78 0.148 0.10 0.27 0.13
428 Esashi 53 0.90 0.138 0.26 0.35 0.17 3.07 0.572 0.44 0.99 0.57 0.56 0.080 0.19 0.23 0.10 1.28 0.220 0.27 0.49 0.21
430 Hakodate 53 1.00 0.146 0.33 0.45 0.16 2.80 0.341 1.23 1.25 0.55 0.49 0.055 0.24 0.27 0.07 1.03 0.142 0.38 0.49 0.16
433 Kucchan 53 2.90 0.217 1.90 1.94 0.35 12.21 1.250 6.46 7.07 1.74 0.98 0.084 0.60 0.62 0.12 1.71 0.130 1.11 1.12 0.22
435 Monbetsu 53 1.25 0.137 0.62 0.66 0.21 3.25 0.297 1.88 1.76 0.61 0.85 0.128 0.26 0.34 0.15 1.62 0.270 0.37 0.58 0.30
440 Hiroo 53 1.90 0.210 0.93 0.94 0.36 6.70 1.119 1.55 2.48 1.25 1.02 0.094 0.58 0.54 0.19 3.10 0.600 0.34 0.98 0.55
512 Ohfunato 51 0.32 0.049 0.10 0.12 0.06 0.86 0.172 0.07 0.22 0.18 0.28 0.037 0.11 0.12 0.06 0.61 0.094 0.18 0.20 0.15
520 Shinjo 53 2.44 0.230 1.38 1.34 0.44 9.69 1.126 4.51 4.68 1.81 1.24 0.155 0.53 0.65 0.17 1.86 0.149 1.18 1.15 0.28
570 Wakamatsu 53 1.24 0.143 0.59 0.63 0.22 3.87 0.580 1.20 1.44 0.80 1.11 0.148 0.43 0.51 0.19 1.61 0.228 0.57 0.69 0.29
574 Fukaura 53 0.97 0.106 0.48 0.45 0.20 3.41 0.480 1.21 1.29 0.73 0.48 0.049 0.25 0.27 0.08 1.25 0.197 0.34 0.54 0.20
575 Aomori 53 2.01 0.172 1.22 1.14 0.38 7.52 0.822 3.74 3.66 1.50 0.87 0.052 0.63 0.57 0.15 1.68 0.173 0.89 0.93 0.27
576 Mutsu 53 1.67 0.260 0.48 0.69 0.29 4.42 0.613 1.61 1.85 0.84 0.96 0.150 0.27 0.40 0.16 1.46 0.190 0.58 0.69 0.24
581 Hachinohe 53 0.96 0.157 0.24 0.32 0.20 2.30 0.424 0.34 0.68 0.45 0.78 0.136 0.15 0.26 0.15 1.14 0.148 0.46 0.48 0.23
582 Akita 53 1.06 0.169 0.28 0.42 0.19 3.66 0.659 0.63 1.21 0.69 0.60 0.081 0.23 0.28 0.10 1.13 0.146 0.46 0.55 0.18
584 Morioka 53 0.79 0.092 0.37 0.39 0.14 2.29 0.345 0.70 0.89 0.44 0.58 0.071 0.25 0.28 0.10 0.94 0.120 0.39 0.45 0.16
585 Miyako 53 0.86 0.139 0.23 0.31 0.18 2.25 0.422 0.31 0.69 0.44 0.76 0.121 0.20 0.28 0.15 1.78 0.292 0.44 0.62 0.36
587 Sakata 53 0.84 0.115 0.31 0.36 0.16 2.69 0.406 0.82 1.01 0.54 0.56 0.053 0.32 0.30 0.11 1.49 0.200 0.57 0.66 0.28
588 Ymagata 53 1.07 0.125 0.49 0.51 0.20 3.16 0.439 1.14 1.26 0.64 0.66 0.083 0.28 0.34 0.10 1.10 0.132 0.49 0.55 0.18
590 Sendai 53 0.38 0.041 0.19 0.18 0.08 0.86 0.113 0.34 0.35 0.20 0.38 0.045 0.17 0.17 0.08 0.78 0.095 0.34 0.33 0.18
592 Ishinomaki 53 0.37 0.041 0.18 0.17 0.08 1.13 0.210 0.16 0.32 0.25 0.37 0.045 0.16 0.16 0.08 1.06 0.197 0.15 0.30 0.23
595 Fukushima 53 0.59 0.082 0.21 0.25 0.11 1.27 0.218 0.27 0.43 0.24 0.55 0.078 0.19 0.23 0.10 0.93 0.136 0.31 0.38 0.18
597 Shirakawa 53 0.71 0.138 0.07 0.22 0.13 1.38 0.296 0.02 0.36 0.25 0.51 0.077 0.15 0.20 0.10 0.94 0.178 0.12 0.31 0.18
598 Onahama 47 0.23 0.040 0.05 0.06 0.06 0.63 0.136 0.01 0.10 0.15 0.23 0.040 0.05 0.06 0.06 0.61 0.130 0.02 0.10 0.15
600 Wajima 53 0.84 0.119 0.29 0.33 0.18 2.90 0.502 0.59 0.90 0.61 0.67 0.095 0.23 0.28 0.14 1.63 0.239 0.53 0.66 0.33
602 Aikawa 53 0.61 0.097 0.16 0.21 0.13 1.49 0.300 0.11 0.38 0.31 0.58 0.098 0.12 0.19 0.12 1.12 0.204 0.18 0.33 0.24
604 Niigata 53 0.95 0.119 0.40 0.38 0.21 3.19 0.525 0.77 1.06 0.67 0.91 0.146 0.24 0.33 0.18 1.95 0.300 0.57 0.78 0.37
605 Kanazawa 53 1.72 0.326 0.22 0.52 0.33 6.70 1.485 -0.13 1.50 1.29 1.14 0.177 0.32 0.43 0.22 2.77 0.478 0.58 0.99 0.50
606 Fushiki 53 2.02 0.382 0.26 0.66 0.36 6.72 1.303 0.73 1.89 1.31 1.12 0.153 0.42 0.52 0.19 2.90 0.468 0.74 1.20 0.48
607 Toyama 53 1.88 0.303 0.48 0.67 0.37 6.51 1.159 1.17 1.91 1.36 1.16 0.153 0.46 0.52 0.21 2.59 0.391 0.79 1.12 0.45
610 Nagano 53 0.71 0.090 0.30 0.32 0.13 1.80 0.344 0.22 0.53 0.34 0.58 0.074 0.24 0.27 0.10 0.81 0.108 0.31 0.35 0.15
612 Takata 53 3.50 0.508 1.16 1.35 0.73 13.90 2.239 3.60 4.81 2.88 1.90 0.246 0.77 0.83 0.37 4.02 0.459 1.91 1.97 0.75
615 Uysunomiya 53 0.33 0.052 0.09 0.10 0.08 0.68 0.135 0.06 0.16 0.16 0.29 0.040 0.11 0.10 0.07 0.59 0.108 0.10 0.15 0.14
616 Fukui 53 2.15 0.416 0.23 0.62 0.41 7.54 1.520 0.55 1.98 1.48 1.26 0.193 0.37 0.50 0.24 2.79 0.379 1.04 1.24 0.51
617 Takayama 53 1.18 0.152 0.48 0.53 0.22 3.27 0.479 1.07 1.18 0.70 0.95 0.135 0.33 0.42 0.16 1.16 0.130 0.56 0.59 0.22
618 Matsumoto 53 0.77 0.142 0.12 0.26 0.14 1.19 0.227 0.15 0.36 0.22 0.57 0.084 0.19 0.24 0.10 0.87 0.136 0.25 0.33 0.17
620 Suwa 52 0.63 0.111 0.12 0.22 0.12 0.98 0.158 0.25 0.35 0.20 0.43 0.053 0.19 0.20 0.08 0.72 0.097 0.27 0.30 0.15
622 Karuizawa 50 0.96 0.159 0.23 0.34 0.18 1.73 0.274 0.47 0.60 0.36 0.70 0.108 0.21 0.29 0.12 0.98 0.148 0.30 0.38 0.19
624 Maebashi 53 0.59 0.139 -0.05 0.11 0.12 1.06 0.250 -0.09 0.19 0.22 0.41 0.082 0.04 0.10 0.09 0.95 0.216 -0.04 0.19 0.20
626 Kumagaya 53 0.51 0.116 -0.03 0.09 0.11 1.11 0.271 -0.14 0.18 0.24 0.40 0.082 0.02 0.09 0.09 0.87 0.193 -0.02 0.17 0.19
629 Mito 53 0.28 0.045 0.07 0.08 0.07 0.60 0.119 0.05 0.12 0.14 0.28 0.045 0.07 0.08 0.07 0.60 0.122 0.04 0.12 0.14
631 Suruga 53 1.87 0.327 0.37 0.56 0.40 6.55 1.299 0.57 1.63 1.37 1.14 0.126 0.56 0.48 0.28 2.90 0.312 1.46 1.23 0.72
632 Gifu 53 0.46 0.068 0.14 0.16 0.10 0.55 0.076 0.20 0.21 0.13 0.46 0.070 0.14 0.16 0.10 0.55 0.077 0.20 0.21 0.13
636 Nagoya 53 0.25 0.043 0.05 0.07 0.06 0.38 0.068 0.07 0.09 0.10 0.25 0.042 0.05 0.07 0.06 0.38 0.068 0.07 0.09 0.10
637 Iida 53 0.69 0.145 0.02 0.19 0.13 1.09 0.226 0.05 0.29 0.22 0.45 0.070 0.13 0.18 0.09 0.81 0.145 0.14 0.26 0.16
638 Kofu 53 0.87 0.214 -0.11 0.16 0.17 1.11 0.250 -0.04 0.25 0.23 0.68 0.154 -0.03 0.15 0.14 0.88 0.175 0.08 0.24 0.19
640 Kawaguchiko 53 1.27 0.252 0.11 0.36 0.24 2.11 0.394 0.30 0.62 0.42 1.02 0.179 0.19 0.33 0.19 1.67 0.308 0.26 0.52 0.32
641 Chichibu 53 0.83 0.184 -0.02 0.19 0.16 1.34 0.249 0.20 0.36 0.29 0.58 0.106 0.09 0.18 0.12 1.12 0.176 0.31 0.35 0.26
648 Choshi 53 0.11 0.031 -0.03 0.01 0.02 0.26 0.079 -0.10 0.02 0.05 0.11 0.031 -0.03 0.01 0.02 0.26 0.079 -0.10 0.02 0.05
649 Ueno 36 0.25 0.039 0.07 0.08 0.06 0.57 0.124 0.01 0.11 0.13 0.25 0.037 0.08 0.08 0.06 0.57 0.124 0.01 0.11 0.13
651 Tsu 53 0.16 0.029 0.02 0.04 0.04 0.44 0.104 -0.04 0.06 0.10 0.16 0.029 0.02 0.04 0.04 0.44 0.104 -0.04 0.06 0.10
653 Irako 37 0.12 0.033 -0.03 0.01 0.03 0.21 0.065 -0.08 0.02 0.04 0.12 0.033 -0.03 0.01 0.03 0.21 0.065 -0.08 0.02 0.04
654 Hamamatsu 45 0.04 0.011 -0.01 0.00 0.01 0.07 0.023 -0.04 0.00 0.02 0.04 0.011 -0.01 0.00 0.01 0.07 0.023 -0.04 0.00 0.02
655 Omaezaki 49 0.02 0.007 -0.01 0.00 0.01 0.03 0.009 -0.02 0.00 0.01 0.02 0.007 -0.01 0.00 0.01 0.03 0.009 -0.02 0.00 0.01
656 Shizuoka 53 0.04 0.011 -0.01 0.00 0.01 0.07 0.022 -0.03 0.00 0.02 0.04 0.011 -0.01 0.00 0.01 0.07 0.022 -0.03 0.00 0.02
657 Mishima 41 0.10 0.022 0.00 0.02 0.03 0.30 0.084 -0.09 0.03 0.06 0.10 0.022 0.00 0.02 0.03 0.30 0.083 -0.09 0.03 0.06
662 Tokyo 53 0.34 0.064 0.04 0.07 0.08 0.89 0.199 -0.03 0.15 0.20 0.30 0.052 0.06 0.07 0.08 0.74 0.153 0.04 0.15 0.17
663 Owase 46 0.05 0.015 -0.02 0.00 0.01 0.04 0.012 -0.02 0.00 0.01 0.05 0.015 -0.02 0.00 0.01 0.04 0.012 -0.02 0.00 0.01
666 Irouzaki 43 0.07 0.022 -0.03 0.00 0.01 0.06 0.019 -0.03 0.00 0.01 0.07 0.022 -0.03 0.00 0.01 0.06 0.019 -0.03 0.00 0.01
668 Ajiro 43 0.11 0.031 -0.03 0.01 0.02 0.24 0.066 -0.06 0.02 0.05 0.11 0.031 -0.03 0.01 0.02 0.24 0.066 -0.06 0.02 0.05
670 Yokohama 53 0.39 0.073 0.06 0.08 0.10 0.84 0.177 0.02 0.16 0.19 0.38 0.071 0.05 0.08 0.09 0.71 0.139 0.07 0.15 0.17
672 Tateyama 38 0.10 0.025 -0.02 0.01 0.02 0.16 0.052 -0.07 0.01 0.03 0.10 0.025 -0.02 0.01 0.02 0.15 0.048 -0.07 0.01 0.03
674 Katsuura 46 0.27 0.081 -0.10 0.02 0.06 0.32 0.092 -0.10 0.03 0.07 0.27 0.081 -0.10 0.02 0.06 0.32 0.092 -0.10 0.03 0.07
675 Ohshima 48 0.26 0.070 -0.07 0.03 0.05 0.90 0.262 -0.31 0.07 0.18 0.26 0.070 -0.07 0.03 0.05 0.90 0.262 -0.31 0.07 0.18
677 Miyakejima 44 0.03 0.011 -0.02 0.00 0.01 0.00 0.000 0.00 0.00 0.00 0.03 0.011 -0.02 0.00 0.01 0.00 0.000 0.00 0.00 0.00
678 Hachijojima 48 0.03 0.010 -0.01 0.00 0.01 0.12 0.041 -0.06 0.01 0.03 0.03 0.010 -0.01 0.00 0.01 0.12 0.041 -0.06 0.01 0.03
682 Chiba 48 0.34 0.076 -0.01 0.06 0.08 0.83 0.193 -0.06 0.12 0.18 0.34 0.076 -0.01 0.06 0.08 0.65 0.137 0.02 0.11 0.16
684 Yokkaichi 31 0.52 0.126 -0.06 0.09 0.10 0.68 0.155 -0.04 0.11 0.15 0.53 0.127 -0.06 0.09 0.10 0.68 0.155 -0.04 0.11 0.15
CHAPTER 5 SNOW LOADS - C5-13 -

Table 5.2.1 (b) Annual maximum snow depth and snow weight (return period: 100 years)
Annual maximum snow depth on Annual maximum snow weight on Annual maximum snow depth on Annual maximum snow weight on
ground ground ground with accumulation for 7 days ground with accumulation for 7 days
Code Station n r=100 1/a b μda σda r=100 1/a b μda σda r=100 1/a b μda σda r=100 1/a b μda σda
(m) (m) (m) (kN/m2) (kN/m2) (kN/m2) (m) (m) (m) (kN/m2) (kN/m2) (kN/m2)
690 Okunikko 53 1.16 0.172 0.37 0.47 0.22 3.46 0.550 0.93 1.16 0.73 0.77 0.101 0.31 0.35 0.14 2.26 0.425 0.31 0.72 0.42
740 Saigo 53 1.09 0.226 0.05 0.29 0.20 2.80 0.611 -0.01 0.68 0.53 0.91 0.177 0.10 0.27 0.18 1.77 0.321 0.29 0.57 0.34
741 Matsue 53 0.95 0.205 0.01 0.25 0.18 2.82 0.660 -0.21 0.60 0.54 0.80 0.165 0.04 0.23 0.15 1.74 0.323 0.25 0.53 0.34
742 Sakai 53 0.95 0.190 0.08 0.27 0.18 3.25 0.745 -0.18 0.70 0.62 0.74 0.130 0.14 0.25 0.14 1.97 0.358 0.32 0.60 0.39
744 Yonago 53 0.88 0.165 0.12 0.29 0.16 2.44 0.507 0.11 0.66 0.45 0.80 0.144 0.13 0.28 0.14 1.75 0.294 0.40 0.61 0.33
746 Tottori 53 1.09 0.143 0.43 0.45 0.23 3.04 0.456 0.94 1.08 0.65 0.98 0.130 0.38 0.41 0.20 1.81 0.165 1.05 0.89 0.41
747 Toyooka 53 1.39 0.195 0.49 0.56 0.28 4.51 0.689 1.34 1.59 0.97 0.89 0.096 0.45 0.45 0.18 1.83 0.146 1.16 1.04 0.38
750 Maizuru 53 0.92 0.142 0.27 0.36 0.18 2.49 0.476 0.29 0.74 0.49 0.68 0.072 0.35 0.33 0.14 1.76 0.292 0.42 0.63 0.34
751 Ibukiyama 28 8.77 0.752 5.31 5.26 1.81 38.08 3.837 20.43 20.89 7.97 7.67 1.048 2.85 3.51 1.44 32.24 5.040 9.05 12.92 6.14
754 Hagi 42 0.28 0.060 0.00 0.06 0.06 1.08 0.273 -0.17 0.15 0.23 0.28 0.060 0.00 0.06 0.06 0.82 0.186 -0.04 0.13 0.18
755 Hamada 48 0.50 0.123 -0.07 0.08 0.10 1.30 0.321 -0.17 0.20 0.27 0.49 0.121 -0.07 0.08 0.10 1.00 0.222 -0.02 0.18 0.22
756 Tsuyama 53 0.41 0.060 0.14 0.16 0.08 0.68 0.128 0.09 0.21 0.14 0.41 0.060 0.14 0.16 0.08 0.65 0.118 0.11 0.21 0.13
759 Kyoto 53 0.17 0.022 0.07 0.06 0.04 0.28 0.052 0.04 0.06 0.07 0.17 0.022 0.07 0.06 0.04 0.28 0.052 0.04 0.06 0.07
761 Hikone 53 0.72 0.101 0.25 0.29 0.15 1.49 0.236 0.40 0.51 0.30 0.64 0.081 0.26 0.28 0.13 1.26 0.178 0.44 0.48 0.26
762 Shimonoseki 53 0.14 0.025 0.03 0.04 0.04 0.47 0.113 -0.05 0.07 0.10 0.14 0.025 0.03 0.04 0.04 0.47 0.113 -0.05 0.07 0.10
765 Hiroshima 53 0.25 0.050 0.01 0.06 0.05 0.31 0.063 0.02 0.06 0.07 0.25 0.050 0.01 0.06 0.05 0.31 0.063 0.02 0.06 0.07
766 Kure 41 0.12 0.027 -0.01 0.02 0.03 0.31 0.095 -0.13 0.02 0.06 0.12 0.027 -0.01 0.02 0.03 0.31 0.095 -0.13 0.02 0.06
767 Fukuyama 43 0.22 0.056 -0.04 0.03 0.05 0.78 0.219 -0.22 0.07 0.16 0.22 0.056 -0.04 0.03 0.05 0.46 0.114 -0.06 0.05 0.10
768 Okayama 53 0.14 0.031 -0.01 0.02 0.03 0.29 0.063 0.00 0.04 0.07 0.14 0.031 -0.01 0.02 0.03 0.29 0.064 -0.01 0.04 0.07
769 Himeji 42 0.21 0.049 -0.02 0.03 0.04 0.22 0.047 0.00 0.03 0.05 0.20 0.049 -0.02 0.03 0.04 0.22 0.047 0.00 0.03 0.05
770 Kobe 53 0.11 0.027 -0.02 0.01 0.02 0.24 0.066 -0.06 0.02 0.05 0.10 0.026 -0.02 0.01 0.02 0.24 0.066 -0.06 0.02 0.05
772 Osaka 53 0.17 0.039 -0.02 0.02 0.04 0.38 0.095 -0.06 0.04 0.08 0.17 0.040 -0.02 0.02 0.04 0.38 0.095 -0.06 0.04 0.08
776 Sumoto 42 0.16 0.033 0.00 0.03 0.04 0.35 0.093 -0.07 0.04 0.07 0.16 0.033 0.00 0.03 0.04 0.35 0.093 -0.07 0.04 0.07
777 Wakayama 53 0.12 0.028 -0.01 0.02 0.03 0.33 0.102 -0.14 0.02 0.07 0.12 0.028 -0.01 0.02 0.03 0.33 0.102 -0.14 0.02 0.07
778 Shionomisaki 49 0.01 0.003 0.00 0.00 0.00 0.00 0.000 0.00 0.00 0.00 0.01 0.003 0.00 0.00 0.00 0.00 0.000 0.00 0.00 0.00
780 Nara 53 0.23 0.041 0.04 0.05 0.05 0.42 0.089 0.01 0.07 0.10 0.22 0.041 0.03 0.05 0.05 0.42 0.089 0.01 0.07 0.10
784 Yamaguchi 48 0.36 0.059 0.09 0.12 0.08 0.53 0.086 0.14 0.16 0.13 0.35 0.057 0.09 0.11 0.08 0.53 0.086 0.14 0.16 0.13
800 Izuhara 48 0.05 0.015 -0.02 0.00 0.01 0.07 0.022 -0.04 0.00 0.02 0.05 0.015 -0.02 0.00 0.01 0.07 0.022 -0.04 0.00 0.02
805 Hirato 41 0.08 0.015 0.01 0.02 0.02 0.36 0.103 -0.11 0.03 0.07 0.08 0.014 0.01 0.02 0.02 0.25 0.064 -0.05 0.03 0.05
807 Fukuoka 53 0.17 0.035 0.01 0.03 0.04 0.47 0.112 -0.04 0.07 0.10 0.17 0.035 0.01 0.03 0.04 0.47 0.112 -0.04 0.07 0.10
809 Iizuka 42 0.29 0.053 0.05 0.08 0.07 1.98 0.573 -0.66 0.20 0.44 0.28 0.049 0.05 0.07 0.06 0.77 0.171 -0.02 0.15 0.16
812 Sasebo 42 0.15 0.035 -0.01 0.02 0.03 0.75 0.222 -0.27 0.06 0.16 0.15 0.035 -0.01 0.02 0.03 0.50 0.141 -0.15 0.04 0.11
813 Saga 53 0.16 0.031 0.02 0.04 0.03 0.31 0.073 -0.03 0.04 0.07 0.16 0.031 0.02 0.04 0.03 0.31 0.073 -0.03 0.04 0.07
814 Hita 42 0.37 0.086 -0.02 0.08 0.07 0.69 0.161 -0.05 0.12 0.15 0.37 0.086 -0.02 0.08 0.07 0.63 0.143 -0.03 0.11 0.13
815 Ohita 53 0.15 0.044 -0.05 0.01 0.03 0.28 0.089 -0.13 0.02 0.06 0.15 0.044 -0.05 0.01 0.03 0.28 0.089 -0.13 0.02 0.06
817 Nagasaki 53 0.17 0.039 -0.01 0.03 0.04 0.29 0.069 -0.03 0.04 0.06 0.17 0.039 -0.01 0.03 0.04 0.29 0.069 -0.03 0.04 0.06
819 Kumamoto 53 0.12 0.029 -0.02 0.02 0.02 0.26 0.082 -0.11 0.02 0.05 0.12 0.029 -0.02 0.02 0.02 0.26 0.082 -0.11 0.02 0.05
821 Asosan 53 1.01 0.209 0.05 0.27 0.20 2.32 0.553 -0.22 0.48 0.47 0.63 0.096 0.19 0.23 0.13 0.93 0.120 0.38 0.37 0.21
822 Nobeoka 40 0.06 0.019 -0.03 0.00 0.01 0.13 0.043 -0.07 0.01 0.03 0.06 0.019 -0.03 0.00 0.01 0.13 0.043 -0.07 0.01 0.03
823 Akune 40 0.36 0.092 -0.07 0.04 0.07 0.63 0.181 -0.20 0.05 0.13 0.36 0.092 -0.07 0.04 0.07 0.63 0.181 -0.20 0.05 0.13
824 Hitoyoshi 40 0.27 0.058 0.00 0.05 0.06 0.37 0.070 0.05 0.07 0.09 0.27 0.058 0.00 0.05 0.06 0.37 0.070 0.05 0.07 0.09
827 Kagoshima 53 0.29 0.069 -0.03 0.04 0.06 0.39 0.090 -0.02 0.05 0.09 0.29 0.069 -0.03 0.04 0.06 0.39 0.090 -0.02 0.05 0.09
829 Miyakonojo 40 0.08 0.021 -0.02 0.01 0.02 0.17 0.050 -0.06 0.01 0.03 0.08 0.021 -0.02 0.01 0.02 0.17 0.050 -0.06 0.01 0.03
830 Miyazaki 53 0.02 0.005 -0.01 0.00 0.00 0.00 0.000 0.00 0.00 0.00 0.02 0.005 -0.01 0.00 0.00 0.00 0.000 0.00 0.00 0.00
831 Makurazaki 43 0.30 0.079 -0.07 0.03 0.06 0.72 0.204 -0.22 0.06 0.15 0.30 0.079 -0.07 0.03 0.06 0.72 0.204 -0.22 0.06 0.15
835 Yuzu 43 0.01 0.003 0.00 0.00 0.00 0.00 0.000 0.00 0.00 0.00 0.01 0.003 0.00 0.00 0.00 0.00 0.000 0.00 0.00 0.00
836 Yakushima 47 0.01 0.005 -0.01 0.00 0.00 0.07 0.022 -0.04 0.00 0.02 0.01 0.005 -0.01 0.00 0.00 0.07 0.022 -0.04 0.00 0.02
838 Ushibuka 42 0.30 0.081 -0.07 0.04 0.06 0.55 0.145 -0.12 0.06 0.12 0.30 0.081 -0.07 0.04 0.06 0.55 0.148 -0.13 0.06 0.11
843 Fukue 48 0.32 0.085 -0.07 0.04 0.07 1.24 0.363 -0.43 0.10 0.26 0.32 0.085 -0.07 0.04 0.07 1.04 0.302 -0.35 0.09 0.23
887 Matsuyama 53 0.14 0.034 -0.02 0.02 0.03 0.29 0.088 -0.12 0.02 0.06 0.14 0.035 -0.02 0.02 0.03 0.29 0.088 -0.12 0.02 0.06
890 Tadotsu 41 0.20 0.055 -0.05 0.02 0.04 0.58 0.159 -0.16 0.05 0.12 0.20 0.055 -0.05 0.02 0.04 0.50 0.139 -0.14 0.05 0.10
891 Takamatsu 53 0.21 0.053 -0.04 0.03 0.04 0.53 0.143 -0.13 0.05 0.11 0.21 0.053 -0.04 0.03 0.04 0.53 0.143 -0.13 0.05 0.11
892 Uwajima 44 0.24 0.043 0.04 0.06 0.06 0.35 0.065 0.05 0.08 0.08 0.24 0.043 0.04 0.06 0.06 0.35 0.065 0.05 0.08 0.08
893 Kochi 53 0.11 0.031 -0.03 0.01 0.02 0.06 0.018 -0.03 0.00 0.01 0.11 0.031 -0.03 0.01 0.02 0.06 0.018 -0.03 0.00 0.01
894 Tsurugisan 26 3.19 0.670 0.11 0.78 0.64 8.22 2.147 -1.66 1.32 1.65 1.94 0.370 0.24 0.54 0.41 2.03 0.384 0.27 0.57 0.44
895 Tokushima 53 0.19 0.044 -0.02 0.03 0.04 0.56 0.162 -0.18 0.05 0.12 0.19 0.044 -0.02 0.03 0.04 0.56 0.162 -0.18 0.05 0.12
897 Sukumo 40 0.22 0.042 0.03 0.05 0.05 0.28 0.054 0.03 0.05 0.07 0.22 0.043 0.03 0.05 0.05 0.28 0.054 0.03 0.05 0.07
898 Shimizu 46 0.03 0.008 0.00 0.01 0.01 0.00 0.000 0.00 0.00 0.00 0.03 0.008 0.00 0.01 0.01 0.00 0.000 0.00 0.00 0.00
899 Murotomisaki 48 0.03 0.010 -0.01 0.00 0.01 0.03 0.009 -0.02 0.00 0.01 0.03 0.010 -0.01 0.00 0.01 0.03 0.009 -0.02 0.00 0.01

5.2.3 Equivalent unit weight for ground snow

Usually snow cover at the annual maximum depth comprises several layers, each of which has
accumulated one after another and with varying properties, from fresh snow to granular snow, as shown
in Fig. 5.2.9. Figure 5.2.8 shows typical examples of the relation between snow depth and the snow
mass, which is defined as the snow weight per unit area. Until the snow depth reaches the annual
maximum, the mass increases in proportion to the snow depth. After that, the mass still increases, and
the reaches the maximum value despite the decrease in depth. Although there is usually a time lag
between the dates of the annual maximum depth and the annual maximum mass, in most cases the
- C5-14 - Commentary on Recommendations for Loads on Buildings

meteorological observation data give only the snow depth. Therefore, we need to use the equivalent unit
weight for the ground snow, which is the ratio of the annual maximum snow mass to the annual
maximum snow depth, in order to evaluate the annual maximum snow mass from the annual maximum
snow depth.

Figure 5.2.8 Relation between snow depth and Figure 5.2.9 Layers of different types of snow
snow mass6) (section of accumulated snow)7)

The statistical properties of the annual maximum snow depth and the annual maximum snow mass
were studied for data measured at 12 meteorological observatories located in heavy snow regions in
Japan. Figure 5.2.10 shows the relation between snow depth 𝑑𝑑(m) and equivalent unit weight for ground
snow 𝑝𝑝(kN/m3) with 100- and 10-year recurrence intervals8). This relation is described through the
following regression:

𝑝𝑝0 = 0.72�𝑑𝑑0 ⁄𝑑𝑑ref + 2.32 (5.2.8)

CHAPTER 5 SNOW LOADS - C5-15 -

Figure 5.2.10 Relation between annual maximum snow depth and equivalent unit snow weight at 12
observatories in Japan8)

In the purview, the snow depth 𝑑𝑑0 in this equation is divided by the reference snow depth 𝑑𝑑ref , which is
1 m, in order to match the units in the equation.
Figure 5.2.11 shows the relation between snow depth 𝑑𝑑 and equivalent unit weight 𝑝𝑝 for data
calculated using the method of snow weight evaluation using the daily precipitations and daily average
air temperatures (mentioned in 5.1), which were observed at 126 meteorological observatories, not only
in heavy snow regions but also in little snow regions. It is shown that Eq. (5.2.8) overestimates the snow
weight at a snow depth of less than about 1 m. Therefore, it is possible to use a smaller equivalent unit
weight for a snow depth of less than approximately 1 m against the equivalent unit weight obtained
through the equation.

Figure 5.2.11 Relation between annual maximum snow depth and equivalent unit snow weight with a
return period of 100 years calculated at 126 observatories in Japan.9)
- C5-16 - Commentary on Recommendations for Loads on Buildings

5.2.4 Environmental coefficient

Snow loads on roofs vary according to various factors owing to environmental conditions. Table 5.2.2
shows environmental factors influencing the depth of snow on a roof.

Table 5.2.2 Environmental factors influencing the depth of snow on a roof

macrofactors distance from the nearest seaside, altitude, topographical
inclination, topographical curvature, etc.
mesofactors trees, neighboring buildings, local topography, etc.
microfactors thermal conditions on the roof surface, roofing materials,
orientation of the roof, etc.

Macrofactors are mostly included in the estimated depth of the ground snow, and are quite variable in
determining the design snow load. Therefore, in this recommendation, only the mesofactors are
considered as the environmental coefficient (𝑘𝑘env).
Figure 5.2.12 shows the measurement results of the snow depth distribution around the city of Sapporo
in Hokkaido. In the snow depth on the ground around the urban area (the point from the coast to the
urban area, the east side of the urban area, and the mountains located on the western side of the city),
there is a tendency for snow depths to be about 10–20% larger than that observed near the
meteorological office. Because the ground snow depth may be larger than the observation point owing
to the influence of the ground surface condition, topography, and other factors, it is necessary to
investigate the deposition characteristics of the planned site and consider it as the environmental
coefficient.
When uneven distributions of wind are generated owing to the influence of the terrain, the snow
particles are blown away at a place with strong winds, and accumulate intensively at a place with weak
winds. In particular, deep snowdrifts will be found where the terrain is hollow. Figure 5.2.13 shows an
actual measurement example of a snowdrift formed by snow blown off the windward side in a dip (road
cut). In the case of a slope gradient of 1:1.5, there was a snowdrift that reached 4.0 m from the upwind
slope to the road. On the other hand, with a slope gradient of 1:3, a snowdrift of 3.7 m at maximum
occurred near the middle of the upwind side slope, and the snowdrift decreased when approaching the
road. As described above, the gradient of the slope greatly affects the location of the snowdrift; however,
it is currently difficult to quantitatively determine the environmental coefficient owing to the local
topography. This should be investigated through wind tunnel experiments, computational fluid
dynamics simulations, field measurements, and so on, depending on the local conditions.
CHAPTER 5 SNOW LOADS - C5-17 -

Meteorological office

Figure 5.2.12 Observed snow depth around Sapporo City10)

(a) Slope gradient of 1:1.5 (b) Slope gradient of 1:3

(March 1969, Tomamae) (March 1980, Tomamae)

Figure 5.2.13 Examples of snowdrift observed in a field survey11)

- C5-18 - Commentary on Recommendations for Loads on Buildings

5.3 Shape Coefficient

According to recent studies, the snow load on a roof depends on the wind speed and temperature.12, 13)
In cold regions, the snow on a roof is easily removed by wind, as shown in Photo 5.3.1 and 5.3.2.
However, in relatively warm regions such as the Hokuriku region, particularly with calm wind
conditions, the snow depth on a roof is similar to that on the ground.14)

Photo 5.3.1 Snow drift on a roof Photo 5.3.2 Snow accumulation on a roof

In this recommendation, the shape coefficient 𝜇𝜇0 is defined as

𝜇𝜇0 = 𝜇𝜇b + 𝜇𝜇d + 𝜇𝜇s , (5.3.1)

where
𝜇𝜇b : basic shape coefficient,
𝜇𝜇d : shape coefficient for an irregular distribution caused by drifting snow,
𝜇𝜇s : shape coefficient for irregular distribution caused by sliding.

This coefficient is only applied to normally shaped buildings. For special forms or large-scale
buildings, the shape coefficient should be determined from special research, such as experiments or
computational fluid dynamics (CFD) simulations.
In Eq. (5.3.1), the snow coverage distribution on a roof is evaluated by calculating the shape
coefficient as the sum of 𝜇𝜇b , 𝜇𝜇d , and 𝜇𝜇s . However, depending on the roof shape, the characteristics of
the roofing material, and the weather conditions of the construction site, the snow coverage distribution
on a roof may be evaluated only based on the basic shape coefficient (𝜇𝜇b ). Furthermore, it can be
evaluated through a symmetrical or asymmetric roof snow distribution pattern, and by combining other
roof shape coefficients.

5.3.1 Basic shape coefficient

The relationship between the ratio of the roof snow depth to the ground snow depth (basic shape
coefficient) and the wind speed is shown in Fig. 5.3.1. In the figure, the solid line shows the former
CHAPTER 5 SNOW LOADS - C5-19 -

recommendation of the AIJ (1986), the pointed dashed line shows the recommendation of the former
Soviet Union15), the dashed line shows the equation proposed by Mihashi et. al. (1988)16), and the dotted
line shows the equation proposed by Tomabechi et. al. (1989)17). Each of these studies has shown that
the basic shape coefficient decreases as the wind speed increases.
Roof snow also depends on the roof shape. According to the wind tunnel tests18), the ratio of the roof
snow depth to the ground snow depth increases for roof slopes of up to 20°, and decreases for roof slopes
of over 20°. By referring to these results, the basic shape coefficient is determined as shown in Fig.5.1.
For roof slopes of larger than 50°, the value for 50° is used.

Figure 5.3.1 Relationship between ratio of roof snow depth to ground snow depth and wind speed

The average wind speed during January and February is calculated in this guideline based on the
average values published by the Japan Meteorological Agency (average value over the past 30 years). In
addition, this average wind speed is converted for the roof height of the building. For a sloping roof, the
average wind speed is obtained at the average elevation and maximum height.
The wind speed 𝑈𝑈z at height 𝑍𝑍 of the roof is expressed through the following equation:
(𝑍𝑍⁄𝑍𝑍GSITE )𝛼𝛼SITE
𝑈𝑈Z = 𝑈𝑈SDP ･ (𝑍𝑍 𝛼𝛼 , (5.3.2)
SDP ⁄𝑍𝑍GSDP ) SDP

where
𝛼𝛼SITE, 𝑍𝑍GSITE (m): parameters that are determined from the wind speed profile at the construction
site,
𝛼𝛼SDP , 𝑍𝑍GSDP (m): parameters that are determined from the wind speed profile at the weather
observation station,
𝑍𝑍SDP (m): anemometer height at the weather observation station,
𝑈𝑈SDP (m/s): wind speed at the height 𝑍𝑍SDP
According to Chapter 6 of these guidelines on the wind load, the parameters that are determined from
- C5-20 - Commentary on Recommendations for Loads on Buildings

the wind speed profile are prescribed depending on the terrain category, as shown in Table 5.3.1.

Table 5.3.1 Parameters determined from wind speed profile for terrain category
terrain Category power-index αSITE, αSDP ZGSITE, ZGSDP

Ⅰ 0.10 250

Ⅱ 0.15 350

Ⅲ 0.20 450

Ⅳ 0.27 550

Ⅴ 0.35 650

5.3.2 Shape coefficient for irregular distribution caused by snow drift

(1) On the troughs of M-shaped roofs, multiple pitched roofs, and multi-span roofs, the snow
accumulates by snow drift, and its depth is usually deeper in the troughs than at the ridge, as shown in
Fig.5.3.219). This phenomenon is also observed in the wind tunnel tests19). Therefore, the shape
coefficient for the irregular distributions caused by drifting snow 𝜇𝜇d is defined for the troughs, whereas
it is zero at the ridges, as shown in Fig. 5.3.3.
(2) On multilevel roofs, the snow accumulation on the lower roof levels increases, especially at the
edge of the high roofs. Figures 5.3.3 and 5.3.4 show the results of the field observations using a
prototype multilevel roof20). The roof snow depth was 1.5 to 2.0 times larger than the ground snow
depth. Such a range of irregular distribution was approximately 2.0 times the distance of the steps
between the lower and upper roofs. The same thing occurs in roofs with parapets and local obstructions,
such as fences and signboards.

Figure 5.3.2 Observation of snow accumulated on multi-span pitched roof (Ottawa)19)

0 0 0 0

μd μd
μd
（a） M-shaped and multiple pitched roofs （b） multispan roof

Figure 5.3.3 𝜇𝜇d in M-shaped, multiple pitched, and multi-span roof

CHAPTER 5 SNOW LOADS - C5-21 -

Wind

Photo 5.3.3 Roof snow distribution caused by snow drift on multilevel roofs20)
Snow depth ratio dr/dg

Snow depth ratio dr/dg

2.5 2.5
Wind Wind
2.0 2.0
1.5 1.5
1.0 1.0
0.5 0.5
0.0 0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Distance x/H Distance x/H
(a) Mean wind speed during snowfall: 2.9m/s, Incremental ground snow depth dg=7cm

2.5 2.5
Snow depth ratio dr/dg

Snow depth ratio dr/dg

Wind Wind
2.0 2.0
1.5 1.5
1.0 1.0
0.5 0.5
0.0 0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Distance x/H Distance x/H

(b) Mean wind speed during snowfall: 2.3m/s, Incremental ground snow depth dg=30cm
Snow depth ratio dr/dg

Snow depth ratio dr/dg

2.5 2.5
Wind Wind
2.0 2.0
1.5 1.5
1.0 1.0
0.5 0.5
0.0 0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Distance x/H Distance x/H
(c) Mean wind speed during snowfall: 1.2m/s, Incremental ground snow depth dg=11cm

Figure 5.3.4 Snow distribution on the lower roof in multilevel roofs20)

- C5-22 - Commentary on Recommendations for Loads on Buildings

5.3.3 Shape coefficient for irregular distribution caused by sliding

The sliding snow on a roof depends on characteristics of the snow, temperature, the roof shape, the
roofing materials, and so on. In a roof with troughs such as M-shaped, multiple pitched, and multi-span
roofs, the roof snow slides along the slope because of gravity. As a typical case, the shape coefficient 𝜇𝜇s
is equal to 𝜇𝜇b , as shown in Fig. 5.3.5. In addition, Fig. 5.3.6 shows the result of field observations using
the prototype of different pitched roofs with the steel sheet roofing21). In a roof slope of 1/10 (5.7°), less
snow sliding occurred. On the other hand, significant snow sliding was observed for a slope of over 2/10
(11.3°). Therefore, for a roof slope of below 1/10, the shape coefficient 𝜇𝜇s can be set to zero. However,
when (2𝜇𝜇b +𝜇𝜇d ) is larger than the roof gap, as shown in Fig. 5.3.7(b), this should be adjusted to the
treatment shown in Fig. 5.3.7(c).

（μs＝μb）
μb μb－μs＝0

μb＋μd＋μs
μb＋μd ＝2μb＋μd
（a） Before sliding （b） After sliding

Figure 5.3.5 Typical case of roof snow sliding

CHAPTER 5 SNOW LOADS - C5-23 -

：Sliding
1.2
Roof snow depth (m)

Roof slope: 1/10（5.7°）

0.8

0.4

0.0
January February March
Roof snow depth (m)

Roof slope: 2/10（11.3°）

0.6

0.4

0.2

0.0
January February March

0.6
Roof snow depth (m)

Roof slope: 4/10（21.8°）

0.4

0.2

0.0
January February March

Figure 5.3.6 Relationship between roof slope and snow sliding21)

（μs＝μb）（μs≠μb）
μb μb－μs＝0 μb－μs

μb＋μd μb＋μd＋μs μb＋μd＋μs

＝2μb＋μd
（a） Before sliding （b） After sliding （μs=μb）（c） Treatment

Figure 5.3.7 Modification of the distribution of roof snow distribution when the snow depth is larger
than the roof gap
- C5-24 - Commentary on Recommendations for Loads on Buildings

5.4 Partial Snow Load on a Roof

The snow load on a roof partially increases because of snowdrift caused by projecting objects, and
snow slides, snow overhangs, cornices, and icicles may also increase the snow load. Although such load
might not be critical for the main structural frame such as girders and columns, it should be considered
in the design of secondary members such as joints, purlins, and rafters. This is essential to determine the
loadings according to the anticipated extent of increases in partial load.
A snowdrift may occur around chimneys, ventilators, and other projections located on roofs. The
deposited snow increases the snow load around such structures, just as with tall parapets. The amount of
snow accumulations will vary with the shape of the projection, air temperature, wind direction, and wind
velocity.
It is particularly important to understand that snow overhangs and cornices can grow to as much as
several meters in length. These snow accumulations mostly consist of granular or compressed snow.
Therefore, it should be considered a concentrated load with a unit weight of about 4 kN/m3.
When ice adhesion arises and prevents the snow on the roof from sliding, it is possible for the sliding
force of the snow on the entire roof to concentrate at the edge of the eaves, possibly causing damage to
the finishing materials. If the sliding force of the snow exceeds the ice adhesion force, the snow on the
roof will give way and slide downward. At this moment, a load equivalent to the ice adhesion force will
act on the eaves.
In principle, the increase in load from the snow sliding off from the higher roofs should first be
prevented. However, if it cannot be prevented, the most adverse conditions should be considered in the
design, and the flight distance, spreading upon impact, and the impact load are significant. According to
the experiments on the impact load of snow-ice blocks, the maximum impact load increases as the unit
weight increases up to 6 kN/m3, whereas beyond this unit weight, it remains to be constant at about 400
kN/m2.
These results were obtained through experiments under a limited range of conditions, but they
indicate that the impact load of a snow-ice block is not negligible. In the instantaneous values of the
impact load, a peak (maximum impact) is reached immediately upon impact, and the snow then begins
behaving as a fluid. The maximum impact value 𝐹𝐹I max immediately upon impact, and the value 𝐹𝐹I for a
state in which a block is treated as a fluid, can be obtained using the equations below:

𝐹𝐹I max = 𝜎𝜎s 𝐴𝐴 (5.4.1)

𝐹𝐹I = 𝜌𝜌s 𝑉𝑉I2 𝐴𝐴 (5.4.2)

where
𝜎𝜎s (N/m2): maximum impact load per unit area,
𝜌𝜌s (kg/m3): snow density,
𝐴𝐴(m2): cross-sectional area of impact,
𝑉𝑉I(m/s): impact velocity.
CHAPTER 5 SNOW LOADS - C5-25 -

5.5 Snow Load on a Roof with Snow Control

5.5.1 Equations for snow load on a roof with snow control
In these recommendations, the snow load design on a roof with control 𝑆𝑆(kN/m2) is defined as

𝑆𝑆 = 𝑆𝑆n 𝜇𝜇n − 𝑆𝑆c , (5.5.1)

where
𝑆𝑆n (kN/m2): characteristic snow load on the ground when the snow load on a roof is controlled,
𝜇𝜇n : shape coefficient when the snow load on a roof is controlled,
𝑆𝑆c (kN/m2): controlled snow load, as defined in 5.5.2.
In this case, 𝜇𝜇n might differ from 𝜇𝜇0 (5.3). However, because 𝑆𝑆n is determined from snow accumulation
for 𝑛𝑛 days, it is difficult to clarify the difference through wind tunnel tests or other estimation methods.
This also depends on the method of snow control. Therefore, in this recommendation, 𝜇𝜇n takes the same
value as 𝜇𝜇0 . When the snow load is controlled, 𝑆𝑆c is calculated as the difference in the initial snow load
remaining on the roof when another heavy snowfall is expected, and the snow load removed by a device
whose performance is guaranteed even during a heavy snowfall. The performance should be evaluated
through experiments, and the value of 𝑆𝑆c should be determined.

5.5.2 Ground snow load with accumulation for 𝒏𝒏 days

(1) Calculation of the characteristic ground snow load with accumulation for 𝑛𝑛 days
1) Estimation of the characteristic snow loads on the ground based on the ground snow depth with
accumulation for 𝑛𝑛 days when using roof snow control
A heavy short-term snowfall produces a critical condition in buildings whose snow loads are usually
controlled by reliable systems. In such cases, the annual maximum snow load for the entire season leads
to an overestimation, and increasing snow depths for certain days need to be statistically estimated. In
these recommendations, the increasing snow depth for 𝑛𝑛 days is considered. The duration 𝑛𝑛 is
determined through careful consideration. For buildings with reliable control systems, the characteristic
snow load on the ground Sn is given by Eq. (5.5.2) as the product of the following factors:

𝑆𝑆n = 𝑘𝑘env 𝑑𝑑n 𝑝𝑝n , (5.5.2)

where

𝑘𝑘env: environmental coefficient, as defined in section 5.2.4,

𝑑𝑑n (m): basic snow depth on the ground when the snow load on the roof is controlled, as defined
in (2) in section 5.5.2,
𝑝𝑝n (kN/m ): equivalent unit weight for ground snow, as defined in (3) in section 5.5.2.
3

2) Estimation of the characteristic snow weights on the ground based on the data of daily precipitation
- C5-26 - Commentary on Recommendations for Loads on Buildings

and daily mean air temperature in 𝑛𝑛 days

In (2) in section 5.2.1, the newly proposed method was shown to be practical and useful for estimating
the ground snow weight. To estimate the characteristic snow weights on the ground, the
above-mentioned method can be easily applied. Figure 5.5.1 shows examples of the accumulated
ground snow weights for 7 days in Sapporo, Kamabuchi, Nagaoka, and Tokamachi using Eq. (5.2.2), in
which the estimated values are compared with the observed values. As shown in the figures, the
estimated values agree fairly well with the observed values.

Figure 5.5.1 Example of observed and estimated snow weight accumulation for 7 days using Eq.
(5.2.2)9)

(2) Basic ground snow depth accumulation for 𝑛𝑛 days

The basic snow depth accumulation for 𝑛𝑛 days on the ground 𝑑𝑑n is defined as the annual maximum
value of snow accumulation for 𝑛𝑛 days with a return period of 100 years for the snow load on a roof with
a reliable control system. The annual maximum increase in intensity of the snow depth is defined as
shown in Fig. 5.5.2. Because of the limited number of meteorological observatories, the basic snow
depth at the construction site is estimated from topographical data.
To clarify the difference between the annual maximum snow depth for the entire season (AMD) and
the annual maximum increase in snow depth for 3 (AMI-3) and 7 (AMI-7) days, two contrasting
examples plotted on Gumbel probability paper are shown in Fig. 5.5.3. In Hokkaido, the values of the
AMD become larger at the end of winter in spite of smaller AMI values because of the low temperature.
CHAPTER 5 SNOW LOADS - C5-27 -

In Gifu, located in the middle of Honshu, however, the AMI values are closer to those of the AMD
because this area is rather warmer and the snow depth increases rather rapidly. A different snow
accumulation mechanism can also be recognized in Fig. 5.5.4. In the figure, types A, B, and C can be
described as having Gumbel, Frechet, and Weibull distribution shapes on the Gumbel probability paper,
which depends on the upper data regression, as shown in Fig. 5.5.3.

Figure 5.5.2 Annual maximum increase in snow Figure 5.5.4 Ratio of AMI-7 to AMD22)
depth for n days22)

Figure 5.5.3 Comparison of AMD, AMI-3, and AMI-7 (Gumbel probability paper, Hazen plot)22)
- C5-28 - Commentary on Recommendations for Loads on Buildings

(3) Equivalent Unit Weight for Ground Snow with Roof Snow Control
When controlling the roof snow load, the snow accumulation for n days is used for the design load.
This value can be used directly when the cumulative snowfall weight during n days is available, and
Table 5.2.1 shows the estimated value of the 7-day accumulated snow weight, but in general, an
equivalent unit snow weight pn multiplied with the n-day accumulated snow depth is obtained. In this
case, because the observed data of the equivalent unit weight 𝑝𝑝n to be multiplied by the n day
accumulated snow depth is insufficient, to be on the safe side, we use the same p0 as the value for the
annual maximum snow depth .
In the old recommendation (1993 edition), the difference in snow depth in n days was used as the
definition of the n day accumulated snow depth. For this reason, a negative value is obtained during a
period in which there is no snowfall and only settlement occurs. At present, observational data of
meteorological offices and AMeDAS observation stations can be obtained on CD-ROM and through the
Internet, and an analysis based on daily snow depth is easy to achieve. In this section, the method based
on the difference in the snow depth on the ground is named “method A,” and the method based on the
daily snow depth is “method B.” On the other hand, it is necessary to use the daily weight observation
for the n day accumulated snow weight. However, a few observations were conducted each day, and
observations at the 12 points in the above-mentioned multiple snowy areas are observed only every few
days except for Nagaoka, Takada, and Tokamachi.
Sakurai and Joh9) discussed using method A with data calculated using an evaluation of the snow
weight based on daily precipitation and daily average air temperatures at 126 meteorological
observatories. Figure 5.5.5 shows the relation between snow depth 𝑑𝑑n and equivalent unit weight 𝑝𝑝n
based on the annual maximum snow depths accumulating for 3 and 7 days with a 100-year recurrence
interval. The values of 𝑝𝑝n shown in Figs. 5.5.5 and 5.5.6 are smaller than those of 𝑝𝑝0 for a snow depth of
larger than 0.5 m in heavy snow regions, as shown in Fig.5.2.10, because the snow density increases
according to the number of days of accumulation. However, they are not similar to the values of 𝑝𝑝0 for
an annual maximum snow depth of less than 0.5 m in low snow regions because the snow melts away
within several days. These tendencies are also demonstrated in Fig. 5.5.7, that is, the equivalent unit
weights 𝑝𝑝n are almost constant for Tokyo and Sendai located on the Pacific Ocean side with relatively
warm winters, but are different for Fukui and Sapporo located on the Sea of Japan side with cold
winters.
CHAPTER 5 SNOW LOADS - C5-29 -

Figure 5.5.5 Relation between annual maximum snow depth accumulating for 3 and 7 days and
equivalent unit snow weight with a return period of 100 years calculated at 126
observatories in Japan9)

Figure 5.5.6 Frequency of equivalent unit snow weight for 1, 3, and 7 day of accumulation9)

Figure 5.5.7 Relation between number of accumulation days and equivalent unit snow weight in certain
cities in Japan9)
- C5-30 - Commentary on Recommendations for Loads on Buildings

The equivalent unit weight for 1, 3, and 7 days are 1.37, 1.41, and 1.57, respectively, which were
obtained as the average values of 126 observation points, as shown in Fig. 5.5.6. Therefore, for safety
purposes, it is considered that the equivalent unit weight for 𝑛𝑛-days of snow accumulation with roof
snow control 𝑝𝑝n can be the same as that for the annual maximum snow accumulation without roof snow
control 𝑝𝑝0 (Eq. (5.2.8)).
On the other hand, Takahashi et al.24) estimated of the n day accumulated snow depth and the n day
accumulated snow weight through method B at AMeDAS observation points using a snow depth meter
for 330 locations nationwide, as shown in Fig. 5.5.8. When comparing the result using a 1-day
accumulated snow depth with a past report on the density of new snow cover (Fig. 5.5.9), it corresponds
approximately. The possibility of extremely dense snow including rain in a region with little snow can
be determined. By overlapping the lines of the minimum snow load 𝑆𝑆min = 0.3 kN/m2, in Fig. 5.5.10, it
can be seen that there is no problem in using this value as the lower limit of the snow load. In an area
with a high snowfall, the equivalent unit snow weight 𝑝𝑝n increases as the number of days of 𝑛𝑛 increases.
At the maximum snow depth per year, the trend of a larger accumulated snow weight tends to increase as
the snow cover increases; however, this tendency has not been seen in the range analyzed for a 14-day
period. The average density for each area is calculated using the snow depth 𝑑𝑑tr , which distinguishes
between the snowy areas indicated in this recommendation and the others shown in Table 5.5.1. This
relationship is shown in Fig. 5.5.11. In this state, because a blank area is formed between the lower limit
and 𝑑𝑑tr , the snow depth 𝑑𝑑c intersecting the lower limit is determined as listed in the rightmost column in
Table 5.5.1.

𝑝𝑝n = 0.33 ln(𝑛𝑛 + 2) + 0.48 (𝑑𝑑 ≥ 𝑑𝑑c ) (5.5.3)

𝑆𝑆min = 0.3 kN/m2 (𝑑𝑑 < 𝑑𝑑c ) (5.5.4)

Therefore, Eq. (5.5.3) can be used as an approximate expression for the equivalent unit snow weight
𝑝𝑝n for the 𝑛𝑛 day accumulated snow depth according to Method B. The estimated value of the snowfall
weight based on this equation is as shown in Fig. 5.5.12.
CHAPTER 5 SNOW LOADS - C5-31 -

Figure 5.5.8 Relationship between n day increment snow depth and equivalent unit weight23)
- C5-32 - Commentary on Recommendations for Loads on Buildings

Table 5.5.1 Unit snow weight of a part regarded as a constant (kN/m3)24)

incremental transfer depth 𝑝𝑝n (𝑑𝑑 ≥ 𝑑𝑑tr ) 𝑝𝑝𝑛𝑛 (𝑑𝑑 < 𝑑𝑑tr ) cross depth
day 𝑛𝑛 𝑑𝑑tr (m) 𝑑𝑑c (m)
average [Link]. average [Link].
1 0.50 0.86 0.32 2.35 1.30 0.35
3 0.57 1.02 0.34 2.51 1.24 0.30
5 0.63 1.12 0.35 2.56 1.24 0.27
7 0.70 1.20 0.37 2.51 1.15 0.25
10 0.80 1.31 0.42 2.48 1.13 0.23
14 0.93 1.42 0.43 2.39 1.07 0.21

Figure 5.5.9 Influence of daily air temperature Figure 5.5.10 One-day accumulated snow
on the density of new snow23) weight and lower limit of snow
weight24)

Figure 5.5.11 Accumulated snow depth in n days Figure 5.5.12 Accumulated snow depth in n days
and equivalent unit snow and ground snow weight24)
weight24)

5.5.3 Snow load with snow control

Controlled snow load, 𝑆𝑆c , is generally determined after field research and experiments investigating
the capacity of sliding or melting devices, where 𝑆𝑆c is the differential between the initial snow load
expected when heavy snow fall starts, and the snow load removed using a device whose performance is
guaranteed even during heavy snow fall.
There are three methods of artificially removing snow from a roof:
CHAPTER 5 SNOW LOADS - C5-33 -

1. Mechanical snow removal (including through manpower)

2. Intentional snow sliding
3. Snow melt
A combination of methods 2 and 3 is also used. In methods 1 and 2, snow is removed after it
accumulates, but it is unclear how much snow has accumulated when the removal starts. In method 3,
there are two possibilities: all of the snow has melted, or some of the snow still remains. Therefore, the
design snow load on a roof should be decided by the planner. Figure 5.5.13 (a) shows the daily snow
depth, and Fig. 5.5.13 (b) shows the snow load on the ground simulated using Eq. (5.2.4).

Figure 5.5.13 Simulations of roof snow control (in the case of Sapporo, 1980-1981).

When snow is removed through manpower, the time lag between the snow load exceeds the design
snow load on the roof, and the removed snow should be considered. If the time lag is long, as shown in
Fig. 5.5.13 (c), the designer should not consider a controlled snow load. Figure 5.5.13 (d) shows a case
in which the snow is removed mechanically with a short delay. As a worst case, we should consider a
heavy snowfall starts when the existing snow load is just over the design load level. Figure 5.5.13 (e)
- C5-34 - Commentary on Recommendations for Loads on Buildings

shows the results for a case in which the reliable snow removal system is available, and for only such a
case can we take the controlled snow load into the design. Figure 5.5.13 (f) simulates a snow melting
system that is operated continuously during the snow season. Its capacity and reliability against heavy
snowfalls should be considered. When the performance of the device is guaranteed for heavy snowfalls,
the controlled snow load 𝑆𝑆c may be reduced from 𝑆𝑆n 𝜇𝜇n , as expressed in Eq. (5.5.1).
According to some studies, a melting system that melts all of the snow on the roof continuously
consumes a lot of energy, and a melting system that only operates after some snow has accumulated is
inefficient because of the snow caving between the snow and roof25, 26). After this trial, a new method
was studied27, 28). With this method, a thin ice layer is created and then melted, thereby inducing snow
sliding and consuming little energy.
With snow control, it is important to ensure a reliable energy supply during a heavy snow fall period.
In Hokkaido, since the 1970s, roads (national route) closed by heavy snow have been reopened after 4
days at most.
CHAPTER 5 SNOW LOADS - C5-35 -

5.6 Other Snow Loads

(1) Side pressure
Side pressure from snow drifts on the outside walls of buildings may be categorized as A or B, as
shown in Fig. 5.6.1. The load for case B might be larger than for case A. The side pressure can be
assumed as follows30):

𝐹𝐹1 = 0.9 ∗ 10−3 �15～20�𝑑𝑑s2 (5.6.1)

where
𝐹𝐹1 (kN/m2): side pressure,
𝑑𝑑s (m): depth from the snow surface.

Figure 5.6.1 Lateral pressure and settlement29) Figure 5.6.2 Observation of

lateral pressure30)

(2) Settlement force

When a building might be buried under snow, snow load caused by sedimentation should be
considered because eaves or braces outside might be broken.
Settlement force of snow cover might be assumed as follows31),32):

𝐹𝐹smax = 0.9 ∗ 1.7(𝑆𝑆max⁄9.8)1.5 (maximum snow depth of up to 3 m) (5.6.2)

𝐹𝐹smax = 1.4 𝑆𝑆max (maximum snow depth of up to 10 m) (5.6.3)

𝐹𝐹s = 9.8 ∗ (−0.132 + 0.782(𝑆𝑆⁄9.8) + 0.387(𝑆𝑆⁄9.8)2 ) (snow depth of up to 5 m) (5.6.4)

where
𝐹𝐹s , 𝐹𝐹smax (kN/m): settlement force, and maximum settlement force during the winter,
𝑆𝑆, 𝑆𝑆max(kN/m2): snow weight per unit area, and maximum snow weight during the winter.

(3) Crown snow and snow accretion

Crown snow is snow accumulating on a small structure like a mushroom. It might grow to over 1 m,
and therefore the crown snow-load should be considered in such cases.
- C5-36 - Commentary on Recommendations for Loads on Buildings

When snow particles intensively hit the structures, snow accretion might be formed on their surface.
In the case of wet snow with strong winds, such phenomenon will be remarkable. The risk of falling
snow adhered to a high rise building should be considered.

(4) Snow load on an external member

In recent years, many buildings have set up eaves on an outer wall surface to control solar radiation.
In case of snowy regions, eaves on a building may be covered with snow. Consequently, snow load and
the risk of falling snow and ice should be considered in such cases.
(5) Blowing snow into balconies or outside corridors
When large amounts of snow blow into balconies or outside corridors, snow load and their affect on
facilities in the area should be considered.

(6) Snowdrift by a neighboring building

The snow depth depends on airflows around a building, and snow load on a roof is therefore
influenced by the neighboring buildings. When snow drifts caused by a neighboring building or trees
might be accumulated on the roofs, such a state should be considered when estimating the snow load. In
such cases, model studies in wind tunnels or CFD are recommended.

(7) Snowdrift by high-rise and/or large building

Strong winds induced by a high-rise building may cause blowing snow and snowdrifts around it. A
newly constructed high-rise building may therefore cause snowdrifts on surrounding buildings and
increase snow loads on other buildings. In such cases, model studies in wind tunnels or CFD are also
recommended.

(8) Ponding of the membrane structure

For membrane structures and large span roofs, ponding phenomenon from rain-on-snow or snow
meltwater should be considered. Roofs of membrane structures shall be designed to accelerate sliding
snow, or snow-melting systems should be installed.
CHAPTER 5 SNOW LOADS - C5-37 -

References
1) Sakurai, S., Joh, O. and Shibata, T.: Estimation of ground snow weight based on daily precipitation
and daily mean air temperature, Snow Engineering: Recent Advances, pp.185-192, Balkema, 1997.
2) Kamimura, S., Umemura, T.: Estimation of daily snow mass on the ground using air temperature and
precipitation data, Second International Conference on Snow Engineering, CRREL Special Report
92-27 pp.157-167, 1992.
3) Takahashi, T., Kawamura, T., Kuramoto, K.: Estimation of ground snow load using snow layer
model, Journal of Structural and Construction Engineering (Transactions of AIJ)，No. 545, pp.
35-41, 2001.
4) Izumi, M., Mihashi, H. and Takahashi, T., Statistical properties of the annual maximum series and a
new approach to estimate the extreme values for long return periods, Proc. of 1st. Int. Conf. on Snow
Engineering, CRREL Special Report 89-6, pp. 25-34, 1989.
5) Takahashi, T., Fukaya, M., Mihashi, H., and Izumi, M.: Influence of altitude and sea area ratio for
geographic distribution of snow depth, Summaries of Technical Papers of Annual Meeting, Vol. B,
AIJ, pp. 219 - 220, 1992.8 (in Japanese).
6) Takahashi, H. and Nakamura, T., Disaster prevention of snow and ice, Hakuashobo, pp.213 - 218,
1986 (in Japanese).
7) Maeda, H.: Fundamental study on estimation of snow load, Transactions of AIJ, No.319, pp.32-37,
1982 (in Japanese with English abstract).
8) Joh, O., Sakurai, S.: Equivalent snow density in heavy snowing area, Journal of Snow Engineering,
JSSE, Vol. 9, No. 2, pp. 112 - 114, 1993.4 (in Japanese)
9) Sakurai, S., Joh, O. and Shibata, T.: Estimation of equivalent density of snow accumulation for short
period, Snow Engineering, Recent Advances and Developments, pp. 143-148, Balkema, 2000.
10) Naruse, R., Aburakawa, H. and Ishikawa, N.: Characteristics on the distribution of snow
accumulation in Sapporo, Hokkaido, Law Temperature Science, Ser. A, 36, pp. 139-153, 1978 (in
Japanese with English Abstract).
11) Ishimoto, K., Takeushi, M., Nohara, T. and Fukuzawa, Y.: Drifting snow capacity of road with
cutting section, Proceedings of 1985 Cold Region Technology Conference, pp.533-538, 1985.
12) Tomabechi, T., Izumi, M. and Endo, A.: A fundamental study on the evaluation method of
roof-snowfall-distributions of buildings, Journal of Structural Engineering (Transactions of AIJ),
Vol. 32B, pp. 49-62, 1986.3 (in Japanese with English abstract).
13) Izumi, M., Mihashi, H., Sasaki, T., Takahashi, T., Matsumura, T.: Fundamental study on estimation
of roof snow loads part 15, Summaries of Technical Papers of Annual Meeting, Vol. B, AIJ, pp.
1405-1406, 1987.10 (in Japanese).
14) Yamada, K. and Matsumoto Y.: Observation of snow cover on real flat roofs in the Hokuriku region,
1982-1989, Summaries of Technical Papers of Annual Meeting, Vol. B, AIJ, pp. 117-118, 1990.10
(in Japanese).
15) ISO 4355, Bases for design of structures - Determination of snow loads on roofs, 1992.
- C5-38 - Commentary on Recommendations for Loads on Buildings

16) Mihashi, H., Takahashi, T. and Izumi, M.: Wind effects on snow loads, Proc. of 1st. Int. Conf. on
Snow Engineering, CRREL Special Report 89-6, pp. 158-167, 1989.
17) Tomabechi, T.: Influence of meteorological conditions for snow depth on roofs, Summaries of
Technical Papers of Annual Meeting, Vol. B, AIJ, pp. 51-52, 1989.10 (in Japanese).
18) Endo, A. and Tomabechi, T.: Wind channel experiment of the forming conditions of the snow depth
on various roof with model snow, Memoirs of the Hokkaido Institute of Technology, No. 11, pp.
163-178, 1983 (in Japanese with English abstract).
19) Taylor, D.A.: Roof snow loads in Canada, Canadian Journal of Civil Engineering, Vol. 7, No. 1, pp.
1-8, 1980.
20) Tsuchiya, M., Tomabechi, T., Hongo, T., and Urda, H.: Field Measurements on Snowdrift on
Two-level Flat Roof Model, Journal of Snow Engineering of Japan, Vol. 18, No. 2, pp. 13-21, 2002
(in Japanese with English abstract)
21) Kobayashi, T., Chiba, T., Tomabechi, T. and Hoshiba, S.: Study on evaluation of snow load on roof
of livestock building, Journal of Agricultural Structures of Japan, Vol. 32, No. 2, pp. 69-77, 2001 (in
Japanese with English abstract)
22) Izumi, M., Mihashi, H. and Takahashi, T.: Statistical properties and regional characteristics of
annual maximum increasing intensity of snow depth, Journal of Structural and Construction
Engineering (Transactions of AIJ), No. 392, pp. 68-77, 1988.10 (in Japanese with English abstract).
23) Takahashi, H. and Nakamura, T.: Prevention of Snow and Ice Disaster, Hakuasyobou, 1986.
24) Takahashi, T.: Study on equivalent unit weight of annual maximum snow accumulation in n-days –
Estimation with snow layer model, Proceedings of Japan Society for Snow Engineering, Vol. 20, pp.
25-26, 2003 (in Japanese).
25) Morino, K., Kobayashi, M., Takebayashi, Y., Kawai, M., Kawashima, M.: A snow melting
experiment on a pneumatic structure (air supported dome) No. 2 a snow melting experiment by a
wind box, Summaries of Technical Papers of Annual Meeting, Planning division, AIJ, pp.789-790,
1984.10 (in Japanese).
26) Nishi, Y., Nishikawa, K., Ishii, H.: Analysis of snow melting process of the air supported structure
-2, Summaries of Technical Papers of Annual Meeting, Vol. D, AIJ, pp. 895-896, 1986.8 (in
Japanese).
27) Ohtsuka, K., Joh, O., Homma, Y., Miyagawa, Y., Masumo, T., and Okada, H.: A Study on the
removal of snow from membrane structures, Proc. of Membrane Structures Association of Japan,
No. 4, pp. 55-68, 1990 (in Japanese with English abstract).
28) Tomabechi, T., Yamaguchi, H., Ito, T., Hoshino, M.: Fundamental study on sliding of snow on the
roof of membrane structure, Journal of Structural and Construction Engineering (Transactions of
AIJ), No. 426, pp. 99-105, 1991 (in Japanese with English abstract).
29) Kubodera, I.: Side pressure of snow on buildings, Kenchiku Gijutsu, No. 384, pp. 203-204, 1983 (in
Japanese).
30) Matsushita, S., Nakajima, H. and Izumi, M.: A study on side pressure of snow, Transactions of the
CHAPTER 5 SNOW LOADS - C5-39 -

AIJ, No. 89 (Summaries of Technical Papers of Annual Meeting) p. 82, 1963.9 (in Japanese).
31) Institute of Scientific Approaches for Fire and Disaster: Part of snow disaster, Overview of regional
disaster prevention data, p. 199, 1986 (in Japanese).
32) Shimomura, T., Yamamoto, M., and Ishihira, S.: Characteristic of snow pressure to horizontal girder,
Civil Engineering Journal, Vol. 28, No. 2, pp. 70-75, 1986 (in Japanese).
- C6-1 -

CHAPTER 6 WIND LOADS

Introduction

Each wind load is determined through a probabilistic-statistical method based on the concept of
“equivalent static wind load,” under the assumption that structural frames and components/cladding
behave elastically in a strong wind.
The mean wind force based on the mean wind speed and the fluctuating wind force based on a
fluctuating flow field usually act on a building. The effect of a fluctuating wind force on a building or
its parts depends not only on the characteristics of the fluctuating wind force itself but also on the object
size and vibration characteristics of the building or its parts. These recommendations evaluate the
maximum loading effects on a building and its parts from a fluctuating wind force through a
probabilistic-statistical method, and calculate the static wind load that provides the equivalent effect.
The design wind load can be obtained from the sum of this equivalent static wind load and the mean
wind load.
A suitable wind load calculation method corresponding to the scale, shape, and vibration
characteristics of the design object is provided here. The wind load is classified into a horizontal wind
load for structural frames, roof wind load for structural frames, and wind load for components/cladding.
Furthermore, the horizontal wind load is classified into the along-wind load, across-wind load, and
torsional wind load. Among them, the along-wind load is calculated from the product of the velocity
pressure, gust effect factor, and projected area. On the other hand, the roof wind load for the structural
frame is evaluated as the sum of the mean and fluctuating roof wind loads.
The estimation procedures of a horizontal wind load for lattice structural frames that stand upright
from the ground, and the wind load on free roofs, differ from those for ordinary buildings owing to the
difference in the acting mechanism of the fluctuating wind forces. For buildings with a large aspect ratio,
the across-wind load and torsional wind load shall be estimated with consideration of the effects of the
vortex shedding.
The present recommendations introduce a wind directionality factor for calculating the design wind
speed for each individual wind direction, thereby making it possible to design a building rationally by
considering its orientation with respect to the wind direction. The topography factor for turbulence
intensity is provided to consider the increase in fluctuating wind load owing to the increase in fluctuating
wind speed. Furthermore, the wind seasonal factor is newly introduced to evaluate the rational design
wind speed for buildings that are used only during a limited period of time.
The recommendations adopt provisions for a combination of wind loads in the across-wind and
along-wind directions, or in torsional and along-wind directions explicitly. Furthermore, a prediction
formula for the response acceleration of the building to evaluate its habitability to vibrations, which is
needed in a performance design, as well as information of a 1-year-recurrence wind speed are provided.
- C6-2 - Recommendations for Loads on Buildings

The revisions of the specifications originally made in the 2015 version may be summarized as
follows:
1. The wind seasonal factor has been introduced to evaluate the rational of the design wind speed
for buildings used only during a limited time period.
2. It has been clearly indicated that computational fluid dynamics (CFD) can be used for evaluating
the design wind speed when considering the effects of topography as well as the wind force and
wind pressure coefficients.
3. The roof wind load has been provided by the sum of the mean and the fluctuating roof wind
loads.
4. The wind force coefficients for buildings with rectangular plans have been specified based on
the aspect ratio rather than on the height.
5. The external wind pressure coefficients for structural frames and the peak external pressure
coefficients for components/cladding have been specified for buildings with hipped roofs, as
well as for buildings with an elliptical plan.
6. Some wind force coefficients and external pressure coefficients have been revised based on
recent studies.
7. The calculation formula of the torsional wind load for structural frames has been revised such
that it is based on the overturning moment at the base of the building.
8. A calculation formula of the wind load for free roofs has been introduced.

Sections 6.1 through 6.4 provide the most basic concepts of the wind load estimation procedure, and
Sections A6.1 through A6.15 provide supplementary provisions.

Notation
Notations used in this chapter are shown here.

Uppercase Letter
A (m2): projected area of building at height Z or of part thereof at length x from its end
AR (m2): subject area of roof beam
AC (m2): subject area of components/cladding
A0 (m2): whole plane area of one face of lattice structure
AF (m2): projected area of one face of lattice structure
B (m): building width
B1 (m): building length in the span direction
B2 (m): building length in the ridge direction
B0, BH (m): width of lattice structure at ground level and height H
BD: background excitation factor for lattice structure
C1, C2, C3: parameters determining topography factor Eg
CHAPTER 6 WIND LOADS - C6-3 -

CD, CR, CF: wind force coefficients

C, CX, CY: wind force coefficients of members
C'L, C'T: rms overturning moment coefficient and rms torsional moment coefficient
Ce: exposure factor, which is generally 1.0 and shall be 1.4 for open terrain with few obstructions
(Category II). When the wind speed is expected to increase owing to the local topography, this
factor shall be increased accordingly.
Cg: overturning moment coefficient in the along-wind direction
C'g: rms overturning moment coefficient in the along-wind direction
Cf : wind force coefficient. For horizontal wind loads, the wind force coefficient CD defined in
A6.2 with kz = 0.9 shall be used. For roof wind loads, the wind force coefficient CR defined in
A6.2 shall be used.
Cpe: external pressure coefficient
Cpe1, Cpe2: external pressure coefficients on windward and leeward walls
Cpi: internal pressure coefficient
C*pi: factor for effect of fluctuating internal pressure
Cr: wind force coefficient at resonance
Ĉ C : peak wind force coefficient
Ĉ pe : peak external pressure coefficient
C’R: rms generalized wind force coefficient of roof beam
D (m): building depth, building diameter, and member diameter
D1, D2 (m): major and minor axis of a section
DB (m): building diameter at the base
Dm (m): building diameter at a height of 2H/3
E: wind speed profile factor
EH: wind speed profile factor at the reference height H
EI: topography factor for the standard deviation of the fluctuating wind speed
Eg: topography factor for the wind speed
EgI: topography factor for the turbulence intensity
Er: exposure factor for flat terrain categories
FD: along-wind force spectral factor
F: wind force spectrum factor
GD: gust effect factor for along-wind load
GR: gust effect factor for estimating the wind loads on the structural frames of free roofs
H (m): reference height
HS (m): height of topography
IT (kgm2): generalized inertial moment of building for torsional vibration
IZ: turbulence intensity at height Z(m)
IZp: turbulence intensity at height Zp(m)
- C6-4 - Recommendations for Loads on Buildings

IrZ: turbulence intensity at height Z for flat terrain categories

L (m): span of roof beam
LS (m): horizontal distance from top of the topography to the point at which the height is half the
topography height
LZ (m): turbulence scale at height Z
M (kg): total building mass
MD (kg): generalized mass of building for along-wind vibrations
ML (kg): generalized mass of building for across-wind vibrations
R: factor expressing correlation of wind pressure of the windward and leeward sides
RD: resonance factor for along-wind vibration
RL: resonance factor for across-wind vibration
RT: resonance factor for torsional vibration
RR: resonance factor for a roof beam
SD: size effect factor
U0 (m/s): basic wind speed
U1 (m/s): one-year recurrence of 10-min mean wind speed at 10 m above the ground over flat and
open terrain
U500 (m/s): 500-year recurrence of 10-min mean wind speed at 10 m above the ground over flat
and open terrain
UH (m/s): design wind speed
U*: non-dimensional wind speed
UtR (m/s): tR-year-recurrence of 10-min mean wind speed at 10 m above ground over flat and open
terrain
U*Lcr, U*Tcr: non-dimensional critical wind speed for aeroelastic instability in across-wind and
torsional directions
Ur (m/s): resonance wind speed
U*T: non-dimensional wind speed for calculating torsional wind load
Ur: resonance wind speed at the mean height of a member
WD (N): along-wind load at height Z
WL (N): across-wind load at height Z
WT (Nm): torsional wind load at height Z
WLC (N): across-wind combination load
Wr (N): wind load at height Z or at length x from its end
W’R (N): fluctuating roof wind load
XS (m): distance from leading edge of topography to construction site
Z (m): height above the ground
Zb, ZG (m): parameters determining exposure factor
CHAPTER 6 WIND LOADS - C6-5 -

Lowercase Letter
aDmax, aLmax, aTmax, (m/s2), (rad/s2): maximum response acceleration in the along-wind, across-
wind and torsional directions at the top of a building
b (m): subject or projected width of a member
f (m), (Hz): rise or frequency
f1 (Hz): smaller frequency of fL and fT
fD, fL, fT (Hz): natural frequency for the first mode in the along-wind, across-wind, and torsional
directions
fR (Hz): natural frequency for first mode of a roof beam
gaD, gaL, gaT: peak factors for response accelerations in the along-wind, across-wind, and torsional
directions
gD, gL, gT: peak factors for wind loads in the along-wind, across-wind, and torsional directions
gR: peak factors for fluctuating roof wind load
h (m): eave height
k1: factor for aspect ratio
k2: factor for surface roughness
k3: factor for D1/D2 for structural frame
k4: factor for end effects
k5: factor for D1/D2 of components/cladding
k6: factor for local negative pressure areas
kC: area reduction factor
kRW: return period conversion factor
kZ: factor for vertical profile for wind pressure coefficients or wind force coefficients
l (m): smaller value between 4H and B, minimum value among 4H, B1, and B2, or member length
la1 (m): smaller value between H and B1
la2 (m): smaller value between H and B2
lb1 (m): smaller value between 4H and B1
lb2 (m): smaller value between 4H and B2
qH (N/m2): velocity pressure at reference height H
qZ (N/m2): velocity pressure at height Z
tR(year): design return period
x (m) : distance from the end of the roof beam or member
y (m) : distance from the leading edge of roof beam normal to the wind direction

Greek Alphabet
α : exponent of power law for wind speed profile
β : exponent of power law for vibration mode
γ : load combination factor
- C6-6 - Recommendations for Loads on Buildings

δL, δT: mass damping parameter

φD, φL, φT : mode correction factor
ζD, ζL, ζT: damping factor for first translational and torsional modes
ζR : damping factor for first mode of a roof beam
ϕ : solidity
λ : mode correction factor of generalized wind force
µ : mode shape
νD (Hz): level crossing rate
θ (°): roof angle, angle of attack to member
θ S(°): inclination of topography
ρ (kg/m3): air density
ρS (kg/m3): building density
ρLT : correlation coefficient between across-wind and torsional vibrations

6.1 General

6.1.1 Scope of application

(1) Target strong wind
Most wind damage to buildings occurs during strong winds. The wind loads specified in the present
recommendations are applied to the design of buildings to prevent failure from strong winds. The strong
winds that occur in Japan are mainly associated with tropical or extratropical cyclones, and down-bursts
or tornados. The former are large-scale phenomena that are spread over about 1000 km in a horizontal
plane, and their characteristics are understood relatively well. However, the maximum wind speed
induced by such a phenomenon is not necessarily picked up at each meteorological observatory owing
to the limited meteorological observation network. Therefore, the minimum value of the basic wind
speed is set to 30 m/s.
Downbursts are gusts of descending airflows caused by severe rainfall in developed cumulonimbus.
Because the scale of these phenomena is quite small, few bursts are picked up by the meteorological
observation network. Tornados are also small-scale phenomena, several hundred meters wide at most,
having a rotational wind with a rapid atmospheric pressure descent. The characteristics of the strong
winds and pressure fluctuations caused by tornados are not well known. Although the number of
occurrences of downbursts and tornados is relatively large, their probability of attacking a particular site
is very small compared with that of tropical or extratropical cyclones. However, the winds caused by
downbursts and tornados are very strong, causing fatal damage to buildings. The present
recommendations focus on strong winds caused by tropical or extratropical cyclones. However, the
minimum wind speed takes into account the influence of tornadoes and downbursts to some degree.
CHAPTER 6 WIND LOADS - C6-7 -

(2) Wind loads on structural frames and wind loads on the components/cladding
The wind loads provided in the recommendations are classified into those for structural frames and
those for components/cladding. The former are used for the design of structural frames such as columns
and beams. The latter are used for the design of finishing and support members as components/cladding
and their joints. The characteristics of the wind loads on the structural frames are different from those
on the components/cladding because there are large differences in their size, dynamic characteristics,
dominant phenomena, and response to wind loading. Wind loads on structural frames are estimated
based on the elastic response of the entire building to fluctuating wind forces. Wind loads on the
components/cladding are estimated based on fluctuating wind forces acting on a small part.
A wind resistant design for the components/cladding has not been appropriately made until now.
However, such a design plays an important role in protecting the interior space from destruction by
strong winds. Therefore, a wind resistant design for components/cladding should be made carefully in
the same manner as that for structural frames.

6.1.2 Estimation principle

(1) Basic value of wind load
When only the quasi-static effect of fluctuating wind loads on the structural response is considered in
the load estimation, it may be sufficient to provide the velocity pressure as the basic value of the wind
loads. However, when dynamic effects such as wind-induced vibrations should be considered, the wind
speed may be more useful for understanding the phenomena than the velocity pressure. Therefore, herein
the 100-year recurrence wind speed specified in A6.1.3 is used as a basic value for evaluating the wind
loads.

(2) Return period conversion factor

The return period conversion factor is used when estimating the design wind speed for a return period
of other than 100 years. The design wind speed is provided by the product of the basic wind speed and
the return period conversion factor. Considering that the return period used in the wind resistant design
is usually 50 to 1000 years, the return period conversion factor is provided by a function of the ratio of
the 500-year recurrence wind speed shown in Fig.A6.2 to the basic wind speed. The formula is derived
from an assumption that the annual maximum wind speed follows the Gumbel distribution.

(3), (4) Classification of wind load

Mean wind forces act on a building. This mean wind force is derived from the mean wind speed,
while the fluctuating wind force is produced by the fluctuating flow field. The effect of the fluctuating
wind force on a building or its parts depends not only on the characteristics of the fluctuating wind force
but also on the size and vibration characteristics of the building or its parts. Therefore, to estimate the
design wind load, it is necessary to evaluate the characteristics of fluctuating wind forces and the
dynamic characteristics of the building.
- C6-8 - Recommendations for Loads on Buildings

The fluctuating wind forces are generally generated through the following factors:
i) wind turbulence (temporal and spatial fluctuation of wind),
ii) vortex generation in building wake,
iii) interaction between building vibration and surrounding air flow.
Figure 6.1.1 shows a schematic illustration of the mechanism of fluctuating wind loads on a building
owing to the turbulence of the approachong flow and the vortex shedding from the building.
Fluctuating wind pressures act on the building surfaces owing to the above-mentioned factors. The
fluctuating wind pressures change both temporally and spatially, and their dynamic characteristics are
not uniform at all positions on the building surface. Therefore, it is more reasonable to evaluate the wind
loads on structural frames based on the overall building behavior and that on the components/cladding
based on the behavior of the individual building parts. For most buildings, the effects of fluctuating wind
forces generated by wind turbulence are predominant. In this case, horizontal wind load on structural
frames in the along-wind direction is important. However, for relatively flexible buildings with a large
aspect ratio, horizontal wind loads on the structural frames in the across-wind and torsional directions
should be considered together with that in the along-wind direction.
Regarding the roof loads, the fluctuating wind forces caused by a flow separation at the leading edge
of the roof becomes more predominant than those by the turbulence of the approaching flow. Therefore,
the roof wind load on the structural frames may be provided in a different way from that for the
horizontal wind load (see Fig. 6.1.2).

a) Fluctuating wind force caused by wind b) Fluctuating wind force caused by vortex
turbulence generation in wake of building

Figure 6.1.1 Fluctuating wind forces based on wind turbulence and vortex generation in the wake of a
building
CHAPTER 6 WIND LOADS - C6-9 -

along-wind load

horizontal wind load across-wind load

wind load on
structural frames torsional wind load
roof wind load
wind load
wind load on
components/cladding

Figure 6.1.2 Classification of wind loads

(5) Vortex-induced vibration and aeroelastic instability

Vortex-induced vibration and aeroelastic instability can occur in the case of flexible buildings or
structural members with very large aspect ratios. Therefore, the criteria for across-wind and torsional
vibrations are provided for buildings with rectangular plans. The criteria for vortex-induced vibrations
are also provided for buildings and structural members with circular sections. When it is expected that
vortex-induced vibration or aeroelastic instability will occur, the structural safety should be confirmed
through wind tunnel tests and so on. Regarding the buildings or structural members with circular sections,
a formula is provided for the wind load caused by vortex-induced vibrations.

(6) Horizontal wind load on structural frames

The maximum loading effect on each part of the building can be estimated through a dynamic
response analysis considering the characteristics of temporal and spatial fluctuating wind pressures and
the dynamic characteristics of the building. The equivalent static wind load producing the maximum
loading effect is given as the design wind load. Regarding the response of the building to strong winds,
the first mode becomes predominant, and the effects of higher frequency modes are not significant for
most buildings. Therefore the horizontal wind load (along-wind load) distribution for structural frames
is assumed to be similar to the mean wind load distribution. In practice, the equivalent wind load is
provided by the product of the mean wind load and gust effect factor, which is defined as the ratio of
the instantaneous maximum value to the mean value of the building response.
By comparison, it is difficult to predict the responses in the across-wind and torsional directions
theoretically in the same manner as that for the along-wind responses. Therefore, prediction formulas
are provided to the wind loads in the across-wind and torsional directions based on the fluctuating
overturning moment in the across-wind direction and the fluctuating torsional moment for the first
vibration mode, respectively, which are obtained from wind tunnel tests.

(7) Roof wind load on structural frames

The characteristics of the wind force acting on the roof are influenced by the features of the fluctuating
wind force caused by a separated flow at the leading edge of the roof. Therefore, the characteristics of
the roof wind load on the structural frames are different from those of the along-wind load. Thus, the
- C6-10 - Recommendations for Loads on Buildings

roof wind load on structural frames cannot be evaluated using the same procedure as that for the along-
wind load. In the present recommendations, the roof wind load on the structural frames is given as the
sum of the mean and fluctuating roof wind loads. The fluctuating roof wind load is evaluated based on
the dynamic behavior of the elastic beams supporting the roof, assuming that the first vibration mode is
dominant.

(8) Wind load on components/cladding

To evaluate the wind loads on the components/cladding, the peak external wind pressure coefficient
and the factor for the effect of the fluctuating internal pressure are provided. The peak wind force
coefficient is calculated as the difference between them. Only the size effect is considered. The
resonance effect is ignored because the natural frequency of the components/cladding is generally so
high that the resonance may not be excited. The wind load on the components/cladding is prescribed as
the maximum positive and negative pressures for each part of the components/cladding regardless of
the wind direction, whereas the wind load on the structural frames is prescribed for the wind direction
normal to the building face. Therefore, for the wind load on the components/cladding, the peak wind
force coefficient or the peak external wind pressure coefficient shall be obtained from wind tunnel tests
or another verification method.

(9) Wind directionality factor

Wind speeds at a construction site have different characteristics in frequency and intensity for each
wind direction owing to the geographical position, influences of a large-scale topography, and so on.
Furthermore, the characteristics of the wind forces acting on a building depend on the wind direction. If
the directional characteristics of both the wind speeds and wind forces acting on a building are
considered, a rational design can be made by appropriately arranging the orientation of the building.
Therefore, a wind directionality factor is introduced for calculating the design wind speed for each wind
direction, which makes it possible to design the building with consideration of wind directionality at the
construction site. To set the wind directionality factor, it is necessary to fully consider the influence of
a typhoon, which is the main cause of strong winds in Japan. However, because the frequency of attack
of typhoons striking a certain point is not very high, it is difficult to quantify the directionality of
probability distributions of strong winds by typhoons only from the records of the Japan Meteorological
Agency for a period of 70 years at most. Therefore, to specify the wind directionality factor, a Monte
Carlo simulation with a typhoon model, and a statistical analysis of meteorological records, are applied
for typhoon winds and non-typhoon winds, respectively.

(10) Wind seasonal factor

Some buildings and structures are used only during certain seasons, and are removed during the other
seasons. In areas where the main cause of strong winds is a typhoon, the design wind speed can be
reduced for buildings and structures that are used only for non-typhoon seasons. The wind seasonal
CHAPTER 6 WIND LOADS - C6-11 -

factor is introduced to evaluate such design wind speeds. In the present recommendations, the wind
seasonal factor for winter (December through March) is specified. The wind seasonal factor cannot be
used together with the wind directional factor simultaneously.

(11) Reference height and velocity pressure

The reference height is generally the mean roof height of the building, as shown in Fig. 6.1.3. The
wind loads are calculated from the velocity pressure at this reference height. The vertical distribution of
the wind load is reflected in the wind force and pressure coefficients. However, the wind load for a
lattice type structure shall be calculated from the velocity pressure at each height.

Figure 6.1.3 Definition of reference height and velocity pressure

(12) Wind load on free roofs

For relatively small free roofs or fences on the ground, the fluctuating wind loads on these structures
are mainly caused by wind speed fluctuations. In such a case, it is thought that the maximum loading
effect occurs at an instant of the maximum wind speed. Thus, the wind load can be estimated by using
a gust effect factor approach, in which the gust effect factor is specified as a function of the turbulence
intensity of an approaching flow.

(13) Combination of wind loads

The wind loads on structural frames are provided for the along-wind, across-wind load, and torsional
directions, independently, where the wind direction is normal to a building face. In practice, however,
these wind loads act on the building not independently but interact with each other. Therefore, these
wind loads shall be combined in the wind resistance design. The methods used to combine these loads
rationally are provided in these recommendations.

(14) Application of CFD to the wind load estimation

When the CFD technique is applied to the evaluation of the wind loads, it is necessary to obtain
accurate numerical results to adopt an appropriate numerical method that can represent the practical
wind characteristics. Because the computations deal with phenomena regarding the engineering scale of
the buildings and their surroundings, the momentum equations and the continuity equation for the
- C6-12 - Recommendations for Loads on Buildings

unsteady flows the incompressible fluids are employed. To avoid the unfavorable effects of numerical
errors on the wind load estimation, it is necessary to select an appropriate discrete approximate solver
and turbulence model adequately representing the wind flow fluctuations.

(15) Effect of neighboring buildings

When groups of two or more tall buildings are constructed in proximity to each other, the wind flow
through the groups may be deformed significantly, generating much more complex effects than is
usually acknowledged, resulting in higher dynamic pressures and motions, especially on neighboring
downstream buildings.

(16) Shielding effect by surrounding topography or buildings

When there are topographical features and buildings that block the wind around a construction site,
wind loads or wind-induced vibrations may be reduced by the shielding effect. In this case, a rational
wind resistant design can be achieved by considering this shielding effect. However, when it is expected
that the area surrounding the construction site will change during the service life of the building, the
shielding effect should not be taken into account. Furthermore, the shielding effect should be
investigated through a careful wind tunnel study or the other suitable verification method because it is
generally complicated and cannot be easily analyzed.

(17) Assessment of building habitability

Building habitability against wind-induced vibration is usually evaluated based on the maximum
response acceleration for a 1-year recurrence wind speed. Hence, a map of a 1-year recurrence wind
speed is provided based on the daily maximum wind speed observed at meteorological stations as well
as using a calculation method for the response acceleration.

(18) Duration of wind speed

When considering the cumulative load effect such as fatigue damage, information on the duration of
the wind speed is required. To evaluate the duration, not only extreme values such as the annual
maximum wind speed targeted at the design wind speeds but also statistics on the occurrence frequency
of the medium and weak winds, as well as the time histories of the strong winds occurring during rare
events, are required. In addition, depending on the buildings, structures, and components to be designed,
vortex-shedding excitation may occur at wind speeds lower than the design wind speed. Indeed, fatigue
damage in a relatively short period of time has been reported. When such cumulative effects of
fluctuating wind loads are considered, the service period and maintenance cycle of the target members
should be taken into account as well.
The procedure for wind load estimation, as summarized above, is shown in Fig. 6.1.4.
CHAPTER 6 WIND LOADS - C6-13 -

Start
A6.1.1 Velocity pressure
A6.1.2 Design wind speed Outline of building
A6.1.3 Basic wind speed
A6.1.4 Return period conversion factor
A6.1.5 Wind directionality factor A6.1 Wind speed and velocity pressure
A6.1.6 Wind seasonal factor
A6.1.7 Wind speed profile factor
A6.1.8 Turbulence intensity and turbulence scale 6.1.3 Buildings to be designed for particular
A6.1.9 Effect of topography wind load or wind induced vibration
(1) across-wind, torsional wind loads
(2) vortex induced vibration, aeroelastic instability Wind tunnel test
Wind tunnel test or CFD A6.12 Effects of neighboring tall buildings or CFD

Wind load on structural frames Wind load on components/cladding

A6.2.1 Procedure for estimating wind force coefficients

A6.2.5 Peak external pressure coefficients

A6.2.2 External wind pressure coefficients
A6.2.6 Factor for the effect of fluctuating
A6.2.3 Internal pressure coefficients
internal pressures
A6.2.4 Wind force coefficients
A6.2.7 Peak wind force coefficients

A6.3 Gust effect factor for A6.4 Fluctuating roof Wind tunnel test
along-wind loads wind loads or CFD

6.2 Horizontal wind loads 6.3 Roof wind loads 6.4 Wind loads on components/cladding

A6.5 Across-wind loads

A6.6 Torsional wind loads

A6.10 Combination of wind loads

A6.7 Horizontal wind loads on lattice structural frames

A6.14 One-year recurrence wind speed
A6.8 Wind loads on free roofs

A6.9 Vortex induced vibration

A6.13 Response acceleration

End

Figure 6.1.4 Procedure for wind load estimation

6.1.3 Buildings for which particular wind loads or wind induced vibrations are taken into account
(1) Buildings for which horizontal wind loads on the structural frames in the across-wind and torsional
directions need to be taken into account
Horizontal wind loads on structural frames imply the along-wind, across-wind, and torsional wind
loads. Both the across-wind and torsional wind loads shall be estimated for wind-sensitive buildings that
satisfy Eq. (6.1). Figure 6.1.5 shows the definition of the wind direction, the three components of the
wind load, and the building shape.
Both across-wind and torsional vibrations are caused mainly by vortices generated in the building’s
wake. These vibrations are not too high for low-rise buildings. However, with an increase in the aspect
ratio caused by the presence of high-rise buildings, more periodic vortices with uniformity in the vertical
direction are shed behind the building, thereby increasing the across-wind and torsional wind forces on
the building. At the same time, the natural frequency of the building decreases and approaches the vortex
- C6-14 - Recommendations for Loads on Buildings

shedding frequency with an increase in building height. As a result, resonance components increase and
building responses become large. In general, as the wind speed increases, the responses in the across-
wind and torsional directions increase more rapidly than the responses in the along-wind direction.
Under normal conditions, the along-wind responses to low wind speeds are larger than the across-wind
responses. However, the across-wind responses to higher wind speeds are larger than the along-wind
responses. The wind speed at which the across-wind response exceeds the along-wind response depends
on the height, shape, and vibration characteristics of the building. The conditions for the aspect ratio
given by Eq. (6.1) have been established through an investigation into the variation of along-wind and
across-wind loads with height H (m) for buildings with a density of 180 kg/m3, assuming that the basic
wind speed is U0 = 40 m/s, the terrain category is II, the natural frequency of the first mode is given by
f1 = 1 /(0.024 H ) (Hz), and the damping factor is 1%. Therefore, it is desirable to estimate the across-
wind and torsional wind loads for buildings with lightweight and small damping factors, even if Eq.
(6.1) is not satisfied.
Furthermore, for buildings with large side ratios and low torsional rigidity, or buildings with large
eccentricity whose natural frequencies in the translational and torsional directions are close to each other,
it is also desirable to estimate the torsional wind loads even if Eq. (6.1) is not satisfied.
The criteria given by Eq. (6.1) was derived for buildings with rectangular plans. They can be applied
to a building with a plan slightly different from rectangular when regarding B (m) and D (m) as the
projected width and depth of the building. When the values of B (m) and D (m) change in the vertical
direction, the wind force acting on the upper part has a larger effect on the building’s response. Therefore,
representative values for the upper part should be used for the computation. Under normal conditions,
the values at a height of (2/3)H can be used in most cases. The computation by Eq.(6.1) using smaller
values of B and D for the upper part yields conservative values of the wind load.

Figure 6.1.5 Definition of load and wind direction

CHAPTER 6 WIND LOADS - C6-15 -

(2) Vortex-induced vibration and aeroelastic instability

Vortex-induced vibration and aeroelastic instability such as galloping and flutter might occur in
buildings with low natural frequency and a large aspect ratio ( H / BD or H / Dm ) as well as in
slender members. The conditions for estimating the aeroelastic instability in both across-wind and
torsional directions for buildings with rectangular plans, as well as the conditions for an estimation of
vortex-induced vibration for buildings with circular plans are given based on the results of wind tunnel
tests and field measurements1-8). The method for estimating the wind loads on buildings with circular
plans during vortex-induced vibration is shown in A6.9. Vortex-induced vibration and aeroelastic
instability may occur in slender buildings with triangular and elliptical plans. When designing these
buildings, attention should be paid to such phenomena.
The first condition required for estimating the aeroelastic instability and vortex-induced vibration is
provided for the aspect ratio ( H / BD or H / Dm ). Aeroelastic instability as well as vortex-induced
vibration seldom occur in buildings with small aspect ratios. In the present recommendations, the aspect
ratio for estimating aeroelastic instability is set to 4 or higher, whereas that for estimating vortex-induced
vibrations is set to 7 or higher. The second condition is provided on the non-dimensional wind speed
( U / f BD or U / fDm ). The occurrence of aeroelastic instability and vortex-induced vibration is
dominated by the non-dimensional wind speed, defined based on the representative width and natural
frequency of the building and the wind speed. The non-dimensional critical wind speed for aeroelastic
instability depends upon the mass-damping parameter, which is defined by the mass ratio between the
building and air and the damping factor of the building, the side ratio of the building, and the turbulence
characteristics of an approaching flow. Thus, the non-dimensional critical wind speed for estimating the
aeroelastic instability of a building with a rectangular plan is provided as a function of these factors. On
the other hand, the non-dimensional wind speed for vortex-induced vibration of a building with a circular
plan is almost independent of these factors. Therefore, the non-dimensional critical wind speed is fixed
at a constant value.
The non-dimensional critical wind speeds specified in Tables 6.1 and 6.2 and Eq. (6.3) correspond
to the experimental values multiplied by a factor of 1/1.1 through 1/1.2. This is because the aeroelastic
instability or vortex-induced vibration occurs within a period shorter than 10 min, which is the
evaluation time of the design wind speed. Furthermore, considering the uncertainty in estimating the
critical wind speed of aeroelastic instability or vortex-induced vibration, the non-dimensional critical
wind speed is multiplied by 0.83 (= 1/1.2). The damping factor of the building is used for computing
the building’s mass damping parameter, and is recommended for evaluating the damping factor to refer
to “Damping in Buildings”9) and the “Guidebook of Recommendations for Loads on Buildings 2.”

6.2 Horizontal Wind Loads on Structural Frames

6.2.1 Scope of application

This section describes the horizontal wind loads on structural frames in the along-wind direction. The
- C6-16 - Recommendations for Loads on Buildings

along-wind loads are generally composed of a mean component caused by the mean wind speed, a quasi-
static component caused by a relatively low-frequency fluctuation, and a resonant component caused by
a fluctuation within the vicinity of the natural frequency. For many buildings, only the first mode is
taken into account as the resonant component. The procedure described in this section can be used to
estimate the equivalent static wind loads producing the maximum structural responses (load effects of
the stress and displacement) using the gust effect factor. The equivalent static wind loads are also divided
into the mean component, quasi-static component, and resonant component. Although the vertical
profiles for these components differ from each other, it was assumed that the profiles are similar to that
of the mean component.

6.2.2 Estimation method

Equation (6.4) for horizontal wind loads is derived from the gust effect factor method, which includes
the effect of the along-wind dynamic response owing to the atmospheric turbulence of the approaching
wind. The gust effect factor is a magnifying rate of the maximum instantaneous value to the mean
building responses. Davenport, who first proposed the gust effect factor, calculated this factor based on
the displacement at the top of a building10). However, in these recommendations, the gust effect factor
based on the overturning moment at the base of a building,11) which can rationally estimate the design
wind loads on the building, has been employed. The projected area is the area of the portion of concern
projected to a plane normal to the wind direction, as shown in Fig. 6.2.1. For wind loads at a unit height
being taken into account, the projected area A(m2) has a certain value of the building width B(m).

Wind

Figure 6.2.1 Projected area A(m2)

6.3 Roof Wind Loads on Structural Frames

6.3.1 Scope of application
Roof wind loads on structural frames should be estimated from the load effects of wind forces that
act on the roof frames. Because the external pressures on the roof are affected strongly by the behavior
CHAPTER 6 WIND LOADS - C6-17 -

of the separated shear layers from the leading edges, and are not directly related to the wind velocity
fluctuations for the along-wind loads, the roof wind loads on the structural frames cannot be evaluated
through the same procedure shown in 6.2. This section describes the procedures to be applied to the roof
frames of buildings with rectangular plans without dominant openings, where the correlation between
fluctuating external pressures and fluctuating internal pressures can be ignored. The procedures
specified in this section cannot be applied to the lightweight roofs with a low stiffness to which a motion-
induced force might be important.
Because a lightweight roof with a low stiffness, such as a hanging roof that satisfies all of the
following conditions, might be aerodynamically unstable12),13), it must be checked whether the
aerodynamic instability does not occur within the design wind speed through wind tunnel tests or CFD.
m
1) <3
ρL
UH
2) > 0.7
f R1 L
3) I H < 0.15
where
m (kg/m2) : roof mass per unit area,
ρ (kg/m3) : air density,
L (m) : span length,
U H (m/s) : design wind speed,
f R1 (Hz) : frequency of first unsymmetrical vibration mode,
IH : turbulence intensity at reference height H(m).

In addition, even if the above conditions are not satisfied, a large amplitude vibration may occur on
large-span lightweight roofs, which seems to weaken the stiffness of the structural frames. In these cases,
wind tunnel tests or CFD must be carried out.

6.3.2 Procedure for estimating wind loads

Roof wind loads on structural frames can be estimated based on the sum of the mean wind load and
equivalent static wind loads, which provides the maximum load effects caused by the dynamic roof
responses. Because the distribution of the equivalent static wind load is not similar to that of the mean
wind load, these loads shall be estimated separately.

6.4 Wind Loads on Components/Cladding

6.4.1 Scope of application

Wind loads on components/cladding are needed to design parts of buildings subjected to the strong
effects of intensive wind pressures, i.e., the finishing of roofs and external walls; supporting members
- C6-18 - Recommendations for Loads on Buildings

such as purlins, furring strips, and studs; roof braces; and tie beams. These wind loads are also applied
to the design of eaves and canopies.

6.4.2 Procedure for estimating wind loads

Wind loads on components/cladding are derived from the difference between the wind pressures
acting on the external and internal surfaces of the building, and are calculated using Eq. (6.6). The peak
wind force coefficients ĈC correspond to the peak values of fluctuating net pressures, defined by the
difference between the external and internal pressures, are given through Eq. (A6.15) for convenience.
For buildings with free-standing canopy roofs, where the top and bottom surfaces are both exposed to
wind, the peak wind force coefficients ĈC are derived directly from the actual peak values of the
pressure differences, as shown in A6.2.7.
The external pressure coefficients provided in the recommendations correspond to the most critical
positive and negative peak pressures on each part of the building irrespective of the wind direction.
Therefore, when the wind loads are calculated by considering the directionality of the wind speeds, the
peak pressure or force coefficients for each wind direction are needed, which should be determined from
appropriate wind tunnel experiments or some other methods15, 16).
The subjected area AC (m2) depends on the subjects to be designed. When designing the finishing of
roofs and external walls, the supported area of the finishing is used, and when designing the supports of
the finishing, the tributary area of the supports is used for the subjected area, AC (m2).

A6.1 Wind speed and velocity pressure

A6.1.1 Velocity pressure

The velocity pressure, which represents the kinetic energy per unit volume of the air flow, is the basic
variable determining the wind load on a building. It corresponds to the increase in pressure from the free
stream (atmospheric ambient static pressure) at the stagnation point on the windward face of the building,
and is defined as (1 / 2)ρU 2 (N/m2), where the air density is ρ (kg/m3) and the wind speed is U(m/s).
It will be sufficient to consider the velocity pressure as the basic value of wind load when the static
effects of the wind are examined. However, it is more appropriate to adopt the wind speed as the basic
value when the dynamic wind effects are under discussion. Thus, the wind speed is adopted in the
recommendations as the basic value for calculating the wind loading. The design velocity pressure, q H
(N/m2), which is based on the design wind speed U H (m/s) at the reference height H (m), is defined in
Eq. (A6.1). In addition, a structural frame such as a lattice shaped tower-type structure requires caution
when calculating the velocity pressure q Z (N/m2) from the wind speed U Z (m/s) at height Z(m)
The air density ρ varies with temperature, ambient pressure, and humidity. However, the influence
of humidity is usually neglected. In these recommendations, the air density is taken as ρ = 1.22 (kg/m3),
which corresponds to the value at a temperature of 15°C and an ambient pressure of 1013 hPa.
CHAPTER 6 WIND LOADS - C6-19 -

A6.1.2 Design wind speed

The wind speed at a construction site is a function of its geographical location, orography, or large-
scale topographic features (e.g., mountain ranges and peninsulas), as well as the ground surface
conditions (e.g., the size and density of obstructions such as buildings and trees) and small-scale
topographic features (e.g., escarpments and hills). The height above the ground level is also a factor.
Among these factors, the geographical location and large-scale topographical features are reflected in
the values of he basic wind speed U 0 and wind directionality factor K D . The influences of the surface
roughness, small-scale topographical features, and elevation are reflected in the wind speed profile
factor E H . In addition, the basic wind load corresponds to the 100-year recurrence wind speed, which
is calculated by applying k rW = 1 . If a different recurrence interval is used for calculating the design
wind speed, the recurrence interval conversion factor k rW can be used.
The wind directionality factor K D makes the design more rational by considering the dependence
of the wind speed, the frequency of occurrence of an extreme wind, and the aerodynamic property on
the wind tracks. The wind directionality factor K D is affected based on the frequency of occurrence
and the typhoon routes, climatological factors, large-scale topographic effects, and so on. If the design
ignores the wind directionality effects, the design wind speed U H can be calculated by applying
KD = 1 .
Seasonal factor K S is introduced to evaluate the rational design wind speeds for the structures
existing only in a particular season of the year and removed in the other seasons when considering the
seasonal change in the wind climates. The seasonal factor in these recommendations is based on an
analysis without considering the wind direction. Therefore, if the wind direction is considered, the
seasonal factor shall not be considered ( K S = 1 ).

A6.1.3 Basic wind speed

The basic wind speed U 0 is the major variable in Eq. (A6.2) for calculating the design wind speed.
The wind speed at a construction site is influenced based on the occurrence of typhoons and monsoons,
the longitude and latitude of their location, and large-scale topographical effects. The basic wind speed
reflects the effects of such factors. The value of the basic wind speed corresponds to the 100-year
recurrence of a 10-min mean wind speed at 10 m above the ground for a flat terrain (terrain category II).
Table A6.2 shows the procedure used for making the basic wind speed map. As the first step of the
procedure, data from different meteorological stations were adjusted or corrected to reduce them to a
common base considering the directional terrain roughness. Then, extreme value analyses were
conducted for mixed wind climates of typhoon and non-typhoon winds. For typhoon winds, a Monte-
Carlo simulation based on a typhoon model was also conducted for each meteorological station in Japan.
Although the analysis was conducted with consideration of the wind directionality effect, the basic wind
speed was considered as a non-directional value. Instead, the wind directionality effect was reflected in
the wind directionality factor, which is defined as the wind speed ratio for a certain wind direction to
the basic wind speed, as defined in A6.1.5.
- C6-20 - Recommendations for Loads on Buildings

The distribution of the basic wind speed is the same as that of the former recommendations (2004
version) because there are no major changes in the analytical method, and statistically robust methods
are applied, such as an extreme value analysis in mixed wind climates as an independent storm, a Monte-
Carlo simulation of a typhoon, and so on. Finally, it will be homogenized such that it is not affected by
one part of the record.
The following conditions are considered for setting the basic wind speed.
1) Data for analysis
Data on the wind speed, wind direction, and anemometer height from the Japan Meteorological
Business Support Center (daily observation climate data from 1961 to 2000, and the observation history
at metrological stations) were used for the analysis. The daily observation climate data from 1961 to
1990 and the Geophysical Review of 1951 to 1999 by the Japan Meteorological Agency were referred
to for modeling the pressure fields and typhoon tracks, respectively. For homogenization of the wind
speed records, data measured by different types of anemometers were corrected to those of propeller
type anemometers.17)
2) Evaluation of directional terrain roughness and homogenization of wind speed
The wind speed records at the meteorological stations were homogenized, that is, they were converted
into data at a height of 10 m over terrain category II by utilizing a method for evaluating the terrain
roughness from the pseudo-gust factor (ratio of daily maximum instantaneous wind speed to daily
maximum wind speed) and height of the measurement point18). The details of the method are as follows.
The pseudo-gust factors were first averaged according to the year and wind direction. The terrain
roughness category, which corresponds to the pseudo-gust factors at the height of the anemometer, was
obtained from profiles of the mean wind speeds (defined in A6.1.7) and turbulence intensity (defined in
A6.1.8). For this calculation, the terrain roughness category was treated as a continuous variable.
Historical changes in the directional terrain roughness were utilized for homogenization of the wind
speed records at meteorological stations and calibration of the wind speeds near the ground surface in
the extreme value analysis and the typhoon model.
3) Extreme value analysis in mixed wind climates
The extreme value analysis for mixed wind climates19) was applied to the extreme wind data generated
by different wind climates, for instance, typhoons and monsoons. In this method, the extreme wind
records were divided into groups and independently fitted by the extreme value distributions, and the
combined distribution was obtained by assuming the independency of each extreme distribution.
Based on the typhoon track data, the records were divided into typhoon and non-typhoon winds, that
is, if the location was within a radius of 500 km from the typhoon center, the wind climate was
considered as a typhoon, and otherwise as a non-typhoon. The wind speed data measured in a typhoon
area were analyzed through a Monte-Carlo simulation based on a typhoon model to obtain the extreme
value distribution, whereas those measured in a non-typhoon area were analyzed using the modified
Jensen & Franck method in which a wind speed of less than the highest value is also included as an
independent storm for analysis. 20)
CHAPTER 6 WIND LOADS - C6-21 -

4) Typhoon simulation technique

In Japan, typhoons are the most common wind climates generating strong winds that need to be taken
into account for a wind resistant design owing to their high wind speeds and large areas of influence.
On average, 28 typhoons occur annually, roughly 10% of which hit Japan. Typhoons occasionally do
not pass near meteorological stations, and thus severe wind damage may occur without high wind speeds
being observed. To improve the instability of the statistical data (sampling error), a typhoon simulation
method was adopted for evaluating the strong wind caused by typhoons.
The results obtained from typhoon and non-typhoon conditions were combined to evaluate the return
period of annual maximum wind speed.
5) Map of basic wind speed
The contour line of a 100-year recurrence wind speed was somewhat complicated even though the
data obtained in 4) had been homogenized according to the surface roughness, wind direction, and other
factors. This may be due to the influences of local topography and structures surrounding the
metrological station, and the applicability of the homogenization models. To remove such local effects,
spatial smoothing was conducted.
In addition, the lower limit of the wind speed was set to 30 m/s, considering the uncertainty weather
factors, which cannot be covered by records from a meteorological station or typhoon model. An
additional evaluation will be necessary for a consideration of the effects of tornedos and downbursts.

A6.1.4 Recurrence period conversion factor

The annual maximum wind speed U tR (m/s) of recurrence period t R (year) are determined
through a monotone increasing function to the recurrence period. These recommendations assume that
this relation follows Gumbel distribution:
t R = [1 − exp{− exp(− a(U tR − b ))}]
−1 (A6.1.1)
where
a (s/m) and b(m/s) are factors depending on the location.
This equation is used to determine the recurrence period conversion factor (the ratio of a tR-year
recurrence wind speed to the basic wind speed) for determining the annual maximum wind speed with
different recurrence periods. Factors a (s/m) and b(m/s) are obtained from 100 and 500 years of
annual maximum wind speeds U 0 (m/s) , U 500 (m/s) , and are listed in Eq. (A6.3).
For application of the recurrence period conversion factor, careful thoughts are necessary before
extrapolation, considering the possibility of difference in the origin of the annual maximum wind speed
with its length of recurrence period. The prediction error equivalent to a 50-year recurrence period
(A6.3) is 5% at most, but increases to 9% at most for a 20-year recurrence period. With these prediction
errors in mind, Eq. (A6.1.1) should be applied to a recurrence period raging from 50 to 1,000 years. For
the 1-year recurrence wind speed used to evaluate the habitability, the values shown in A6.14, shall be
used together with the wind direction characteristics.
- C6-22 - Recommendations for Loads on Buildings

A6.1.5 Wind directionality factor

It is difficult to obtain the extreme value distribution of wind speeds for each wind direction because
the frequency of strong winds that are used to examine the safety of the structure is very low, and the
meteorological station in Japan does not classify the record of annual maximum wind speed based on
the wind direction. Thus, the typhoon effect should be considered when the wind directionality factor is
determined. In these recommendations, the Monte-Carlo simulation for typhoon winds and a statistical
analysis of the non-typhoon observation data were conducted to obtain the wind directionality factor.
There are two types of wind directionality factors. One defines a wind directionality factor that
changes with direction, as shown in the British standards (BS6399.2)21) and Australia/New Zealand
22) 23)
standards (AS/NZS 1170.2) (cyclone-prone regions excluded). The other defines a constant
reduction coefficient regardless of the wind direction, as in the ASCE 24) standards. For the latter, it is
difficult in practice to reflect on the directional design wind speeds.
In the recommendations, to achieve a reasonable wind resistant design, the wind directionality factor
was defined for each direction, as with the former type.
The wind directionality factor was provided uner the assumption that the wind load is calculated
according to the following procedure.
(1) When the aerodynamic shape factors for each wind direction are known
The wind directionality factor K D , which is used to evaluate the wind loads on the structural frames
and components/cladding for a particular wind direction, shall take the same value as that for the cardinal
direction whose 45° sector includes the wind direction under consideration. In this case, the wind tunnel
experiments should be conducted to measure the directional characteristics of the aerodynamic shape
factors of the structure.
(2) When the aerodynamic shape factors in A6.2 are used
1) When assessing the wind loads on structural frames, two conditions are considered, namely, whether
or not the aerodynamic shape factors depend on the wind direction.
a) When the aerodynamic shape factors are dependent on the wind direction
Four wind directions that coincide with the principal coordinate axis of the structure should be
considered. If the wind direction is within a 22.5° sector centered at one of the eight cardinal
directions, the value of the wind directionality factor K D for this direction should be adopted, as
shown in Fig. A6.1.1(a). If the wind direction is outside the 22.5° sector, the larger of the two nearest
cardinal direction values should be adopted, as shown in Fig. A6.1.1(b). For the lattice structures,
the effect of an inclined wind on the wind force coefficient can be considered directly, and thus the
same measures as for the above rectangular cylinders are adopted for the eight directions (4-leg
square plan) and six directions (3-leg triangular plan).
b) When the aerodynamic shape factors are independent of the wind directions.
For a structure that has a circular plan, the wind directionality factor K D shall take the same
value as for the cardinal direction whose 45° sector includes the wind direction.
2) When assessing the wind loads on the cladding according to the peak wind pressure coefficients in
CHAPTER 6 WIND LOADS - C6-23 -

A6.2, those obtained under the condition of K D = 1 should be used for the design because the
maximum peak pressure coefficient irrespective of the wind direction is shown in the recommendations.
The wind directionality factors for the eight cardinal directions shown in Table A6.1 were originally
obtained for 16 directions (22.5° sector). When the 16 directional values are converted into eight cardinal
directional values, the values are determined to be the maximum of those for the relevant direction and
its two neighboring directions. Therefore, the value for a given direction represents the influence of a
67.5° sector centered on that direction. For a building with a rectangular plan, the wind force coefficients
for the wind directions normal to the building faces are given through the recommendations. When the
wind direction considered is at an intermediate position between two cardinal directions shown in the
table, the greater value for the two neighboring directions is adopted. This means that the value considers
the influence from a 112.5° sector. In addition, considering that the current observation networks may
not always catch the maximum wind speed, the lower limit of the wind directionality factor is given as
0.85.
The case in which the wind directionality effects are not considered corresponds to the condition when
the wind directionality factors are equal to unity for all directions. This results in a conservative
estimation of the design wind load compared with the condition in which the wind directionality effects
are considered.
Whether or not the wind directionality effects are considered corresponds to whether or not the wind
directionality factors are adopted. As shown in Table A6.1, the wind directionality factors are less than
unity, and are defined as values for evaluating 100-year recurrence wind loads. It is possible to make a
more rational design by considering the orientation of the building plan from the viewpoint of the wind
directionality factor. In other words, the wind loads are overestimated if the wind directionality factor
is not considered. However, the amount of this overestimation depends on the orientation of the building,
as well as its shape.
wind direction
N KD=0.9
0.9
NW NE
0.85 0.95

W E
1.0 0.85

SW SE
0.95 0.85
S
0.9

(a) When the wind direction falls within a 22.5° sector, as shown in Table A6.1
- C6-24 - Recommendations for Loads on Buildings

wind direction
N
larger value of 0.9 and 0.95
0.9
KD = 0.95
NW NE
0.85 0.95

W E
1.0 0.85

SW SE
0.95 0.85
S
0.9

(b) When the wind direction does not fall within a 22.5° sector, as shown in Table A6.1
Figure A6.1.1 Selection of the wind directionality factor (when using the wind force coefficient of
buildings with rectangular horizontal sections defined in these recommendations)

When the wind directionality effects are considered, the wind loads will be smaller than those
predicted through a conventional method, in which the wind directionality is not taken into account
because the wind directionality factor is less than unity, the designers should be conscious of the fact
that the safety level decreases when the wind directionality factor is utilized.
The wind directionality factors defined in these recommendations are valid only for locations near
major metrological stations. The wind directionality factor defined in Table A6.1 can be applied to
construction sites near the metrological stations, but cannot be applied to construction sites far from the
metrological stations and influenced through a large-scale topography. For these situations, special
consideration should be taken, for instance, by not using the wind directionality factors, i.e., by setting
K D = 1 . Moreover, when applying seasonal factors specified in A6.1.6, the wind directionality factor
should not be used ( K D = 1 ) because it varies greatly depending on the season.

A6.1.6 Wind seasonal factor

For estimating the wind load for a specific season, it is necessary to define the design wind speed with
the equal probability of exceedance (mean recurrence interval) in order to ensure the safety equal to the
long-lived structure in consideration of the wind speed characteristics of the target season.
The wind seasonal factor is defined as the ratio of the 100-year recurrence seasonal wind speed to the
100-year recurrence basic wind speed U0. The wind seasonal factor is applied to the winter season (4
months from December to March), which is necessary to estimate the combination of wind and snow
loads. The practical design wind speed of a temporary structure built only in winter can be provided by
multiplying the 100-year recurrence basic wind speed U0 by the wind seasonal factor KS at the
construction site.
When the wind seasonal factor is adopted, the wind directionally factor shall not be considered, i.e.,
CHAPTER 6 WIND LOADS - C6-25 -

KD=1. The wind seasonal factor of another season (spring, summer, or autumn) can be estimated through
the following steps: 1) homogenize the observation data, 2) statistically analyze the extreme values, 3)
evaluate the 100-year recurrence seasonal wind speed, and 4) calculate the ratio of the 100-year
recurrence seasonal wind speed to the 100-year recurrence basic wind speed U0 (Fig. A6.1). According
to a similar analysis25), it was reported that the wind seasonal factor for the period (from June to October)
in which the typhoons prevail is less than 1 in a northern area, such as Hokkaido, but is nearly 1 in the
western Kanto region.

A6.1.7 Wind speed profile factor

(1) Effects of terrain roughness and topography on wind speed profile
The wind speed near the ground varies with the terrain roughness, e.g., buildings and trees, and the
topography. The friction force from the terrain roughness and the concentration or blockage effects from
the topography influence the atmospheric boundary layer from the ground to the gradient height. In the
recommendations, the influence of the surface roughness on the wind speed profile over flat terrain is
expressed through Er, whereas the influence of small-scale topographical features is represented through
Eg.
(2) Wind speed profile over flat terrain
Terrain roughness causes a gradual decrease in wind speed toward the ground. The domain that is
influenced by the terrain roughness is called the boundary layer, where the wind speed profile changes
with the category of terrain roughness. The boundary layer depth increases with the fetch length, which
means that the wind speed profile extends to a higher elevation downstream. In addition, the boundary
layer tends to develop faster when the terrain is rougher.
For a fully developed boundary layer, the wind speed profile can be represented by a power or
logarithmic law. The following power law is adopted in the recommendations:
Z α
U Z = U Z0 ( ) , (A6.1.2)
Z0
where Uz (m/s) is the mean wind speed at height Z(m), UZ0 (m/s) is the mean wind speed at height Z0,
and α is the power law exponent.
It has been realized from many observation data that the power law exponent becomes greater as the
terrain becomes rougher.
However, it is rare for the terrain roughness to be uniform over a long fetch. The roughness conditions
usually vary. When the terrain roughness changes suddenly, a new boundary layer develops according
to the new terrain roughness, which gradually propagates with the elevation and fetch, such that wind
speeds above this new boundary layer remain unchanged after the roughness changes. Thus, the wind
speed profile corresponding to the new roughness conditions cannot be applied to a high elevation. This
tendency is particularly obvious when the wind flows from the sea to the city center, where the roughness
changes suddenly from smooth to rough. After a fetch of approximately 3 km (or 40 H ), the new
- C6-26 - Recommendations for Loads on Buildings

boundary layer is considered fully developed. Hence, in the recommendations, the roughness condition
in a region of smaller than 40 H and 3 km upstream from a construction site is considered when the
roughness category, shown in Table A6.2, is to be determined.
The influence of terrain roughness decreases at higher elevations. In these recommendations, it is
assumed that the design wind speed at Z G is not influenced by the terrain roughness, and is considered
constant for convenience sake. However, this does not mean that the wind speeds are really constant
above height Z G . Because the boundary layer depth becomes greater when the terrain roughness
increases, Z G is assumed to increase with the terrain category, as shown in Table A6.3. However,
Z G is defined only for a utilization of the power law for different terrain categories because the velocity
profile is actually unknown in detail at higher elevations. This is different from the boundary layer depth.
CFD studies on the wind speed profile in urban areas show that the wind speed below a certain height
Z b does not follow the power law when the ratio of building plan area to the regional area is over a few
percent, as shown in Fig. A6.1.2. The wind speed profile here is complex owing to nearby buildings.
For heights below Z b , the wind speed at Z b is usually at maximum, and thus the wind speeds in this
region are assumed to be equal to that at Z b , which is defined in Table A6.3, in order to arrive at a safer
design. For heights above Z b , the power law can approximate the mean wind speed profile.

height

Figure A6.1.2 Mean wind speed profile in urban areas

Figure 6.1.3 shows an example of the mean wind speed profiles measured in natural wind26), in which
the wind speed profiles measured simultaneously at coastal and inland locations are compared. As
mentioned before, the wind speed near the ground decelerates owing to the inland terrain roughness. As
a result, there is significant difference between the wind speed profiles within the two locations.
The exposure factor E r of a flat terrain, shown in 2) of A6.1.7(2), is defined while including the
above considerations. Figure A6.1.4 shows Er for each terrain category. The exposure factor is the
ratio of wind speed at a given height Z for each terrain category to the wind speed at 10 m over terrain
roughness category II.
CHAPTER 6 WIND LOADS - C6-27 -

Mean wind speed (m/s)

Figure A6.1.3 Example of mean wind speed profiles measured simultaneously at the coast of Tokyo
bay and a suburban residential area 12 km inland26)

Figure A6.1.4 Exposure factor E r

Figure A6.1.5 shows an example of terrain roughness categories.

Terrain category I represents open sea or lake, or unobstructed coastal areas on land.
Terrain category II is defined as terrain with scattered obstructions up to 10 m high. Rural areas are
representative. Low-rise building areas also belong to this category, if the building area ratio (total
- C6-28 - Recommendations for Loads on Buildings

building area divided by the regional area) is less than 10.0%.

Terrain category III is characterized by closely spaced obstructions up to 10 m high, or by sparsely
spaced medium-rise buildings of four to nine stories tall. Suburban residential areas, manufacturing
districts, and wooded fields are typical of this category. Regions where the building area ratio is 10–
20%, or the building area ratio is larger than 10%, while a high-rise building ratio (area of buildings
higher than four stories divided by the total building area) of less than 30% belongs to this category. The
example in Fig. A6.1.5(c) is an area with a building area ratio of 30% and a high-rise building ratio of
5–20%.
CHAPTER 6 WIND LOADS - C6-29 -

(a) Terrain category I (b) Terrain category II

(c) Terrain category III (d) Terrain category IV

(e) Terrain category V

Figure A6.1.5 Example of surface roughness (photographs provided by Kindai Aero, Inc.)

Terrain category IV is mainly where there are many four to nine stories buildings. Central regions of
local cities are typical of this category. Areas with a building area ratio larger than 20%, and a high-rise
building ratio larger than 30% belong to this category.
In terrain category V, tall buildings of ten stories or more are densely distributed. Central regions of
large cities such as Tokyo and Osaka belong to this category.
In an area where the building purpose, floor area ratio, and building coverage ratio are the same, the
- C6-30 - Recommendations for Loads on Buildings

terrain can usually be considered uniform. Typically, in a wide area around a construction site, the terrain
roughness is not usually identical. It is common for several terrain categories to co-exist. When the
terrain roughness changes downstream, a new boundary layer gradually develops, and the developing
process depends on whether the change is from smooth to rough or rough to smooth. Figure A6.1.6
illustrates approximately the development of a new boundary layer with a terrain roughness change from
smooth to rough. When the terrain roughness changes from smooth to rough, the new boundary layer
develops slowly, and thus the fully developed boundary layer over the new roughness cannot be
anticipated if the fetch downstream is insufficiently long. As a result, a wind speed profile corresponding
to the new roughness category cannot be adopted. Thus, if there is a change in terrain roughness from
smooth to rough within a distance of less than 40H and 3km(the smaller of 40H and 3km) from the
construction site, the terrain category at the upstream region before the roughness change will be adopted
as the terrain category for the construction site.

developing internal
boundary layer

Smooth Rough 3 ~ 5km

Figure A6.1.6 Development process of a new boundary layer when the terrain roughness changes from
smooth to rough

In determining the terrain category for a given wind direction, the upwind area inside a 45° sector
within a distance of less than 40 H and 3 km from the construction site will be counted.
When there is a change in terrain roughness upwind of the construction site, a weighting average of
the wind speed profile on the roughness and the fetch distance is calculated in AS/NZS 1170.2 to
determine the exposure factor.
However, in the recommendations, the overall terrain roughness in the upwind sector is adopted as
the terrain category in this direction if there is no sudden change in roughness. In general, the wind load
will be overestimated when a smoother surface roughness category is utilized.
For an urban area centered on a railway station, larger buildings are closely spaced near the station.
Figure A6.1.7 shows an example of how to determine the terrain category if a construction site is near
a railway station, in which the roughness downstream changes from smooth to rough. In this case, where
there is a sudden change in roughness within a distance of less than 40 H and 3 km upwind of the
construction site, the smoother terrain category upwind before the change in terrain roughness shall be
selected.
CHAPTER 6 WIND LOADS - C6-31 -

Wind

Category I

Category III
smaller of
40H and 3km
The terrain roughness
in this wind direction
should be recognized as
category I.

Figure A6.1.7 Selection of terrain category (with change in terrain roughness from smooth to rough)

If the terrain roughness changes from rough to smooth, the terrain category after the terrain roughness
change is selected. However, if there is a smooth area in a rough area, e.g., a park in a downtown area,
it is sometimes necessary to consider the acceleration of wind speed near the ground downstream.
In general, careful consideration should be given in the determination of the terrain category owing
to arbitrariness. If the wind tunnel tests or CFD reproduces the surroundings of the construction site as
a model, it defines a terrain category in accordance with the conditions of the windward areas of the
area reproduced by the model.

A6.1.8 Turbulence intensity and turbulence scale

The natural wind speed fluctuates with time. The wind speed U (t ) at a point, shown in Fig.A6.1.8,
can be separated into the mean wind speed component U and fluctuating wind speed component u (t )
in the longitudinal direction, as well as v(t ) and w(t ) in the crosswind directions. Usually, the
longitudinal fluctuating wind speed component u (t ) is important for the building design, and thus only
the characteristics of u (t ) are defined in the recommendations. For long-span structures such as bridges
and for tall slender buildings, the vertical and lateral fluctuating wind-speed components w(t ) and v(t )
are also occasionally important.

Figure A6.1.8 Mean wind and turbulence components

- C6-32 - Recommendations for Loads on Buildings

(1) Turbulence intensity

1) On flat terrain
A fluctuation in wind speed can be expressed quantitatively using a statistical approach. Turbulence
intensity I indicates the turbulence level, and is defined in the following equation as the ratio of
standard deviation of the fluctuating component σ u to the mean wind speed U .
σu
I= (A6.1.3)
U
Turbulence is generated by friction on the ground and drag on the surface obstacles, and is influenced
by the terrain roughness. Figure A6.1.9 shows the turbulence intensity observed in natural wind27), along
with the recommended values calculated from Eq. (A6.7). The figure shows the value (0.26) of the
normalized difference between the guideline value and the standard deviation of the observation results
with the guideline value. Small turbulence intensity data of terrain category I include data on bridges.

300
Height above ground Z(m)

100

50
30

5
3
0.05 0.1 0.2 0.5 0.05 0.1 0.2 0.5 0.05 0.1 0.2 0.5 0.05 0.1 0.2 0.5 0.05 0.1 0.2 0.5
Turbulence intensity I Turbulence intensity I Turbulence intensity I Turbulence intensity I Turbulence intensity I
rZ rZ rZ rZ rZ

Category I Category II Category III Category IV Category V

Figure A6.1.9 Observed turbulence intensity and recommended value

The turbulence intensity I Z at height Z above the ground, is defined in Eq. (A6.7), in which the
turbulence intensity I rZ above a flat terrain expressed in Eq.(A6.11), and the topography factor EgI ,
shown in Tables A6.6 and A6.7, are considered separately.

(2) Power spectral density

The power spectral density reflects the contribution to the turbulence energy at each frequency. In the
recommendations, a von Karman type power spectrum, expressed through Eq. (A6.1.4), is employed to
express the power spectral density of the fluctuating component of the wind speed u (t ) .

4σ u2 ( L / U )
Fu ( f ) = (A6.1.4)
{1 + 70.8( fL / U ) 2 }5 / 6

where
CHAPTER 6 WIND LOADS - C6-33 -

f : frequency,
σ u : standard deviation of fluctuating component of wind speed u (t ) ,
U : mean wind speed,
L : turbulence scale.
(3) Turbulence scale
Equation (A6.8) is used as the turbulence scale LZ of the wind speed fluctuation u (t ) at height Z .
The turbulence scale is an important parameter in the power spectrum, expressed in Eq. (A6.1.4), and
is the averaging length scale of the turbulence vortices. Figure A6.1.10 shows an example of a profile
of the turbulence scale27), which can be expressed in Eq. (A6.8) according to the terrain categories. The
figure shows the value (0.34) of the normalized difference between the guideline value and the standard
deviation of the observation results with the guideline value.

300
Height above ground Z(m)

100

50
30

5
3
30 50 100 300 500 30 50 100 300 500 30 50 100 300 500 30 50 100 300 500 30 50 100 300 500
LZ (m) LZ (m) LZ (m) LZ (m) LZ (m)

Category I Category II Category III Category IV Category V

Figure A6.1.10 Observation of turbulence scale of wind speed fluctuation u (t )

(4) Co-coherence
Co-coherence of the wind speed fluctuation Ru ( f , rz , ry ) is evaluated using Eq. (A6.1.5). It
expresses quantitatively the frequency-dependent spatial correlation of the wind speed fluctuation.
 f k 2r 2 + k 2r 2 
Ru ( f , rz , ry ) = exp − 
z z y y
(A6.1.5)
 U 
 

where
f : frequency,
rz , ry : distance between two points in the vertical and horizontal directions,
k z , k y : decaying factors reflecting the degree of spatial correlation of wind speed in the vertical
and horizontal directions,
- C6-34 - Recommendations for Loads on Buildings

U : mean wind speed averaged at two points.

It has been shown through observation that the decay factor is between 5 and 10.

A6.1.9 Effects of topography

The topography factors for the mean wind speed and turbulence intensity (Eg and EgI) are shown in (1)
and (2) below. In addition, important points in simulating the effect of the topography on the wind
through CFD are shown in (3).
Parts (1) and (2) below are specified from the results of the wind tunnel experiments on two-
dimensional escarpments and a ridge-shaped topography with different angles28-30). These experiments
were conducted such that the wind passes at right angles over escarpments or a ridge-shaped topography,
which had a smooth surface with a height of several tens to 100 m, in an airflow corresponding to the
terrain category Ⅱ.
Although a two-dimensional topography does not exist, and the wind does not always pass at right
angles over escarpments or a ridge-shaped topography, if the measurement of the topography is several
times longer than the height of the target building, it is possible to adopt the topography factor of this
section. Experimental measurements and numerical calculations show that speedup ratios for a two-
dimensional topography tend to be higher than those for a three-dimensional topography31). Thus,
application of Eq. (A6.9) to a simple three-dimensional topography can be a conservative estimation.
There is a possibility that the wind will intensify by an effect that is not taken into account in this
section (e.g. when the wind blows along an area with a complicated geography such as a valley between
mountains). Accordingly, when the target building is built on a complex topography, it is recommended
to evaluate the effects of the topography on the wind through an appropriate wind tunnel experiment or
CFD.
(1) Topography factor
When airflow passes through escarpments or a ridge-shaped topography with a large inclination angle,
as show in Fig. A6.1.11, the flow is blocked at the front of the escarpment, and the mean wind speed
decreases. Then, the flow starts to accelerate uphill, resulting in a mean wind speed larger than that over
a flat terrain from the middle of the windward slope to the top of the topographic feature. Further, the
wind speed near the ground on the leeward side from the hilltop is smaller than that over a flat terrain
because of the separation region of the hilltop. However, if the slope is insufficiently large, the mean
wind speed will be larger than that over a flat terrain over a long region windward and leeward of the
hill top.
CHAPTER 6 WIND LOADS - C6-35 -

U U+ΔU
+ΔU
UU ΔU
ΔU

Thin solid curve and thick solid

U
ZZ curve are for the mean wind speed
over flat terrain and escarpments
ZZ EgE＝(U +ΔU)/U respectively.
g=(U+ΔU)/U

Figure A6.1.11 Change of mean wind speed over an escarpment

Although the wind speed near the ground windward of the escarpment and in a separation region on
the leeward side from the hilltop is smaller than that over a flat terrain, the topography factor in these
regions is defined as 1 in the recommendations because only the acceleration of the wind speed is
considered29). However, ｗhen estimating by a wind tunnel experiment or CFD, the topography factor
may be made less than 1.0.
The topography factor can be defined as 1 if the inclination angle of the slope is less than 7.5 degrees.
However, when it becomes for a relatively large topography factor, it is desirable to consider the
“Guidebook of Recommendations for Loads on Buildings 2.”
Tables A6.4 and A6.5 show the values of the parameters in Eq. (A6.9) for escarpments and a ridge-
shaped topography determined from the results of wind tunnel experiments.
Owing to the complicated calculation of this formula, the calculation results for several examples are
shown in Fig. A6.1.12 as a reference.
In addition, for a particular location and slope angle, not shown in Tables A6.4 or A6.5, the
topography factor can be obtained through linear interpolation. For the interpolation, for each of the two
inclination angles and the point closet to the desired inclination angle, the preliminary coefficients of
the wind speed are obtained, and linear interpolation is applied to them.
- C6-36 - Recommendations for Loads on Buildings

(a) Escarpments

(b) Ridge-shaped topography

Figure A6.1.12 Calculation results of example topography factor

(2) Topography factor for turbulence intensity

Not only the mean wind speed, but also the wind speed fluctuation, is influenced by the topography.
Particularly in a region where the mean wind speed decreases, there is an obvious increase in the
standard deviation of the wind speed fluctuating component u(t) (fluctuating wind speed, hereafter)
compared to that on a flat terrain, as shown in Fig. A6.1.13. The variations in mean and fluctuating wind
speed are closely related. The location of the maximum fluctuating wind speed generally corresponds
to the location in which the vertical gradient of the mean wind speed is at maximum. The region where
the fluctuating wind speed is greater than the flat terrain counterpart is generally located inside the
separation region where the mean wind speed is smaller than that on flat terrain.
CHAPTER 6 WIND LOADS - C6-37 -

EI EI
0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1

-3 -2 -1 0 1 2 3 4 5 6 -3 -2 -1 0 1 2 3 4 5 6 7 8
Xs/Hs Mean wind speed Xs/Hs Mean wind speed
decreasing region decreasing region
(a) Escarpment (b) Ridge-shaped topography
(inclination angle of 60°) (inclination angle of 30°)
Figure A6.1.13 Topography factor for fluctuating wind speed.
The symbols are for experimental results, and the thick solid lines are for Eq. (A6.10).

In this section, the topography factor for the turbulence intensity is defined as the ratio of the
topography factor for a fluctuating wind speed to the topography factor for the mean wind speed. The
topography factor for a fluctuating wind speed is defined in Eq. (A6.12), in which the values of the
parameters other than C1, C2, and C3 are identical to those in Eq. (A6.9) for the topography factor for
the mean wind speed. The topography factors of the mean wind speed and the fluctuating speed are
defined as greater than 1 without considering the decrease in mean wind speed and fluctuating wind
speed owing to the topography effects29). However, where the topography factor for a fluctuating wind
speed is smaller than that for the mean wind speed, the topography factor for the turbulence intensity
will be less than 1.
The fluctuating wind speed near the ground in the leeward slope of the escarpments or ridge-shaped
topography becomes greater. In these regions, the mean wind speed is smaller, which results in the
maximum instantaneous wind speed being less than that for a flat terrain in this area. Because the
decrease in the mean wind speed is not considered, the maximum instantaneous wind speed and wind
load is possibly overestimated in the region if only the topography factor of the fluctuating wind speed
is fitted to the experimental data. To reduce this possible overestimation, the actual topography factor
for the fluctuating wind speed (<1), where the topography factor for the mean wind speed is 1, was
utilized for the whole region with deceleration, as shown in Fig. A6.1.13.
For a particular slope and the location of an escarpment or a ridge-shaped topography, not shown in
Tables A6.6 and A6.7, the topography factor of the fluctuating wind speed can be obtained through
linear interpolation. In this case, the interpolation should not be performed for the individual parameters
shown in Tables A6.6 and A6.7, as with the topography factor of the mean wind speed. In addition,
when the slope angle of the topographic feature is less than 7.5°, it is not necessary to consider the
topography factor of the turbulence intensity because the fluctuating wind speed has almost zero
influence from the topography. As with the topography factor of the mean wind speed, some calculation
examples of the topography factor of the turbulence intensity are shown in Fig. A6.1.14.
- C6-38 - Recommendations for Loads on Buildings

Figure A6.1.13 shows the topography factor of the fluctuating wind speed calculated from Eq. (A6.12)
as a solid line. It agrees well with the experimental data at any position and slope on the escarpment.
However, although the topography factor for a fluctuating wind speed does not match well with Eq.
(A6.12) for a ridge-shaped topography because of the complex variation of the fluctuating wind speed,
the consistency with the experimental data is good when the topography factor of the mean wind speed
is greater than 1.

(a) Escarpment

(b) Ridge-shaped topography

Figure A6.1.14 Example calculation results of topography factor of fluctuating wind speed

(3) Estimation through CFD for the terrain effects on wind speed
With regard to the topography factors of the mean wind speed and the turbulent intensity through the
terrain effect, which are prescribed in the present recommendations, their values are determined
experimentally under the conditions of ground roughness category II for rural areas covered by grasses
or low trees. Therefore, in the case of more roughened surface covered by dense trees, these factors are
CHAPTER 6 WIND LOADS - C6-39 -

expected to be given through CFD. CFD can generally provide parameters and conditions fitting the
actual state during the computational process on the terrain model, and has an advantage in avoiding an
excessive wind load. Such a new technique requires careful application, and in this session, significant
points are therefore shown for an appropriate numerical simulation.
The statistics predicted through CFD required for the terrain problems are the mean and the
fluctuating speeds based on the time history data. Additionally, using these data, the power spectra,
spatial correlation, turbulence scale, and turbulence intensity can be estimated. In this case, the CFD has
to deal with wind as a turbulent flow, and correctly extract the fluctuating components of the wind load.
Then, a large-eddy simulation (LES) is selected because it can be used to apply a time-dependent
computation. However, another Reynolds averaged Navier-Stokes (RANS) type of turbulence model
can also be utilized, as far as this model can reproduce the turbulence statistics such as the mean flow
or turbulence fluctuation for complicated flows including a separated shear flow over very steeply sloped
terrain. The standard RANS-type models can possibly be applied to the terrain problem in limited cases
because the flow over terrain does not generate so large separation compared to the problem for the flow
around buildings. In the process of its application, although above model may be employed under careful
judgment from an engineering perspective, it is necessary to confirm that the solution is not affected by
inherently existing defects of the model that have often been indicated in academic studies32). On the
other hand, the improved model can be introduced, and a nonlinear model33) as an example can predict
each component of turbulence individually.
To apply the CFD technique in the evaluation of the terrain effects, the following tasks to be solved
can be displayed. For the numerical model of an actual mountainous area, the terrain has a continuous
undulation area horizontally and occupies a boundary altitude-changing ground surface in all parts of
the circumference of the computational domain. As a result, the blockage ratio tends to increase and the
vertical size of computation domain should be determined by taking into consideration the appropriate
depth of the boundary layer. The flow near the sidewall boundary of the computational domain is
characterized based on the shape of the terrain near the sidewall boundary. Generation of such near-side
boundary flows cannot be achieved through the boundary conditions. In general, although a free slip
condition is imposed, this boundary condition forces the wind direction along the sidewall boundary.
The effects of the sidewall should be removed based on the inconsistency of the wind direction near the
side. The inflow boundary condition also needs to be checked. It is necessary to fix how large a region
is required in front of the main object of the terrain to determine the wind profile and wind speed
distribution in an arbitrary location. For the ground surface conditions, the terrain in Japan usually has
various types of roughness patterns such as grasses, bushes, and trees. For such a surface roughness, the
computation needs to deal with the boundary treatment based on the roughness effects of the vegetation.
Currently, several techniques can be provided such as coverage of the blocks, the wall function based
on the roughness length, the canopy model, and a hybrid method with the RANS model.
To examine the predictive accuracy of the CFD technique based on the LES, a comparison will be
made between the numerical results and the results of a wind tunnel test for simple terrain. For the wind
- C6-40 - Recommendations for Loads on Buildings

flow over a relatively steep sinusoidal hill, the roughness effect has been investigated, and the resulting
accuracy of the mean and fluctuating velocities has been clarified. Here, the simulation imposes an
inflow condition prescribed by the wind profile of the turbulent boundary layer over a flat surface with
the same level of roughness.
The numerical results by a generalized coordinate grid system using the roughness blocks34) and an
unstructured grid system with a layer mesh near the ground using the wall function based on the
roughness length35) are shown in Figs. A6.1.15 and A6.1.16. The mean and fluctuating wind speeds are
non-dimensionalized by the uniform flow velocity above the boundary layer or mean velocity at the
height of a hill. For the mean wind speed, both numerical results show good agreement with the wind
tunnel experiment results. Concretely speaking, the locations of the reattachment and the sizes of the
recirculating flow behind the hill are consistent with each other. For the fluctuating wind speed, the
sufficient accuracy over the hill including the separation region can be confirmed. However, by checking
the details, the results of the generalized coordinate system show a higher level of consistency, the reason
for which is that the results by an unstructured grid are influenced by the numerical dissipation, and it
is then necessary to check this important point when selecting the numerical discrete method.
Considering the fact that an unstructured grid is usually used for an open-source or commercial code,
there is a higher possibility to adopt a stable numerical scheme through the default setting. The setup of
the input data used for the computation should be carefully examined.
For the velocity field over a roughened surface, the amplification factor tends to increase compared
to a smooth surface. It should be noted that the amplification factor in the present recommendations
where a smooth surface is assumed does not necessarily provide a safer value. On a rough surface, the
velocity at the accelerated location decreases by the drag effect, but increases above the location owing
to the continuity.

(a) Mean wind speed (b) Fluctuating wind speed

Figure A6.1.15 Vertical profiles of the non-dimensionalized mean and fluctuating wind speeds over a
simple hill
(generalized coordinate grid system, covered by roughness blocks)34)
CHAPTER 6 WIND LOADS - C6-41 -

5 5
LES LES
Exp Exp
4 4

3 3

Z/H
Z/H

2 2

1 1

0 0
-2.50H -1.25H 0.00H 1.25H 2.50H 3.75H 5.00H -2.50H -1.25H 0.00H 1.25H 2.50H 3.75H 5.00H
0.0 X /H 0.0 0.4 X/H
U/UH1.0 σU/UH

(a) Mean wind speed (b) Fluctuating wind speed

Figure A6.1.16 Vertical profiles of the non-dimensionalized mean and fluctuating wind speeds over a
simple hill
(unstructured grid system, layer mesh near the ground using the wall function based on the roughness
length)35)

A6.2 Wind force coefficients and wind pressure coefficients

A6.2.1 Procedure for estimating wind force coefficients

The wind force coefficients and wind pressure coefficients depend on the building shape, building
surface condition, terrain condition, and local topography at the construction site. Therefore, they should
be determined from wind tunnel tests or CFD that properly simulate the full-scale conditions. However,
the coefficients for buildings with regular shapes (rectangular, circular, or elliptical sections) can be
estimated from the procedure described in this section. The coefficients are divided into two categories,
one for the design of structural frames and the other for the design of building components/cladding,
because the wind effects on the structural frames and components/claddings are quite different from
each other.
The peak wind force coefficients and peak wind pressure coefficients are generally defined in terms
of the velocity pressure 𝑞𝑞𝐻𝐻 (N/m2) evaluated at the reference height 𝐻𝐻 (m). For lattice structures and
members, the wind force coefficients are defined in terms of the velocity pressure 𝑞𝑞𝑍𝑍 (N/m2) evaluated
at height 𝑍𝑍 (m), where the members under consideration are placed.
The flow field around a building changes with the aspect ratio of the building. For a building with a
rectangular section, √𝐵𝐵𝐵𝐵 is used as the reference length to define the aspect ratio. The aspect ratio
𝐻𝐻/√𝐵𝐵𝐵𝐵 is generally large for tall buildings. However, the majority of buildings are lower with
𝐻𝐻⁄√𝐵𝐵𝐵𝐵 ≤ 2. Therefore, two different procedures are provided to estimate the wind force coefficients
for buildings with 𝐻𝐻⁄√𝐵𝐵𝐵𝐵 ≤ 2 and those with 𝐻𝐻/√𝐵𝐵𝐵𝐵 > 2.
The sign of the wind pressure coefficient indicates the direction of the pressure on the surface or
member; positive values indicate the pressures acting toward the surface, and negative values are
pressures acting away from the surface (suction). In the case of a curved roof, the direction of the wind
pressure varies with the location, as shown in Fig. A6.2.1. The wind forces on buildings and structures
- C6-42 - Recommendations for Loads on Buildings

are the vector sum of the forces calculated from the pressures acting on surfaces such as walls and roofs
or on structural members.
The wind force coefficients 𝐶𝐶D for estimating the horizontal wind loads on the structural frames are
generally given based on the difference between the wind pressure coefficients, 𝐶𝐶pe1 and 𝐶𝐶pe2, on the
windward and leeward faces, as shown in Eq. (A6.13), the exception of which is for buildings with
circular or elliptic sections, where the resultant wind force coefficients are provided. Similarly, the wind
force coefficients 𝐶𝐶R for estimating the roof wind loads on structural frames are generally given based
on the difference between the external and internal pressure coefficients, 𝐶𝐶pe and 𝐶𝐶pi, on the roof, as
shown in Eq. (A6.14), with the exception for open roofs. The wind pressure coefficients are space- and
time-averaged values. The averaging area depends on the building shape. The averaging duration is 10
min. The wind force coefficients 𝐶𝐶D for estimating the horizontal wind loads on the lattice structures
are given as a function of the solidity φ. The wind force can also be calculated using the wind force
coefficients for the individual members provided in A6.2.4 (3).
The peak wind force coefficients 𝐶𝐶̂c for the design of the components/cladding are generally given
based on the difference between the peak external pressure coefficient 𝐶𝐶̂pe and the factor 𝐶𝐶pi
∗
for the
effect of the fluctuating internal pressures, except for open roofs, in which the value of 𝐶𝐶̂c is provided.
The values of 𝐶𝐶̂pe (and 𝐶𝐶̂c in an open roof case) are determined from the maximum positive and
∗
negative peak values irrespective of the wind direction. Note that factor 𝐶𝐶pi for the effect of the
fluctuating internal pressures is not the actual peak internal pressure coefficient 𝐶𝐶̂pi but an equivalent
value producing the peak wind force coefficient 𝐶𝐶̂c when it is combined with the peak external pressure
coefficient 𝐶𝐶̂pe .
The wind force coefficients and wind pressure coefficients given in this chapter are all for isolated
buildings, and are obtained from the results of wind tunnel tests. When nearby buildings are expected
to influence the wind forces and pressures, it is necessary to carry out wind tunnel tests, CFD, or other
particular studies to determine the coefficients15, 16).

Figure A6.2.1 External wind pressure on a building with a vaulted roof in a wind parallel to the gable
walls
CHAPTER 6 WIND LOADS - C6-43 -

A6.2.2 External wind pressure coefficients for structural frames

As described in the previous section, the external wind pressure coefficients for structural frames are
divided into two classifications using the aspect ratio, which is the ratio of the reference height H (m) to
the reference length. The classifications are an aspect ratio of greater than 2 or less than or equal to 2.
The square root of a product of the length of each side BD is used for the reference length of the
building to calculate the aspect ratio for a rectangular section. The outside diameter D (m) is used in the
case of a circular section. The outside diameter of the minor axis D2 (m) is used in the case of an elliptical
section. Moreover, the white painted arrow in the figures indicates the wind direction.

(1) External wind pressure coefficients Cpe for buildings with rectangular sections and aspect ratios
H / BD of greater than 2
Buildings with aspect ratios H / BD of greater than 2 are mainly high-rise buildings. The
coefficients of the wall shown in Table A6.8 are estimated from the results of the wind tunnel test39) on
high-rise buildings with rectangular sections.
External wind pressure coefficients on the windward and leeward walls of buildings with rectangular
sections have the following features:
1) External wind pressure coefficients on windward walls are nearly proportional to the velocity pressure
of the approaching flow, except for areas near the top and bottom of the building.
2) External wind pressure coefficients on leeward walls are negative and almost independent of the
height.
Based on these features, the vertical distribution of external wind pressure coefficients on the
windward walls are assumed to be proportional to the factor for the vertical profile (kZ), whereas those
on the leeward walls are assumed to have a constant negative value regardless of the height. The external
wind pressure coefficients on the leeward walls decrease with an increase in the side ratio D/B. This
feature is related to the behavior of the separated shear layer from the windward edge, and is reflected
in the value of Cpe2. The aspect ratio H / BD of high-rise buildings is within the range of 2 to 8 in
most cases. In this range, the effect of H / BD on the wind pressure coefficients is not significant.
Therefore, the pressure coefficients are provided independently of H / BD .
The external wind pressure coefficients on roofs are determined from the results of various wind
tunnel tests22, 40) as well as on the provisions of international codes and standards. Although flat-roofed
buildings have parapets in many cases, their effects on the pressure coefficients are not considered herein.
A reduction factor for the external pressure coefficients on roofs with parapets is provided in the
Eurocode40).

(2) External wind pressure coefficients Cpe for buildings with rectangular sections and aspect ratios
H / BD of less than or equal to 2
1) Buildings with flat, gable, mono-sloped, and hip roofs
Medium- or low-rise buildings with a relatively small aspect ratio have various roof shapes. The
- C6-44 - Recommendations for Loads on Buildings

coefficients in this section are estimated from the results of wind tunnel tests41–44) conducted on buildings
with rectangular sections. The roof shapes under consideration are flat, gable, mono-sloped, and hipped.
When the roof angle is less than or equal to 10°, the roof can be regarded as a flat roof.
The roof and walls are divided into several zones, and the external pressure coefficients for these
zones are provided in Table A.6.9(1) as a function of the building configuration parameters (width B
(m), reference height H (m), depth D (m), and roof angle θ (°)).
The external pressure coefficient for each zone is estimated from the spatially averaged pressure over
the zone for a range of wind directions, the center of which is normal to the wall. Both positive and
negative values are provided for the external pressure coefficient for zones RU, Ra, and Rb because the
pressure coefficient becomes both positive and negative owing to a small change in experimental
conditions. It is necessary to combine these values with those for the other zones when the stresses in
the members are calculated.
The wind load of the safe side is given only by setting the negative external pressure coefficient,
assuming the positive external pressure coefficient of the flat roof as “not necessary to evaluate” because
the wind force coefficient CR on a roof has a negative value when the internal pressure coefficient Cpi
of a closed type building is considered.
The net wind forces on the windward eaves become quite large because negative pressures act on the
top surface and positive pressures act on the bottom surface of the eaves.
According to the Australian/New Zealand standards22), the external pressure coefficients on the
bottom surface of the eaves are approximately equal to those of the wall just below the eaves.
2) Buildings with vaulted roofs
The external pressure coefficient for a building with a curved surface generally depends on the shape
and surface roughness of the building, the flow conditions, and the Reynolds number. Buildings with
vaulted roofs, however, are immersed in very turbulent flows. Furthermore, such buildings have walls
in most cases, and therefore the flow tends to separate at the windward edge. These features suggest that
the external pressure coefficients on vaulted roofs are less sensitive to the surface roughness and
Reynolds number than those on circular cylindrical structures, as shown in A6.2.4. The external pressure
coefficients Cpe in Table A6.9(2) are determined from the results of wind tunnel tests45, 46) that focus on
medium-scale buildings in urban areas. The effects of the surface roughness are not considered during
the test. The building shape parameters in the table vary with wind direction because of the difference
between the flow patterns of the two wind directions.
For a wind normal to the gable wall (wind direction W1), the parameter is represented by the rise/width
ratio f / B and the eaves-height/width ratio h / B . However, for a wind parallel to the gable wall
(wind direction W2) it is represented by the rise/depth ratio f / D and the eaves-height/depth ratio
h / D . The zone definitions (Ra, Rb, Rc) regarding the external pressure coefficients are the same. For
wind direction W1, the definition of a zone is similar to that for flat, gable, and mono-sloped roofs. For
wind direction W2, however, the definition is similar to that for spherical domes.
The external pressure coefficient corresponds to the area-averaged value, and the design wind load is
CHAPTER 6 WIND LOADS - C6-45 -

assumed to be constant over each zone. When h / B = 0 and f / B = 0 , or when h / D = 0 and

f / D = 0 , the roof level coincides with the ground level. The coefficients for these cases, which have
no physical meaning, are provided to make the interpolation possible.
The external pressure coefficients on the walls are determined in the same way as for buildings with
flat, gable, mono-sloped, and hip roofs.

(3) External pressure coefficients Cpe for spherical domes

In the same manner as for buildings with vaulted roofs, the external pressure coefficients for spherical
domes are determined from the results of a wind tunnel test47).
Because the contour lines of the mean pressure coefficients on a spherical dome are almost
perpendicular to the wind direction, the dome surface is divided into four zones (Ra through Rd), and the
external pressure coefficient Cpe for each zone is given by spatially averaging the mean external pressure
coefficient over the zone. The building shape parameters for each zone are the rise/span ratio f / D
and the eaves-height/span ratio h / D . Linear interpolation can be used for values not given in Table
A6.9(3). Both positive and negative values of Cpe are provided for zone Ra, and are the same as for gable
roofs, as shown in Table A6.9(1). It is necessary to consider both values with the wind load estimation.
The values of h / D = 0 and f / D = 0 in Table A6.9(3) are provided for an interpolation.
The wind force coefficients for walls can be obtained from the wind force coefficients for buildings
with circular sections, as shown in Table A6.12, by substituting h for H .
It is possible to refer to studies on cylindrical tanks48) for the distributions of the external pressure
coefficients Cpe when h / D is less than 1 and f / D is relatively small.

A6.2.3 Internal pressure coefficients for structural frames

(1) Determination of internal pressure coefficients based on a generation mechanism of internal pressure
The characteristics of the internal pressures are determined through the following factors:
a) distribution of external pressures,
b) openings and gaps in building envelope,
c) internal volume of the building,
d) openings and gaps in internal partitions,
e) operation of air-conditioners,
f) distortion of walls and/or roofs,
g) air temperature,
h) damage to building envelope.
In general, buildings have many gaps and openings, such as ventilation openings, in their envelopes.
Air leaks through these gaps and openings owing to differences between the external and internal
pressures. The internal pressure is determined by applying the mass conservation principle to the air in
an internal volume. For instance, a dominant opening in the windward wall may produce positive
internal pressures, whereas one in a side or leeward wall may produce negative internal pressures.
- C6-46 - Recommendations for Loads on Buildings

Moreover, the fluctuation of internal pressure and its characteristics depend on the relationship between
the size of the gaps or openings, as well as the internal volume of the building. In this section, the internal
pressure coefficients for buildings without a dominant opening are provided based on the results of a
series of computations, in which it is assumed that the internal pressures are significantly influenced by
factors a) and b) mentioned above. That is, the coefficients of 0 or -0.4 in A6.2.3 are provided based on
the calculations49, 50) of the mean internal pressures for various building configurations, assuming that
the gaps and openings are uniformly distributed over the external walls and the internal pressure is
caused by external pressures acting at the locations of the gaps and openings. Therefore, the use of the
above coefficients is not permitted for non-uniformly distributed gaps or openings. By considering a
combination of the external pressure coefficient and other load effects, the coefficient that provides the
most conservative load effect should be selected from the values of 0 and -0.4.
When the influence of other factors is assumed to be significant, it should be taken into account for
evaluating the internal pressure coefficient. For instance, when the internal volume is divided by airtight
partitions, the influence of factor d) is significant. When powerful air-conditioners are in operation, the
influence of factor e) is significant. In buildings with flexible roofs and/or walls, such as membrane
structures, the influence of factor f) is significant. When glass windows on the windward face are broken
by wind-borne debris in strong winds, the internal pressure is suddenly increased by winds blowing into
the building, which often results in failures of the roof structures. In such cases, factor h) should be
appropriately considered.

(2) Consideration of dominant opening

A dominant opening is defined as an opening with an adequately larger area than the sum of other
gap areas on the same wall, the existence of which can significantly influence the characteristics of the
internal pressure. When a dominant opening should be considered on a wall, it is necessary to use not
the internal pressure coefficient recommended in A6.2.3, but another value determined from results of
the wind tunnel experiment or numerical simulation, in which the opening condition such as the size
and location should be appropriately considered. In general, several possible cases in which the existence
of a dominant opening on a wall is considered can be as follows:
a) A dominant opening is designed and installed as the building envelope.
b) Heavy-weighted overhead doors, which are ordinarily open and take time to be closed during a
storm, are installed in a large specialized building such as a distribution center.
c) Significant damage to window glass induced by flying debris is considered.
The condition of case a) is the same as a building with a dominant opening regulated in Notification
No. 1454 of the Ministry of Construction (May 31, 2000). On the other hand, it is generally difficult for
structural designers to determine whether or not the condition of the enclosed building should be applied
to cases b) or c). When they determine that the situation corresponding to cases b) or c) is applied for
the design of an enclosed building, the actual overhead door or window glass is supposed to be “a
dominant opening,” and the internal pressure coefficient can then be determined according to the
CHAPTER 6 WIND LOADS - C6-47 -

location and size of the supposed opening.

In general, a dominant opening located in the windward wall may produce positive internal pressures,
whereas one located in a side or leeward wall may produce negative internal pressures. When the size
of the dominant opening is large, the external pressure can have a significant influence on the
characteristics of the internal pressure, which is thought to be dependent on the opening area ratio of the
dominant opening to the other gaps, as shown in the following equation38, 50).
A1 + a1
r= , (A6.2.1)
a 2 + a3 + a 4
where r is opening area ratio of the dominant opening, A1 is the sum of areas of the dominant openings
(m2), a1 is the sum of areas of gaps and openings located in the wall with dominant openings (m2), and
a2, a3, and a4 are the sum of areas of gaps and openings located in each wall without dominant openings
(m2), respectively.
For the determination of values a1, a2, a3, and a4, it is possible to evaluate the actual areas located in
each wall through air control testing using a fan. However, it should be noted that the pressure difference
obtained by such a test corresponds to the value under the condition of relatively weak wind. Although
technical information related to the pressure difference under a high wind is significantly inadequate,
informative data related to the ratio of the sum of the gap areas to the total area of the building envelope
of low-rise buildings, namely, approximately 0.36% in low airtight conditions, 0.18% in middle airtight
conditions, and 0.08% in high airtight condition, respectively, are provided 51).
If the opening area ratio of the dominant opening r, is calculated from the relationship between the
dominant opening’s size and sum of other gap sizes, it can provide a reasonable estimate of the internal
pressure coefficient. Figures A6.2.2 and A6.2.3 show examples of the relation between the opening area
ratio and internal pressure (mean and fluctuating) coefficients. The vertical axis indicates the ratio of
the “internal” wind pressure coefficient to the “external” wind pressure coefficient. These results
illustrate that in the case of the opening area ratio of more than 10, the internal pressure coefficient tends
to be almost the same as external pressure coefficient of the wall with the dominant opening.

Table A6.2.1 Model configuration

M1 M2 M3 M4 M5 M6
Height (m) 100 50 50 50 40 20
Width (m) 50 50 50 12.5 12.5 10
Depth (m) 50 50 12.5 50 50 10
- C6-48 - Recommendations for Loads on Buildings

平均内圧係数/平均外圧係数

変動内圧係数/変動外圧係数
1.0 1.0

0.5 0.8
pressure ratio

pressure ratio
0.6
0.0
M1 * M4 0.4 M1 * M4
-0.5 M2 M5 M2 M5
0.2 M3 + M6
M3 + M6
-1.0 0.0
0.1 1 10 100 1000 0.1 1 10 100 1000
開口面積比ｒ
opening area ratio r 開口面積比ｒ
opening area ratio r

(a) mean internal wind pressure coefficient (b) fluctuating internal wind pressure coefficient
Figure A6.2.2 Relation between the opening area ratio r and the internal pressure coefficient38, 50)
(case of a dominant opening located in windward wall)

変動内圧係数/変動外圧係数
1.0 1.0
平均内圧係数/平均外圧係数

0.8 0.8
pressure ratio

pressure ratio

0.6 0.6
M1 * M4 M1 * M4
0.4 M2 M5 0.4 M2 M5
0.2 M3 + M6 M3 + M6
0.2
0.0
0.0
0.1 1 10 100 1000
0.1 1 10 100 1000
開口面積比ｒ
opening area ratio r 開口面積比ｒ
opening area ratio r

(a) mean internal wind pressure coefficient (b) fluctuating internal wind pressure coefficient
Figure A6.2.3 Relation between the opening area ratio r and the internal pressure coefficient38, 50)
(case of a dominant opening located in a side wall)

A6.2.4 Wind force coefficients for design of structural frames

(1) Wind force coefficients CD for buildings with circular and elliptic sections
The wind force coefficients for cylinders are affected based on the Reynolds number, flow condition,
aspect ratio H / D2 , surface roughness of the cylinders, and other factors. Figure A6.2.4 shows the
variation of the drag coefficient CD on a two-dimensional smooth cylinder in a uniform flow with
Reynolds number Re ( = UD /ν , where U , D and ν are the wind speed, cylinder diameter, and
kinematic viscosity coefficient of a flow, respectively)52). For wind, the Reynolds number is
approximately given by Re ≈ 7UD × 104. It can be seen from Fig. A6.2.4 that CD decreases from 1.0
to 0.2 with Re within the range of 2.8 × 105 to 3.5 × 105. On the other hand, CD increases along
with Re from 3.5 × 105. Here, CD becomes approximately 0.5 with Re around 4 × 106. The
flow around a cylinder is usually classified into four regimes, i.e., subcritical, critical, supercritical, and
transcritical, as shown in Fig. A6.2.4. Because the Reynolds number of the flow around buildings in
strong winds is in the transcritical regime, the provision of CD in A6.2.4(1) is based on the values of
the drag coefficients in this regime.
CHAPTER 6 WIND LOADS - C6-49 -

Subcritical Critical Subercritical Transcritical

1.5

1
p (bar)
1
CD

2
6
0.5 15
30
51

0
4 5 6 7
10 10 10 10
Reynolds number Re

Figure A6.2.4 Plots of drag coefficient CD on a two-dimensional cylinder with a very smooth surface
as a function of the Reynolds number Re 52)

The aspect ratio and surface roughness of the cylinder also affect the drag coefficient. In particular,
the effect of the surface roughness is significant in the transcritical regime. This is the case for the flow
around an elliptic cylinder. The drag coefficient CD for a building with elliptical sections (major axis
D1 (m), and minor axis D2 (m)) targets the wind direction normal to the major axis with a maximization
of the wind force. The drag coefficient is defined as Eq. (A.6.2.2).
wind force with the wind direction normal to major axis
CD = (A6.2.2)
q H D1 H
where qH is the velocity pressure at the reference height H .
In Table A6.12, the effects of aspect ratio H / D2 and surface roughness are represented by k1 and
k 2 , respectively53).
The elliptic cylinder is less sensitive to the Reynolds number than a circular cylinder. However, it is
known that the drag coefficient changes based on the effects of D1 / D2 54). The effects of D1 / D2 are
represented by k3 .
The external pressure coefficients Cpe on roofs of buildings with a circular section are given in Table
A6.9(3) assuming that f / D = 0 and h / D = 1 . However, the external pressure coefficients Cpe on roof
of buildings with an elliptical section should be validated through wind tunnel tests or CFD.

(2) Wind force coefficients CR for free roofs with a rectangular base
For free roofs where strong winds can flow under a roof, highly fluctuating pressures act on both the
top and bottom surfaces. It is reasonable to evaluate the net wind force coefficients directly, not from
- C6-50 - Recommendations for Loads on Buildings

the wind pressure coefficients on the top and bottom surfaces, because the correlation between the
fluctuating wind pressures on both surfaces is higher than that for enclosed buildings. The wind force
coefficients in Table A6.11 can be used for only small-scale buildings where the reference height 𝐻𝐻 is
approximately 10 m or less because the coefficients are determined from the results of wind tunnel tests
on free roofs with 𝐻𝐻 ≤ 10 m.
The wind force coefficients shown in Table A6.11(1) and A6.11(2) were determined by referring to
the results of wind tunnel tests55, 56) and the standards of foreign countries22, 24). The wind tunnel tests are
on measurement for wind pressures and wind forces acting on pitched (𝜃𝜃 > 0° ), troughed (𝜃𝜃 < 0° ), and
mono-sloped free roofs. Although the most tested roof pitch 𝜃𝜃 in previous studies is limited to the
range of |𝜃𝜃| < 30° , the roof pitch is limited to the range of |𝜃𝜃| < 20° . This is consistent with a roof
pitch range of |𝜃𝜃| < 20° for the peak wind force coefficients on claddings to the free roofs shown in
A6.2.7(1). The wind force coefficients for a roof pitch range of are shown in
55, 56)
elsewhere . Combinations 1 and 2 of the wind forces shown in Table A6.11 are provided under
consideration of the two load cases where the load effects induced by fluctuating wind pressures
(focusing on the axial force of columns supporting a roof) are in critical states, which shall be verified
for both combinations. Further, the wind force coefficients shown in the table are regulated for a clear
flow case where there are no obstructions under a roof. The flow pattern around a roof is significantly
affected by obstructions under it. If there are any obstructions whose blockage ratio are larger than
approximately 50%, the wind pressure on the bottom surface may increase significantly, resulting in a
significant increase in the net wind force on the roof. In such cases, it is necessary to evaluate the wind
force coefficients from wind tunnel tests and other measures.

(3) Wind force coefficients CD for lattice structures

The size of an individual lattice structure member is generally much smaller than the width of the
structure, and the members are arranged symmetrically. Therefore, it is assumed that the only wind force
acting on a plane of a structure is drag. The total drag can be estimated as the summation of the all drag
on each member of the structure. Because the flow around a member depends only on the characteristics
of the local flow around it, the drag is proportional to the velocity pressure at the height of the member.
Based on these features, the following two methods are often used for estimating the wind force on
lattice structures. One is to multiply the wind force coefficient, given as a function of the solidity ϕ of
the plane, by the projected area of the plane. The other method57) is to sum up the wind forces on all
members, which is given by the product of the wind force coefficient CD of each member and its
projected area. For either method, the solidity ϕ should be small. In these recommendations, the
former method is used, and the wind force coefficient CD is provided only for 0.6 ≥ ϕ .
The wind force coefficient is represented as a function of the solidity ϕ , the plan of the structure,
and the cross section of the member. The solidity ϕ is defined as the ratio of the projected area AF
(m2) of the plane to the entire plane area A0 = (Bh ) (m2) of the structure. The value of ϕ is calculated
for each panel of the lattice structure when the wind direction is normal to the plane. In the calculation,
CHAPTER 6 WIND LOADS - C6-51 -

the areas of the leeward lattice members and the appurtenances are not included. The wind forces on the
appurtenances can be estimated from the provisioning of CD for the members (Table A6.14) or from
wind tunnel tests or CFD, and are added to the wind force on the structure.
Table A6.12 provides the wind force coefficients CD for lattice structures with square and
triangular section shapes, which consist of angles or circular pipes. The wind force coefficient CD for
a triangular shape in a section is the same for the two wind directions shown in the table. When the
members are circular pipes, the wind force coefficients CD for the members are affected based on the
Reynolds number. The provisions are based on the value in the subcritical Reynolds number regime. In
strong winds, the value of CD may become smaller than that given in the provisions owing to the effect
of the Reynolds number. However, this effect is not considered here.
When the section of the structure and/or the cross-section of the member are different from those in
Table A6.12, the wind loads on the structure can be estimated by using the wind force coefficients of
the members given in Table A6.14 together with the local velocity pressure. However, the solidity ϕ
of the structure is required to be less than 0.6.

(4) Wind force coefficients CD for fences on the ground

Wind force coefficients CD for fences on the ground are defined as a function of the solidity ϕ in
the same manner as those for lattice structures. The value of CD for ϕ = 0 in Table A6.13 is
introduced to obtain intermediate values of CD for 0 < ϕ < 0.2 . The wind load for a fence can be
calculated according to the simplified procedure using CD and the projected area A (m2), which is
defined as the whole area multiplied by ϕ .

(5) Wind force coefficients CX and CY for members

The wind force coefficients CX and CY for the members are determined from the wind tunnel
tests58) with two-dimensional models in a smooth flow. The values of CX and CY can be applied to
line-like members of less than approximately 50 cm wide, but should not be applied to ordinary
buildings. In some cases, the value of CX in the across-wind direction becomes relatively large when
the wind direction deviates only a little from the normal direction. In such cases, two values of CX
(± 0.6) are provided in Table A6.14.
The wind force coefficients for the components may also be used for calculating the wind loads on
lattice structures, together with the local velocity pressure qZ (N/m2) at height Z (m) of the member
under consideration. The wind loads on a component in each direction is given by the product of qZ
(N/m2), CX or CY , (1 + 3Iz ) and bl ( blϕ for nets) (m2), where I Z is the turbulence intensity at
height Z (m) (see Eq. (A6.7)).

A6.2.5 Peak external pressure coefficients for components/cladding

(1) Peak external pressure coefficients Ĉpe for buildings with rectangular sections and aspect ratios
H/ B1 B2 of greater than 2
- C6-52 - Recommendations for Loads on Buildings

Peak external pressure coefficients for components/cladding correspond to the most critical positive
and negative peak pressure coefficients irrespective of the wind direction. Positive pressures occur on
the windward walls, and their characteristics are affected by the vertical profile of the approaching flow.
On the other hand, negative pressures (suctions) occur on the side and leeward walls, and their
characteristics are not significantly affected by the vertical profile of the approaching flow; that is, the
vertical distribution is nearly uniform. Large negative pressures occur near the windward edges of the
sidewalls owing to a flow separation from the windward edge.
The peak external pressure coefficients provided in Table A6.15 are determined from the results of
wind tunnel tests36, 37, 39, 59). The positive external peak pressure coefficients are given by the product of
the external pressure coefficients influenced by the profile of the mean wind speed and the gust effect
factor influenced by the profile of the turbulence intensity. Therefore, the coefficients are affected by
the terrain category around the building. On the other hand, negative external peak pressures, directly
specified in the table, are almost independent of the terrain category.
For tall buildings with recessed or chamfered corners, the negative peak pressures are influenced by
the size of the recess or chamfer. The values of Ĉpe in the recessed or chamfered corners are also
determined from the results of the wind tunnel tests38, 59). The values in Table A6.15 can also be used
for buildings with more than one recessed or chamfered corner, although the figure in the table shows a
building with only one.
The peak external pressure coefficients for the roofs are provided only for flat roofs with rectangular
sections. For diagonal wind directions, very large suctions are induced near the windward corners owing
to the generation of conical vortices. However, the large suction zone is limited to a relatively small
area60). Therefore, the use of such large peak pressure coefficients for large components may
overestimate the design wind loads. To consider the tributary area AC (m2) of the components/cladding
in zone Rc, an area reduction factor kC for the roofs is introduced. The value of AC (m2) should be
determined appropriately considering the practical area supported by each supporting member,
particularly when the tributary area is relatively large and the rigidity is relatively low.
The provisions are applicable to buildings with H / B1 ≤ 8 , because the values are based on wind
tunnel tests on such buildings.

(2) Peak external pressure coefficients Ĉpe for buildings with rectangular sections and aspect ratios
H / B1 B2 of less than or equal to 2
1) Buildings with flat, gable, mono-sloped, and hip roofs
For estimating the peak pressure coefficients for components/cladding of buildings with H / B1 B2 ≤
2, the tributary area is assumed to be 1 m2 as a typical value. Positive peak external pressure coefficients
are given as a function of the turbulence intensity because the pressures depend significantly on the
turbulence of the approaching flow. The positive peak external pressure coefficient on a roof is evaluated
using the positive external pressure coefficient C pe for zone Ru in Table A 6.9(1). If no positive value
of C pe is provided for small roof angles, it is not necessary to evaluate the positive wind pressures.
CHAPTER 6 WIND LOADS - C6-53 -

Negative peak external pressure coefficients in the edge and corner regions are significantly influenced
by vortices related to flow separation at the edge61). Therefore, in the previous recommendations (2004),
it was assumed that the negative peak in the external pressure coefficients on a roof do not depend on
the turbulence intensity of an approaching flow. The values were specified based on wind tunnel tests
using a turbulent boundary layer with a power law exponent of α = 0.25 and turbulence intensity of IH
= 0.17. However, recent studies on low-rise buildings, such as residential houses, have indicated that
the negative peak wind pressure coefficients are influenced by the turbulence intensity62). In the present
recommendations (2015), the values for the zones except Wb and Rf are modified into those for IH =
0.26, which corresponds to Terrain Category III. The negative peak pressure coefficients Ĉ pe in the
edge and corner zones of a roof and wall are given approximately through the following equations:

 (1 + 17.5 I H )Cpe （wall）

Cˆ pe =  (A6.2.3)
 (1 + 12.0 I H )Cpe （roof）

where IH is the turbulence intensity at the mean roof height, and C pe is the mean external pressure
coefficient. The values of Ĉ pe for IH = 0.26 and 0.17 were computed from Eq. (A6.2.3), and the ratio
of these two values was used as a correction factor for modifying the negative peak pressure coefficients.
High suctions (negative pressures) are induced in the edge and corner regions of walls and roofs, the
widths of which are affected by the building dimensions, such as the height, width, and depth.
For gable roofs, very high suctions are induced near the corners (zone Rb) when the roof angle is as
small as θ ≤ 10°, and in the ridge corner (zones Rb and Rg) when θ ≈ 20 °. For mono-sloped roofs with
θ ≥10°, very high suctions are induced near the higher eaves corners (zone Rd); these suctions are larger,
and the high suction area is wider, than those for gable roofs. However, such phenomenon was not
observed when θ < 10°, and the negative peak pressure coefficients are similar to those for flat roofs.
The peak external pressure coefficients on hip roofs are specified based on the results of wind tunnel
tests with models of typical detached houses44). In the tests, the turbulence intensity IH at the mean roof
height was about 0.25, which corresponded to Terrain Category III. Very high suctions are induced in
zone Rb when θ ≈ 10° and in zones Rb and Re when θ ≈ 20° . In such high suction zones, the wind
load can be reduced using the area reduction factor kC when the load burdened area AC is greater than 1
m2 in the same manner as for gable roofs60, 61). The specifications cannot be applied to buildings with
large B2/B1 ratios, because such buildings were not considered in the wind tunnel tests.

2) Buildings with vaulted roofs

The peak external pressure coefficients Ĉpe are determined from the results of the wind-tunnel
tests46), focusing on medium-scale buildings in urban areas, in which the h / B1 ratio is varied from zero
to 0.7, and the f / B1 ratio from 0.1 to 0.4. When the f / B1 ratio is small, the corner and edge regions
of a roof are significantly affected by the vortex generation, as in a flat roof case. This results in larger
peak suctions in zones R a and Rd . When the f / B1 ratio is relatively large, large peak suctions are
induced in zone Rd for winds nearly perpendicular to the gable edge, and in zone Rc for winds nearly
perpendicular to the eaves.
- C6-54 - Recommendations for Loads on Buildings

Taking these wind pressure features into account, the roof is divided into several zones, and positive
and negative peak external pressure coefficients are provided for them, as shown in Table A6.16(2).
When the f / B1 ratio is lower than 0.1, the roof is subjected to higher suctions similar to gable and
mono-sloped roofs. Therefore, it is not necessary to evaluate the positive peak external pressure
coefficients. The values for walls can be determined from Table A6.16(1).

(3) Peak external pressure coefficients Ĉpe for buildings with circular sections and spherical domes
The peak external pressure coefficients in Table A6.16(3) are determined from the results of the wind
tunnel tests47). External pressures on the domes fluctuate significantly owing to the effects of turbulence
of an approaching flow as well as through a vortex generation. Therefore, both positive and negative
peak pressure coefficients are provided. Because the geometry of a spherical dome is axisymmetric, it
is divided into three zones ( R a , R b , and R c ) by coaxial circles. When the rise/span ratio ( f / D) is
small, external pressures with a negative peak become larger in magnitude near the windward edge
(zone R a ) owing to the flow separation at the windward edge. On the other hand, when the f / D
ratio is large, external pressures with a large positive peak are induced near the windward edge owing
to the direct influence of the approaching flow. Therefore, the external pressure coefficients with a
positive peak for zone R a are provided as a function of the turbulence intensity I H at the reference
height H of the approaching flow when f / D ≥ 0.2 .

(4) Peak external pressure coefficients Ĉpe for buildings with circular and elliptical sections
For buildings with circular and elliptical sections, the external pressure coefficient with the maximum
positive peak occurs at the stagnation point on the windward face, whereas the external pressure
coefficient with the maximum negative peak occurs near the point of the maximum negative mean of
the external pressure63). The vertical distribution of pressure coefficients with a positive peak depends
strongly on the mean velocity profile of the approaching flow in the same manner as that for buildings
with rectangular sections. On the other hand, external pressure coefficients with a negative peak for
buildings with circular sections are influenced by the aspect ratios and surface roughness of the buildings.
Factor k1 considers the effect of the aspect ratio, whereas factor k 2 considers the effect of the surface
roughness in the transcritical Reynolds number regime. The external pressure coefficients with a
negative peak become larger in magnitude near the top of the building because of the flow separation
from the top (i.e., the end effect). Factor k 4 considers this effect. In the buildings with elliptical
sections, it is known that the separation points change based on the effect of the side ratio D1 / D2 , and
the local negative pressures appear in wide zones54, 64). Therefore, factor k5 considers the effect of the
side ratio D1 / D2 , and factor k 6 considers the effect of the local negative pressure areas. The values
in Table A6.17 are applicable to buildings with H / D2 ≤ 8 and 1 ≤ D1 / D2 ≤ 3 because the provision
is based on wind tunnel tests using such models.
Only negative peak pressure coefficients are considered for roofs of buildings with circular sections.
The values of Ĉpe for domes with f / D = 0 provided in Table A6.16(3) can be used. However, there
CHAPTER 6 WIND LOADS - C6-55 -

have been no examples of negative peak pressure coefficients on roofs of buildings with elliptic sections,
and verification through wind tunnel tests or CFD is expected.

A6.2.6 Factor for effect of fluctuating internal pressures

(1) Basic concept
The peak wind force coefficients for components/cladding shall be determined from the maximum
instantaneous values, both positive and negative, of the pressure difference between the exterior and
interior surfaces. However, there are very few data on these pressure differences. In the present
recommendations (2015), it is assumed that the peak wind force coefficient ĈC is represented through
Eq. (A.6.15) because the peak external pressure coefficients Ĉpe are usually obtained from wind tunnel
tests, and a large amount of data for Ĉpe is available.
Figure A6.2.5 shows a schematic illustration of the fluctuating external and internal pressures. The
frequency of the internal pressure fluctuations is lower than that of the external pressure fluctuations,
and the peak external and internal pressures are not induced simultaneously. Factor Cpi* for the effect
of fluctuating internal pressures in Eq. (A6.15) does not represent the peak internal pressure coefficient
itself but an equivalent value providing the actual peak wind force when combined with the peak external
pressure coefficient Ĉpe . The value of ĈC is evaluated from a series of computations for the peak
wind force coefficients using wind tunnel data on Ĉpe for various building configurations. The
following assumptions are made in the computations49):
1) Gaps and openings in the external walls are uniformly distributed, and the internal pressures are
generated from the external pressures at the locations of the gaps and openings.
2) The fluctuating internal and external pressures are independent of each other.
Recently, numerical simulations on fluctuating internal wind pressures have been conducted by
considering the gaps distributed in partitions and external walls65),66). For example, peak internal wind
pressure coefficients of approximately zero through -0.6 are obtained for the case of a high-rise building
with highly air-tightened external walls and low air-tightened partition walls. These are consistent with
values regulated in A6.2.6.
- C6-56 - Recommendations for Loads on Buildings

Peak value of wind force Factor for effect of fluctuating

internal pressures
Peak value of external wind pressure
Wind force and pressures

風力
Wind force
外圧
External wind pressure
内圧
Internal wind pressure
0 time

Peak value of internal wind pressure

Fig. A6.2.5 Example of fluctuating external and internal pressures acting on components/cladding

(2) Treatment of dominant opening

As described in A6.2.3, an area with large openings or an envelope damaged by wind-borne debris
can be much larger than that of usually distributed gaps in a building. The characteristics of internal
wind pressure in such cases differ from those of enclosed buildings. For a building with a dominant
opening, it is necessary to properly evaluate the peak wind force coefficient using experimental or
numerical data related to the external and internal wind pressures because it is not permitted to use
internal wind pressure coefficients of zero and -0.550).
For the case of a building model with a dominant opening on the windward wall or sidewall, Fig.
A6.2.6 illustrates an example of how to evaluate the factor for the effect of fluctuating internal
pressures38),50), in which the opening area ratio of the dominant opening defined in Eq. (A6.2.1) is applied
to the experimental result. For the case of r equal to or less than 0.5, the factor is zero through -0.5,
which means that the case is consistent with that of an enclosed building. On the other hand, for the case
of r of more than 0.5, the factor is different from that of an enclosed building. Some regulations and
standards in other nations involve internal wind pressure coefficients for buildings with a dominant
opening on the wall or glass potentially subjected to wind-borne debris 22, 67).
fluctuating internal pressures

fluctuating internal pressures

2.5 0.0
2.0
Factor for effect of

Factor for effect of

-0.5
1.5
1.0 -1.0
0.5
-1.5
0.0
-0.5 -2.0
0.1 1 10 100 1000 0.1 1 10 100 1000
r=0.5 r=0.5
opening area ratio r opening area ratio r
Note) Each symbol indicates a difference in the conditions of the model configuration

Figure A6.2.6 Relation between opening area ratio r and internal wind pressure
CHAPTER 6 WIND LOADS - C6-57 -

A6.2.7 Peak wind force coefficients for components/cladding

For free roofs, it is necessary to directly evaluate the wind force represented by the pressure difference
between the top and bottom surfaces. The provisioning of the peak wind force coefficients are based on
previous wind tunnel tests for the most critical peak wind force irrespective of the wind direction68) and
other country standards22, 24). However, the models used in wind tunnel tests have a small roof angle,
and therefore the range of the roof angle is limited to −20° to 20°. When the roof angle is relatively large,
large peak wind forces are induced along the roof edges as well as along the ridge, because large suctions
are induced by conical vortices on either the top or bottom surfaces of the roof. The roof is divided into
several zones, and positive and negative peak wind force coefficients are provided for each zone as a
function of roof angle θ.
When any obstruction whose blockage ratio is larger than approximately 50% is placed under the roof,
it is necessary to evaluate the peak wind force coefficients from appropriate wind tunnel tests, CFD, and
other measures.

A6.3 Gust effect factor on along-wind loads on structural frames

(1) Fundamental consideration
In the present recommendation, gust effect factor GD is based on the overturning moment, as described
through the following equation.
M D max M D + g Dσ MD g σ
GD = = = 1 + D MD (A6.3.1)
MD MD MD
where MDmax (Nm), M D (Nm), σMD (Nm), and gD are the maximum value, mean value, and rms of the
overturning moment at the base of the building, and the peak factor, respectively. In addition, MDmax and
σMD involve the load effect owing to the dynamic response of the building. If σMD(Nm) is expressed as
a composition of the background component σMDQ (Nm) and resonance component σMDR (Nm), Eq.
(A6.3.1) becomes as follows.
g D σ MDQ
2
+ σ MDR
2
σ MDQ πf S ( f )
GD = 1 + ≈ 1+ gD 1 + φ D2 D MD 2 D , (A6.3.2)
MD MD 4ς Dσ MDQ
where SMD(fD) is the power spectral density of the overturning moment at the natural frequency for the
first mode fD (Hz), φD is the mode correction factor, and ζD is the damping factor for the first mode. In
addition, σMDR (Nm) is considered for only the first mode vibration, and σMDR (Nm) is the inertia force
through vibration, as described in the following equation.
σ MDR = ∫0 σ a ( Z ) m( Z ) ZdZ = σ a ( H ) ∫0 µ ( Z ) m( Z ) ZdZ
H H
(A6.3.3)
where σa(Z)，m(Z)，and μ(Z) are the rms of the acceleration at height Z(m), mass per unit height, and
vibration mode, respectively.

The parameters of Eq. (A6.3.2) are expressed through the aerodynamic force coefficients as follows.
M D = qH BH 2CMD (A6.3.4)
- C6-58 - Recommendations for Loads on Buildings

′
σ MDQ = q H BH 2 C MD (A6.3.5)
* *
f D S MD ( f D ) f S CMD ( f )
= D D
(A6.3.6)
σ MDQ
2
′2
C MD
where CMD is the overturning moment coefficient, C’MD is the rms overturning moment coefficient, and
SCMD(f*D) is the power spectrum of the overturning moment coefficient at non-dimensional frequency
f*D. If these equations are taken into consideration, Eq. (A6.3.2) becomes the following:
′
C MD πf * S ( f * )
GD ≈ 1 + g D 1 + φ D2 D CMD 2 D . (A6.3.7)
C MD ′
4ς D C CMD
Additionally, in this formula, the non-dimensional frequency is defined through the turbulence scale
LH(m), where f*D = fDLH/UH; however, the building width B(m) is usually used under f*D=fDB/UH during
wind tunnel tests .
If the values of CMD, C’MD, and SCMD(f*D) at non-dimensional frequency f*D are defined through
properly conducted wind tunnel tests, the gust loading effect factor can be calculated using Eq. (A6.3.7).
(2) Wind force model
The wind force model is based on the assumption that the wind velocity fluctuation is directly changed
into the wind pressure on the wall of the building69). In this model, the mean wind velocity, turbulence
intensity, power spectral density of the wind velocity, and co-coherence are described through Eqs.
(A6.2), (A6.6), (A6.8), (A6.1.4), and (A6.1.5), respectively. Additionally, the wind force coefficient CD
is expressed by the difference in wind pressure coefficient CPA of the windward side and the wind
pressure coefficient (constant) CPB of the leeward side, as described through the following equation.
2α
Z
C D = C PA   − C PB (A6.3.8)
H
Here, CMD，C’MD, and SCMD(f*D) are expressed using the parameter of the recommendation equations,
as follows.
C MD = C H C g (A6.3.9)
′ = C H Cg′
C MD (A6.3.10)
f S CMD ( f ) = C ′ FD
*
D
*
D
2
MD
(A6.3.11)
where CH is the wind force coefficient at the top of the building, Cg is a factor relevant to the overturning
moment in the along-wind direction, C’g is a factor relevant to the rms overturning moment in the along-
wind direction, and FD is a spectral factor of the windward force. The spectral factor of wind velocity F,
size effect factor SD, and correlation coefficient of the wind pressure on the windward and leeward sides
R are considered for FD.
Characteristics of the overturning moment expressed through Eqs.(A6.3.9)−(A6.3.11) are shown in
Fig. A6.3.1 in comparison with those obtained from the wind tunnel tests in the case of H / BD = 4 .
The recommended values of the overturning moment and rms overturning moment are slightly greater
than the test values, and the spectral density is mostly in agreement with the test values.
CHAPTER 6 WIND LOADS - C6-59 -

1.5 1.5
10 -2

Recommendation vale/test value

1.0 1.0
10 -3

fSCMD(f)
terrain
地表面粗度区分 category 地表面粗度区分
terrain category
0.5 0.5 10 -4
II II 実験値
test
III III 指針値
recommendation
IV IV
0.0 0.0 10 -5 -3
0 1 2 3 0 1 2 3 10 10 -2 10 -1 10 0

Side ratio D/B Side ratio D/B fB/UH

(a) mean overturning moment (b) rms overturning moment (c) power spectral density of
ratio ratio over turning moment

Figure A6.3.1 Along-wind force in comparison with those obtained from the wind tunnel tests70)

(3) Fluctuating component of overturning moment

When the vibration mode is μ=Z/H, the relation between the spectrum of the overturning moment
owing to the wind force SMD(f), and the spectrum of the overturning moment owing to the load effect
by vibration S’MD(f), is expressed through the following equation.
′ ( f ) = χ m ( f ) S MD ( f ) ,
2
S MD (A6.3.12)
where χ m ( f ) 2
is the mechanical admittance, as expressed through the following equation.
1
χm ( f ) =
2
(A6.3.13)
{1 − ( f / f ) }
+ 4ς D ( f / f D ) 2
D
2 2

The variance of overturning moment owing to the load effect through vibration σ2MD (N2m2) is the
integral of Eq. (A6.3.12), and the variance consists of background component σMDQ (Nm) and resonance
component σMDR (Nm), as expressed through the following equation.
∞
σ MD
2
= ∫0 S MD
′ ( f )df ≈ σ MDQ
2
+ σ MDR
2

∞ ∞ π f S (f ) (A6.3.14)
∫0 SMD ( f )df + ∫0 SMD ( f D ) χ m ( f ) df =
= σ MDQ + D MD D
2 2

4ς D
In this equation, the resonance component is estimated approximately as a response to the wind
force power spectral density SMD(fD) at the natural frequency fD (Hz).
Therefore, the overturning moment for the maximum load effect is expressed through the following
equation.
M Dmax = M D + g D σ MDQ
2
+ σ MDR
2
, (A6.3.15)
where gD is expressed through the following equation based on the theory of a stationary stochastic
process.
0.577
g D = 2 ln(ν D T ) + ≈ 2 ln(ν D T ) + 1.2 , (A6.3.16)
2 ln(ν D T )
where T (s) is time of the evaluation and νD(Hz) is level crossing rate calculated from the power
- C6-60 - Recommendations for Loads on Buildings

spectral density, as in the following equation.

∞

νD =
∫0
′ ( f )df
f 2 S MD
≈ fD
RD
, (A6.3.17)
∞
1 + RD
∫ 0
′ ( f )df
S MD
where RD is the resonance factor for the along-wind vibration.
Additionally, in certain foreign wind loading standards, MDmax(Nm) is expressed through the
following equation. 71)
M Dmax = M D + g Q2 σ MDQ
2
+ g R2 σ MDR
2
(A6.3.18)
where gQ is the peak factor of the background component (= 3.4), and gR is the peak factor of the
resonance component calculated from Eq. (A6.3.16) as νD = fD (Hz). In this equation, the background
component and the resonance component are distinguishable.
(4) Vertical distribution of equivalent static wind loads
With the gust effect factor method, the vertical distribution of the wind loads is given based on the
mean wind loads multiplied by the gust effect factor. These wind loads are an approximate value based
on the assumption that the vibration mode is close to the mean wind load distribution, and the building
has a uniform density. Actually, the mean, background, and resonance components of the wind load
distribution are different. The mean component is expressed through Eq. (A6.3.8), and the resonance
component in the case of a uniform mass distribution in height is expressed through Eq. (A6.3.3).
Therefore, if the vertical distribution of the building mass is remarkably uneven, the resonance
component should be estimated carefully. In this case, the distribution of the resonance component for
the fundamental vibration mode can be estimated from the following equation.
WD = WD + WDQ
2
+ WDR
2
(A6.3.19)
where
WD = q H C D A
C g′
WDQ = g DQ q H C D A
Cg

WDR = α Dmax µ (Z )m(Z )

A
B
in which
WD ， WDQ ，and WDR (N) are the mean, background, and resonance component of the wind
loads, respectively. In addition,
2
aDmax(cm/s ): the maximum acceleration at the top of building, as defined in A6.13.2,
gDQ: the peak factor of the background component.
In the present recommendation, it is assumed that the background component has a similar
distribution as the mean component. The following methods may also be used.
1) Shear force or overturning moment at a certain building height may be obtained from the integral of
the pressure on an area over the height. 22)
2) The load distribution can be defined through the load response correlation (LRC) formula72).
CHAPTER 6 WIND LOADS - C6-61 -

(5) Example calculation of gust effect factor

Figure A6.3.2 shows the variation in gust effect factors for each terrain category and each building
1
height for H/B = 4, B/D=1, f D = , ζ D =0.01, and U0=35m/s. The gust effect factors become
0.024 H
larger with the terrain category and smaller with the building height.

3.8
terrain category
3.6
3.4 Ⅴ
V
Gust effect factor GD

Ⅳ
IV
3.2 Ⅲ
III
3.0 Ⅱ
II
Ⅰ
I
2.8
2.6
2.4
2.2
2.0
00 50 100
100 150 200
200 250 300
300
H (m)

Figure A6.3.2 Gust effect factor

(6) Gust effect factor for rigid structures

If the following conditions are applicable, the resonance component can be neglected for a relatively
small scale and rigid structure.
i) H ≤ 15m
H
ii) ≤ B ≤ 30m
2
In this case, the gust effect factor can be calculated using the following equation, under the
assumption of RD = 0, gD = 4.
C g′
GD = 1 + 4 (A6.3.20)
Cg
(7) Gust effect factor for a fence on the ground and members
For a fence on the ground, the wind force coefficients of which are given in Table A6.13, and for the
members of structures whose wind force coefficients are given in Table A6.14, the quasi-static wind
loads are evaluated using the same method as for free roofs with a rectangular base given in section
A6.8.
The gust effect factor GD is calculated using Eq. (A6.3.21) for a fence on the ground, or using Eq.
(A6.3.22) for the members of the structures.
G D = (1 + 3I H ) 2 (A6.3.21)
G D = (1 + 3I Z ) 2
(A6.3.22)
where
I H : turbulence intensity at the top of the fence H (m),
- C6-62 - Recommendations for Loads on Buildings

I Z : turbulence intensity at the height of the member Z(m).

(8) Mode shape correction factor
In the present recommendation, the mode shape correction factor φD can be used in calculating the
gust effect factor, the across-wind loads, and the torsional wind loads for a conventional building, as
described in A.6.3, A6.5.2, and A6.6.2, respectively. For the basic conditions of the calculation, the
vertical distribution of mass per height of a building over the ground is nearly uniform, and the basic
first translation mode shape is represented in the formula µ(Z) = (Z/H)β; in addition, the calculated value
using the basic first translation mode can be corrected using the mode shape correction factor φD. The
mode shape correction factor for the wind loads is the product of the correction factor λ for a generalized
wind force and the correction factor for a generalized mass or generalized inertial moment.
The value of λ is defined through a simple approximation of the wind tunnel test results of
rectangular section structures (D/B = 0.2–0.5) and its mode power index β = 0.2–4. Therefore, when the
mode shape correction factor is used, be careful about value range of D/B and β .70)
In the case of a vertical distribution of mass per height of a building over the ground, m(Z) is not
∫0 m(Z )(Z / H ) dZ for β = 1, and the generalized
H 2
regarded as a constant, the generalized mass MD1=
mass MD = ∫0 m(Z )(Z / H ) dZ for µ(Z) = (Z/H),β are given in the present recommendations to enable
H 2β

a correction for a generalized mass and generalized inertial moment.

(9) Mode shape correction for base-isolated structure
As shown in Fig. A6.3.3, the mode shape µ(Z) of a base-isolated structure is represented using the
mode displacement µ0 at the base isolator in the following equation.

𝑍𝑍
𝜇𝜇(𝑍𝑍) = 𝜇𝜇0 + �1 − 𝜇𝜇0 � (A6.3.23)
𝐻𝐻

Figure A6.3.3 shows an example approximation of the mode shape of a base-isolated structure. The
value of β is approximated with µ0 using Eq. (A6.3.25).

μ(Z) = (Z/H)β (A6.3.24)

𝛽𝛽 = 0.51𝜇𝜇20 − 1.5𝜇𝜇0 + 0.99 (A6.3.25)

In this case, the difference between the mode value and its approximation affects the generalized
mass and generalized wind force. However, through trial analyses based on the wind tunnel test results
with ζ = 0.02, it was confirmed that the response analysis using a mode correct factor φD provides
applicable results under the condition of an applicable value range of D/B and β for λ.
The typical nonlinear characteristic model of base isolation layer is a bi-liner model. It should be
noted that the response analysis results using the mode shape of Eq. (A6.3.23) or Eq. (A6.3.24) is a
result of an equivalent liner analysis. To evaluate the accurately yielded range of wind response of a
base isolation system, a nonlinear time history response analysis is recommended.
CHAPTER 6 WIND LOADS - C6-63 -

z=H 1 1 1
ββ=0.85
=0.85 β
β=0.37
=0.37 ββ=0.19
=0.19
z/H z/H z/H
0.8
μμ0=0.1
=0.1 μμ0=0.5
=0.5 μμ0=0.7
=0.7
0 0.8 0 0.8 0

µ0=0 0.6 0.6 0.6

µ0=1
0.4 0.4 0.4

0.2 0.2 0.2

z=0 0 0 0
Base isolation layer 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
µ µ µ

Figure A6.3.3 Base-isolated Figure A6.3.4 Example of base-isolated mode

mode approximation

A6.4 Fluctuating Roof Wind Loads on Structural Frames

A6.4.1 Scope of application
The procedure described in this section applies to the equivalent static wind load in consideration of
a vertical forced roof vibration at a design wind speed of lower than the critical wind speed for
aeroelastic instability.
Along-wind vibration is caused by turbulence in natural wind, but the vertical vibration of a roof is
caused not only by the wind turbulence but also through the separation of shear layers from the leading
edges. Although there have been many studies with regard to the behavior of separated shear layers from
the leading edges, unclear points remain. Furthermore, because the behavior of separated shear layers is
greatly affected by the roof shape, it is difficult on the whole to theoretically estimate the vertical
vibrations of a roof in the same manner as the along-wind vibrations. Focusing on the first mode of a
flat roof beam, an estimation equation for the roof load has been derived from data on the generalized
wind force coefficients obtained from the results of the wind tunnel tests.

A6.4.2 Procedure for estimating wind loads

Because the resonance effects caused by the wind force on a roof beam cannot be neglected, the roof
wind loads on the structural frames are calculated from Eq. (A6.17). The characteristics of the wind
force acting on a roof greatly depend on the shape and height of the roof. Equation (A6.17) is derived
from the generalized wind force coefficients for the first vibration mode obtained through the wind
tunnel test for a flat roof supported by beams with a rectangular section, as shown in Fig. A6.4.1.
The generalized wind force coefficients are calculated from an external pressure coefficient
database73) of flat roofs with rectangular sections when the wind direction is normal to the wall. There
are 12 types of buildings to be considered, one with a width of 16 m, three with depths of 16, 24, and
40 m, and four types with reference heights of 4, 8, 12, and 16 m. There are four types of roof beams to
be considered, leading edge and center roof beams normal to the wind direction, and edge and center
- C6-64 - Recommendations for Loads on Buildings

roof beams parallel to the wind direction.

The generalized fluctuating wind force coefficients C’R,ext calculated from only the external pressure
can be approximated through a summation of the functions with H/D and H/B as variables. A
comparison of the approximate values of C’R,ext with the experimental values is shown in Fig. A6.4.2 for
each type of roof beam. In the present recommendations, for roof beam parallel to the wind direction,
because the difference in wind load characteristics between the edge roof beams and center roof beams
is small, the approximate expression is not dependent on the position on the roof. For a roof beam normal
to the wind direction, the difference in wind load characteristics between the leading edge roof beams
and center roof beams is large. Therefore, the approximate expression of C’R,ext is expressed through a
function of y/D using values of the leading edge (y/D = 0.1) and center (y/D = 0.5). The fluctuating
internal pressure coefficients are calculated to be 0.10 under the assumption that gaps are uniformly
located on the wall using the external pressure coefficients database73) of flat roofs with rectangular
sections. The generalized fluctuating internal pressure coefficients are calculated as 0.07 by multiplying
the fluctuating internal pressure coefficients by the first mode under the assumption that the external
pressure and the internal pressure are uncorrelated. In this way, the generalized fluctuating wind force
coefficients C’R in Eq. (A6.17) are calculated.

D B
B D
L L

x x y

roof beam roof beam

πx
first mode of roof beam: sin
L
(a) For roof beam parallel to the wind direction (b) For roof beam normal to the wind direction
Figure A6.4.1 Roof beams to be applied and first mode of the roof beam
CHAPTER 6 WIND LOADS - C6-65 -

leading edge roof beam normal to the wind direction

center roof beam normal to the wind direction
edge roof beam parallel to the wind direction
center roof beam parallel to the wind direction
0.3
C'R,ext (Approximation formula)

0.2

0.1

0
0 0.1 0.2 0.3
C'R,ext (Experimental value)
Figure A6.4.2 Comparison of experimental value and approximation formula for C’R,ext

Analysis examples of the power spectral density of the generalized fluctuating wind force CR,ext are
shown in Fig. A6.4.3. When non-dimensional frequency fB/UH is taken as the horizontal axis of the
figure, the influence of the reference height H (m) is small, and has almost the same shape.

0.1
fS(f)/σ2

0.01 H=4m
H=8m
H=12m
H=16m

0.001
0.01 0.1 1 10
fB/UH

Figure A6.4.3 Power spectral density of CR,ext

(D/B=1.5, leading edge roof beams normal to the wind direction)
- C6-66 - Recommendations for Loads on Buildings

The power spectral density of CR,ext is approximated by FR in Eq. (A6.17), and variables r1 through r6
in the equation of FR are calculated for each D/B. For roof beam normal to the wind direction, the power
spectral density of CR,ext can be represented if the variables r2 and r6 are quadratic functions of D/B, as
shown in Fig. A6.4.4, and the other variables are constants. For a roof beam parallel to the wind direction,
the power spectral density of CR,ext can be approximately represented if variables r1 through r6 are
constants regardless of D/B.
Figure A6.4.5 compares the power spectral density of CR,ext calculated from the results of wind tunnel
tests with FR. It is understood that both spectral densities approximately agree.

2.5 0.10

2.0 0.08

1.5 0.06
r2

1.0 0.04

0.5 0.02

0.0 0.00
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
D/B D/B
Figure A6.4.4 Relation between variables r2, r6 and D/B (roof beams normal to the wind direction)

1 1

0.1 0.1
fS(f)/σ2

fS(f)/σ2

0.01 0.01
leading edge(Exp.) edge(Exp.)
center(Exp.) center(Exp.)
FR FR
0.001 0.001
0.01 0.1 1 10 0.01 0.1 1 10
fB/UH fB/UH

(a) For roof beam normal to the wind direction (b) For roof beam parallel to the wind direction
Figure A6.4.5 Comparison of power spectral density of CR,ext and variable FR (D/B=1)

Even if the shape and structural characteristics of the roof are somewhat different from a flat roof
with a rectangular section, and if the properties of the approaching flow are relatively similar, the scale
effect is not significantly different from the case of a flat roof calculated herein. Therefore, in the case
CHAPTER 6 WIND LOADS - C6-67 -

of a roof where the resonance effect caused by the wind force can be ignored, that is, when Eq. (A6.4.1)
is satisfied, if the roof gradient is 30° or less, the fluctuating roof wind loads can be calculated through
Eq. (A6.4.2).
fR H
> 1.5 (A6.4.1)
UH

WR' = 2.4q H AR (A6.4.2)

where
f R (Hz) : natural frequency for the first mode of the roof beam,
H (m) : reference height as defined in 6.1.2(11),
U H (m/s) : design wind speed as defined in A6.1.2,
'
W R (N) : fluctuating roof wind load,
2
qH (N/m ) : design velocity pressure as defined in A6.1.1,
AR (m2) : subject area of the beams.

A6.5 Across-Wind Loads on Structural Frames

A6.5.1 Scope of application

The procedure described in this section applies to the equivalent static wind load in consideration of
the across-wind forced vibration at a design wind speed lower than the critical wind speed for vortex-
induced vibration or an unstable aeroelastic oscillation.
Along-wind vibration is caused by turbulence in natural wind, but across-wind vibration is caused not
only by wind turbulence but also by the vortex shedding in the wake of the building. Although there is
a lot of research with regard to the behavior of vortices in the wake of a building, unclear points remain.
Furthermore, because the behavior of the vortices is greatly affected by building shape, it is difficult on
the whole to estimate the across-wind vibrations theoretically in the same manner as for along-wind
vibrations. Thus, an estimation equation for an across-wind load in consideration of the first vibration
mode of a building has been derived from data on the across-wind fluctuating overturning moment
obtained through wind tunnel tests. The subjects for this estimation equation are structures with a
rectangular section (side ratio of D/B = 0.2–5) for which a large amount of experimental data have been
obtained. By taking into account the fact that the experimental data for buildings with an aspect ratio
H / BD exceeding 6 are insufficient, the scope of the application is limited to aspect ratios of 6 or
less. Using data on an across-wind fluctuating overturning moment of buildings with sectional shapes
other than the rectangular section obtained from the results of wind tunnel tests or CFD, the across-wind
wind loads can be calculated using the method indicated in the recommendations.

A6.5.2 Procedure for estimating wind loads

(1) Concept of wind load estimation
- C6-68 - Recommendations for Loads on Buildings

Because a fundamental mode usually predominates in across-wind vibration, across-wind loads are
calculated using the spectral modal method considering only the first translational mode in the same
manner as for along-wind loads. The across-wind load is calculated from the non-resonance component
and the resonance component of the response. For the non-resonance component, the vertical profile of
the fluctuating across-wind force is assumed to have a linear mode shape Z/H and the magnitude of the
fluctuating wind force is determined to agree with the fluctuating overturning moment. The resonance
component is estimated from the inertia force from vibrations, and the vertical profile of the across-wind
load is determined using φL in A6.5.2 so as to be proportional to the first translational mode µ(Z)=(Z/H)β.
It is recommended that the damping factor be estimated with reference to “Damping in buildings” 9) and
the “Guideline of Recommendations for Loads on Buildings 1”91).

(2) Modeling of overturning moment

The overturning moment varies with the building shape and wind characteristics; however, in the
subjective scope, the side ratio D/B has the greatest effect on the overturning moment, and the effects
of other parameters are slight. Therefore, in the recommendations, the fluctuating overturning moment
is set as a function of only the side ratio D/B of a building based on the wind tunnel test data. Here,
C’L(=σL/(qHBH2), where σL is the standard deviation) in equation (A6.18) is the fluctuating overturning
moment coefficient for an across-wind, which is shown in Fig. A6.5.1 as a function of the side ratio D/B.
It is possible to use the fluctuating overturning moment coefficient for the across-wind obtained from
the results of a wind tunnel test or CFD.
Here, FL is a normalized power spectral density of the fluctuating overturning moment for the across-
wind, which is shown in Fig. A6.5.2 as a function of non-dimensional frequency fLB/UH and the side
ratio D/B. It is possible to use the power spectral density of the fluctuating overturning moment for the
across-wind obtained from the results of a wind tunnel test or CDF.
CHAPTER 6 WIND LOADS - C6-69 -

0.4

Flucuating orverturning moment coefficientCL’

CL’ ×1.15

0.3

0.2

CL’ ×0.85

0.1 CL’

: Experiment

0
0 1 2 3 4 5 6
Side ratio (D/B)

Figure A6.5.1 Fluctuating overturning moment coefficient in across-wind direction

10 10

1 1

D/B D/B
0.2 2
FL
FL

0.1 0.1
0.5 3
0.8 4
1 5
0.01 0.01

0.001 0.001
0.01 0.1 1 0.01 0.1 1
fLB/UH fLB/UH

Figure A6.5.2 Power spectral density of fluctuating overturning moment in across-wind direction

(3) Buildings with a circular section

Across-wind responses of buildings with sectional shapes other than rectangular sections can be
estimated through the same concept. This section details buildings with a circular section. The parameter
- C6-70 - Recommendations for Loads on Buildings

values used in A6.5.2 need to be set to C’L = 0.06, m = 1, κ1 = 0.9, fS1 = 0.15UH/B, and β1 = 0.2. These
parameter values correspond to the supercritical Reynolds number (UHD ≥ 6(m2/s)).

A6.6 Torsional Wind Loads on Structural Frames

A6.6.1 Scope of application

The procedure described in this section applies to the equivalent static wind load in consideration of
the torsional vibration with a design wind speed lower than the non-dimensional critical wind speed for
a vortex-induced vibration or an unstable aeroelastic oscillation.
Torsional vibration is induced through the torsional moment of the building, which is caused by an
asymmetrical wind pressure distribution on the windward and leeward faces. This is due to both wind
turbulence and the vortex shedding in the building’s wake. The torsional moment induced by these wind
forces is subject to the effects of the building shape and the characteristics of the wind. Therefore, the
method for assessing the torsional wind load is derived from the fluctuating torsional moment data
obtained from wind tunnel tests in the same way as for the across-wind direction.74) Subjects for this
estimation equation are buildings with rectangular sections (side ratio D/B = 0.2–5) and an aspect ratio
H / BD of 6 or less, for which a large amount of experimental data has been obtained.
Furthermore, using data on the torsional moment of buildings with sectional shapes other than
rectangular sections obtained from the results of the wind tunnel tests or CFD, the torsional wind loads
can be calculated using the method indicated in these recommendations.

A6.6.2 Procedure for estimating wind loads

(1) Concept of wind load estimation
Basically, the estimation method for a torsional wind load is the same approach as for an across-wind
load. A torsional-wind load is calculated from the non-resonance and resonance components of the
response. For the non-resonance component, the vertical profile of the fluctuating torsional moment is
assumed as linear mode shape Z/H, and the magnitude of the fluctuating torsional moment is determined
to agree with the generalized fluctuating torsional moment at the base of the building. The resonant
component is estimated to be the inertia force owing to the vibrations, and the vertical profile is
determined to be proportional to the first translational mode µ(Z)=(Z/H)β. Buildings with an eccentric
factor (eccentric distance/radius of rotation) of 0.15 or less for which any effect of eccentricity can be
ignored are subject to the estimation equation. The wind load on a building for which the eccentricity
cannot be ignored needs to be calculated based on the results of wind tunnel tests or CFD. It is
recommended for the damping factor to be estimated with reference to “Damping in buildings”9) and
“Guideline of Recommendations for Loads on Buildings 1”12).

(2) Modeling of torsional moment

Here, C’T(=σT/(qHB2H), where σT is the standard deviation) in Eq. (A6.19) is the fluctuating torsional
CHAPTER 6 WIND LOADS - C6-71 -

moment coefficient, which is shown in Fig. A6.6.1, as a function of the side ratio D/B. It is possible to
use the fluctuating torsional moment coefficient obtained from the results of the wind tunnel tests or
CFD.
In addition, FT is the normalized power spectral density of the fluctuating torsional moment, which is
shown in Fig. A6.6.2, as a function of non-dimensional frequency fTB/UH, the side ratio D/B, and
turbulent intensity IH. It is possible to use the power spectral density of the fluctuating torsional moment
obtained from the results of a wind tunnel test or CFD.
Fluctuating torsional moment

1.0

● : Experiment
0.8
coefficient

0.6

0.4
CT'
0.2

0.0
0 1 2 3 4 5
D/B

Figure A6.6.1 Fluctuating torsional moment coefficient

(Dotted line, CT' on A6.6.2; black circle, experimental value)
FT
FT

fTB/UH fTB/UH

(a) D/B=0.5 (b) D/B=1

Figure A6.6.2 Power spectral density of fluctuating torsional moment
(Solid line, FT on A6.6.2; circle, experimental value)
- C6-72 - Recommendations for Loads on Buildings

A6.7 Horizontal Wind Loads on Lattice Structural Frames

A6.7.1 Scope of application

This procedure was prepared to estimate the horizontal wind loads on lattice structures built directly
on the ground, and whose members all have sufficiently small sections in comparison with the width of
the structure for the flow field around a member to be dominated by the local wind speed. The procedure
for estimating the wind loads on lattice structures is basically the same as that described for horizontal
wind loads on buildings in 6.2, and can be applied to lattice structures of varying widths and solidity
ratios in the vertical direction. In addition, the effects of accessory ladders are considered through an
evaluation of the wind force coefficients of those obtained from wind tunnel tests, CFD, and other
measures, or through the application of wind force coefficients for members given in A6.2.4(5).

A6.7.2 Procedure used for estimating wind loads

Horizontal wind loads are estimated using a gust effect factor method72). The wind loads are calculated
from the local design velocity pressure because lattice structures often have varying widths and solidity
ratios in the vertical direction.
The projected area in Eq. (A6.20) is the total projected area of all elements on one face normal to the
wind. The area per panel is typically calculated.

A6.7.3 Gust effect factor

In deriving Eq. (A6.21), the gust effect factor is assumed as follows:
i) Solidity ratios in the vertical direction are uniform, that is, the wind force coefficients of
each panel are uniform.
ii) The fundamental vibration mode shape can be given through Eq. (A6.7.1), where β = 2 ,
and higher modes are neglected.
β
Z
µ (Z ) =   (A6.7.1)
H
According to the above assumptions, the peak response xmax,Z at height Z is given as a function
of the generalized stiffness of the fundamental mode by
q H C D HB0 2 I H µ ( Z )
xmax,Z = g D BD (1 + RD ) . (A6.7.2)
K 0.95 + α + β
However, the mean response X Z at height Z is given by
qH CD H  B0 B − BH 
XZ =  − 0 µ (Z ) , (A6.7.3)
K  1 + 2α + β 2 + 2α + β 
where q H and I H are the velocity pressure and turbulence intensity, respectively, at height H, and
α is the exponent of the power law in the wind speed profile. In addition, g D , RD , and BD are the
peak factor, the resonance factor, and the background excitation factor, respectively, and B0 and BH
in Fig. A6.7.1 are the width of the lattice structure at the ground level and the width at height H ,
CHAPTER 6 WIND LOADS - C6-73 -

respectively.
The gust effect factor is given through Eq. (A6.21).
Furthermore, φ D given by Eq. (A6.21) is the mode shape correction factor. This corrects the gust
effect factor for an along-wind load on a lattice structure according to its vibration mode, and can be
used in calculating the gust effect factor if the first mode shape function is different from
µ ( Z ) = ( Z / H ) 2 and the mass per height of a lattice structure is not regarded as almost constant. The
mode shape correction factor can be applied with β ranging from 1 to 3.5 for a lattice structure when
the mode shape function can be approximated through the function µ ( Z ) = ( Z / H ) β .
The mode shape correction factor φ D can be derived by multiplying the correct factor λL given by
Eq. (A6.7.4) of the generalized wind force by the correct factor M D 2 / M D of the generalized mass of
a lattice structure. In addition, λL given by Eq. (A6.7.4) is the ratio of the resonance component of the
generalized wind force for its first vibration mode (the power index of vibration mode β ) to the
resonance component for the reference vibration mode shape ( β = 2 ). This is to deal with the lattices
of varying widths in the vertical direction.

  
− 0.3 (β − 2) + 1.4(1 − 0.4 ln β )
BH
λL =  0.5 (A6.7.4)
 B0  

The mode shape correction factor φ D can be set to 1 if the vibration mode shape agrees with the
reference vibration mode shape ( β = 2 ), regardless of the vertical distribution of mass per unit height
of a lattice structure over the ground.

BH
H
B0

B0
Figure A6.7.1 Definition of B0 , BH , and H

A6.8 Wind Loads on Free Roofs

A6.8.1 Scope of application
This section defines the procedures for estimating the wind loads on structural frames of free roofs
with mono slope, pitched, and troughed configurations that have rectangular plan shapes. Because the
specifications of the design wind load are based on the results of wind tunnel tests for relatively small
rigid free roofs, they cannot be applied to flexible roofs in which a wind-induced deformation and
- C6-74 - Recommendations for Loads on Buildings

vibration significantly affect the wind forces. It can be applied to free roofs that satisfy the geometric
conditions described in Table A6.11.
The wind force coefficients are regulated for a clear flow case when there are no obstructions under
the roof. The flow pattern around a roof is significantly affected by such obstructions. If there are any
obstructions whose blockage ratio is larger than approximately 50%, the wind pressure on the bottom
surface may increase significantly, resulting in a significant increase in the net wind force on the roof.
In such a case, it is necessary to evaluate the wind force coefficients based on wind tunnel tests, CFD,
and other measures.

A6.8.2 Procedure for estimating wind loads

The design wind loads on the structural frames of relatively small rigid buildings can be estimated
based on a quasi-steady assumption, in which the peak gust induces the maximum load effect on the
buildings. Therefore, the wind load is given by Eq. (A6.22). In this structure, focus is on the axial forces
induced by dynamic wind loading as the load effect for evaluating the wind force coefficients CR,
assuming that the roof is supported by four corner columns55). That is, the wind force coefficients CR,
presented in Table A6.11, are not the time-averaged wind force coefficients but some “equivalent wind
force coefficients” that provide the maximum load effects when combined with the gust effect factor
GR. Two sets of wind force coefficients on the windward and leeward halves of the roof are provided in
Table A6.11, one of which corresponds to the maximum tension, and the other correspond to the
maximum compression induced in the columns. In the structural design, the design wind loads for these
two sets of wind force coefficients are calculated for each wind direction and are combined with the
other loads.

A6.8.3 Gust effect factor

Based on a quasi-steady assumption, the gust effect factor GR for estimating the wind loads on
structural frames of free roofs is approximated based on the square of gust factor Gu defined through the
ratio of the peak gust to the mean wind speed. The peak gust may be given by the following equation:

Uˆ H = U H + g u σ u = U H (1 + g u I H ) = G uU H , (A6.8.1)

where Û H (m/s), U H (m/s), and σ u (m/s) represent the peak gust, mean wind speed, and standard
deviation of the wind speed fluctuation at height H, respectively, and g u and G u represent the peak
and gust factors. The values of g u and G u depend on the averaging time for evaluating the peak
values. Considering that the specification is applied to relatively small free roofs, the peak factor, g u
=3, is assumed, based on the results of a previous study55). The gust effect factor GR is approximated
through the square of Gu.
CHAPTER 6 WIND LOADS - C6-75 -

A6.9 Vortex Induced Vibration

A6.9.1 Scope of application

This section describes vortex-induced vibration that can occur in tall slender buildings, chimneys,
and structural components with circular sections.

A6.9.2 Vortex induced vibration and resulting wind load on buildings with circular sections
Shear layers separated from windward corners of both sides of buildings roll up alternately to shed
into a wake and form Karman vortex streets behind the buildings. According to the alternate shedding,
the periodic fluctuating wind loads act on the buildings in the across-wind direction. When the natural
frequency of the building coincides with the vortex shedding frequency, the vibration of the building
can be resonant with the periodic fluctuating wind loads, causing the building to vibrate at large
amplitude in the across-wind direction. This is vortex-induced vibration, which is a problem for many
structures, particularly chimneys.
The critical wind speed of the resonance is larger than the design wind speed for most buildings, and
thus these phenomena are not normally important. However, because the critical wind speed is smaller
than design wind speed for very slender buildings with a small natural frequency and damping, such as
steel chimneys, tall buildings, and building components, the effect of vortex-induced vibration should
be checked carefully during the wind resistance design stage.
A large amount of research has been conducted on vortex-induced vibration, and a number of methods
have been developed in the past decade for estimating the vibration amplitude and its equivalent static
wind loads, particularly for structures with circular sections. The equivalent static wind loads described
in the recommendations are based on the spectral modal method in which the Strouhal number of vortex
shedding is 0.2, and the power spectrum of the fluctuating wind loads depends on the vibration
amplitude8) and Reynolds number.
The effects of structural density, damping, and Reynolds number are included in the resonant wind
force coefficient C r , which is shown in Table A6.20 for three categories of Reynolds number region
and for two types of structures with various density and damping. The rows in the table show the effect
of the Reynolds number, that is, U r Dm < 3 is the subcritical region, 3 ≤ U r Dm < 6 is the critical
region, and 6 ≤ U r Dm is super/trance critical Reynolds number region. The value of ρ s ζ L in Table
A6.20 reflects the amplitude under the resonant condition, and ρ s ζ L < 5 corresponds to a large
amplitude, and ρ s ζ L ≥ 5 corresponds to a small amplitude.

A6.9.3 Vortex-induced vibration and resulting wind load on building members with circular
sections
The occurrence of vortex-induced vibration of building members with a circular section can be checked
through Eq. (A6.26). Most design wind speeds for members such as those of truss towers are larger than
the critical wind speed, and thus the effect of a vortex-induced vibration should be checked carefully. In
- C6-76 - Recommendations for Loads on Buildings

particular, the vibration amplitude can be very large for members such as steel pipes whose mass and
damping are small. The equivalent static wind loads described in Eq. (A6.27) are introduced in the sub-
critical Reynolds number region based on wind tunnel tests76). The equation is applicable for various
boundary conditions at the ends of the members.

A6.10 Combination of Wind Loads

A6.10.1 Scope of application

This section defines the combination of horizontal wind loads and roof wind loads on structural
frames for buildings with a rectangular section. These wind loads are evaluated separately, but this does
not mean that each wind load acts on the building independently. However, the maximum wind loads
do not occur at the same time. Therefore, if they are applied to the building at the same time, the
combination of wind loads overestimates the actual loads. This section shows the formula for a
combination of wind loads considering the correlations of the wind force and response. The formula is
divided in two ways: for buildings with and without an aspect ratio of H / BD < 3 . A combination of
horizontal and roof wind loads is also described.

A6.10.2 Combination of horizontal wind loads

(1) Buildings with aspect ratio H / BD < 3
Buildings with an aspect ratio of H / BD < 3 have a small resonance component. For such cases,
it is considered that a wind load of γ times the windward loads act in the across-wind direction, as
shown in Fig. A6.10.1. According to the stress calculation results of columns in buildings less than 80
m high, γ is defined as a function of D/B. Therefore, an approximate equation of γ is described
through Eq. (A6.30)77).
D

γWD
WD
B
wind Item show the difference of
−γWD calculation conditions

section of building

Figure A6.10.1 Windward load and combined Figure A6.10.2 Relation between side ratio
load for across wind direction (D/B) and combination factor

(2) Buildings with aspect ratio of H / BD ≥ 3

Buildings with aspect ratio of H / BD ≥ 3 have a large resonance component. For such cases, it is
assumed that the response probability is expressed through a normal distribution. If the overturning
moments in two directions, Mx and My, are expressed through a 2-dimensional normal distribution, the
equivalence line of probability becomes an elliptical line using the correlation coefficient of the response,
CHAPTER 6 WIND LOADS - C6-77 -

ρ , as shown in Fig. A6.10.3. Every point on the elliptical line (solid line) can be considered as a load
combination, but it is impractical to consider a large number. Therefore, load combinations can be
defined as the apexes of an octagon enveloping an oval. In other words, y-direction overturning moment
Myc, which should be combined with maximum x-direction overturning moment Mxmax, is defined
through the following equation using the mean y-direction overturning moment M y and the maximum
fluctuating component of the y-direction overturning moment mymax78).
(
M yc = M y + m ymax 2 + 2 ρ − 1 ) (A6.10.1)
Table A6.21 shows the combination of loads according to the upper equation considering the
following characteristics of the along-wind, across-wind, and torsional wind loads.
・Co-coherence (correlation coefficient for each frequency) is negligible between the along-wind
force and across-wind force, and between the along-wind force and torsional wind force. Therefore,
ρ = 0 because the co-coherence of the response is negligible.
・Because the co-coherence between the across-wind force and torsional wind force is not zero, the
absolute value of the correlation coefficient of response ρ LT , shown in Table A6.22, is defined
through calculations based on wind tunnel tests.
The value of ρ LT is calculated using a time history response analysis method79) under the conditions
in which the damping factors for the across-wind vibration and torsional vibration are within the range
of 0.01 to 0.05, and the building has no coupling vibration mode. Therefore, if the damping factor is not
in the prescribed range of 0.01 to 0.05, or if the building’s vibration mode is significantly coupled, it is
necessary to carry out special research.

point A

considered point of
combination load

Figure A6.10.3 Schema of load combination in consideration of response correlation

- C6-78 - Recommendations for Loads on Buildings

A6.10.3 Combination of horizontal and roof wind loads

The combination of horizontal and roof wind loads can be considered theoretically, as in A6.10.2 or
A6.10.3. However, because the relation between horizontal and roof wind loads is very complex, it is
supposed that horizontal and roof wind loads act at the same time77).

A6.11 Wind load estimation using CFD

To evaluate the time-averaged and fluctuating components of velocities as well as pressures that are
required to obtain the wind loading on buildings, the incompressible Navier-Stokes equations are
employed as the governing equations. However, such a nonlinear type of equation cannot be generally
applied to obtain an analytical solution, and therefore the approximate solutions can be obtained as
numerical values using the discrete equations for computer use, which are induced through a finite
difference method or finite volume method. However, because the approximate solutions always include
certain errors in the numerical process, the solutions obtained may have a large difference from the
correct solutions according to the numerical conditions. Therefore, it is necessary to take into
consideration the numerical manner or numerical increment on the discretization for deriving the
appropriate approximate solutions. Discrete equations implicitly have two types of numerical errors as
numerical viscosity, which causes a damping of the amplitude and numerical dispersion, thereby
generating a difference in phase velocity. In addition, numerical errors appear during the temporal
processes as well as the spatial processes. Then it is necessary to check these errors, considering the
relation of all terms of the governing equations. In the case of a positive value for the numerical viscosity,
the fluctuation of the numerical solution is suppressed compared to the correct solution. In the case of a
dispersive error, the fluctuation becomes unstable as a result of the phase lag of each wave number
component. The effect of numerical errors becomes larger as the discretized increment increases. The
numerical viscosity is often utilized in the engineering field because a stable solution can be always
obtained, and the numerical solution never shows divergence even for too large a spatial increment.
However, because this scheme tends to generate a solution much different from the correct one, this
mumerical scheme should be treated carefully. In the case of a central difference scheme, a dispersive
error is generated. Because a Navier-Stokes equation has a physical viscous term, its viscous effect
suppresses the dispersive error and can avoid numerical instability. However, as the level of the
dispersive error is determined based on the size of the spatial increment, the spatial increment should be
sufficiently small to obtain the viscous effect appropriately. Namely, in the case of high Reynolds
numbers, the small viscous effect requires a finer mesh, and the computational cost becomes higher. In
wind tunnel tests, the breadth of a building model equal to 10-1m and wind velocity equal to 101m/s, the
Reynolds number is on the order of 104. For an actual building in a strong wind, the Reynolds number
is around 107-8. A simulation is required to represent vortices from the largest scale to the smallest scale
where the turbulent energy dissipates owing to the viscous effect. Its range depends on the Reynolds
numbers. Current high-end super computers have difficulty simulating the wind flow acting on actual
buildings.
CHAPTER 6 WIND LOADS - C6-79 -

As shown above, when the CFD technique is applied to the wind load estimation for buildings, a
turbulence modeling technique is needed. The direct numerical method of the governing equations used
here requires a relatively very small resolution to represent the turbulent motions with a wide range of
scales. As a result, the number of the discrete points may increase, and the computational load may
become significantly larger. To simulate the turbulent flow at a reasonable computation cost, turbulent
models are generally used to substitute the motion with a small scale and thereby reduce the
computational cost. Based on the concept of turbulence modeling, fluid motions with a small scale are
replaced with the effects of the turbulent viscosity, and the dissipation effect is estimated from the
physical fluid motion. Accordingly, when introducing this modeling technique, the magnitude of the
turbulent viscosity determines the type of the solutions, namely, steady or unsteady. Modeling using a
large turbulent viscosity where the steady solution is usually obtained can easily avoid the difficulty
related with a numerical instability, and is widely used in practical problems. However, the estimation
of the wind loading supposes the usage of a numerical model that can reproduce an unsteady solution.
Even if an unsteady solution is obtained, the selection of a numerical scheme for the computational
technique generates numerical viscosity and cannot provide the appropriate effect of turbulence
modeling on the solution. The viscous effect strongly suppresses the fluctuation in the high-frequency
region in general. To capture the fluctuations within the frequency region required for a wind loading
estimation, it is very important to take care in the treatment of both the turbulence modeling and the
numerical discretization. Because the turbulence modeling used for an evaluation of the wind loading
should judge the presence of numerically unfavorable effects, it is desirable to obtain an unsteady
solution having a wide frequency band. Accordingly, a large eddy simulation (LES) is currently the
most appropriate model. In the case of LES, complicated flows to be targeted in estimating the wind
load are not subject to modeling, and are directly solved. There is a significant advantage with such an
approach in that it avoids having to check the adequency of the modeling to the various types of flows.
A Reynolds averaged Navier-Stokes (RANS) type of turbulence modeling, which is frequently utilized
in engineering field,, provides averaged solutions. However, it should be used carefully even when
utilizing only the averaged values. In the case of RANS, many assumptions are used for formulating the
model. Accordingly in computing the flow around buildings or over complicated terrain, it is generally
difficult to achieve their general versatility because of various flows such as impinging and separating
flows. When RANS is used, it is therefore necessary to grasp the characteristics of the model and
confirm the accuracy and applicability in advance.
- C6-80 - Recommendations for Loads on Buildings

Figure A6.11.1 Schematic of wind load estimation using CFD78),79,80).

The application methods shown are concrete examples of using CFD for wind loading estimations.
First, the construction of numerical model is considered for a case in which the approaching wind attacks
a specified building (Fig. A6.11.1). For the wind tunnel tests conducted thus far, using an experimental
model that reproduces the specified building and its surrounding ground condition within a range of
about 600 m, the wind force and response measurements were carried out for a design wind speed profile
forming over rough ground. When constructing a numerical model in the same manner and providing
the time history of the wind velocity profile at an inflow boundary based on the category of the surface
roughness, an LES analysis can be applied to the CFD for this example. However, considering the recent
state of the preparation of GIS data, the wind load estimation may be conducted using a model broadly
reproducing the upstream domain. In this case, the wind profile at the inflow boundary is given by
determining the level of surface roughness at same location and the wind blowing toward the specified
building. For CFD, the computational cost with the numerical model that reproduces the state of the
upwind surface roughness is almost the same as that for the generation of inflow turbulence, and thus it
does not become a heavy burden. It is possible for an analysis to produce the approaching flow correctly
for a sufficiently upwind domain length and to impose this flow on the specified building. The present
AIJ recommendations recommend that the smallest roughness effect be selected from various roughness
categories for an area of less than 3 km upwind of the building site and distance of 40 times the building
height. With the CFD, such an overestimation for the wind velocity may be avoided in certain cases.
As the next example, concerning the setup of the design wind velocity, the CFD simulations for the
actual terrain effect are shown. However, the following points must be made clear when constructing a
CHAPTER 6 WIND LOADS - C6-81 -

numerical model of actual terrain. For small terrain, a model of several tens to hundreds of meters in
height is considered. It should be examined whether the relatively large-scale flow considered in this
case can be treated in the same manner as the engineering-scale unsteady flows around a building. The
applicability of the governing equations used for this problem has to be confirmed. It is necessary to pay
attention to the method used to set up the inflow boundary conditions. In addition, the questions occur
as follows: how much area is required for the upwind region which determines the wind field at the
building site, and which location should be selected to represent a flat surface where a reasonable wind
profile can be imposed. The method used to determine the wind velocity profile and the velocity value
at the inflow position should be established. An analytical example for actual terrain is shown in Fig.
A.6.11.2. In this case, considering that the location of the inflow is located at sea, the inflow condition
is given as a smooth-wall turbulent boundary layer flow, and the magnitude of the wind velocity is set
to the values obtained through the simulation using a meso-scale meteorological model at the height of
the top of the boundary layer. Regarding the surface roughness on actual terrain, each roughness level
is determined separately at each location. Concretely, at the bottom boundary, the wind velocity profile
represented by the log law with the roughness length (0.1 to 0.8 m), the friction drag based on the
vegetation canopy, and the wind profile on a smooth surface is imposed. The results obtained by the
LES are shown with observational data in (c) and (d). The numerical results based on the canopy model
and observational data are in good agreement with each other at the measurement point of the E-pole.
At the measurement point of the W-pole where the separation occurs, the numerical results show a small
difference from the observational data but are improved when considering the local estimation of the
roughness length of the surface aspects. The total terrain such as a ridge or valley generates three-
dimensional flows dependent on the wind direction, whereas the LES results show agreement with the
observational data at two locations. On the other hand, according to (b), when the sampling point of the
results is changed slightly, some cases show a significant change in the velocity. In the case of a wind
loading estimation using CFD, considering the sensitivity, the numerical results should be sufficiently
checked by surveying the velocity field near the specified point.

NNW

E-pole

W-pole
Radius
=9km
Radius
=7km

(a) Site and location of field measurements (b) Mean wind velocity distribution at a 60 m height
Figure A6.11.2 Numerical example of wind flow over actual terrain81).
- C6-82 - Recommendations for Loads on Buildings

(c) E-pole (d) W-pole

Figure A6.11.2 Numerical example of wind flow over actual terrain81) (cont.)

A6.12 Effects of Neighboring Tall Buildings

When groups of two or more tall buildings are constructed in proximity, the fluid flow through the
group may be significantly deformed and have a much more complex nature than that of an isolated
building, resulting in enhanced dynamic pressures and motions, particularly in neighboring downstream
structures. Therefore, a study of interference among closely located tall buildings is an important
problem, not only in a wind-resistant design but even in minimizing wind-motion discomfort to building
occupants. Wake-induced oscillation in a downstream building is considered to be affected by
interference from upstream buildings of various sizes placed at various locations, as well as by the
turbulence of incident flows. Figure A6.12.184),85) shows the contours of the increase or decrease in
ratios for the maximum along/across-wind responses of the downstream building exposed to
interference from an upstream building at various locations to those of an isolated building where the
maximum responses, including the mean deflection, are estimated at near the design wind speeds of 60
m/s using a modal-spectrum method. The contours are illustrated for an identical pair of square tall
buildings with an aspect ratio of H / = 4, where two coordinate axes are normalized by the reference
building width.
The interference ratio in the across-wind direction is usually larger than those in the along-wind
direction. Positions producing interference ratio contours higher than 1.2 are generally restricted to
regions of 12 in the x-direction and 6 in the y-direction for the across-wind direction.
When a slender building is located upwind of a building in terrain Category II, the downwind building
shows large response owing to regular vortex shedding from the upwind building. The tendency
becomes clear, particularly in the across-wind direction.
When the flat terrain subcategories increase from terrain Category II to Category IV, the dynamic
responses of the downstream building are relatively independent of the interference effect86).
CHAPTER 6 WIND LOADS - C6-83 -

y
6 BD
upstream building
4
1.1 1.1

1.2 2

x 1.0 1.0
12 BD 6 4 2
0.8
downstream building

(a) Terrain category II, along-wind direction (b) Terrain category II, across-wind direction

(c) Terrain category IV, across-wind direction

Figure A6.12.1 Contours of the response ratios

A6.13 Response Acceleration

A6.13.1 Scope of application

This section defines the maximum along-wind response acceleration for ordinary buildings, the
maximum across-wind response acceleration for buildings with rectangular sections satisfying the
conditions of A6.5.1, and the maximum torsional response acceleration for buildings with rectangular
sections satisfying the conditions of A6.6.1.
Each formula is applied for only the first vibration mode. If a building has a large dynamic response in
higher modes, or a partial vibration, other specialized researches should be carried out.

A6.13.2 Maximum along-wind response acceleration

The standard deviation of the generalized response acceleration σaD(m/s2) is given through the
following equation.
2
 ∞ χm ( f ) 
σ aD =

∫
0
S g ( f )(2πf ) 4
K g2
df  ,

(A6.13.1)

where Sg(f) is the power spectral density of the generalized wind force, χ m ( f )
2
is the mechanical
admittance of a building as described in Eq. (A6.3.13), f is frequency (Hz), and Kg is the generalized
stiffness as described in Eq. (A6.13.2).
- C6-84 - Recommendations for Loads on Buildings

K g = M D (2πf D ) 2 , (A6.13.2)

where MD is generalized mass. Because the resonant component is dominant in terms of acceleration,
Sg(f) can be substituted with white noise having a power spectral density at natural frequency fD, as
described in Eq. (A.6.13.3).
FD
S g ( f D ) = (q H BHCH C 'g λ ) 2 , (A6.13.3)
fD
where FD is the along-wind force spectral factor, as shown in A6.3. Substituting (A6.13.2) and (A6.13.3)
for Eq. (A6.13.1), Eq. (A6.13.4) is derived.
RD
σ aD = q H BHCH C 'g λ (A6.13.4)
MD
Furthermore, the maximum acceleration at height Z of a building is calculated from σaD multiplied by
the peak factor and vibration mode. Equation (A6.31) estimates the maximum acceleration at the top of
the building height H(µ(H)=1). Because the natural frequency component of a building is dominant in
terms of acceleration, the level crossing rate νD(Hz) for calculating the peak factor is approximated based
on the natural frequency fD. When the fluctuating overturning moment coefficient for along-wind C’MD
at the base of a building, and the power spectral density of the fluctuating overturning moment SCMD(f*D)
at the non-dimensional frequency f*D, are estimated from the appropriate wind tunnel experiment or
CFD,, the maximum response acceleration for the along-wind direction at the top of the building is
calculated using the following equation.

qH g aD BHC 'MD λ πf D* S CMD ( f D* )

aDmax = (A6.13.5)
MD 4ς DC '2MD

A6.13.3 Maximum across-wind response acceleration

Equation (A6.32) consists of coefficients according to the across-wind direction as a development in
the along-wind direction, as shown in A6.13.2.

A6.13.4 Maximum torsional response acceleration

Equation (A6.33) of the torsional response acceleration has almost the same development in the along-
wind direction, as shown in A6.13.2. It includes a coefficient of 0.6, which transfers from a fluctuating
torsional moment coefficient to a generalized fluctuating torsional moment coefficient. A generalized
inertial moment IT(kgm2) is used with regard to the generalized mass of a building.
When the angular acceleration (rad/s2) is calculated through Eq. (A6.33), the maximum acceleration
admax(m/s2) caused by torsion is calculated from Eq. (A6.13.6) by multiplying a distance d(m) from the
center of the figure.
admax = aT max d (A6.13.6)
CHAPTER 6 WIND LOADS - C6-85 -

A6.14 One-Year Recurrence Wind Speed

The 1-year recurrence wind speed U1H is used to calculate the acceleration of the wind response for
an evaluation of the habitability, as defined in Eq. (A6.34).
Figure A6.6 shows a smoothed wind speed map based on the 1-year recurrence wind speed at the
meteorological offices, from which the wind speed U1 at any location can be estimated. The 1-year
recurrence wind speeds at the meteorological offices are estimated based on the daily-maximum wind
speed data regardless of the wind directions collected from 2001 to 2010. On the other hand, because
the wind response characteristics of buildings vary depending on the wind direction, the wind speed
under which the same acceleration is obtained may also vary depending on the wind direction. Therefore,
if the characteristic of the wind direction, that is, the frequency of exceedance of each wind speed, can
be understood, a reasonable design is possible. This characteristic within the range of the wind speed
used to evaluate the habitability is generally clarified. However, because the wind direction factor in
A6.1.5 should be used for a 100-year recurrence wind speed, it is not possible to use in this case.
When the maximum acceleration amax is approximated through wind tunnel tests or CFD as a function
of wind speed U, shown in Eq. (A6.14.1), the return period tamax for maximum acceleration amax is
calculated through Eq. (A6.14.2). The probability on the right side of Eq. (A6.14.2) is expressed as the
total sum of the occurrence probability of the wind speed in every 16 azimuths shown in Eq. (A6.14.3).
amax = f (U ) (A6.14.1)
1
t a max = (A6.14.2)
1 − Fa (amax )

{ }
16
Fa (amax ) = ∑ pi FU f i −1 (amax ) (A6.14.3)
i =1

where
Fa (amax ) : probability of the maximum acceleration not exceeding amax,
pi : occurrence frequency for wind direction i,
{ (a )}
FU f i −1
max : probability of the wind speed not exceeding the wind speed at which the
maximum acceleration is equal to amax for wind direction i.
The occurrence frequency pi for each wind direction, and parameters ai and bi in Eq. (A6.14.4), which
are the parameters used to calculate the right side of Eq. (A6.14.3), are shown in Table A6.14.1. These
parameters are estimated based on the daily maximum wind speed in 30 cities. For the estimation, the
least squares method was applied for the data in Naha where typhoons frequently hit, whereas Gumbel’s
moment method was applied for the data on other cities. Note that these parameters ai and bi should be
used for a return period of less than 1 year.

FU (U i ) = exp[− exp{− ai (U i − bi )}] (A6.14.4)

where
Ui (m/s) : 10-min mean wind speed at 10 m above the ground corresponding to the flat terrain
category of II, defined in Table A6.2 for wind direction i,
ai(s/m), bi(m/s) : parameters estimated based on the daily maximum wind speed for wind direction i.
- C6-86 - Recommendations for Loads on Buildings

Table A6.14.1 Parameters ai and bi and occurrence frequency pi for each wind direction
Asahikawa Sapporo Aomori Akita Sendai
a b pi(%) a b pi(%) a b pi(%) a B pi(%) a b pi(%)
NNE 0.50 4.48 3.4 1.06 4.41 0.2 0.93 3.90 4.9 0.73 5.45 0.2 0.63 5.03 1.1
NE 0.38 6.00 1.4 0.69 4.73 0.1 0.77 4.50 4.6 - - 0.0 1.05 4.31 0.4
ENE 0.85 3.02 0.3 0.77 3.55 0.3 0.65 5.99 4.6 0.32 6.95 0.1 0.80 4.82 0.4
E 0.92 2.29 0.6 1.02 4.35 1.5 0.75 5.96 7.6 1.60 5.27 0.2 0.95 4.97 0.7
ESE 0.86 2.66 0.3 0.85 5.25 2.7 0.78 6.53 0.9 0.51 7.14 7.1 0.67 5.04 1.3
SE 0.60 3.72 3.3 0.56 6.82 7.2 0.57 7.28 0.6 0.64 6.54 15.8 1.13 4.29 20.0
SSE 0.52 5.39 17.7 0.40 8.76 19.8 0.47 5.43 0.6 1.33 5.71 0.2 0.75 4.84 13.5
S 0.45 5.46 2.7 0.32 8.25 6.4 0.97 3.14 0.5 3.02 4.71 0.1 0.80 4.90 6.1
SSW 0.46 6.30 3.4 0.42 7.64 2.4 0.48 6.14 3.2 0.51 9.46 2.0 0.58 6.23 1.4
SW 0.50 6.68 3.2 0.32 8.77 1.5 0.46 7.09 11.8 0.44 7.29 10.5 0.68 5.62 0.8
WSW 0.42 6.95 15.9 0.41 8.20 2.3 0.45 7.78 12.5 0.39 6.90 12.7 0.42 8.76 1.5
W 0.60 5.66 21.9 0.62 8.78 2.2 0.44 9.79 12.7 0.35 7.62 20.3 0.36 9.37 7.9
WNW 0.61 5.04 12.4 0.52 9.31 5.6 0.55 8.22 14.1 0.34 9.74 9.1 0.38 10.0 15.6
NW 0.81 4.62 2.9 0.41 9.85 20.3 0.70 5.68 7.9 0.43 9.79 9.0 0.39 8.45 8.0
NNW 0.74 5.29 5.7 0.52 7.16 24.8 0.91 4.20 6.6 0.58 7.83 4.5 0.53 6.18 9.0
N 0.67 5.03 4.9 0.81 4.76 2.7 0.91 3.58 7.0 0.73 6.08 8.0 0.73 5.15 12.3
Niigata Kanazawa Utsunomiya Maebashi Tokyo
a b pi(%) a b pi(%) a b pi(%) a b pi(%) a b pi(%)
NNE 0.57 6.33 4.4 0.65 5.79 3.9 0.57 5.07 20.7 - - 0.0 0.70 6.69 4.6
NE 0.95 4.38 1.7 0.68 5.27 3.6 0.89 4.56 8.3 - - 0.0 0.70 6.24 6.0
ENE 0.95 4.50 1.4 0.88 5.45 13.7 1.22 4.43 1.8 0.77 4.95 0.1 0.72 6.54 8.0
E 1.00 4.19 0.5 0.89 4.87 9.5 1.03 3.89 3.1 0.40 5.57 0.7 0.86 5.94 4.6
ESE 0.49 5.42 1.3 0.59 3.68 0.6 1.35 3.75 8.1 0.81 5.42 24.5 0.68 6.01 4.1
SE 0.57 6.13 14.9 0.82 3.89 0.5 1.19 4.15 9.7 0.98 5.08 8.6 1.10 5.71 3.7
SSE 0.85 4.10 5.5 2.32 3.22 1.1 1.00 4.54 9.0 1.11 3.98 2.7 0.95 5.89 4.5
S 0.77 5.02 2.1 0.27 5.48 0.5 0.73 5.02 6.8 0.92 4.55 1.1 0.46 6.82 8.9
SSW 0.84 5.45 4.4 0.37 9.21 9.3 0.82 4.51 6.6 0.67 3.22 0.3 0.51 7.85 4.9
SW 0.41 5.67 3.2 0.40 9.79 9.7 0.74 4.30 2.7 1.12 4.35 0.3 0.46 7.69 11.5
WSW 0.36 9.44 12.4 0.38 9.47 9.5 0.45 6.00 1.9 0.53 4.75 1.5 0.45 6.99 0.4
W 0.35 9.63 13.2 0.31 8.13 14.6 0.51 8.30 2.3 0.37 7.02 1.9 0.36 8.86 0.3
WNW 0.35 8.57 8.1 0.35 7.63 4.0 0.43 7.60 2.2 0.36 7.33 3.9 0.44 6.40 0.8
NW 0.36 8.64 8.8 0.38 6.74 5.8 0.48 6.35 1.2 0.46 6.66 32.6 0.40 8.44 9.5
NNW 0.50 6.90 10.2 0.54 4.93 4.4 0.43 6.47 3.6 0.45 9.30 17.7 0.51 7.45 18.8
N 0.79 6.31 8.0 0.60 5.55 9.2 0.54 4.80 12.1 0.49 11.3 4.2 0.56 7.01 9.4
Chiba Yokohama Shizuoka Hamamatsu Nagoya
a b pi(%) a b pi(%) a b pi(%) a b pi(%) a b pi(%)
NNE 0.57 7.22 4.5 0.65 7.74 1.7 0.72 4.88 2.8 0.78 5.02 0.5 0.80 4.00 1.3
NE 0.65 7.20 9.8 0.33 7.75 0.1 0.69 6.20 12.2 1.10 3.90 3.7 2.20 3.28 0.7
ENE 0.94 5.83 9.9 0.80 7.55 1.9 0.83 6.16 20.1 0.47 7.11 6.3 1.52 2.52 0.9
E 1.01 5.56 2.1 0.98 6.19 11.5 0.67 5.90 2.4 0.56 7.18 7.9 1.11 5.21 0.1
ESE 1.12 4.51 6.2 0.51 5.57 1.3 0.78 4.92 3.6 0.73 6.65 2.2 0.60 5.08 0.5
SE 0.93 4.62 9.1 0.72 6.63 0.8 2.08 3.75 1.0 0.80 6.07 6.6 0.62 6.27 3.9
SSE 0.61 4.97 4.3 0.67 6.09 6.0 0.87 6.29 7.0 1.00 5.27 6.7 0.54 7.19 11.7
S 0.66 6.16 2.1 0.75 6.53 5.0 0.81 6.52 16.9 0.96 5.26 2.6 1.22 5.58 11.4
SSW 0.33 10.8 3.8 0.40 9.90 6.8 0.45 8.05 2.9 1.21 6.75 0.1 1.08 5.29 4.2
SW 0.36 9.83 12.1 0.35 9.40 15.0 0.45 9.47 13.0 0.74 6.22 3.1 0.87 5.20 1.2
WSW 0.55 5.69 11.5 0.36 9.23 4.8 0.47 10.1 3.8 0.70 6.54 14.3 1.04 4.39 1.2
W 0.47 5.04 1.1 0.28 8.56 0.6 0.46 10.3 6.2 0.52 7.98 11.9 0.69 6.46 2.5
WNW 0.50 7.28 1.3 0.23 8.75 0.2 0.34 6.97 4.0 0.49 9.88 27.6 0.58 7.16 20.2
NW 0.42 8.37 8.1 0.25 10.2 0.7 0.59 4.70 1.4 0.39 8.80 5.4 0.48 7.39 17.5
NNW 0.44 6.51 13.7 0.42 8.71 1.4 1.12 4.74 2.1 0.51 4.95 1.0 0.49 5.56 13.0
N 0.70 6.74 0.4 0.47 7.90 42.2 0.53 3.85 0.7 0.66 4.66 0.2 0.73 4.17 9.7
CHAPTER 6 WIND LOADS - C6-87 -

Table A6.14.1 (continued) Parameters ai and bi and occurrence frequency pi for each wind direction
Kyoto Osaka Kobe Wakayama Okayama
a b pi(%) a b pi(%) a b pi(%) a b pi(%) a b pi(%)
NNE 0.67 4.10 4.1 0.91 6.94 4.7 0.67 6.41 1.5 0.65 6.44 8.1 0.62 5.57 1.2
NE 1.14 4.27 4.5 0.59 6.07 14.6 0.55 5.66 5.6 0.87 5.20 2.8 1.01 3.91 3.8
ENE 0.82 5.65 7.1 0.51 6.28 3.8 0.53 5.67 15.1 1.81 4.36 11.1 1.14 4.08 6.7
E 0.75 5.53 6.2 0.59 6.49 1.1 0.81 4.70 2.4 1.48 4.21 1.9 0.60 5.35 6.0
ESE 0.78 5.23 2.8 0.58 5.00 0.1 1.24 3.59 0.6 1.55 4.61 0.2 0.48 5.91 5.5
SE 0.81 4.89 1.6 0.78 4.25 0.8 1.68 3.83 0.8 0.83 8.39 0.3 1.16 4.59 8.7
SSE 0.83 4.33 3.8 0.82 4.23 0.8 2.29 3.15 0.1 0.40 8.64 0.7 1.15 4.15 3.6
S 0.81 5.20 11.1 0.34 6.30 0.2 2.65 4.61 0.4 0.31 9.11 6.0 1.26 3.99 4.7
SSW 0.69 6.25 12.2 0.35 9.05 1.8 0.62 6.01 7.3 0.47 7.40 8.9 1.06 4.51 4.1
SW 0.61 6.97 3.1 0.76 6.35 18.5 1.10 6.72 9.2 1.01 5.53 8.4 0.65 5.58 12.2
WSW 0.54 7.71 2.2 0.58 7.25 18.2 0.63 6.83 16.1 1.10 4.01 11.0 0.52 6.07 5.6
W 0.50 7.22 3.3 0.50 7.50 19.9 0.43 7.87 9.9 0.46 7.34 4.6 0.50 7.65 6.2
WNW 0.58 7.06 6.2 0.44 8.11 3.7 0.49 8.47 4.5 0.49 8.39 6.1 0.51 7.31 6.2
NW 0.46 8.72 2.6 0.67 8.28 4.6 0.66 6.43 2.8 0.60 6.98 8.5 0.58 6.61 5.8
NNW 0.60 7.06 11.3 0.67 7.96 2.3 0.53 7.24 9.9 0.74 5.74 8.6 0.57 5.65 10.6
N 0.77 5.74 18.0 0.77 7.04 4.9 0.70 5.88 13.6 0.73 6.22 12.8 0.57 5.86 9.2
Matsue Hiroshima Takamatsu Kochi Matsuyama
a b pi(%) a b pi(%) a b pi(%) a b pi(%) a b pi(%)
NNE 0.86 5.56 1.9 1.28 4.70 27.9 0.62 4.34 2.8 0.69 8.14 5.9 0.69 7.10 3.7
NE 0.63 5.94 8.2 0.63 5.69 0.9 0.61 4.12 1.1 0.67 6.77 2.4 0.75 6.64 3.0
ENE 0.63 6.79 9.4 0.60 6.72 0.2 1.10 4.46 10.1 0.72 5.77 4.7 0.99 4.90 3.5
E 0.73 6.18 10.0 1.39 7.34 0.1 1.08 4.31 13.5 0.51 4.82 4.5 1.19 3.66 3.5
ESE 1.06 4.65 6.1 1.08 5.01 0.2 0.66 4.35 9.9 0.41 5.98 4.7 1.13 3.54 6.4
SE 0.77 6.61 1.3 0.70 6.29 0.2 1.98 3.32 0.6 1.86 4.78 21.2 0.63 4.86 2.6
SSE 0.75 4.48 0.4 0.53 5.07 0.8 1.62 4.19 0.3 1.74 5.16 7.0 0.56 5.57 4.0
S 0.58 6.52 0.4 0.53 4.95 7.7 1.05 3.58 0.2 1.37 5.26 7.6 0.52 6.43 2.7
SSW 0.31 8.59 2.1 0.77 6.21 7.2 0.42 5.53 0.5 0.68 5.08 3.8 0.76 5.15 2.2
SW 0.63 7.48 1.2 1.14 4.62 10.9 0.80 4.61 2.1 0.67 5.77 2.4 0.87 6.48 4.9
WSW 0.41 8.07 13.9 0.52 8.26 0.9 0.52 6.37 19.4 0.61 5.84 4.7 0.75 6.31 5.3
W 0.39 7.95 27.0 0.63 7.41 3.8 0.46 7.86 18.0 0.50 5.35 20.0 0.58 6.97 13.6
WNW 0.42 6.79 6.4 0.60 8.01 3.1 0.49 7.66 2.7 0.63 4.18 5.6 0.56 6.88 18.0
NW 0.64 5.77 9.6 0.63 7.40 0.9 0.58 7.08 2.8 0.61 7.01 3.0 0.85 5.97 16.3
NNW 0.71 6.47 1.7 0.88 6.85 3.3 0.86 4.51 10.5 0.65 8.71 0.9 0.73 5.46 7.6
N 0.77 5.49 0.6 0.90 5.08 31.8 1.05 4.67 5.7 0.47 9.44 1.6 0.63 7.41 2.6
Fukuoka Oita Kumamoto Kagoshima Naha
a b pi(%) a b pi(%) a b pi(%) a b pi(%) a b pi(%)
NNE 0.74 5.56 2.2 0.85 5.16 9.0 0.79 3.50 3.9 0.65 5.95 8.5 0.88 7.42 11.3
NE 0.76 5.62 1.9 0.70 5.16 6.7 0.53 6.73 1.8 0.59 6.13 8.3 0.59 5.92 2.2
ENE 0.92 6.19 0.7 1.46 4.37 9.6 0.62 6.60 2.2 0.19 12.6 0.3 0.49 4.44 7.6
E 0.40 5.04 0.4 1.26 4.32 2.3 0.55 6.31 2.7 0.43 8.18 0.7 0.27 -1.50 9.7
ESE 0.59 6.35 0.5 0.26 7.14 0.7 0.60 5.30 3.6 0.30 6.65 1.5 0.25 -1.14 9.4
SE 0.64 6.01 8.2 0.69 6.45 0.9 0.61 5.05 1.4 0.55 5.79 5.5 0.20 -1.93 7.0
SSE 0.36 8.06 4.3 0.48 6.37 7.2 0.40 4.24 0.9 0.51 5.14 10.0 0.46 6.49 4.2
S 0.36 9.19 2.7 0.76 4.69 9.7 0.50 5.59 2.0 1.08 4.45 4.5 0.59 7.29 4.7
SSW 0.55 8.51 3.5 0.52 5.88 3.4 0.44 5.44 3.9 0.43 8.67 0.9 0.65 5.19 8.1
SW 0.43 8.29 0.5 0.47 8.63 2.0 0.67 5.94 18.6 0.60 7.14 2.5 0.21 -2.18 3.6
WSW 0.42 8.24 0.4 0.49 7.99 1.6 0.71 6.01 12.5 0.67 7.14 5.5 0.32 -0.71 2.8
W 0.49 9.19 2.8 0.59 9.32 7.0 0.55 6.08 3.4 0.98 5.59 10.4 0.12 -9.25 1.1
WNW 0.52 9.62 5.4 0.47 8.93 2.5 0.43 7.14 4.6 0.57 7.15 13.9 0.22 3.51 1.1
NW 0.57 7.67 3.1 0.59 8.25 11.8 0.61 5.84 13.8 0.52 7.60 6.1 0.63 5.48 1.8
NNW 0.65 6.31 32.1 0.66 7.15 18.5 0.66 5.32 17.8 0.70 5.74 17.9 0.24 -0.92 8.3
N 0.61 6.28 31.4 1.18 3.78 7.0 0.67 5.18 6.8 0.74 5.35 3.5 0.49 6.19 17.2
- C6-88 - Recommendations for Loads on Buildings

A6.15 Duration time of wind load

In these recommendations, the wind loads are given as expectation values of the maximum load
effects corresponding to design return periods. In general, buildings are designed elastically in a wind-
resistant design by confirming the safety under a static wind load (i.e., equivalent static load). Therefore,
it is rare to conduct a fatigue design considering the fluctuating wind forces and their duration times.
On the other hand, the numbers of base-isolated and vibration-controlled buildings using energy
dissipation devices have increased in recent years. In some cases, a yielding of the energy dissipation
materials (such as a steel bar damper) occurs under loads lower than the design wind loads. Therefore,
it is important to estimate the wind speed levels and their duration times in the fatigue design under
fluctuating wind loads. Furthermore, if necessary, it is important to make a maintenance plan including
an exchange of members88).
For the main structural frames such as columns and beams, which did not yield under the design wind
loads, it is necessary to confirm the safety against fatigue failure during the service period of a building
under loads lower than the design wind loads. This is because the wind loads repeat for a long time,
unlike seismic loads89).
For the design of components/cladding, it is necessary to consider a shorter periodical component
than for the structural frames. Therefore, if components/cladding do not have sufficient strength, there
is a possibility of fatigue failure under a single high-wind event (e.g., single severe typhoon event). It is
then important to confirm the fatigue safety for components/cladding. Further, the occurrence of vortex
excitation under the design wind speed increases the degree of fatigue damage on a tower-like structure
such as chimney and rod-like members of a steel tower. Therefore, when the occurrence of vortex
excitation is assumed, it is necessary to check the fatigue damage considering the wind speeds and their
duration times.
As described above, to ensure the safety against building fatigue owing to wind forces that are
repeated for a long period of time, it is necessary to evaluate the accumulated fatigue damage under one
strong wind occurrence and the entire building service duration. Within this time, the designers need to
determine the wind speeds and their duration appropriately. However, such information is not well
organized. In this section, for strong wind and equivalent durations, the “Guidelines for Wind-resistant
Design of Base-isolated Buildings”88) are introduced as a reference.
In these guidelines, the equivalent duration is introduced as that of the maximum wind speeds causing
equivalent fatigue damage on a member, based on the relationships between the maximum wind speeds
(normally equal to the design wind speeds) and accumulated fatigue damage (or total energy) acting on
the member88).
CHAPTER 6 WIND LOADS - C6-89 -

References

1) Document for wind-resistant design of Buildings (Part 2) - Dynamic behavior and load evaluation of
wind, AIJ, 1991 (in Japanese)
2) Nishimura, H. and Taniike, Y.: Aeroelastic instability of high-rise building in a turbulent boundary
layer, Journal of Structural and Construction Engineering AIJ, No. 456, pp. 31-37, 1994 (in Japanese)
3) Nishimura, H. and Taniike, Y.: Aeroelastic instability of high-rise building in a turbulent boundary
layer, Journal of Structural and Construction Engineering AIJ, No. 482, pp. 27-32, 1996 (in Japanese)
4) Katagiri, J., Ohkuma, T., Marukawa, H. and Shimomura, S.: Motion-induced wind forces acting on
rectangular high-rise buildings with side ratio of 2 － Characteristics of motion-induced wind
forces during across-wind and torsional vibrations, Journal of Structural and Construction
Engineering AIJ, No. 534, pp. 25-32, 2000 (in Japanese)
5) Katagiri, J., Ohkuma, T. and Marukawa, H.: Prediction formula of critical wind speed of torsional
aerodynamic unstable vibratin, AIJ Journal of technology and design, No. 15, 65-70, 2002 (in
Japanese)
6) Fujimoto, M. and Ohkuma, T: A theoretical study on evaluation of wind-induced vibrations of towers
with a circular cross section, Part I, Transactions of Architectural Institute of Japan, No. 185, – C6-
78 – Recommendations for Loads on Buildings, pp. 37-44, 1971 (in Japanese)
7) Ohkuma, T: A theoretical study on evaluation of wind-induced vibrations of towers with a circular
cross section, Part II, Transactions of Architectural Institute of Japan, No. 187, pp. 59-67, 1971 (in
Japanese)
8) Ohkuma, T: A theoretical study on evaluation of wind-induced vibrations of towers with a circular
cross section, Part III, Transactions of Architectural Institute of Japan, No. 188, pp. 25-32, 1971 (in
Japanese)
9) Damping in Buildings, AIJ, 2000 (in Japanese)
10) Davenport, A.G.: The application of statistical concept to the wind loading of structures, Proc. Inst.
Civil Engrs., Vol. 19, pp. 449-472, 1961
11) Zhou, Y. and Kareem, A.: Gust loading factor: New model, Jourrnal of Structural Engineering, Vol.
127(2), pp. 168-175, 2001
12) Ohkuma, T. and Marukawa, H.: Mechanism of aeroelastically unstable vibration of large span roof,
Journal of Wind Engineering, JAWE, No. 42, pp. 35-42, 1990 (in Japanese)
13) Matsumoto, T.: Wind tunnel study of the self-excited oscillation of one-way type suspension roofs
in a smooth flow, Journal of Structural and Construction Engineering AIJ, No. 384. pp. 90-96, 1988
(in Japanese)
14) Ito, S., Okuda, Y. and Kikitsu, H.: Study on peak wind force coefficients on gable roof tiles using
wind tunnel model with underside and ridge tiles, Proceedings of 22th National Symposium on Wind
Engineering, pp. 161-166, 2012 (in Japanese)
15) The Building Center of Japan: Guide book for engineers on wind tunnel test for buildings, 2008 (in
Japanese)
16) Architectural Institute of Japan: Guide for numerical prediction of wind loads on building, 2005 (in
Japanese)
17) Ishihara, T., Hibi, K., Kato, H., Ohtake, K. and Matsui. M.: A database of annual maximum wind
speed and corrections anemometers in Japan, Journal of Wind Engineering, JAWE, No. 92, pp. 5-54,
2002 (in Japanese)
- C6-90 - Recommendations for Loads on Buildings

18) Ohtake, K. and Tamura, Y.: Study on estimation of flat terrain for each wind direction, Summaries
of Technical Papers of Annual Meeting Architectural Institute of Japan, Structures I, pp. 119-120,
2002 (in Japanese)
19) Gines, I. and Vickery, B. J.: Extreme wind speeds in mixed wind climate, Journal of Wind
Engineering and Industrial Aerodynamics, Vol. 2, pp. 331-344, 1978
20) Cook, N. J.: Improving the Gumbel analysis by using M-th highest extremes, Wind and Structures,
Vol. 1, No. 1, pp. 25-42, 1998
21) British Standard, Loading for buildings, Part 2. Code of practice for wind loads, BS6399-2, 1997
22) AS/NZS 1170.2:2011, Australian/New Zealand Standard, Structural design actions, Part 2: Wind
actions, 2011
23) Melbourne, W. H.: Desibning for directionality, 1st Workshop on Wind Engineering and Industrial
Aerodynamics, Highett, Victoria, 1984
24) ASCE 7-98: Minimum design loads for buildings and other structures, Revision of ANSI/ASCE 7-
95, 2000
25) Sarukawa, A., Matsui, M., Okuda, Y., Otake, K., Nakamura, O., Katsumura, A. and Akahoshi, A.:
Study on diminution of design wind speed for specific seasons, Summaries of Technical Papers of
Annual Meeting Architectural Institute of Japan, B-1, pp. 79-80, 2013 (in Japanese)
26) Suda, K., Sasaki, A., Ishibashi, R., Fujii, K., Hibi, K., Maruyama, T., Iwatani, Y. and Tamura, Y.:
Observation of wind speed profiles in Tokyo city area using doppler sodars, Proceedings of 16th
National Symposium on Wind Engineering, pp. 13-18, 2000 (in Japanese)
27) Sadaki, R., Ohtake, K., Goto, S., Mtsui, M., Mtsuyama, T, and Miyashita, K.: Profile of turbulence
intensity and turbulence scale of wind based on observation data, AIJ Journal of Technology and
Design, Vol. 21, No. 48, pp. 477-480, 2015 (in Japanese)
28) Kondo, K., Kawai, H. and Kawaguchi, A.: Topographic multipliers for mean and fluctuating wind
velocities around up-slope cliffs, Summaries of Technical Papers of Annual Meeting Architectural
Institute of Japan, B-1, pp. 105-106, 2001 (in Japanese)
29) Kawai, H. and Kondo, K.: Topographic multiplier around micro-topography – In case of two
dimensional escarpment and hill, Summaries of Technical Papers of Annual Meeting Architectural
Institute of Japan, B-1, pp. 103-104, 2003 (in Japanese)
30) Tsuchiya, M., Kondo, K. and Sanada, S.: Effects of micro-topography on design wind velocity –
Characteristics of wind velocity amplification around various topographies, Summaries of Technical
Papers of Annual Meeting Architectural Institute of Japan, B-1, pp. 119-120, 1999 (in Japanese)
31) Meng, Y. and Hibi, K.: An experimental study of turbulent boundary layer over steep hills,
Proceedings of 15th National Symposium on Wind Engineering, pp. 61-66, 1998 (in Japanese)
32) Kajishima, T.: Numerical simulation of turbulent flows, Yokendo, 1999
33) Shih, T. H., et al.: Comp. Methods Appl. Mech. Eng., Vol. 125, pp. 287-302, 1995
34) Ono, Y. and Tamura, T.: Application of LES to the high wind turbulence around a steep hill with
surface roughness, Journal of Structural and Construction Engineering AIJ, No. 606, pp. 73-80, 2006
(in Japanese)
35) Kishida, T. and Tamura, T.: CFD of wind flow around simple hill considering the effect of surface
roughness, 27th CFD Symposium, B09-4, 2013 (in Japanese)
36) Ohtake, K.: Peak wind pressure coefficients for cladding of a tall building - Part 1 Characteristics
of peak pressure, Summaries of Technical Papers of Annual Meeting Architectural Institute of Japan,
Structures I, 2000, pp. 193-194 (in Japanese)
CHAPTER 6 WIND LOADS - C6-91 -

37) Ohtake, K.: Peak wind pressure coefficients for cladding of a tall building - Part 2 Stretch of Peak
Wind Pressure, Summaries of Technical Papers of Annual Meeting Architectural Institute of Japan,
Structures I, 2001, pp. 143-144 (in Japanese)
38) Architectural Institute of Japan (AIJ): Manual of cladding wind resistance evaluation for designers
and engineers, 2013 (in Japanese)
39) Nishimura, H., Asami, Y., Takamori K. and Okeya M.: A wind tunnel study of fluctuating pressures
on buildings – Part 4 Pressure coefficient and drag coefficient, Summaries of Technical Papers of
Annual Meeting Architectural Institute of JAPAN, Structures I, 1992, pp. 57-58 (in Japanese)
40) Eurocode ENV 1991-2-4: 1997
41) Kamei, I. and Maruta, E.: Wind tunnel test for evaluating wind pressure coefficients of buildings
with a gable roof – Part 2, Summaries of Technical Papers of Annual Meeting Architectural Institute
of Japan, Structures, 1981, pp. 1041-1042 (in Japanese)
42) Kanda, M. and Maruta, E.: Study on design wind pressure coefficient of low rise buildings with a
flat roof or a gable roof – Part 3 Averaging wind pressure coefficient and wind direction, Summaries
of Technical Papers of Annual Meeting Architectural Institute of Japan, Structures I, 1992, pp. 119-
120 (in Japanese)
43) Ueda, H., Tamura, Y. and Fujii, K.: Effect of turbulence of approaching wind on mean wind
pressures acting on flat roofs – Part 1 Study on characteristics of wind pressure acting on flat roofs,
Journal of Structural and Construction Engineering, No. 425, pp. 91-99, 1991 (in Japanese)
44) Terazaki, H., Katsumura, A., Uematsu, Y., Ohtake, K., Okuda, Y., Kikuchi, H., Noda, H., Masuyama,
Y., Yamamoto, M. and Yoshida, A.: Wind force coefficient and gust loading factor for roof and eave,
Journal of Wind Engineering, JAWE, No. 129, 2011, pp. 343-361 (in Japanese)
45) Ueda, H., Hagura, H. and Oda, H.: Characteristics of stress generated by wind pressures and wind
loads acting on stiff two-dimensional arches supporting a barrel roof, Journal of Structural and
Construction Engineering, AIJ, No. 496, pp. 29-35, 1997 (in Japanese)
46) Kikuchi, T., Ueda, H. and Hibi, K.: Characteristics of wind pressures acting on the curved roofs,
Summaries of Technical Papers of Annual Meeting Architectural Institute of Japan, Structures I,
2003, pp. 147-148 (in Japanese)
47) Noguchi, M. and Uematsu, Y.: Design wind pressure coefficients for spherical domes, Journal of
Wind Engineering, JAWE, No. 95, 2003, pp. 177-178 (in Japanese)
48) Yasunaga, J., Koo, C. and Uematsu, Y.: Wind loads for designing cylindrical storage tanks – Part 2
Wind force model with consideration of the buckling behavior under wind loading, Journal of Wind
Engineering, JAWE, No. 132, 2012, pp. 79-92 (in Japanese)
49) Chino, N. and Okada, H.: Wind-induced internal pressures in buildings – Part 1 Mean internal
pressures, Journal of Wind Engineers, No. 56, pp. 11-20, 1993 (in Japanese)
50) Architectural Institute of Japan (AIJ): Data for wind resistant design of buildings – Wind resistant
design and wind resistance evaluation of buildings, pp. 3-6-3-9, 2008 (in Japanese)
51) Ueda, H. and Hibi, K.: Simulation of internal pressures in low-rise buildings induced by wind,
Journal of Structural and Construction Engineering AIJ, No. 622, pp. 65-72, 2007 (in Japanese)
52) Schewe, G.: On the force fluctuations acting on a circular cylinder in crossflow from subcritical up
to transcritical Reynolds numbers, Journal of Fluid Mechanics, Vol. 133, pp. 265~285, 1983
53) Uematsu, Y., Yamada, M.: Aerodynamic forces on circular cylinders of finite height, Journal of
Wind Engineering and Industrial Aerodynamics, Vol. 51, pp. 249-265, 1994
54) Noguchi, M.: Study of fluctuating wind pressures on elliptical cylinder structures, Doctoral
- C6-92 - Recommendations for Loads on Buildings

Dissertation, University of Tokyo, 2009

55) Uematsu, Y., Iizumi, E. and Stathopoulos, T.: Wind loads on free-standing canopy roofs – Part 2
Wind force coefficients for the design of main force resisting systems, Journal of Wind Engineering,
JAWE, Vol. 31, No. 2 (No. 107), April 2006, pp. 35-49 (in Japanese)
56) Uematsu, Y., Iizumi, E. and Stathopoulos, T.: Wind loads on free-standing canopy roofs – Part 3
Validity and application of the proposed wind force coefficients, Journal of Wind Engineering,
JAWE, Vol. 31, No. 4 (No. 109), October 2006, pp. 115-122 (in Japanese)
57) JEC: Design standards on structures for transmissions, JEC-127, 1979 (in Japanese)
58) Nishimura, H.: Aerodynamic characteristics of shapes yielding stagnated flow, GBRC, No. 106, pp.
19-24, 2002
59) Katagiri, J., Kawabata, S., Niihori, Y. and Nakamura, O.: Pressure characteristics of rectangular
cylinders with cut corner, Summaries of Technical Papers of Annual Meeting Architectural Institute
of Japan, Structures I, 1992, pp. 47-48 (in Japanese)
60) Maruta, E., Ueda, H. and Kanda, M.: Local wind pressure on gable roofs, Summaries of Technical
Papers of Annual Meeting College of Industrial Technology Nihon University, 1991, pp. 33-36 (in
Japanese)
61) Uematsu, Y.: Peak gust pressures acting on low-rise building roofs, Proceedings of the 8th East
Asia-Pacific Conference on Structural Engineering and Construction, Singapore, 2001
62) Uematsu, Y., Isyumov, N.: Peak gust pressures acting on the roof and wall edges of low-rise building,
Proceedings of the 8th East Asia-Pacific Conference on Structural Engineering and Construction,
Singapore, 2001
63) Uematsu, Y. and Yamada, M.: Fluctuating wind pressures on buildings and structures of circular
cross-section at high Reynolds numbers, Proceedings of the 9th International Conference on Wind
Engineering, New Delhi, 1995, pp. 358-368
64) Ono, Y., Tamura, T. and Kataoka, H.: LES analysis for maximum negative pressure on an elliptic
structure, Journal of Structural and Construction Engineering AIJ, No. 652, pp. 1081-1087, 2010 (in
Japanese)
65) Ueda, H., Hibi, K. and Kikuchi, H.: Wind-induced internal pressures in low-rise buildings with
partition walls, Journal of Structural and Construction Engineering AIJ, Vol. 75, No. 648, pp. 261-
268, 2010 (in Japanese)
66) Ueda, H., Hibi, K. and Kikuchi, H.: The characteristics of wind-induced internal pressures in high-
rise buildings, Journal of Structural and Construction Engineering AIJ, Vol. 77, No. 676, pp. 833-
842, 2012 (in Japanese)
67) ISO4354: International standard, 1997
68) Uematsu, Y., Iizumi, E. and Stathopoulos, T.: Wind loads on free-standing canopy roofs – Part 1
Peak wind force coefficients for the design of cladding, Journal of Wind Engineering, JAWE, No.
105, 2005, pp. 91-102 (in Japanese)
69) Asami, Y. and Nakamura, O.: A proposal for alongwind load model, Summaries of Technical Papers
of Annual Meeting Architectural Institute of Japan, Structures I, pp. 195-196, 2002 (in Japanese)
70) Asami, Y., Kondo, K. and Hibi, K.: Experimental research of aerodynamic force on rectangular
prism, Journal of Wind Engineering, JAWE, No. 91, pp. 83-88, 2002 (in Japanese)
71) ASCE: Minimum design loads for buildings and other structures, ASCE7-98. 1998
72) Holmes, J. D.: Effective static load distributions in wind engineering, Journal of Wind Engineering
and Industrial Aerodynamics, Vol. 90, pp. 91-109, 2002
CHAPTER 6 WIND LOADS - C6-93 -

73) TPU Aerodynamic Database. [Link]

74) Katsumura, A., Kawai, H., Terazaki, H., Tamura, Y. and Fugo, Y.: Modeling of wind-induced
torsional moment on tall buildings, Proceedings of 23th National Symposium on Wind Engineering,
pp. 421-426, 2014 (in Japanese)
75) Marukawa, H., Tamura, Y., Sanada, S. and Nakamura, O.: Lift and across-wind response of a 200
m concrete chimney, Journal of Wind Engineering, JAWE, No. 19, pp. 37-52, 1984 (in Japanese)
76) Tamura, Y. and Amano, A.: Vortex induced vibration of circular cylinder – Part Ⅲ Model for
vortex induced oscillations of continuous systems, Transactions of Architectural Institute of Japan,
No. 337, pp. 65-72, 1984 (in Japanese)
77) Hibi, K., Tamura, Y. and Kikuchi, H: Peak normal stresses and wind load combinations of middle-
rise buildings, Summaries of Technical Papers of Annual Meeting, AIJ, Structures I, pp. 113-114,
2003 (in Japanese)
78) Asami, Y: Combination method for wind loads on high-rise buildings, Proceedings of 16th National
Symposium on Wind Engineering, pp. 531-534, 2000 (in Japanese)
79) Somekawa, D., Kawai, H., Nishimura, H: Expermental study on the correlation coefficient for
combination of wind loads, Summaries of Technical Papers of Annual Meeting, AIJ, Structures I,
pp. 149-150, 2014 (in Japanese)
80) Yoshikawa, M. and Tamura, T.: Evaluation of fluctuating wind pressures on 3D square cylinder by
unstructured-grid LES – Formulation of LES using unstructured grid system for wind-resistant
design of buildings (Par1 1), Journal of Structural and Construction Engineering AIJ, Vol. 78, No.
687, pp. 913-921, 2013 (in Japanese)
81) Kataoka, H. and Tamura, T.: Study on the relationship between roughness parameters and vertical
wind velocity profiles over an urban area by LES, Journal of Structural and Construction Engineering
AIJ, Vol. 77, No. 678, pp. 1203-1210, 2012 (in Japanese)
82) Nozu, T., Tamura, T., Kishida, T. and Ktsumura, A.: Combined model of structured and unstructured
grid system for wind pressure estimation of a tall buildings, 6th European and African Conference on
Wind Engineering, pp. 1-4, 2013
83) Maruyama, Y., Tamura, T. and Kishida, T.: LES of wind flow around an actual complex terrain –
A study of turbulence affeted by terrain, Annual Meeting of The Japan Society of Fluid Mechanics,
2013 (in Japanese)
84) Takamori, K., Nishimura, H., Taniike, Y., Okazaki, M., and Taniguchi, T.: Aerodynamic
interference effect between twin tall buildings – Part 1 Square sectioned buildings, Summaries of
Technical Papers of Annual Meeting, AIJ, Structures I, pp. 249-250, 2000 (in Japanese)
85) Takamori, K., Nishimura, H., Taniike, Y. and Okazaki, M.: Aerodynamic interference effects
between twin tall buildings – Part 1 Study of influences on the wind load, Summaries of Technical
Papers of Annual Meeting, AIJ, Structures I, pp. 173-174, 2001 (in Japanese)
86) Taniike, Y.: Interference effects between tall buildings with square section in a high-turbulent
boundary layer, Proceedings of 10th National Symposium on Wind Engineering, pp. 247-252, 1988
(in Japanese)
87) Guidelines for the evaluation of habitability to building vibration, AIJ, 2004 (in Japanese)
88) Guidelines for Wind-resistant Design of Base-isolated Buildings, JSSI, 2012 (in Japanese)
89) Fatigue Design Recommendations for Steel Structures, JSSC, 2012 (in Japanese)
90) Structural Design Concepts for Earthquake and Wind, AIJ, 1999 (in Japanese)
91) Guidelines of Recommendations for Loads on Buildings 1, AIJ, 2016 (in Japanese)
- C7-1 -

CHAPTER 7 SEISMIC LOADS

Introduction
When earthquakes occur, a building undergoes dynamic motion. This is because the building is
subjected to inertia forces that act in the opposite direction to the acceleration of earthquake excita-
tions. These inertia forces, called seismic loads, are usually dealt with by assuming forces external to
the building. Because both earthquake motions and inertia forces to buildings are time varying and
direction dependent, seismic loads are not constant in terms of time and space. In designing buildings,
the maximum story shear force is considered to be the most influential, and therefore the seismic loads
in this chapter are the static loads used to give the maximum story shear force for each story, i.e., the
equivalent static seismic loads. Time histories of earthquake motions are also used to analyze high-rise
buildings, and their members and contents for a seismic design. The earthquake motions for a dynamic
design are called design earthquake motions. In the previous recommendations, only the equivalent
static seismic loads were considered to be seismic loads. In a wider sense, both equivalent static seis-
mic loads and design earthquake motions as the time histories are included in the seismic loads con-
sidered in this chapter. In ISO/TC98, which deals with “bases for design of structures”, the term “ac-
tion” is used instead of “load”, and includes both the load as an external force and various influences
that may cause deformations to the structures. In the future, the term “action” may take the place of
“load”.
Section 7.1 deals with the fundamentals of estimating seismic loads. Section 7.2 is concerned with
the calculation of seismic loads that provides equivalent static seismic loads for a building above the
ground level. Section 7.3 discusses the design earthquake motions in the time histories used for dy-
namic analyses. The overall framework and fundamental items and equations are described in each
section.
The specific value of each factor is given in the main text when it is possible to give a normative
value, or is provided in the commentary when only an informative value is possible.
Section 7.2 defines each seismic load as the story shear force for each story, and uses the equivalent
static seismic loads for a limit state design, allowable stress design, and ultimate strength design. As in
the 1993 recommendations, the equivalent static seismic loads are expressed as the story shear forces
in the building that can be calculated through a response spectrum analysis. The analytical model is
fundamentally the multi- degree-of-freedom (MDOF) model (also called a “lumped mass model”) that
takes into account the soil-structure interaction considering sway and rocking motions. The maximum
story shear force for each mode is calculated from the natural frequency and from the modes based on
the eigenvalue analysis of the model, estimating the damping factor for each mode. The seismic loads
are the maximum story shear forces that can be calculated from the shear force of each mode. The
fixed based model can be employed if the building is on firm soil, and thus the soil-structure interac-
tion may be neglected.
Because the response spectra are affected by the characteristics of the earthquake motions and
- C7-2 - Recommendations for Loads on Buildings

buildings, they are expressed as functions of many parameters, representing input motions to buildings
or soil-structure systems.
The response spectra defined for the engineering bedrock are used as inputs to the soil-structure
system, or the response spectra that consider the amplification effect (including nonlinear characteris-
tics) of the surface soil and soil-structure interaction are used as inputs of the building.
The seismic loads that take into account the inelastic response of buildings during strong earthquake
motions include the reduction factor related to the ductility and response deformation, which are de-
fined considering the nonlinear characteristics of the building and its limit deformation. This reduction
factor was given as a unique value for each direction of the building plan in the previous recommen-
dations, but is given for each story and in each direction in the current recommendations. Because it is
possible to provide the same value for all stories, the concept of a reduction factor is the same as in the
previous recommendations.
For relatively regular buildings, a simplified method can be used to evaluate the equivalent static
seismic loads, without performing a dynamic response analysis. In this case, it is also recommended
for the natural period and damping factor to be evaluated by considering the soil-structure interaction.
For irregular buildings in terms of plan, elevation, or both, as well as very important buildings and
buildings with long spans, a dynamic analysis is conducted using design earthquake motions, as dis-
cussed in Section 7.3. It is also recommended for an equivalent static analysis to be used in a comple-
mentary manner.
Section 7.3 deals with design earthquake motions used for a dynamic analysis, which were included
in the commentary of the previous recommendations. This is because a dynamic analysis has become
prevalent in the design and technical development when evaluating design earthquake motions with
reliable accuracy. After the 2011 off the Pacific Coast of Tohoku Earthquake, it has become much
more important to evaluate the characteristics of earthquake motions in both the period and time do-
mains and to evaluate the responses of the building structures, non-structural members, furniture, and
equipment, in order to design safe buildings and prevent future disasters1),2). There are two methods
used to evaluate design earthquake motions. The first method is to make design earthquake motions
that fit the response spectrum, as prescribed in Sec. 7.2. The second is to make design earthquake mo-
tions based on scenario earthquakes that take into account various conditions of the construction site.
The first method is to make design earthquake motions that fit the response spectrum on the engineer-
ing bedrock or at the location of the input level of the building. The second method is to generate de-
sign earthquake motions based on several scenario earthquakes, considering the seismic source effects,
propagation path of the seismic waves, and soil conditions of the construction sites.

Notations: Notations used in the text are as follows:

Roman upper-case notations
AEi : maximum floor response acceleration at the i-th story of the building
Af : area of the foundation base
CHAPTER 7 SEISMIC LOADS - C7-3 -

E : seismic load effect

FNS : seismic load on the nonstructural components or equipment
Ga(ω) : acceleration power spectrum at the ground level or base of the foundation
GaE(ω) : acceleration power spectrum at the engineering bedrock
H(ω) : transfer function of acceleration
HGS(ω) : soil amplification function representing the amplification characteristics
of the surface soil
HSSI(ω) : adjustment function representing the soil-structure interaction
QE : live load for a seismic load
S0(T, ζ) : normalized acceleration response spectrum at the engineering bedrock
Sa0(T, ζ) : basic acceleration response spectrum at the engineering bedrock
SaE(T j, ζ j) : acceleration response spectrum at the engineering bedrock
Sa(T j, ζ j) : acceleration response spectrum at the base of the foundation
or at the ground level
T : period
T1 : fundamental natural period of the building
TG : predominant period of soil above the engineering bedrock
Tc, Tc’ : corner periods related to the predominant period of the engineering bedrock
Td : duration for spectral conversion
Tf : fundamental natural period of the building fixed at the base
Tj : j-th natural period of the building
Tr : natural period of rocking assuming the building is rigid
Ts : natural period of sway assuming the building is rigid
VEi : seismic shear force of the i-th story of a building
Wi : equivalent weight amount at the i-th story of a building
Wi j : equivalent weight amount at the i-th story for the j-th natural mode
Vs : shear wave velocity
Roman lower-case notations
a0 : basic peak acceleration at the engineering bedrock
g : acceleration from gravity
g(η, ω) : function representing embedded foundation
i : the story number or imaginary unit
jc : maximum number of natural modes considered
k : factor representing statistical seismic hazard at the site
k1, k2 : factors determined depending on the number of stories
(or height or fundamental natural period T1)
kDi : reduction factor related to ductility of the building
kFi : amplification factor owing to irregularities of the building
- C7-4 - Recommendations for Loads on Buildings

kG : acceleration response ratio of the nonstructural elements

or equipment
kR0 : acceleration response ratio for the period in which
the acceleration response level is constant
kVEi : seismic shear force distribution factor for the i-th story
kp(T j, ζ j) : peak factor of the acceleration response
of a single-degree-of-freedom system
kRE(T) : return period conversion factor
n : total number of stories of the building
tR : return period
wi, wk : weights at the i-th and k-th stories of the building
wNS : weight of the nonstructural elements and equipment
Greek notations
αG : impedance ratio from the engineering bedrock to the surface soil
αi : normalized weight
βj : participation factor of the j-th natural mode
δd : normalized depth of the foundation
ζ : damping factor
ζ1 : damping factor for the fundamental mode of the building
ζG : damping factor of soil above the engineering bedrock
ζf : damping factor for the fundamental mode of the building fixed at the base
ζj, ζk : j-th and k-th mode damping factors
ζs : damping factor of sway assuming the building is rigid
ζr : damping factor of rocking assuming the building is rigid
η : ratio between embedment depth of foundation d and equivalent width l
µm : adjustment factor for single-degree-of-freedom assumption
of multi-degree-of-freedom system
ρjk : correlation factor for the j-th and k-th natural modes
σa(T j, ζ j) : root mean square of the acceleration response
of a single-degree-of-freedom system
φk j : j-th natural mode shape at the k-th story
ω : circular frequency
ωj : j-th natural circular frequency of the building
- C7-5 -

7.1 Principles of Defining Seismic Loads

7.1.1 Component factors of seismic loads

As represented through a probabilistic seismic hazard evaluation, various uncertain factors of influ-
ence concerning earthquake motions are modeled based on probability theory, and are formulated
against the background of the probabilistic method. Then, the seismic load can be set for buildings
considering both the spectral characteristics of earthquake motion and the dynamic property of the
buildings.
For ordinary buildings, an equivalent static load is calculated using a response spectrum method
described in Sec. 7.2, and is applied for a static stress analysis (this series of procedures may be re-
ferred to as the equivalent static analysis). The response spectrum method is basically applicable only
for elastic structures, but can be used to approximately estimate elasto-plastic structures with uniform
plasticity within the buildings. In general, the influence of a seismic load in the vertical direction and
the resulting vertical vibration cannot be ignored when the influence of irregularity of the ground and
bedrock surfaces is large. However, in some ordinary buildings, the influence of vertical vibration may
be ignored in construction sites with horizontally layered soil deposits where the horizontal seismic
loading becomes dominant. On the other hand, for irregular buildings in plane or elevation, buildings
with a large span frame, particularly important buildings, and buildings where it is difficult to apply
the response spectrum method, as shown in a) through h) below, dynamic analyses (earthquake re-
sponse analyses) with design ground motions given as the acceleration time history, described in Sec.
7.3, should be implemented to verify the seismic safety. Note that the static analysis in Sec. 7.2 is also
preferably carried out to grasp the required safety level even when a dynamic analysis is implemented.

a) a building with abrupt change in horizontal stiffness and strength in height and possible dam-
age concentration in a particular story,
b) a building with plane irregularity in mass and stiffness and significant torsional response,
c) a building for which the vertical vibration cannot be ignored owing to its large- spanned beams
and/or long cantilevers,
d) a non-multi-story and special shaped building, such as shells or spatial structures,
e) a tall or large sized building with great importance,
f) a building that contains telecommunication devices and valuable content, and whose structural
dynamic response must be estimated,
g) a building that contains dangerous materials, and whose failure may greatly affect the sur-
roundings, and
h) a building with special devices such as base-isolation systems.
- C7-6 - Recommendations for Loads on Buildings

Table 7.1.1 compares the recommendations for the seismic load effects proposed in the current
(2015) and previous (2004) versions of the Recommendations for Loads on Buildings. As can be seen
from this table, the available methods for evaluating the seismic load effects have been improved in
the current recommendations.
Based on a standard probabilistic seismic hazard map, the equivalent static force procedure evalu-
ates the acceleration response spectrum, 𝑆𝑆a (𝑇𝑇𝑗𝑗 , 𝜁𝜁𝑗𝑗 ), which is input to the building, by taking account
the spectral property, amplification of wave motion through shallow ground deposits above the bed-
rock, and the effect of soil-structure interaction. As the standard seismic hazard map, it is possible to
select either the peak ground acceleration (PGA) or the uniform hazard spectrum of the acceleration
response value which is defined on the outcropped engineering bedrock. Once 𝑆𝑆a (𝑇𝑇𝑗𝑗 , 𝜁𝜁𝑗𝑗 ) is evaluated,
the seismic load can be calculated as the seismic design story-shear force, which is obtained by multi-
plying the following basic story-shear force with a reduction factor related to building ductility and an
amplification factor from building irregularities. The basic story-shear force shall be calculated
through the following two procedures. One procedure is a response spectrum method in which an ei-
genvalue analysis of a building vibration model is first implemented and a superposition of each mod-
al response based on the response spectrum is then made to determine the seismic load. Another pro-
cedure is a simplified method without performing an eigenvalue analysis. When the vibration modes
higher than the first mode are significant, it is desirable to perform an eigenvalue analysis.
Regarding the design earthquake ground motion for dynamic analyses, it has been first adopted in
the main text since the previous version (2004) of the recommendations in addition to the seismic load
used for the equivalent static force procedure until then. In the current recommendations (2015), the
building response evaluation from the dynamic analyses has also been incorporated in the main text.
CHAPTER 7 SEISMIC LOADS - C7-7 -

Table 7.1.1 Comparison of current and previous recommendations

Effects Current (2015) Previous (2004)
Ground Statistical char- Adopt publicly released evaluation 𝑎𝑎0 : PGA on the outcropped en-
motion acteristics of results as a standard seismic hazard gineering bedrock with 𝑉𝑉S =
property in ground motion map with latest information, and add 400m/s equivalent, for
equivalent intensity a new ground motion index of 100-year return period based on
static force 𝑆𝑆a0 (𝑇𝑇, 0.05), uniform hazard spec- a standard probabilistic seismic
procedure trum of acceleration response value, hazard map.
in addition to 𝑎𝑎0 in previous version
(2004).
Wave amplifica- 𝐻𝐻GS (𝜔𝜔) in previous version (2004), 𝐻𝐻GS (𝜔𝜔): Transfer function based
tion characteris- and added an idea of dealing with on a 1D shear wave propagation
tics in surface liquefaction and lateral flow of sur- theory, for which a simplified
layers face soil deposits. formula is also provided.
SSI (kinematic 𝐻𝐻SSI (𝜔𝜔): Modification function representing embedment effect of a
interaction) building.
Response spec- 𝑆𝑆a (𝑇𝑇𝑗𝑗 , 𝜁𝜁𝑗𝑗 )
trum
Dynamic SSI (inertial in- (1) Sway-rocking (SR) model when performing an eigenvalue analysis.
property of teraction) (2) Modification of 𝑇𝑇𝟏𝟏 , 𝜁𝜁𝟏𝟏 when not performing an eigenvalue analy-
buildings in sis.
equivalent Distribution of (1) Superposition of eigenmodes (SRSS, CQC) when performing ei-
static force seismic sto- genvalue analysis.
procedure ry-shear forces (2) Empirical distribution (𝑘𝑘Vi ) of shear forces when not performing
eigenvalue analysis.
Effect of inelas- For an estimate of maximum response 𝑘𝑘Di : Reduction factor from plas-
tic response displacement, the limit strength cal- tic energy absorption.
culation method and the energy bal-
ance method, in addition to FE-
MA356 method in previous version
(2004).
Effect from ir- Add case study in the US, and for 𝑘𝑘Fi : Amplification factor from
regularity of wooden building in Japan, in addition irregularities of the building.
building to that in previous version (2004).
- C7-8 - Recommendations for Loads on Buildings

Table 7.1.1 (continued)

Design Generation (1) Generation method compatible with the design response spectrum.
earthquake methods of de- (2) Generation method based on the scenario earthquakes according to
motion and sign earthquake the local site conditions and requirements of the building.
response motions
evaluation Generation of Add empirical time history prediction Fundamental concept setting,
in dynamic design earth- methods of the long-period earth- classification of site-related
analysis quake motions quake motions, in addition to those in earthquakes to be considered in
based on the previous version (2004). design, and classification of
scenario earth- generation methods of design
quakes earthquake motions based on
scenario earthquakes.
Evaluation of Fundamental concept setting, and not considered
building re- classifications for structural model of
sponse building and ground, dynamic analy-
sis method, and nonlinear ground
model.

7.1.2 Idealization of building and soil

(1) Analytical model

An actual building is very complex, but a building is often idealized into an adequate mathematical
model to evaluate the seismic load. A lumped mass structural model is often used since most of the
mass is concentrated on each floor of the building.
In a response analysis, the building should be idealized such that the forces and deformations of
each part induced by earthquakes can be appropriately evaluated. Among the analytical models, there
are equivalent shear models where the restoring characteristics of a story is expressed by the shear
element, a flexure-shear model where the flexure component and shear component of vibration can be
treated separately, a frame model where the mass is allocated at each joint of the structural members
with their restoring characteristics, and a soil-foundation-structure model where the soil is treated as a
finite element. These models should be selected by considering the characteristics of the buildings of
concern because the application limit of each model is not identical.
The mass concentration on the floor level is assumed to move horizontally and independently with
disregard to the motion in another orthogonal direction. If the mass and stiffness of a building are not
evenly distributed, two orthogonal horizontal motions are not independent of each other and rotational
CHAPTER 7 SEISMIC LOADS - C7-9 -

(torsional) motion is excited. Therefore, three degrees of freedom (two horizontal and torsion) should
be considered for each story. When a building is irregular and the torsional vibration is dominant, a
response analysis with three-dimensional frame models using design earthquake motions as two direc-
tional simultaneous inputs should be performed.
When the floor diaphragm is insufficiently stiff to assume that the diaphragm is rigid, a model that
connects each rigid diaphragm through flexible members should be used. Because the vertical motion
of a building is usually small, it is often neglected. For buildings with long span beams or long span
cantilever beams, not only horizontal motions but also vertical motions should be considered, and a
dynamic model that takes the vertical motion into account is often used. In this case, live loads should
also be appropriately specified. For planarly large buildings, the effects of a phase difference of the
ground motion should be examined.
When the soil-structure interaction effect is small such as high-rise buildings without embedding on
firm ground, the building can be modeled as a fixed based model, as illustrated in Fig.7.1.1a. On the
other hand, the effects of the soil-structure interaction may not be ignored for low-rise buildings, pla-
narly large buildings, or buildings with an embedment.
If the soil-structure interaction effect cannot be neglected, the so called sway-rocking (SR) model,
which can represent the sway and rocking motions of a foundation, should be used to take into account
the soil-structure interaction effect, as shown in Fig.7.1.1b. This SR model is the basic analysis model
referred in the main text.
There are two methods used to include the soil-structure interaction effect, one is a dynamic sub-
structure method such as an SR model, and the other is an integrated method, as shown in Fig.7.1.1c.
The former uses dynamic soil spring (impedance) and input motion of the foundation (driving force)
for a response analysis of a building, and the latter analyzes the response of the building and soil sim-
ultaneously by modeling the soil as a finite element. In the latter case, the procedure used to evaluate a
seismic load is simple, but the analytical models tend to be large and require an increasingly longer
CPU time. The details of this are discussed elesewhere3). Note that the treatment of the soil-structure
interaction effect is discussed in Sec. 7.2.2 and (4) of Sec. 7.2.4, and dynamic analyses are discussed
in Sec. 7.3.4.
It is well known that the soil conditions affect the earthquake damage, and that many seismic design
codes include the soil profile types. In many cases, however, only the seismic design forces are ad-
justed or the design spectra are modified. To consider the dynamic characteristics of the surface soil
layers, it is necessary to model the soil layers to be analyzed. In this case, the earthquake motion is
determined at the so-called engineering bedrock, and how the amplitude and frequency contents are
affected through the soil layers is analyzed. The earthquake motion is then determined at the ground
surface or at the bottom of the foundation. The engineering bedrock is chosen as the stiff soil layer that
is deeper than the layer supporting the foundation. The stiff soil layer for engineering purposes is the
layer whose shear wave velocity is approximately 400 m/s or greater.
All of the above analytical models should have an appropriate damping. The damping factor de-
- C7-10 - Recommendations for Loads on Buildings

pends on the structural type and materials, as well as on the stiffness of the foundation and soil. The
general feature of damping is indicated in the commentary of (2) in Sec. 7.2.1.

(2) Analysis method

For the equivalent static load, the structural model is analyzed using the response spectrum method
considering the reduction factor described in Sec. 7.2.5 on the plastic deformation capability.
For such dynamic systems with torsional motion, the response spectrum method can also be applied.
However, the assumption that the maximum of each mode appears independently may not be appro-
priate, and the square root of sum of squares (SRSS) rule to estimate the maximum response cannot be
used. In such cases, the complete quadratic combination (CQC) rule is used when considering the cor-
relations between the eigen-modes of the vibration.
For dynamic analyses using design earthquake motion, soil and buildings are analyzed through an
equivalent linear or non-linear method, according to the degree of non-linearity. An equivalent linear
method can be applied to soil whose strain is between 0.1% and 1%4), and a non-linear method is ap-
plied when a larger strain is anticipated. An equivalent linear or non-linear method is applied to build-
ings. Although a small deformation can be assumed in the response analysis of a building, a finite de-
formation must be employed for treating a building collapse or large span structures5).

1. Ground surface
2. Basement and foundation
3. Spring for sway
4. Spring for rocking
5. Ground

a. Fixed base b. Interaction model c. Interaction model

model (Sway-rocking model) (FEM model)
Figure 7.1.1 Models from the viewpoint of soil structure interaction
(Dashpots representing the damping are omitted.)

(3) Live load for seismic load

The live load for a seismic load is the weight of the objects per unit floor area that are fixed to the
building. Because the motion of people inside a building differs from the building itself, a live load
CHAPTER 7 SEISMIC LOADS - C7-11 -

caused by people is not included in a seismic load. Only the objects fixed to the building are included,
because they move along with the building, and objects that may slide during an earthquake are not
included. However it is difficult to decide whether an object in a building is fixed. Therefore, the av-
erage weights of fixed objects for the room types in Table 4.1 are shown in Table 7.1, and can be used
as the live load for a seismic load. When a vertical vibration is considered, the appropriate live loads
should be determined. The background for Table 7.1 is provided in Sec.4.4 Recommendations for
Loads on Buildings (1993). An example of calculating the object weight from another study is shown
in Fig. 4.4.16), in which a tributary is the area of influence of the floor. The objects considered are the
same as described in Sec. 4.2. Therefore, if there are some objects that are not considered in Sec. 4.2,
the load in Table 7.1 should be increased. For a live load consisting mainly of people, the live load
used for a seismic load is set to 300 N/m2 when considering fixed chairs, furniture, and other such ob-
jects. Each value in the table has been determined by looking into Table 4.2.5. The weight of fixed
chairs was obtained through a simulation using weight data listed in a catalogue.
Takanashi, et al.7) showed that equivalent acceleration response can be reduced in a short period for
sliding objects in a building during an earthquake. To consider the effect of sliding objects in a seismic
design, a method is proposed that uses the equivalent live load where the weight of the sliding objects
is excluded. During a long period, objects may not slide, but the live load caused by people can be ex-
cluded. Therefore, a method to reduce the live load for use in a seismic load can be used for all peri-
ods.
From the above discussions, when objects slide during an earthquake, the live load may be reduced
considering the situations and conditions of the objects.

7.2 Estimation of Equivalent Static Seismic Loads

7.2.1 Methods for estimating seismic load

In Eq. (7.1), the maximum shear force of the i-th story for the j-th natural mode is given as an equivalent
static force. Spectral acceleration at the basement or ground surface 𝑆𝑆a �𝑇𝑇1, ζ1 �, where 𝑇𝑇1 (unit, s) and ζ1
are the natural period of the first mode and the damping factor, respectively, is estimated using Eq. (7.11)
by considering the regional characteristics of a seismic hazard (see 7.2.4). The seismic shear force 𝑉𝑉E𝑖𝑖 is
estimated by taking in consideration the reduction factor related to ductility 𝑘𝑘D𝑖𝑖 , the amplification factor
according to the irregularities of the building 𝑘𝑘F𝑖𝑖 , and the equivalent weight considering the vibration
characteristics of a multi-story building 𝑊𝑊𝑖𝑖 .
The equivalent weight 𝑊𝑊𝑖𝑖 is equal to the weight of the building for a single-story building. For an esti-
mation of 𝑊𝑊𝑖𝑖 , it is recommended to utilize an eigenvalue analysis that can take higher vibration modes into
consideration (see 7.2.3(1)). A simplified method such as the Ai distribution is not precluded (see 7.3.2(2)).
Eq. (7.1) also introduces a factor related to ductility 𝑘𝑘D𝑖𝑖 and a factor according to the irregularities of
the building 𝑘𝑘F𝑖𝑖 , respectively. The reduction factor related to ductility 𝑘𝑘D𝑖𝑖 is 1 for the serviceability limit
state. Here, 𝑘𝑘D𝑖𝑖 is estimated through a proper consideration of the corresponding deformation limit for a
- C7-12 - Recommendations for Loads on Buildings

safety limit state. The amplification factor according to the irregularities of the building 𝑘𝑘F𝑖𝑖 is 1 for a reg-
ular shaped building (see 7.2.6).

1) Natural period
In general, natural periods of the first and higher elastic vibration modes are estimated based on an ei-
genvalue analysis. Statistical relations used to estimate these parameters are shown as references.
The relation between the natural period of the first vibration mode 𝑇𝑇1(s) and building height ℎ(m)8) is
given as follows:
𝑇𝑇1 = (0.02 + 0.01𝛼𝛼h )ℎ, (7.2.1)
where ℎ：height of the building from ground surface (m)，
𝛼𝛼h : ratio of height which is made mainly from steel to total height ℎ.
The relationships as a function of height based on measurement9) are as follows:
𝑇𝑇1 = 0.015ℎ (reinforced concrete structure),
(7.2.2)
𝑇𝑇1 = 0.02ℎ (steel structure)
The relationships as a function of the number of stories 𝑛𝑛9) are as follows:
𝑇𝑇1 = 0.049𝑛𝑛 (reinforced concrete structure),
(7.2.3)
𝑇𝑇1 = 0.082𝑛𝑛 (steel structure)
The relationships between the natural period of the first vibration mode and those for higher vibration
modes9) are given as follows:

𝑇𝑇2 = 0.33𝑇𝑇1
𝑇𝑇3 = 0.18𝑇𝑇1 (7.2.4)
𝑇𝑇4 = 0.13𝑇𝑇1

2) Damping factor
Damping is a phenomenon in which the amplitude decreases with time, and the degree of decrease is
determined by the damping coefficient. The larger the damping is, the smaller the response tends to be
when buildings are subjected to earthquake excitation.
The damping within a building is caused by a transformation into thermal energy caused by the friction
of connections and deformation of the members, additional damping due to the installation of a damper,
and dispersion of the wave energy to soil. It should be noted that damping caused by energy consumption
from inelastic deformation of the members is not included in the damping factor ζ1 of Eq. (7.1), because it
is considered as a factor 𝑘𝑘D𝑖𝑖 (see 7.2.5).
The damping in the ground is caused by soil viscosity, dispersion of wave energy to the adjacent ground,
inelastic hysteresis, and so on. The damping factor is inversely proportional to the frequency for the Max-
well model where a spring and dashpot are applied in parallel. Hysteretic damping does not depend on the
frequency10).
In a dynamic analysis of a building, the damping coefficient proportional to the stiffness given by Eq.
CHAPTER 7 SEISMIC LOADS - C7-13 -

(7.2.5a) is frequently used. In this case, the damping factor becomes proportional to the frequency, and
gives a higher damping for higher modes. A Rayleigh type damping coefficient is represented as a combi-
nation of mass proportional damping and stiffness proportional damping, as indicated in Eq. (7.2.5b).
[𝑐𝑐] = 𝑏𝑏1 [𝑘𝑘], (7.2.5a)
[𝑐𝑐] = 𝑏𝑏0 [𝑚𝑚] + 𝑏𝑏1 [𝑘𝑘], (7.2.5b)
where [𝑚𝑚] and [𝑘𝑘] are a mass matrix and stiffness matrix, respectively, and 𝑏𝑏0 and 𝑏𝑏1 are constants.
It is known that several observations do not show a tendency in which the damping factors for higher
modes are larger than those for the fundamental mode. Therefore, the damping factors are considered to be
overestimated for high-rise buildings if the damping coefficient is assumed to be proportional to the stiff-
ness9), 11).
Conventionally, damping factors for different structures are as follows:
Reinforced concrete structure ζf = 0.02 − 0.04
� (7.2.6)
Steel structure ζf = 0.01 − 0.03
These values are for the internal damping of a structure and do not include dissipation damping to the
ground. The values also depend on the strain level and type of structure.
A damping factor of 5% is conventionally used for a seismic hazard with respect to the spectral accelera-
tion. Several formulae are proposed to obtain the spectral acceleration for different damping factors, in-
cluding the following12):
1.5
𝑆𝑆a (𝑇𝑇1 , ζ) = 𝑆𝑆 (𝑇𝑇 , 0.05) (7.2.7)
1 + 10ζ a 1
A complex eigenvalue analysis is employed to estimate the mode damping factors if different damping
characteristics are assigned to different components. The following equation13) can also be used:
T
�∅𝑗𝑗 � �𝐾𝐾ζ ��∅𝑗𝑗 �
ζ𝑗𝑗 = T
(7.2.8)
�∅𝑗𝑗 � [𝐾𝐾]�∅𝑗𝑗 �

where �∅𝑗𝑗 �, [𝐾𝐾], and �𝐾𝐾ζ � are the mode vector for the j-th vibration mode, stiffness matrix, and stiffness
matrix multiplied by the damping factors for each component, respectively.

3) Notes on allowable stress design

It is assumed in the recommendations that a seismic load is estimated depending on the performance re-
quirements of the building. It is sometimes regarded that the allowable stress design be used for checking
the serviceability limit state, and that a lateral load-carrying capacity calculation be applied for checking
safety limit state.
It is often the case that a lateral load-carrying capacity calculation is not required, with the exception of
buildings specified through article 20(2) of the Building Standard Law of Japan. In such cases, the compat-
ibility of the requirements corresponding to the safety limit state is verified by the design based on the al-
lowable stress design. Therefore, a design seismic load for such allowable stress design must not be esti-
mated for the return period corresponding to serviceability limit state, but instead estimated based on the
- C7-14 - Recommendations for Loads on Buildings

return period corresponding to the safety limit state. The seismic load is then reduced depending on the
margin within the plastic range, i.e., the difference between the criteria of the serviceability and safety lim-
its. A multiplication of 0.2 (0.3 in the case of steel structures) can be used for the reduction corresponding
to the margin. This note is critical if the contributing earthquake source is different depending on the return
period of the seismic hazard, e.g., in cases in which a building is located close to an active fault. It should
also be noted that an important building must not be designed using a procedure described herein, but in-
stead designed based on a rigorous method.

4) Seismic load on non-structural elements and equipment 𝐹𝐹NS

The seismic load on non-structural elements such as ceiling and equipment 𝐹𝐹NS is estimated as fol-
lows14):
𝑘𝑘NS 𝐴𝐴E𝑖𝑖
𝐹𝐹NS = 𝑤𝑤NS (7.2.9)
𝑔𝑔
where
𝑘𝑘NS : amplification of acceleration within non-structural elements or equipment,
𝐴𝐴E𝑖𝑖 : acceleration floor response at the 𝑖𝑖-th floor where non-structural elements or equipment are
located,
𝑤𝑤NS : weight of non-structural elements or equipment.
The acceleration floor response at the 𝑖𝑖-th floor 𝐴𝐴E𝑖𝑖 in Eq. (7.2.9) is estimated similarly as Eq. (7.1)
(see 7.2.3). If non-structural elements or equipment behave elastically, and the effects of irregularity can be
ignored, 𝐴𝐴E𝑖𝑖 can be estimated as follows:

𝑗𝑗 2
𝐴𝐴E𝑖𝑖 = �∑𝑗𝑗=1
𝑐𝑐
�𝛽𝛽𝑗𝑗 ∅𝑖𝑖𝑖𝑖 𝑆𝑆a �𝑇𝑇𝑗𝑗 , 𝜁𝜁𝑗𝑗 �� , (7.2.10)

where 𝑗𝑗𝑐𝑐 : the maximum number of natural modes considered (𝑗𝑗𝑐𝑐 ≥ 3),
𝛽𝛽𝑗𝑗 : participation factor for the j-th natural mode,
∅𝑖𝑖𝑖𝑖 : j-th natural mode shape of the i-th story,
𝑆𝑆a �𝑇𝑇𝑗𝑗 , 𝜁𝜁𝑗𝑗 �: spectral acceleration for the for the j-th vibration mode (natural period 𝑇𝑇𝑗𝑗 and damping
ratio ζ𝑗𝑗 ).
Eq. (7.2.10) is called the square-root-of-sum-of-square (SRSS) method when assuming that the peak re-
sponse for each vibration mode does not always occur simultaneously.
Amplification of acceleration from non-structural elements and equipment 𝑘𝑘NS can be unity if they are
fixed appropriately to the floor, and the natural periods of non-structural elements or equipment 𝑇𝑇NS are
different from those of the building. If the natural periods of non-structural elements and equipment 𝑇𝑇NS
are close to those of the building, the resonance effects need to be taken into consideration. The value of
factor 𝑘𝑘NS also depends on the damping characteristics of non-structural elements and equipment. Here,
𝑘𝑘NS for a ceiling is specified in Article 39 (3) of the Building Standard Law of Japan 15) as follows:
CHAPTER 7 SEISMIC LOADS - C7-15 -

𝑗𝑗𝑐𝑐 2
𝛽𝛽𝑗𝑗 ∅𝑖𝑖𝑖𝑖 𝑆𝑆a �𝑇𝑇𝑗𝑗 , 𝜁𝜁𝑗𝑗 �
𝑘𝑘NS = �� 𝑘𝑘NS𝑗𝑗 � (7.2.11)
𝐴𝐴E𝑖𝑖
j=1
3
1 + 5�𝑇𝑇NS ⁄𝑇𝑇𝑗𝑗 � �𝑇𝑇NS ≤ 𝑇𝑇𝑗𝑗 �
𝑘𝑘NS𝑗𝑗 = � 3 (7.2.12)
6�𝑇𝑇𝑗𝑗 ⁄𝑇𝑇NS � �𝑇𝑇𝑗𝑗 ≤ 𝑇𝑇NS �
When the natural periods for different modes are expected to be close to each other, 𝐴𝐴E𝑖𝑖 is calculated as
follows:
𝑗𝑗𝑐𝑐 𝑗𝑗𝑐𝑐

𝐴𝐴E𝑖𝑖 = �� 𝛽𝛽𝑗𝑗 ∅𝑖𝑖𝑖𝑖 𝑆𝑆a �𝑇𝑇𝑗𝑗 , 𝜁𝜁𝑗𝑗 � ∙ 𝜌𝜌𝑗𝑗𝑗𝑗 ∙ 𝛽𝛽𝑘𝑘 ∅𝑖𝑖𝑖𝑖 𝑆𝑆a (𝑇𝑇𝑘𝑘 , 𝜁𝜁𝑘𝑘 ) (7.2.13)
𝑗𝑗=1 𝑘𝑘=1

where 𝜌𝜌𝑗𝑗𝑗𝑗 is the correlation factor for the j-th and k-th natural modes given as follows:

8�ζ𝑗𝑗 ζ𝑘𝑘 �ζ𝑗𝑗 − 𝑟𝑟𝑗𝑗𝑗𝑗 ζ𝑘𝑘 � 𝑟𝑟𝑗𝑗𝑗𝑗 3⁄2

𝜌𝜌𝑗𝑗𝑗𝑗 = (7.2.14)
2
�1 − 𝑟𝑟𝑗𝑗𝑗𝑗 � + 4ζ𝑗𝑗 ζ𝑘𝑘 𝑟𝑟𝑗𝑗𝑗𝑗 �1 − 𝑟𝑟𝑗𝑗𝑗𝑗 2� + 4 �ζ𝑗𝑗 2 − ζ𝑘𝑘 2 � 𝑟𝑟𝑗𝑗𝑗𝑗 2
where 𝑟𝑟𝑗𝑗𝑗𝑗 is the natural frequency ratio of the j-th and k-th modes (𝑟𝑟𝑗𝑗𝑗𝑗 = 𝜔𝜔𝑗𝑗 ⁄𝜔𝜔𝑘𝑘 = 𝑇𝑇𝑘𝑘 ⁄𝑇𝑇𝑗𝑗 ), and ζ𝑗𝑗 and ζ𝑘𝑘
are damping factors for the j-th and k-th modes, respectively. This method is called the complete quadratic
combination (CQC) method.

5) Load combination of seismic loads for two orthogonal horizontal components16)

The recommendations assume that seismic loads are estimated for each structural axis independently. It
is sometimes desirable to consider a combination of seismic loads for two orthogonal horizontal compo-
nents, such as an axial load on the corner columns and the torsional moment of the building with
two-directional eccentricity. Let two horizontal components of the seismic load be 𝐸𝐸x and 𝐸𝐸y according
to the orthogonal axes x-y. It is recommended to use the following quadratic combination to obtain the
combined seismic load 𝐸𝐸 as follows:

𝐸𝐸 = �𝐸𝐸x 2 + 2𝜀𝜀𝐸𝐸𝑥𝑥 𝐸𝐸y + 𝐸𝐸y 2 (7.2.15)

While factor 𝜀𝜀 can be from −1 to 1 (𝜀𝜀 = 0 indicates the SRSS method), it is empirically taken as 0 to 0.3.
The maximum of Eq. (7.2.16) may be used instead:
𝐸𝐸 = 𝐸𝐸x + 𝜆𝜆𝐸𝐸y
(7.2.16)
𝐸𝐸 = 𝜆𝜆𝐸𝐸x + 𝐸𝐸y
The value of 𝜆𝜆 may be taken as 0.3 to 0.5.

7.2.2 Natural period and damping considering soil-structure interaction

1) Eigenvalue analysis considering soil-structure interaction
When a building is located on firm ground, it is appropriate to assume that the foundation is also fixed on
- C7-16 - Recommendations for Loads on Buildings

firm ground. However, when this assumption does not hold, it is necessary to apply an analytical model
considering a deformation of the soil. This model is called an sway-rocking model (SR model) considering
the horizontal motion (sway) and rotational motion about the horizontal axis (rocking) from soil defor-
mation. Either a complex eigenvalue analysis or an eigenvalue analysis for an undamped model can be ap-
plied to estimate the natural periods and associated damping factors for this SR model.

2) Natural period and damping factor estimated using equivalent model

The equivalent single degree of freedom model (SDOF) of the MDOF can be used by considering the
sway and rocking motions, called an equivalent SR model (Fig.7.2.1) 17). Let the mass, stiffness, damping
factor, and height of the equivalent SDOF model be 𝑚𝑚𝑒𝑒 , 𝑘𝑘𝑒𝑒 , 𝑐𝑐𝑒𝑒 and ℎ𝑒𝑒 respectively. Then, the total
stiffness k t is estimated as follows:
2
1 1 1 he
= + + (7.2.17)
kt ke k s k r

where 𝑘𝑘𝑠𝑠 and 𝑘𝑘𝑠𝑠 are the stiffness of the sway spring and rocking spring, respectively. Multiplying me
and using T = 2π m / k gives the following relation:

Tt2 = Te2 + Ts2 + Tr2 (7.2.18)

me m e he2
where Ts = 2π 、 Tr = 2π .
ks kr

Equation (7.2) is derived based on the fact that the natural period Te for the equivalent SDOF model is
identical to that of a building fixed to the ground T f . It is possible to apply Eq. (7.2) in either case when
T f is given by an appropriate formula without an eigenvalue analysis, or is obtained from an eigenvalue
analysis.
To obtain the damping factor of the equivalent SR model, let us consider the strain and absorbed energy
E and ∆E during a cycle of vibration during period T1 . The damping factor is expressed as follows10):

Figure 7.2.1 Equivalent SR model

CHAPTER 7 SEISMIC LOADS - C7-17 -

1 ∆E
ζ= (7.2.19)
4π E
The total displacement is as follows:
x t = x e + x s + x r = x e + x s + θhe (7.2.20)
The strain energy is as follows:

E=
1
2
( ) 1
k e x e2 + k s x s2 + k r θ 2 = k t xt2
2
(7.2.21)

The absorbed energy is as follows:

(
∆E = πω1 c e x e2 + c s x s2 + c r θ 2 ) , ω1 = 2π / T1 (7.2.22)

where 𝑥𝑥𝑒𝑒 , 𝑥𝑥𝑠𝑠 , and 𝑥𝑥𝑟𝑟 are the relative displacement of a building owing to building deformation, sway,
and rocking motions, respectively. Substituting them into Eq. (7.2.19), the total damping factor ζ t be-
comes the following:

πω1 (c e x e2 + c s x s2 + c r θ 2 )
ζt = (7.2.23)
2πk t x t2

Using the following relationship,

ce cs cr
ζe = ζs = ζr = (7.2.24)
2 me k e 2 me k s 2 m e he2 k r

2 2 2
x e k t  Te  x s k t  Ts  x r he2 k t  Tr 
= =  = =  = =   (7.2.25)
x t k e  T1  x t k s  T1  xt kr  T1 

the damping factor ζ1 for the first mode in the SR model is calculated through Eq. (7.3).

3) Stiffness and damping for sway and rocking

For flat, embedded, and pile foundations, the stiffness K and damping coefficient C for sway and
rocking motions are obtained from the relationship between the force and displacement at the base of the
iωt iωt
superstructure18). When a foundation is subjected to a force, P = Po e , the displacement U = U o e is
obtained, and the complex impedance function K d (ω) is given as follows
:
P
K d (ω) = = K (ω) + iK ' (ω) (7.2.26)
U

where K (ω) and K ' (ω) are the real and imaginary parts of an impedance function, respectively.
The impedance function of a foundation in multilayered soil with rigidity, and damping becomes an
irregular function in terms of frequency, as indicated in the dotted lines in Fig.7.2.2. When the effects of
sway and rocking are expressed by a spring with stiffness k and a dashpot with damping coefficient c as in
Fig.7.2.1, the stiffness and damping coefficient are obtained as follows:
- C7-18 - Recommendations for Loads on Buildings

K ' (ω1 )
k = K (0) c= (7.2.27)
ω1

Equation (7.2.27) indicates that k is assumed to be identical to the static stiffness and c is determined such
that the damping force is consistent during the fundamental period, as shown in Fig.7.2.2.
The impedance function (or the stiffness and damping) of the sway and rocking motions are evaluated by
applying the method based on a) the wave propagation theory, b) finite element method, or c) discrete
spring type method18). The characteristics of the method a) is that the wave propagation in semi-infinite soil
is evaluated rigorously, and those of the method b) is that the foundation shape is modeled in detail, while
those of the method c) is practical.

Figure 7.2.2 Impedance in multi-layered soil (dynamic stiffness)

7.2.3 Methods for calculating the equivalent weight

(1) Procedure using eigenvalue analysis

When the maximum response considering all vibration modes is evaluated from the maximum response
of each mode, the sum of the maximum response of each mode gives an overestimate because the maxi-
mum response for each mode does not always occur simultaneously. Therefore, the maximum response
considering all vibration modes is evaluated as the square root of sum of squares (SRSS) of the maximum
response for each mode. This is an underlying theory of Eq. (7.4). Because the fundamental mode is the
most influential and the higher mode is less influential in the above equation, the sum need not be from the
fundamental mode to the highest mode. The recommendations propose applying the sum from the funda-
mental mode to the 𝑗𝑗𝑐𝑐 -th mode, where 𝑗𝑗𝑐𝑐 can usually be 3 at most.
For buildings with fewer irregularities in their distributions of stiffness and mass, the maximum response
considering all vibration modes can be evaluated using the SRSS. However, when torsional vibration can-
CHAPTER 7 SEISMIC LOADS - C7-19 -

not be neglected for buildings with larger eccentricities, the eigenvalues for adjacent modes are close to
each other. In such case, Eq. (7.6) should be used instead of Eq. (7.1). This equation is called the complete
quadratic combination (CQC). In this case, 𝑗𝑗𝑐𝑐 should be sufficiently large to include all influential modes.

(2) Procedure without eigenvalue analysis

The values of factors 𝑘𝑘1 and 𝑘𝑘2 depend on the characteristics of the building and ground motion19), 20).
It must be noted that Eq. (7.9) does not provide a good approximation of the story shear for buildings with
an abrupt change in mass and stiffness in elevation, such as a tower-like structure located on a lower build-
ing. Equation (7.9) is applicable for a case in which the weight of the lower building is half that of the tow-
er-like structure or less. Using 𝑘𝑘1 = 𝑘𝑘2 = 2𝑇𝑇1⁄(1 + 3𝑇𝑇1 ) in Eq. (7.9), the so-called Ai distribution of the
Building Standard Law of Japan is given as follows21):
1 2𝑇𝑇1
𝑘𝑘V𝑖𝑖 = 𝐴𝐴𝑖𝑖 = 1 + � − 𝛼𝛼𝑖𝑖 � (7.2.28)
�𝛼𝛼𝑖𝑖 1 + 3𝑇𝑇1

7.2.4 Acceleration response spectrum

(1) Acceleration response spectrum at the base of the foundation (or at the ground level)
1) Basis of determination of acceleration response spectrum
This section illustrates how to calculate the acceleration response spectrum of input ground motion to the
structural model (an SSI model and a fixed-base model), described in Sec. 7.2.1.
Determination of the spectral acceleration requires an adequate consideration of ground properties af-
fecting the earthquake motions, vibrational properties of the building, and other characteristics. Figures
7.2.3 and 7.2.4 show a schematic relationship of these factors. Seismic loads are determined such that these
factors are taken into consideration as realistically as possible. To do so, available information and a data-
base related to the site characteristics and building must be fully utilized. The analysis method should be
carefully selected depending on the importance of the building as well as the amount and quality of the
available information.
As shown in Table 7.2.1, the spectral acceleration of an input ground motion to a building is estimated by
combining the modification functions as a function of the frequency representing effects of the spectral
characteristics of ground motion in engineering bedrock, the amplification of seismic waves within the soil
above the engineering bedrock, and the soil-structure interaction (effects of the size and shape of the foun-
dation and embedment on the input ground motion). Technical guidebooks and recommendations published
by the Architectural Institute of Japan 3), 18) can be used to estimate these modification functions. Engineer-
ing bedrock is defined as a layer located deeper than the foundation and that has sufficient stiffness and
thickness. Conventionally, in Japan, engineering bedrock is defined as a layer with a shear wave velocity of
400 m/s or larger.
The standard procedure used to determine the acceleration response spectrum is recommended. However,
- C7-20 - Recommendations for Loads on Buildings

when the seismic load on a building is estimated approximately, or a detailed procedure cannot be used be-
cause the amount of data for the standard procedure is insufficient, a simplified procedure (see 7.2.4(3) and
the right-side column of Table 7.2.1) can also be used. Note that the accuracy of the simplified procedure is
not as good as the detailed version.
An estimation of the spectral acceleration should be based on the site-specific characteristics of the
ground motion. A transfer function representing the wave amplification 𝐻𝐻GS (𝜔𝜔) of Eq. (7.14) represents
the site-specific characteristics of the wave amplification. Note that characteristics of the wave amplifica-
tion change depending on the degree of nonlinearity of the soil. Uncertainty in the modeling of these char-
acteristics should be taken into consideration. Wave amplification from the seismic bedrock (a layer with a
shear wave velocity of about 3 km/s or larger) to the engineering bedrock can be important when a building
with a fundamental natural period of 1 s or longer is located on deep sedimentary layers such as the Kanto
Plane, the Nobi Plane, or the Osaka Plane22-24). The spectral acceleration shown in the recommendations
(see 7.2.4(2)-1)) do not include these effects and should be discussed for future updates. In such a case, the
method specified in 7.2.4 (2)-2) can be supplementary used in combination as a conservative estimate.

Soil Amplification
（Eng. Bedrock to Surface）
Eng. Bedrock
Soil Amplification
（Seismic Bedrock to Eng. Bedrock）
Seismic
Bedrock
Propagation path
characteristics

Fault（Source characteristics）

Figure 7.2.3 Factors affecting seismic load

Seismic Load Effect

Soil-Structure Building Vibration

= Source
Characteristics
× Propagation Path
Characteristics
× Amplification
due to Soil
× Interaction
× Characteristics

Seismic hazard map Soil amplification Function for Vibration mode,

function soil-structure natural period,
interaction damping ratio

Figure 7.2.4 Factors affecting seismic load considered in the recommendations

Figure 7.2.5 illustrates the composition of the seismic load and relationship among the factors and pro-
cedures used to obtain the seismic load on a building. The recommendations recommend utilizing a method
of equivalent static load using a transformation between the response spectra and power spectra. However,
CHAPTER 7 SEISMIC LOADS - C7-21 -

an appropriate method can be selected depending on the applicability of each method as well as on the ca-
pability of the user.

Table 7.2.1 Factors affecting the spectral acceleration for an input ground motion to the structural model
(an SSI model and a fixed-base model)
Standard procedure
Detailed procedure Simplified procedure
Modification (recommended)
factor Required Required Required
method method method
information information information
Approxima- Size and shape Size and shape Approxima- Size and
SSI tion using Eq. of foundation of foundation tion using Eq.
shape of
(effective HSSI(ω) (7.18) and embed- and embed- (7.18) foundation
input motion) ment ment and embed-
ment
2D or 3D
1D shear Layer compo- 1D shear wave Solutions to Predominant
wave propa-
wave propa- sition shear propagation Eqs. (7.16) period,
gation anal-
gation analy- wave velocity, analysis con- and (7.17) for damping fac-
yses assum-
sis consider- strain depend- sidering strain shear wave tor and im-
ing rigid
Wave ampli- ing strain ency of stiff- dependency of propagation pedance ratio
HGS(ω) foundation
fication dependency ness and soil stiffness through of soil
of soil stiff- damping of and damping two-layered
ness and layers above factor model
damping fac- the engineering
tor bedrock

Equivalent seismic load Dynamic load [7.3]

Equivalent seismic Response time

load Vei [7.2.1] history
Structural
response SRSS／CQC Response analysis of
characteristics [7.2.3] building and foundation
Power spectrum Acc. response Sa
Ga [7.2.4(1)] [7.2.4(1)]
Soil-structure
Function for SＳＩ
interaction [7.2.4(4)]
characteristics
Power spectrum Acc. response Time history

Soil
Soil amplification Response analysis for soil
amplification function [7.2.4(3)]
characteristics
Acc. response
Power spectrum Time history
SaE [7.2.4(5)]
GaE [7.2.4(1)]
Simulation of ground
Source motions for design
characteristics
Seismic hazard map
Scenario earthquakes
[7.2.4(5)]

Figure 7.2.5 Composition of seismic load and factors affecting response spectra of each model ([ ]
indicates where each component is described in the recommendations)
- C7-22 - Recommendations for Loads on Buildings

2) Power spectral representation of ground motion

A Fourier analysis dictates that a time-varying process can be expressed in terms of the superposition of
various sinusoidal waves with different amplitudes. On the other hand, in the field of earthquake engineer-
ing, a response spectrum is often used to represent the ground motion properties. The response spectrum
represents the frequency characteristics of ground motions through the maximum elastic response of a sin-
gle-degree-of-freedom system with a specified period and damping factor, and its representation has been
adopted for engineering purposes. However, it has a shortcoming in that it cannot be decomposed into sev-
eral factors affecting the maximum response of an SDOF, as shown in Fig.7.2.4.
On the other hand, the power spectra (power spectral density or power spectral density function) does
express the time-variant characteristics of the square amplitude of the mean squared time-varying process
along the frequency range, which is mathematically easy to handle in the form of transfer functions within
the frequency range.
As shown in Fig.7.2.4, the input ground motion of a building can be estimated by multiplying the power
spectrum of the engineering bedrock with a wave amplification function and soil-structure interaction func-
tion. Furthermore, a power spectrum of the input ground motion can be directly used for a dynamic re-
sponse analysis within the frequency domain.
For a response analysis of a building, the ground motion as an input of the building is specified in the
form of the response spectra, which familiarly is also adopted in the recommendations (see Eq. (7.1)).
Therefore, it is necessary to transform the power spectrum into the corresponding response spectrum, and
the transformed response spectrum can be used in the various typical ways.

3) Transformation between response and power spectra

Figure 7.2.5 illustrates the composition of the seismic load and the relationship among the various fac-
tors. The input ground motion of a building, along with the ground motion on the engineering bedrock, is
characterized in the form of the response spectra. Therefore, a transformation from the response spectra to
the power spectra, and vice versa, is required.
When a time history of the ground motion is available, it is straightforward to obtain the corresponding
response and power spectra. When estimating the response spectra from the power spectra, the maximum
response of a single-degree-of-freedom system can be stochastically estimated with sufficient accuracy by
assuming the main part of the ground motion to be a stationary process with a finite duration. Several
methods are available for the transformation. A method for a spectral transformation based on the first ex-
cursion theory of a stationary random process is used in the recommendations.
The transformation from power spectra to response spectra is called “a forward transformation” hereaf-
ter, and the reverse is called “a backward transformation.”
The probability density function of the maximum response of a single-degree-of-freedom system can be
estimated from the power spectra and duration of ground motion.
Here,𝑘𝑘p (𝑇𝑇, ζ) is called the peak factor of the acceleration response of a single-degree-of-freedom sys-
tem, where it is defined as a ratio of the acceleration response spectrum 𝑆𝑆a (𝑇𝑇, ζ) to the standard deviation
CHAPTER 7 SEISMIC LOADS - C7-23 -

of the acceleration response of a single-degree-of-freedom system, 𝜎𝜎a (𝑇𝑇, ζ). The expected value of peak
factor 𝑘𝑘p (𝑇𝑇, ζ) is a function of the duration and non-exceedance probability. The peak factor 𝑘𝑘p (𝑇𝑇, ζ)
can be assumed to be 3.0 for ordinary ground motions.
The details of the transformation between the response and power spectra are described in the “Guide-
book of Recommendations for Loads on Building” and elsewhere 25).

(2) Acceleration response spectrum of the engineering bedrock

The acceleration response spectrum of the engineering bedrock in a return period of r years can be
evaluated through a uniform hazard spectrum (UHS), or the spectrum evaluated from a normalized acceler-
ation response spectrum. UHS can also be determined from the results of seismic hazard analysis released
from a public agency such as the National Research Institute for Earth Science and Disaster Prevention
(NIED) and The Earthquake Research Committee of the Headquarters for Earthquake Research Promotion.
The UHS released from the NIED is shown in 7.2.4(5).
The UHS should be set with reference to the latest result because seismic hazard results are updated pe-
riodically by the latest information regarding the seismicity and seismic hazard analysis method. For exist-
ing buildings designed through the recommendations, it is necessary to refer to the latest results of hazard
analysis and re-evaluate the seismic load as necessary. When using the values in the recommendations, it is
necessary to consider that they can be updated in the future.

(3) Soil amplification function

The soil amplification function 𝐻𝐻GS (𝜔𝜔) used to express the amplification effect of seismic motion
through surface soil deposits is assumed to be a transfer function, which is the theoretical solution of a
one-dimensional shear wave propagation analysis (equivalent linear analysis)26). The transfer function
is a frequency-dependent ratio of the free field motion at the ground surface or the base of the building
foundation to the exposed surface motion in the engineering bedrock. Note that, because the transfer
function is a complex function, the square of its absolute value is the definition of the soil amplifica-
tion function. The equivalent linear analysis is based on the incident time history waveform compati-
ble with the above-mentioned acceleration response spectrum of the engineering bedrock and on the
soil data taking the nonlinearity into consideration based on the strain-dependent soil properties of the
shear modulus and the damping factor. Note that the engineering bedrock should be set to a level
deeper than the upper level of the bearing stratum with sufficient stiffness and thickness, and its shear
wave velocity might be expected to be equal to or greater than 400 m/s.
Alternatively, one can substitute the simplified formula instead of the above-mentioned transfer
function considering the soil nonlinearity, if it is difficult to implement an equivalent linear analysis, or
if the soil amplification can be obtained simply, because the soil type of the site can be easily deter-
mined without much information. The simplified formula, as shown below, takes into account not the
soil nonlinearity but the hysteretic damping, and is an algebraic expression corresponding to the solu-
tion to the soil model with two layers above and below the internal surface of the soil deposits and en-
- C7-24 - Recommendations for Loads on Buildings

gineering bedrock. In a one-dimensional shear wave propagation analysis, the seismic motion verti-
cally propagating in horizontally layered soil deposits is expressed as a weighted sum of upward and
downward waves with harmonic oscillations at a point. Upward waves are equal to downward waves
in amplitude at a free field or an exposed ground surface. Then, for a two-layer soil model, a simpli-
fied formula of the transfer function is derived and given by the following27, 28):

1
𝐻𝐻GS (𝜔𝜔) = , (7.2.29)
cos 𝐴𝐴 + i𝛼𝛼G sin 𝐴𝐴

𝜔𝜔𝑇𝑇G
𝐴𝐴 = , (7.2.30)
4�1 + 2𝑖𝑖𝜁𝜁G

where 𝛼𝛼G represents the impedance ratio from the engineering bedrock to the surface soil, 𝑇𝑇G and
𝜁𝜁G represent respectively the predominant period in seconds and the damping factor of soil above the
engineering bedrock, and 𝑖𝑖 = √−1. As far as this simplified formula is concerned, as a reference val-
ue, Table 7.2.2 shows the soil-type classification and the parameters of 𝑇𝑇G , 𝜁𝜁G , and 𝛼𝛼G for each soil
type, based on the previous 2004 version of the Recommendations for Loads on Buildings. However,
it is desirable to readily use the site-specific value of the parameters instead of the reference value
shown in the table, if the site exploration data are available for the parameter setting. Figure 7.2.7
shows the absolute value of the transfer function corresponding to each soil type. Note that, in this
figure, the abscissa does not have a circular frequency 𝜔𝜔(rad/s) but period 𝑇𝑇(𝑠𝑠)[= 2𝜋𝜋⁄𝜔𝜔].
On the other hand, if it is necessary to consider the larger ground deformation owing to the liquefac-
tion of the ground and the resulting lateral flow of the liquefied ground, and it is expected that the am-
plification effect of the seismic motion is beyond the coverage of the equivalent linear analysis, then
measures against the above larger ground deformation need to be taken by using a more appropriate
analysis that be applicable to the liquefaction and lateral flow. The ground amplification function
𝐻𝐻𝐺𝐺𝐺𝐺 (𝜔𝜔) can then be set to the frequency transfer function obtained through the analysis used to take
the countermeasures.

Table 7.2.2 Soil types and parameters for simplified formula

Soil Type 𝑇𝑇G (s) 𝜁𝜁G 𝛼𝛼G
Type I : Hard (Standard) 0.22 0.05 0.50
Type II : Loose Diluvial or Dense Alluvial 0.37 0.05 0.35
Type III : Soft 0.56 0.05 0.20
CHAPTER 7 SEISMIC LOADS - C7-25 -

5
Soil Amplification Function
Type I

Type II

Type III

0.5
0.1 1 5
T (s)

Figure 7.2.7 Absolute value of the transfer function for each soil type through the simplified formula

Although the above equivalent linear analysis is based on a method modeling of the ground as a
one-phase system of saturated soil (total stress analysis method), there is a better method that treats the
ground as a two-phase system of the pore fluid and soil skeleton (an effective stress analysis method),
where the total stress is decomposed into the pore water pressure and effective stress. Recently, an ef-
fective stress analysis has been used as a more appropriate analysis method 31). Therefore, when con-
sidering the larger ground deformation owing to liquefaction and the resulting lateral flow, it is desira-
ble to take measures for the ground through an effective stress analysis using the constitutive law de-
scribing the cyclic soil stress-strain relationship. After the above countermeasures are taken, the
ground amplification function 𝐻𝐻𝐺𝐺𝐺𝐺 (𝜔𝜔) may be set to the frequency transfer function obtained from
the results of the effective stress analysis.

(4) Adjustment function of soil-structure interaction

The behavior of an embedded foundation during an earthquake is different in general than the behavior
of a free field ground without a foundation. The behavior of a foundation under a massless assumption is
one of the effects of a dynamic soil-structure interaction, which is called foundation input motion. For an
evaluation of the seismic load, the effect can be considered an adjustment factor for an acceleration re-
sponse spectrum. The input motion is different from the free field motion because the free field motion is
not constant along the depth, and the displacement of soil is restricted by the foundation, which is assumed
to be rigid. This concept is explained through Fig.7.2.8.
- C7-26 - Recommendations for Loads on Buildings

Figure 7.2.8 Schematic view of embedded foundation

When the free field motion is Ug(z), the resultant force Fc restricted by the foundation becomes

Fc = K bsU g (d ) + k ws (z )U g ( z )dz ,
d
∫ 0
(7.2.29)

where Kbs and kws(z) are springs at the base and side wall, respectively, including the wave dissipation ef-
fect. The foundation input motion Ufh is obtained by releasing the force as follows:
Fc
U fh =
Kf
(7.2.30)

K f = K bs + ∫ k ws (z )dz
d
0

Assuming kws to be constant, the input motion is expressed as follows.

K bsU g (d ) + K wsU gw
U fh =
K bs + K ws
(7.2.31)
1 d
d ∫0
K ws = dk ws , U gw = U g ( z )dz

From Eq. (7.2.33) the input motion is regarded as motion averaging Ug(z) with a weighting soil spring.
Because Ug(z) is less than UGL, Ufh becomes less than UGL. The input motion during a long period is
equivalent with the free field motion because the wavelength is long.
A soil structure interaction effect was investigated using the records observed during the 1995 Hyogoken
Nambu Earthquake32). The relationship between maximum acceleration and velocity among the records in a
free field and at the foundation is shown in Fig.7.2.9. The white circles are the pure results recorded, and
the free field motion shown with the black circles is inversely evaluated from the analyses. The factor η in
buildings A and B are around 0.3 and 0.5, respectively. The average ratios of Ufh to UGL obtained from the
least squares method were 0.7 for acceleration and 0.9 for velocity. These results show that, compared to
UGL, within a long period, Ufh is slightly smaller, and the difference within a short period becomes smaller.
An analysis was conducted to investigate the characteristics of the input motions on the frequency33).
There are several analytical methods33, 34). The method applied is the thin layer method35). Parameter η is
assumed to be 1/4 and 1/2. The surrounding soil has Vs = Vsb and 0.5Vsb. The amplitude ratios of Ufh to UGL
are shown in Fig.7.2.10, where the circular frequency ω of the horizontal axis is normalized through the
frequency ω d (= πV s / 2 / d ) , which is expressed as δ d . The rotational motion caused by the phase differ-
CHAPTER 7 SEISMIC LOADS - C7-27 -

ence of free field motion along the depth is shown in Fig.7.2.11. The trough and peak of around 1 and 2 of
δ d are influenced mainly by the free field motion at the bottom Ug(d).

From the observed results in Fig.7.2.9 and the analytical results in Fig.7.2.10, the adjustment factor
H SSI (ω) is defined as follows:

 1 
2  : δ d ≤ 1
2
U fh 1 + 2ηδ d 
H SSI (ω) =
2
=  (7.2.32)
U GL  1 : δ d > 1
1 + 2η 

( )
where η = d / l ， δ d = ω / ω d , ω d = (πV s ) / (2d ) ，l is the equivalent foundation width = A f , d is the

embedment depth of the foundation, and Vs is the shear wave velocity of the soil adjacent to the sidewall.
Examples of H SSI (ω) are shown in Fig.7.2.12. The rotational input motion should be considered when
it is significant.

(a) Peak acceleration (b) Peak velocity

Figure 7.2.9 Iteration effects of observed records during Hyogoken Nambu earthquake32)
- C7-28 - Recommendations for Loads on Buildings

Figure 7.2.10 Horizontal foundation input motion Figure 7.2.11 Rotational foundation input motion

Figure 7.2.12 Adjustment function of embedded foundation

(5) Basic acceleration response spectrum

The basic acceleration response spectrum is given through a seismic hazard analysis, which is open to
the public. Uniform hazard spectra of major cities provided by the National Research Institute for Earth-
quake Science and Disaster Resilience36) are shown in Fig.7.2.13 as an example of public information.
Uniform hazard spectra are the acceleration response spectra of 5% damping (given at 17 natural periods
from 0.05 to 5.0 s) on the reference ground (Vs = 292 m/s) corresponding to an exceedance probability of
39%, 10%, 5%, and 2% in 50 years as of January 2007. In the recommendations, such spectra are regarded
as 100-, 500-, 1,000-, and 2,500-year return period response spectra. Standard seismic hazard maps and
500-year return period (10% exceedance probability during a 50-year period) acceleration response spectra
are shown in Figs. 7.2.14 and 7.2.15, respectively.
CHAPTER 7 SEISMIC LOADS - C7-29 -

1) Seismic hazard and the analytical method

The basic parameter used to determine the seismic load of a building is the strength of the strong ground
motions at the construction site during the lifetime of the building. It is more appropriate to estimate the
maximum ground motion stochastically than to determine it deterministically because the strength of the
strong ground motions in the future depends on various factors with large randomness, such as the seismic
source characteristic, the wave path characteristic of the seismic waves, and the site soil characteristics.
The seismic hazard (seismic risk) represents stochastically the strength of the seismic motion on the
ground or the seismic bedrock caused by future earthquakes, and the peak ground acceleration or peak
ground velocity with a 100-year return period is used as a measure of the seismic hazard.
One of initial studies on the seismic hazard is the “Kawasumi hazard map” (1951), and some analytical
methods have been developed using different models of seismic sources and data based on this research17).

20 Response acceleration (m/s2) 80

Response acceleration (m/s2)

h=5% h=5%
15 60

10 40

2500years
52500years 20
1000years
1000years 500years
500years 100years
100years
0 0
0.05 0.1 0.5 1.0 5.0 0.05 0.1 0.5 1.0 5.0
period(s) period(s)
(a) Sapporo (b) Kushiro
20 20
Response acceleration (m/s2)

Response acceleration (m/s2)

h=5% h=5%
15 15

10 10
2500years 2500years
1000years 1000years
500years
5 5 500years
100years 100years

0 0
0.05 0.1 0.5 1.0 5.0 0.05 0.1 0.5 1.0 5.0
period(s) period(s)
(c) Sendai (d) Tokyo
Figure 7.2.13 Uniform hazard spectrum Sa0(T, 0.05) and SaE(T, 0.05) in ten cities
- C7-30 - Recommendations for Loads on Buildings

80 40

Response acceleration (m/s2)

h=5% h=5%
60 30

40 20
2500years 2500years
1000years 1000years
20500years 10
500years
100years 100years

0 0
0.05 0.1 0.5 1.0 5.0 0.05 0.1 0.5 1.0 5.0
period(s) period(s)
(e) Shizuoka (f) Nagoya
20 40
Response acceleration (m/s2)
Response acceleration (m/s2)

h=5% h=5%
15 30

10 20

2500years 2500years
51000years 10
1000years
500years
500years
100years
100years
0 0
0.05 0.1 0.5 1.0 5.0 0.05 0.1 0.5 1.0 5.0
period(s) period(s)
(g) Kanazawa (h) Osaka
20 20
Response acceleration (m/s2)

Response acceleration (m/s2)

h=5% h=5%
15 15

10 10
2500years

51000years
500years
52500years
100years 1000years
500years
100years
0 0
0.05 0.1 0.5 1.0 5.0 0.05 0.1 0.5 1.0 5.0
period(s) period(s)
(i) Hiroshima (j) Fukuoka
Figure 7.2.13 Uniform hazard spectrum Sa0(T, 0.05) and SaE(T, 0.05) in ten cities (cont.)
CHAPTER 7 SEISMIC LOADS - C7-31 -

Figure 7.2.14 Seismic hazard map: basic acceleration a0 (m/s2)

(peak horizontal acceleration at the surface of the
engineering bedrock with 100-year return period)
Note) Thick contour lines are drawn every 2 m/s2 and a thin curve every 1 m/s2.
- C7-32 - Recommendations for Loads on Buildings

Figure 7.2.15 Seismic hazard map: acceleration a500 (m/s2)

(peak horizontal acceleration at the surface of the
engineering bedrock with a 500-year return period)
Note) Thick contour lines are drawn every 4 m/s2 and a thin curve every 2 m/s2.

a) Seismic hazard analysis using earthquake catalogue

The past seismic data (occurrence date and time, location, and magnitude) in Japan have been collected
in an earthquake catalogue (database) for more than 1,000 years. The recent data within 100 years are based
on seismographic records, but the other data were estimated through earthquake damage described in vari-
ous documents.
The strength of earthquake motion at a certain point can be estimated using seismic data from the earth-
quake catalogue and the appropriate attenuation relationship (distance attenuation curve). It is preferable to
consider the randomness of the attenuation relationship because it is an empirical relationship based on
CHAPTER 7 SEISMIC LOADS - C7-33 -

numerous observation records, and is composed of only a few factors such as the magnitude and epicentral
or hypocentral distance. Although the seismic hazard can be estimated from the distribution function ap-
plied to the sample distribution of earthquake motions, the applied distribution function is not sensitive to
large values of the sample distribution.
A more reasonable and improved method is to use an extreme value distribution composed of the annual
maximum values. The extreme value distributions that are commonly used are the Gumbel distribution,
Fréchet distribution, and Weibull distribution37), which satisfy the stability condition in which the shape of
the distribution does not change with the size of the group or the distribution proposed by Kanda41), which
has the minimum and maximum limits.
To use the extreme value distribution, it is necessary to extend the period of the earthquake catalogue
long enough to reflect the influence of a large earthquake within a short distance because the recurrence
period of a large earthquake is several thousand years or more. However, setting the long period of the
earthquake catalogue is usually difficult because the reliability of old earthquake data is low.
These results are significantly different regardless of whether a large earthquake is contained in the
earthquake catalogue. This problem can be avoided by using a seismotectonic zone, which has the same
characteristic of earthquake occurrence. This analytical method (b value model) is based on research by
Cornell39) and uses the magnitude distribution, which is basically the exponential distribution based on the
Gutenberg and Richter law, as well as the occurrence model of large earthquakes, which is generally as-
sumed to be a Poisson process. The seismic hazard is estimated through a probability calculation using data
on the maximum magnitude, b value, average earthquake occurrence rate, and an appropriate attenuation
relationship. This method can easily take into account the randomness of the distance attenuation relation-
ship.

b) Seismic hazard analysis using earthquake data from the plate boundary and active fault
It is well known that large earthquakes occur repeatedly in the plate boundary or active fault regions. The
characteristic occurrence of earthquakes in such regions shows that large earthquakes occur repeatedly
during almost the same recurrence period. Such characteristic suggests that the earthquake model with the
same scale earthquakes is more appropriate than the b value model with different scale earthquakes. Re-
cently, a seismic hazard analysis using such earthquake model was improved as data on the position of the
active fault, shape, magnitude of the earthquake, and recurrence interval have accumulated.
The problem with this method is that the reliability of the data from the plate boundary regions or the ac-
tive fault regions in Japan are not the same. Moreover, the reliability of the average displacement rate is
low because the recurrence interval of large earthquakes is several thousand years or more.

c) Seismic hazard analysis in the current recommendations

In the current recommendations, probabilistic seismic hazard maps using a uniform hazard spectrum
and peak acceleration as measures of ground motions, shown in Appendix 2 of the Technical Notes of the
National Research Institute for Earth Science and Disaster Prevention No. 336, are used. In the maps, the
- C7-34 - Recommendations for Loads on Buildings

attenuation relationship by Kanno et al. (2016) was used for 1 km mesh zones in Japan, and acceleration
response spectra of 5% damping (given at 17 natural periods of 0.05 to 5.0 s) and peak accelerations on the
reference ground (Vs = 292 m/s), corresponding to the exceedance probability of 39%, 10%, 5%, and 2%
in 50 years as of January 2007, are calculated. The seismic activity model follows that of the National
Seismic Hazard Maps for Japan in 200741).

2) Conditional mean spectrum

Because a uniform hazard spectrum gives spectral values corresponding to an exceedance probability for
all natural periods, the shape of the spectrum is different from that of the observed ground motion when the
spectrum is used as the target. Baker42) proposed an evaluation method for the mean spectrum, which pro-
vides the spectral value corresponding to the exceedance probability for a certain natural period.
Let T* denote a period of interest, where the spectral value Sa(T*) is obtained from a seismic hazard map.
The period of interest T* is often equal to the first-mode period of the structure of interest. The mean values
of magnitude M, epicentral distance R, lnSa(T*), and standard deviation σ ln Sa (T *) are also obtained from
the seismic hazard map. The deviation from the mean value ε(T*) is given as follows:
ln Sa (T *) − µ ln Sa ( M , R, T *)
ε (T *) = . (7.2.33)
σ ln Sa (T *)

The mean value of the deviation at period Ti, given ε(T*) is obtained as follows:
µ ε (T ) ε (T *) = ρ (Ti , T *)ε (T *) .
i
(7.2.34)

The correlation coefficient between ε(Ti) and ε(T*), ρ (Ti , T *) , can be given as follows:

π  Tmin  Tmax 
ρ (Ti , T *) = 1 − cos −  0.359 + 0.163I (T ln  ln ,
0.189  Tmin 
min < 0.189 )
(7.2.35)
2 
where, Ti and T* are between 0.05 and 5 s. 𝐼𝐼(𝑇𝑇𝑚𝑚𝑚𝑚𝑚𝑚 <0.189) = 1 if Tmin<0.189 s, and is equal to zero other-
wise.
The mean value of lnSa(Ti), given lnSa(T*), is obtained as follows:

µ ln Sa (Ti ) ln Sa (T ) = µ ln Sa ( M , R, Ti ) + ρ (Ti , T )ε (T *)σ ln Sa (Ti ) . (7.2.36)

(6) Return period conversion factor

1) Return period conversion factor
The return period conversion factor krE(T) can be obtained using the basic acceleration response spectrum
S a 0 (T ,0.05) , and the acceleration response spectrum for a 500-year return period S a 500 (T ,0.05) if the
type of probabilistic distribution is assumed. When a random variable X follows the Fréchet distribution,
CDF FX(x) is given by
CHAPTER 7 SEISMIC LOADS - C7-35 -

ν
FX ( x) = exp{−( ) k } , (7.2.37)
x
where ν and k are parameters of the distribution. Taking the natural logarithm of both sides twice and ap-
plying a rearrangement provides

ln{− ln FX ( x)} = −k (ln x − lnν ) . (7.2.38)

When the probability of exceedance is 1 − F X ( x) << 1, and thus 1 − FX ( x) ≅ − ln FX ( x) , the equation

above can be approximated as

ln{1 − FX ( x)} = −k (ln x − ln v) . (7.2.39)

This equation means that the probability of exceedance 1 − F X ( x) can be approximated using a line at
the double-logarithmic scale. Because 1 − F X ( x) = 1 / r for return period r，the slope of line k is given by
the following equation using the acceleration response spectrum for a 100-year return period S a 0 (T ,0.05)
and 500-year return period S a 500 (T ,0.05) :

ln 500 − ln100
k (T ) = . (7.2.40)
ln S a 500 (T ,0.05) − ln S a 0 (T ,0.05)

Let us define the acceleration response spectrum for an r-year return period S ar (T ,0.05) , and thus the
conversion factor of the return period k rE (T ) is S ar (T ,0.05) / S a 0 (T ,0.05) . Replacing 500 with r and
S a 500 (T ,0.05) with S ar (T ,0.05) in the equation above, and applying a rearrangement, gives the follow-
ing equation.

k rE (T ) = (r / 100)1 / k (T ) . (7.2.41)

If the peak acceleration is used instead of the acceleration response spectrum, the conversion factor of
the return period can be obtained in the same manner. Because the conversion factor of return period
k rE (T ) described above is obtained through a linear approximation of the seismic hazard curve between
values corresponding to 100- and 500-year return periods, the farther the return period is from the range,
the larger the error becomes.
Because 1 − F X ( x) = 1 / r for return period r，the slope of line k is given using S a 1 (T ,0.05) of the
r1-year return period and S a 2 (T ,0.05) of the r2-year return period as follows.
ln r2 − ln r1
k (T ) = (7.2.42)
ln S a 2 (T ,0.05) − ln S a1 (T ,0.05)

Let us define the acceleration response spectrum for an r-year return period S ar (T ,0.05) , and thus the
conversion factor of return period k rE (T ) is S ar (T ,0.05) / S a1 (T ,0.05) . Replacing S a 2 (T ,0.05) with
S ar (T ,0.05) in the equation above and applying a rearrangement gives the following equation.

k rE (T ) = (r / r1 )1 / k (T ) (7.2.43)

If we obtain the conversion factor of the return period directly from a seismic hazard curve, such error
- C7-36 - Recommendations for Loads on Buildings

will decrease. Comparisons of a seismic hazard curve and a linear approximation for Tokyo and Osaka are
shown in Fig.7.2.16.

1.E+00 1.E+00

1.E-01 1.E-01
Probability of exceedance

Probability of exceedance
1.E-02 1.E-02

1.E-03 1.E-03
Hazard curve Hazard curve

Linear Linear
approximation approximation

1.E-04 1.E-04
0.1 1 10 0.1 1 10
PGA(m/s2) PGA(cm/s2)

a) Tokyo b) Osaka
Figure 7.2.16 Comparison between seismic hazard curve and linear approximation

2) Failure probability of structure

The seismic load effect S and resistance of structure R are both random variables and can be modeled
through a probabilistic distribution. Although the probability of failure of a structure can be evaluated using
these probabilistic distributions, the calculation is usually complicated. Cornell43) showed that the probabil-
ity of failure of a structure Pf is evaluated by following an explicit equation when the probabilistic distribu-
tion of resistance of the structure is lognormal and a seismic hazard curve can be approximated through a
line on a double logarithmic scale.
1
( kδ R ) 2
Pf ≈ H (aˆ R )e 2 (7.2.44)

where H(x) is the seismic hazard curve， â R is a median of structural resistance, and δ R is the
lognormal standard deviation of the structural resistance. To achieve the failure probability of a structure of
less than P0, â R should be as follows:
1 1
kδ R2 kδ R2
aˆ R ≥ H −1 ( P0 )e 2 = a p0e 2 (7.2.45)
CHAPTER 7 SEISMIC LOADS - C7-37 -

1 2
Table 7.2.3 Values of correction factor 𝑒𝑒 2𝑘𝑘𝛿𝛿𝑅𝑅
δR
k
0.0 0.2 0.4 0.6
1.0 1.00 1.02 1.08 1.20
2.0 1.00 1.04 1.17 1.43
3.0 1.00 1.06 1.27 1.72

where a p 0 is the peak ground acceleration corresponding to the failure probability P0 in the seismic haz-
1 2
ard curve. The value of 𝑒𝑒 2𝑘𝑘𝛿𝛿𝑅𝑅 is considered to be a correction factor for peak acceleration to determine the
1 2
design resistance of a structure. The values of 𝑒𝑒 2𝑘𝑘𝛿𝛿𝑅𝑅 are shown in Table 7.2.3 for k and δ R 44). The steeper
the slope of a seismic hazard curve becomes, and the larger the variation of the structural resistance, the
larger the correction factor.

(7) Comparison with previous recommendations (2004)

1) Basic peak acceleration in previous recommendations

Large earthquake occurrence zones assuming earthquake model with the same scale earthquakes and
background source zones assuming the b value model are used to obtain the basic peak acceleration also in
the previous recommendations (2004) based on the seismic source model compiled by Annaka45). The at-
tenuation relationship proposed by Annaka and Yashiro (2000)46) is used, which shows the predicted medi-
an of peak acceleration on the seismic bedrock in Japan.

2) Comparison of attenuation relations

Attenuation relations of the acceleration response spectrum and peak acceleration by Kanno et al. (2006)
are adopted in the current recommendations. The attenuation relations by Kanno et al. are defined on the
ground, where the average shear wave velocity from the ground to a 30 m depth is equivalent to 292 m/s.
Annaka, in contrast, used bedrock motions estimated through the one-dimensional wave propagation theory
for observation points whose surface ground shear wave velocity is lower than 300 m/s, or bedrock motions
as they are for other observation points. A comparison of the attenuation relations is shown in Fig.7.2.17.
The values obtained by dividing the PGA of Si and Midorikawa’s attenuation relation47) (inter-plate) by
ground amplification factor of 1.4 are also shown for comparison. The earthquake magnitude is 7.0 and the
focal depth is 30 km. In a comparison between attenuation relation by Kanno et al. and that by Annaka, the
former gives slightly larger values for a short distance and slightly smaller values for a long distance than
the latter, although the difference is small.

3) Comparison of peak accelerations in six cities

A comparison of the peak accelerations on the engineering bedrock in six cities between the current
recommendations and the previous recommendations (2004) is shown in Table 7.2.4. The values of 100 and
- C7-38 - Recommendations for Loads on Buildings

500 year return period of the previous recommendations are the values in Table 7.2.5 of the previous rec-
ommendations. Values of 1,000- and 2,500-year return periods of the current recommendations are also
shown for reference.
For the 100-year return period in Table 7.2.4, the values of the current recommendations are 1.2 to 2.2
times larger than those of the previous recommendations, and the values for Tokyo, Nagoya, and Osaka are
about twice of those of the previous recommendations. For the 500-year return period, the values of the
current recommendations for cities except Sapporo are larger than those of the previous recommendations,
and the ratios are within the range of 1.2 to 1.4, except for Nagoya, where the ratio is 1.8. The source mod-
els, the evaluation method of earthquake occurrence, and the data used for the evaluation differ between the
current and previous recommendations. A stochastic process based on the Brownian Passage Time model
for earthquake recurrence, along with a Poison process, was used for the earthquake occurrence in the cur-
rent recommendations, whereas a Poison process alone was used for the earthquake recurrence in the pre-
vious recommendations. The longer the time after the previous occurrence becomes, the larger the proba-
bility of the earthquake occurrence in the Brownian Passage Time model for earthquake recurrence. There
are cases in which the peak acceleration is larger than in the previous recommendations for regions close to
such seismic zones.

10
Annaka's Eq.

Kanno's Eq.

Si & Midorikawa's
Eq./1.4

1
Peak ground acc. (m/s2)

0.1

0.01
10 100 1000
Epicentral Distance (km)

Figure 7.2.17 Comparison of attenuation relations (magnitude, 7.0; depth, 30 km)

CHAPTER 7 SEISMIC LOADS - C7-39 -

Table 7.2.4 Comparison of accelerations (m/s2) at engineering bedrock in six cities

Acc. for 100-year Acc. for 500-year Acc. for Acc. for Building Standard Law
return period (m/s2) return period (m/s2) 1000 2500 （Limit capacity design）
Current Prev. Current Prev. year year Rare Ex- Seismic
City Recom. Recom. Recom. Recom return return Earth- tremely zoning
(2015) (2004) (2015) (2004) period period quake. rare factor
(m/s2) (m/s2) Earth-
quake.
Sapporo 0.97 0.83 1.73 2.03 2.30 3.31 0.58 2.9 0.9
Sendai 2.98 1.62 4.48 3.42 5.27 6.40 0.64 3.2 1.0
Tokyo 2.95 1.89 4.32 3.41 5.01 5.96 0.64 3.2 1.0
Nagoya 4.50 2.09 6.95 3.86 8.14 9.70 0.64 3.2 1.0
Osaka 3.26 1.91 5.28 3.78 6.46 8.30 0.64 3.2 1.0
Fukuoka 0.95 0.64 1.76 1.43 2.33 3.37 0.51 2.6 0.8

7.2.5 Factor to reduce story shear force according to plastic deformation capacity and response dis-
placement

(1) Reduction factor related to ductility

The seismic shear of the i-th story of a building is estimated using reduction factor kDi related to the duc-
tility of the building. However, for an evaluation within the elastic range of a building, this reduction factor
is not considered.
48)
The structural characteristic factor Ds of the current Building Standard Law of Japan , and the R factor
49)
of the US regulations , are equivalent to this reduction factor kDi. The equal energy rule or equal dis-
50)
placement rule can be applied.
Moreover, when kDi is considered in relation with an input earthquake motion, the time histories of the
near-field earthquakes recorded during the 1994 Northridge Earthquake and the 1995 Hyogo-ken Nambu
Earthquake exhibit a pulse-like wave form with very high amplitude. The efficiency of hysteretic damping
becomes less than in the case of a far-field earthquake, because the inelastic response owing to these time
histories tends to increase in one direction. When there is a fault in the near-field region where a high
probability of activity is expected in the near future, it is recommended that the seismic shear force for the
design should be higher than that predicted using the method described in the recommendations.

1) Structural characteristic factor Ds of the current building code regulations

The structural characteristic factor Ds used in the current regulations of the Building Standard Law of
Japan is calculated according to the ductility of the building. Here, Ds is calculated using the ductility
factors and shear force ratios of structural elements constituting a building, such as columns, beams,
shear walls, and braces. In addition, Ds is assigned a value of 0.25∼0.5 for steel and
steel-reinforced-concrete buildings, and 0.3∼0.55 for reinforced concrete buildings. A detailed descrip-
48)
tion of Ds can be obtained elsewhere .
- C7-40 - Recommendations for Loads on Buildings

2) Reduction factor based on equal energy rule or equal displacement rule

The equal energy rule or the equal displacement rule based on the maximum response relationship be-
tween elastic and elastoplastic systems having the same initial periods may be used. The strength reduction
factor Rµ and the maximum response displacement dp are expressed as follows.
For short-period systems (T<0.5s), the equal energy rule can be applied, and the following equation is
obtained.

Ve μ
Rμ= = 2μ −1 , d p= de (7.2.44)
Vy 2μ −1
For long-period systems (T>0.5 s), the equal displacement rule can be applied and the following equa-
tion is obtained.

Ve
Rμ= =μ d p= d e
Vy , (7.2.45)

where Ve is the maximum elastic force response, Vy is the yield strength of an elastoplastic system, de is
the maximum elastic displacement response, and the ductility factor µ is the ratio of the maximum ac-
ceptable displacement of the inelastic system to the yield displacement. It should be noted that the two de-
sign criteria are not always conservative and may vary owing to input earthquake motions.

7.2.6 Factor to increase story shear force according to irregularity of building

When a building having structural irregularities in plan and elevation is subjected to a strong ground
mo-tion, the damage may concentrate on a specific portion of the building. This is also pointed out in
reports of past earthquake damage. The concentrated damage leads to lesser seismic resistance than the
uniformly distributed damage.
The design seismic shear for these building is increased by the amplification factor kFi related to the ir-
regularities of the building. Furthermore, for buildings where large concentrated damage is expected, the
seismic performance will be evaluated through a time history response analysis using a suitable analytical
model, as described in Sec. 7.1.1.

(1) Fes in Ministry of Construction notification

48)
In the current building code , Fes is used as an amplification factor for the required story capacity
based on the irregularities in plan and elevation of the building. Here, Fes is obtained by multiplying Fs
by Fe. The assigned value of Fs is 1.0∼2.0, as shown in Fig.7.2.18 and Fe is 1.0∼1.5. In addition, Fs is cal-
culated according to the ratio of story stiffness distribution in elevation (Rs), Fe is calculated according to
the ratio of eccentricity in plan (Re), and Rs and Re are calculated based on the elastic stiffness of structur-
al elements of the building.
Because the damage from strong earthquake motion is caused by inelastic behavior, an index related to
CHAPTER 7 SEISMIC LOADS - C7-41 -

strength is more suitable in representing th e damage concentration than an index based on elastic
stiffness. There have been other guidelines and studies using the index related to strength; however, they
have not been proposed in a form that can be directly used as kFi. Here, Fes, which has generally common
values, may be used as kFi.

Fs Fe
2.0 2.0

1.5 1.5

1.0 1.0

0.5 0.5

Rs Re
0.0 0.0
0.0 0.2 0.4 0.6 0.8 1.0 0.00 0.15 0.30 0.45

Figure 7.2.18 Assigned values of Fs and Fe

7.3 Design Earthquake Motions and Building Response Evaluation

7.3.1 Fundamental concept

For a building whose dynamic behavior should be precisely captured, such as a very important
building, the design earthquake motions are generated according to the concept described in this sec-
tion, and a dynamic response analysis is carried out for the appropriate model by taking into account
the soil-structure interaction and dynamic properties of the building.
The procedures used to generate the design earthquake motions are classified into two types as fol-
lows:
• generation of design earthquake motions that are compatible with the design response
spectrum,
• generation of design earthquake motions based on scenario earthquakes considering local
site conditions and design requirements of the building.
It has been recognized that the main purposes of a building design are both to save human lives and
protect the property. For the latter purpose, the maximum probable earthquake motion of at least one
time during the design service life of the building is defined, and most of the main structures should
stay within the elastic limit, with no structural damage allowed. For the former purpose, the maximum
credible earthquake motion from an engineering viewpoint is defined at the site, and no harmful dam-
age to the building is allowed. In practice, to design buildings of importance, design earthquake mo-
tions are first determined, and dynamic earthquake response analyses are then conducted to ensure the
building safety against a seismic hazard.
- C7-42 - Recommendations for Loads on Buildings

The characteristics of earthquake ground motions depend on the seismic source, the propagation
path of seismic waves, and the local site conditions. Fig.7.3.1 shows a schematic diagram of these
factors. Recently, significant knowledge in seismology, earthquake engineering, geology, and other
areas has been achieved, which provide useful information for evaluating earthquake motions. The
design earthquake motion should reflect the natural phenomena as much as possible in that, which
may affect a building in the future. Therefore, it should be evaluated based on up-to-date knowledge
and a consistent philosophy, not far from scientific basis and within acceptable engineering limits.
From the above-mentioned perspective, this chapter proposes a rational methodology for evaluating
site-specific design earthquake motions based on site-specific scenario earthquakes.
Figure 7.3.2 shows the methodology for evaluating design earthquake motion51, 52). The procedures
used to generate design earthquake motions are classified into two types: a) the generation of a design
earthquake motion that is compatible with the target spectrum (see Section 7.3.2), and b) the genera-
tion of a design earthquake motion based on a scenario earthquake considering the local site condi-
tions and properties of the building (see Section 7.3.3).

Ground
地表面 level Surface
表面波（Wave (Loveayl
Love波・R Wave, Rayleigh Wave)
eigh波）
Surface
表層地盤 Ground Non-linearity,
地盤非線形化・液状化 Liquefaction

Body実体波
Wave P Wave, S Wave Density
Ｐ波速度・Ｓ波速度・密度
増幅 Long-period
長周期化
Motion
Ascending
上昇波（Ｅ） (E) Descending
下降波（Ｆ） (F)
Reflection 工学基盤
反射 Engineering
Bedrock 増幅
Amplification
Fault
断層破壊 Rupture Refraction
屈折
(Slip)
（滑り）
重複反射 Seismic
地震基盤 Bedrock 基盤波

Body Wave (P Wave, S Wave)

実体波（Ｐ波・Ｓ波）
Internal Damping Initial 主要動Ｓ波
初動Ｐ波
初動Ｐ波主要動Ｓ波 Later
Main 後続動表面波
後続動表面波
Earthquake
地震 Geometrical
内部減衰・幾何減衰 Damping Part Part Part (Coda)
(Source Fault)
（震源断層） P Wave S Wave Surface Wave

Source震源・伝播のモデル化
and Propagation Modeling Attenuation of Earthquake Motion
地震動の距離減衰 Amplification
表層地盤による増幅in Surface Layers
Epicentral Distance Site Attenuation
Acceleration Response Spectra
Amplitude of Earthquake Ground Motions

震央距離建設地震源近傍に適用可

地地 Soft Soil
震 Relations
能な距離減衰式震 Long-period
Depth 動 Hard Rock
震源深さ動 Components
Hypocentral
震源距離の
Ground の Short-period 軟らかい地盤
are
Hypocenter 最一般的な
加 Components （長周期卓越）
震源 Distance Motion
地盤での Dominant
Magnitude
マグニチュード大速 are
硬い地盤
Records
Point
点震源モデルSource Model 加観測記録 Dominant
度（短周期卓越）
速応
度答
断 Seismic Moment
Width

地震モーメント振ス
層幅増
ペ
幅 Rupture
破壊伝播ク
幅
面的効果 Near-fault
震源近傍での
断層長さ Length ト Amplification
増
Fault Model
巨視的断層モデル Records
振幅の頭打ちル幅
Ground Motion Bedrock
基盤
背岩盤での
Records
景
領 at Rock Sites
観測記録
域 Structure
地下構造
Asperity,
主破壊領域 Background Ｑ値
Q Value Distance
距離 Period
周期
Heterogeneous
不均質断層モデル Fault Model

Figure 7.3.1 Schematic diagram of the characteristics of earthquake ground motions:

the seismic source, propagation path of seismic waves, and the local site
conditions
CHAPTER 7 SEISMIC LOADS - C7-43 -

The target response spectrum is defined either

The Target Spectrum: Construction Site Condition Building Condition

the Acceleration Response
Spectrum Established in
Section 7.2
Geographic Environment, State of Building Use,
Seismic Environment, Structural Characteristics,
Open Engineering Bedrock Function for Daily Lives,
Design Service Life

Study on Building-design-related Earthquakes:

Location, Magnitude, Time Interval, Fault Parameters, etc.
Simulated Earthquake
Ground Motion Corre-
sponding to
the Target Spectrum Scenario Earthquake

Evaluation of Earthquake Ground Motion

by Using the Most Adequate Technique According to
the Quality and Quantity of the Available Information

Design Earthquake Ground

Motion Design Earthquake Ground Motion

Flowchart for Section 7.3.2 Flowchart for Section 7.3.3

Figure 7.3.2 Methodology for evaluating design earthquake motions

7.3.2 Design earthquake motions fit to the design response spectrum

The design earthquake motions are generated as simulated earthquake motions (time history) that
are compatible with the design response spectrum described in Section 7.2 as the target response spec-
trum.
• at the free surface of the engineering bedrock, or
• at the location of the input motion in the idealized model of the building.
In the latter case, the target response spectrum is evaluated considering the soil conditions between
the engineering bedrock and the defined location of the target response spectrum.
First, a time history is generated through the application of a Fourier Transform to the Fourier am-
plitudes and Fourier phases, which are assumed in advance. When the Fourier phases are assumed to
be random, the design envelope function is applied to the time history. The Fourier amplitudes are
modified iteratively to make the response spectrum of the time history compatible with the design re-
sponse spectrum (the target spectrum), and the design earthquake motion is generated.
- C7-44 - Recommendations for Loads on Buildings

7.3.3 Design earthquake motions based on scenario earthquakes

As shown in the right-hand side of Fig.7.3.2, the procedure used to set the design earthquake motion
based on the scenario earthquake sources is composed of the following three steps: assumption of
scenario earthquake sources that need to be considered, evaluation of earthquake motions at the con-
struction site resulting from the scenario earthquakes, and setting the design earthquake motions60, 61).
The design lifetime should be determined based on negotiations with the building owner, the building
use, daily functions, and laws related to depreciation. The earthquake source, the propagation path of
seismic waves, and the local site conditions should be examined as much as possible to evaluate the
design earthquake motions. Here, the design motion is defined at the free surface on the bearing stra-
tum that supports the building (i.e., the engineering bedrock).
Figure 7.3.3 shows a schematic describing the classification of the earthquake sources. First, the
earthquake sources considered herein are classified based on knowledge in the areas of seismology,
earthquake engineering, and geology, among others. The earthquake sources under consideration are
classified into the following four types: inter-plate earthquakes, oceanic intra-plate (out-
er-rise)/intra-slab earthquakes, continental intra-plate (crustal) earthquakes, and non-predetermined
intra-plate (crustal) earthquakes just beneath the site. The location, magnitude, and recurrence time of
each earthquake are determined by considering the plate tectonics, historical earthquakes, active faults,
seismotectonics, microearthquakes, layered structure, and so on53–65).
Next, the characteristics of the earthquake source, the propagation path of seismic waves, and the
local site conditions are taken into account, and earthquake motions induced through scenario earth-
quakes are evaluated at the site. Figure 7.3.4 and Table 7.3.1 show a flowchart and description of typ-
ical methods used to evaluate the time histories of earthquake ground motions66–70). Here, several
evaluation methods are considered: the use of strong motion records of scenario earthquakes at or near
the site, a theoretical approach based on fault models and elastic waves, an empirical or statistical
Green’s function approach based on fault models, a hybrid approach (i.e., combination of a theoretical
approach for longer period range, and a semi-empirical approach using Green’s function for shorter
period range), and an empirical approach using the standard spectrum and standard duration of earth-
quake motions or by directly estimating the time histories71).
The evaluation methodology described in Fig.7.3.4 and Table 7.3.1 adopts, in an integrated manner,
the most adequate evaluation method according to the quality and quantity of the available information,
because the feasibility and reliability of such an evaluation depend highly on the available information.
The methods shown in Fig.7.3.4 are aligned from left to right based on the quality and quantity of the
data necessary to adopt the method. Using the adopted method, the time histories of the earthquake
motions are evaluated.
Finally, the design earthquake motions are set through judgments of the evaluated earthquake mo-
tions. The most adequate of the following three methods is selected: using all of the evaluated motions,
using the representative evaluated motions among the evaluated motions, or generating new design
earthquake motions using the target spectrum and the duration based on the characteristics of the
CHAPTER 7 SEISMIC LOADS - C7-45 -

evaluated motions.
Continental Intra-plate Earthquakes Trench/Trough
( Crustal Earthquakes on Land and at sea)
Normal Fault Reverse
Fault

Right Lateral
Strike-slip Fault

Left Lateral
Strike-slip Fault
Movement of the Movement of the
Continental Plate, Oceanic Plate,
where the Japan subducting
islands lie beneath the Japan
islands

Inter-plate
Earthquakes
( Plate boundary )

Oceanic Intra-plate Thin arrows denote the

Earthquakes in the direction of the relative
Subducting/Subducted Plates slip on the fault

Figure 7.3.3 Classification of earthquake sources

Table 7.3.1 Modeling of source, path, and site effects for each method
to evaluate earthquake ground motions
Method to Evaluate Earthquake Path Site
Earthquake Source
Ground Motions (Propagation) (Local)

Strong Motion Record of

Record Record Record
Scenario Earthquake

Theoretical Method Theory Theory Theory or Statistics

Theory & Theory &

Hybrid Method Theory & Statistics
Record or Statistics Record or Statistics

Empirical Green’s
Theory & Record Record Record
Function Method

Statistical Green’s
Theory & Statistics Statistics Theory or Statistics
Function Method

Empirical Method Statistics Statistics Statistics

- C7-46 - Recommendations for Loads on Buildings

Assume a Scenario Earthquake

NO
YES
Records of Assumed Scenario Earthquake Exists
at or near the Construction Site
YES NO
Consider the Fault Model of
Assumed Scenario Earthquake NO
YES

Information of
Subsurface Structure Consider
is Sufficient enough NO Inner Fault
to Model the Parameters NO
Propagation and Site YES
Effects at the
Construction Site for Model Sufficient
the whole Target Propagation and Number of
Period Range Site Effects for Records to
the Long Period Evaluate the
YES Range and Use NO Propagation NO
Records for and Site
Short Period Effects for the
Range whole Target
YES Period Range
YES
Model Irregular Consider
Subsurface Several
Structure NO Rupture Cases NO
YES YES
Strong Theoretical Theoretical Hybrid Hybrid Empirical Statistical Empirical Empirical
Motion Method for Method for Green’s Method Green’s Green’s Long Period Time
Record of Irregular Horizontally Function Function Function Time History
Scenario Medium Layered Method Method Method History Est. Estimation
Earthquake Medium Method Method

Evaluation Evaluation Evaluation

is is is
Sufficient NO Sufficient NO Sufficient NO
YES YES YES

Time History of Earthquake Ground Motion

Quality and Quantity of Information Used for Strong Motion Evaluation = Capability to Explain the Phenomena
High Low
Practical Usefulness of Long Period Ground Motion Evaluation
High Low High Low
Practical Usefulness of Broadband Ground Motion Evaluation
High Low Low High Low Low
Practical Usefulness of Short Period Ground Motion Evaluation
High Low High Low High
Practical Usefulness of Spatial Ground Motion Evaluation
Low Moderate Moderate Low Moderate High High

Figure 7.3.4 Flowchart and explanation of methods used to evaluate earthquake ground motions
CHAPTER 7 SEISMIC LOADS - C7-47 -

7.3.4 Earthquake response analyses

In an earthquake response analysis, the earthquake response of a building corresponding to the ca-
pacity of each component is estimated, namely, the story shear force, the maximum member defor-
mation, maximum member stress, maximum inter-story drift, and maximum floor response, among
others. The possibility of P-Δ effects are also taken into consideration. Low-cycle fatigue of the
damping device, energy input to a structural member, and soil liquefaction should also be estimated
using an earthquake response analysis. The application of an earthquake response analysis is not lim-
ited to structural members but can also be applied to non-structural elements and equipment installed
in a building. Although the seismic load on non-structural elements and equipment is often estimated
based on an equivalent static approach, the use of an earthquake response analysis is preferable when
it exhibits clear frequency dependence in the response.
An earthquake response analysis is useful when considering the frequency and temporal character-
istic of input earthquake motions, as well as complicated dynamic characteristics of a building, which
is difficult to model through an equivalent static approach. On the other hand, a seismic load estimated
using an earthquake response analysis is quite sensitive to the assumption of the input earthquake mo-
tion and analysis model. Because earthquake response analyses of buildings are conducted prior to
construction, the building model for a response analysis in the design phase is not the same as an actu-
al building. A building model for a response analysis is developed and used when considering this fact.
Bias in the response predicted through response analyses can also be generated owing to a simplifica-
tion of the model, a modeling of the material properties, an approximation of the building geometry,
an assumption of the boundary conditions, discretization errors, rounding errors, errors in displaying
the results, and human errors, among other factors. Bias in the results needs to be decreased within an
allowable limit by introducing measures for quality assurance. Measures for verification and valida-
tion are introduced depending on the quality of the analysis required. The quality required for an anal-
ysis can differ depending on the objective of the earthquake response analysis. A peer review should
also be conducted by professionals in an appropriate discipline for quality assurance of the seismic
load estimated72, 73). For traceability, the results of a peer review need to be documented.
An equivalent static seismic load can be important to examine the appropriateness of a seismic load
estimated through seismic response analyses from different perspectives. It is also effective to specify
the required minimum seismic load applied to each member based on the equivalent static seismic
load.

(1) Model for earthquake response analyses

The following characteristics are taken into consideration when developing a response analysis
model.
a) Soil-structure interaction
b) Vibration characteristics such as vibration mode, natural period, and damping factor
c) Hysteretic restoring force characteristics of the building and soil
- C7-48 - Recommendations for Loads on Buildings

d) Non-structural elements
e) Torsional characteristics of the building
f) Shear and bending characteristics of the building
g) Floor stiffness of the building

Procedures for developing a response analysis model can reference the recommendations published
by the Architectural Institute of Japane.g., 74).

1) Soil
Amplification of earthquake motion from the engineering bedrock (or seismic bedrock) to the
ground level (or bottom of the foundation) is estimated. Parameters for the development of a soil
model are based on on-site tests, such as a standard penetration test and/or PS logging, and laboratory
tests. Uncertainty of the parameters is taken into consideration when developing a model. Amplifica-
tion of earthquake motion is estimated using one of the following approaches75).
a) Multiple reflection model
b) Multi-degree-of-freedom (lumped mass) model
c) Lattice (lumped mass) model
d) Finite element model or thin layer element model
One of the following analyses is employed depending on the level of soil strain.
a) Linear analysis
b) Equivalent linear analysis
c) Nonlinear response analysis

2) Building and foundation

One of the following models is employed for the building and foundation.
a) Equivalent single-degree-of-freedom model
b) Multi-degree-of-freedom model
c) Two- or three-dimensional model
d) Finite element model
One of the following analyses is employed depending on the level of building strain.
a) Linear analysis
b) Equivalent linear analysis
c) Nonlinear response analysis

3) Combined model
A full model combining the soil, building, and foundation can be used if necessary.
CHAPTER 7 SEISMIC LOADS - C7-49 -

(2) Estimation of seismic loads based on seismic response analyses

Earthquake motions used in a seismic response analysis are determined based on the procedures
specified in Sec. 7.3.2 and/or Sec. 7.3.3. Two components of ground motion, or moreover, three com-
ponents of ground motion including an up-down component, are better for use as inputs to a seismic
response analysis instead of a single component of earthquake motion when a 3D building model is
used for a response analysis.
When estimating a seismic load based on a seismic response analysis, it should be assumed that
variability is an inherent characteristic of a seismic response analysis owing to variability in the phase
characteristics of the input earthquake motion, for example. A number of earthquake response analyses
for different ground motion parameters, such as the phase, have been conducted and considered for an
estimation of seismic load.
When earthquake motions are synthesized using a ground motion simulation considering the fault
rupture process, certain parameters such as the characteristics of strong motion generation area
(SMGA) and location of the hypocenter, length of the fault, and simultaneous rupture of multiple
faults remain arbitrary. These arbitrary properties add more uncertainty in the predicted earthquake
motions. A sensitivity analysis should be conducted to quantify the effects of these uncertainties on the
seismic load.
When examining the fatigue of structural members such as vibration control devices, the effects of
multiple occurring earthquakes, which are expected during its lifetime, are considered instead of a sin-
gle earthquake.
Earthquake motions for estimating the seismic load conform to the performance requirements of a
building (see Chapter 2). This can be checked through a comparison of the equivalent static load esti-
mated in Sec. 7.2, as well as the load factor obtained through a procedure described in Chapter 2.
When the return period of the design earthquake motions is specified, the return period of these earth-
quake motions should be compared with that of the design return period.

(3) Seismic loads from a flow and deformation of the soil

Examples of a seismic load from a flow and deformation of the soil are as follows:
• Forced displacement, subsidence, and differential settlements owing to liquefaction
• Impact of rock and/or soil caused by a slope failure
• Fault displacement beneath the building76)

References
1) Architectural Institute of Japan: Preliminary Reconnaissance Report of the 2011 Tohoku-Chiho
Taiheiyo-Oki Earthquake, 2011.7 (in Japanese)
2) Architectural Institute of Japan: Recommendations for Aseismic Design and Construction of
Nonstructural Elements, 2003.1 (in Japanese)
3) Architectural Institute of Japan: Seismic Response Analysis and Design of Buildings Consider-
- C7-50 - Recommendations for Loads on Buildings

ing Dynamic Soil-Structure Interaction, 2006.2 (in Japanese)

4) The Japanese Geotechnical Society: Process and Commentary of Soil Test, pp. 691-695, 2006
(in Japanese)
5) Architectural Institute of Japan: Structural design concepts for earthquake and wind - current
status and outlook -, 2006 (in Japanese)
6) Architectural Institute of Japan: Recommendations for Loads on Buildings, p. 160, 1993 (in
Japanese)
7) Takanashi, K. and Gao, X.: Earthquake Resistant Design of Single Story Frame with Sliding
Floor Load, Summaries of technical papers of Annual Meeting Architectural Institute of Japan.
B, Structures I, pp. 47-48, 1989.10 (in Japanese)
8) Building Guidance Division, Housing Bureau, Ministry of Land, Infrastructure, Transport and
Tourism, Japan Conference of Building Officials and The Building Center of Japan: : Commen-
tary for Technical Codes of Building Structures, 2001.3 (in Japanese) .
9) Architectural Institute of Japan: Damping in Buildings, 2000 (in Japanese)
10) Ohsaki, Y.: Vibration Theory for Buildings, Shokokusha Publishing Co. Ltd., 1996 (in Japanese)
11) Architectural Institute of Japan: Damping and Response Control of Shell and Spatial Structures,
2008 (in Japanese)
12) Umemura, H, Osawa, Y. and Matsushima, Y.: Research on Response Spectra for Various Soil
Profiles, Annual Report of Synthetic Research Laboratory, No. 24, Vo. 2, 1966 (in Japanese)
13) Kawashima, K. Nagashima, H. and Iwasaki, H., Civil Engineering Journal, 35-5, pp. 62-67,
1993 (in Japanese)
14) International Organization for Standardization: ISO 13033 Bases for design of structures -
Loads, forces and other actions - Seismic actions on nonstructural components for building ap-
plications, 2013.8.
15) Building Performance Standardization Association, 2013 (in Japanese)
16) International Organization for Standardization: ISO 3010 Bases for design of structures - Seis-
mic actions on structures, 2001.12.
17) Architectural Institute of Japan: Seismic Loading - State of the Art and Future Developments,
1987.11 (in Japanese)
18) Architectural Institute of Japan: An Introduction to Dynamic Soil-Structure Interaction, 1996.4
(in Japanese)
19) Ishiyama, Y.: Seismic code and structural dynamics, Sanwa Shoseki Ltd., 2008 (in Japanese)
20) Ishiyama, Y.: Distribution of Seismic Shear and the Base Shear Coefficient for Various Types of
Buildings, Journal of Structural and Construction Engineering (Transactions of AIJ), No. 439,
pp. 65–72, 1992.9 (in Japanese)
21) Ishiyama, Y.: Birth of Ai Distribution and Its Related Matters, Proceedings of Annual Meeting
of Architectural Institute of Japan, pp. 183–184, 1992.8 (in Japanese)
22) Satoh, T., Okawa, I., Nishikawa, T., Sato, T., and Seki, M.: Prediction of Waveforms of
CHAPTER 7 SEISMIC LOADS - C7-51 -

Long-Period Ground Motions for Hypothetical Earthquakes using Empirical Regression Rela-
tions of Response Spectra and Phase Spectra, Journal of Structural and Construction Engineer-
ing (Transactions of AIJ), Vol. 75, No. 649, pp. 521-530, 2010 (in Japanese with English ab-
stract)
23) Yuzawa, Y. and Kudo, K.: Shakeability Distribution for Long-period ( 1-15 sec ) Ground Mo-
tion in Japan, Journal of Japan Association for Earthquake Engineering, Vol. 11, No. 3, pp.
21-39, 2011
24) Itoi, T. and Takada, T.: Empirical Correction of Ground Motion Prediction Equations of Re-
sponse Spectra at Rock Sites for Near-Field Earthquakes Considering Amplification Effect
within Deep Subsurface Structure, Journal of Japan Association for Earthquake Engineering,
Vol. 12, No. 1, pp. 43-61, 2012
25) Osaki, T. and Takada, T.: Transformation between Spectrum Density Function and Response
Spectrum Based on First Passage Theory - on Draft of AIJ Recommendations for Loads on
Buildings, Structural Engineering Symposium, 2003.4 (in Japanese)
26) Schnabel, P.B., Lysmer, J. and Seed, H.B.: SHAKE - A Computer Program for Earthquake Re-
sponse Analysis of Horizontally Layered Sites, Report No. EERC 72-12, Earthquake Engineer-
ing Research Center, University of California, Berkeley, 1972.12
27) Shibata, A.: Dynamic Analysis of Earthquake Resistant Structures, Tohoku University Press,
2010.
28) Ohsaki, Y.: Dynamic Characteristics and One-Dimensional Linear Amplification Theories of
Soil Deposits, Research Report 82–01, Dept. of Architecture, Faculty of Eng., Univ. of Tokyo,
1982.3
29) Japan Road Association: Speciﬁcations for Highway Bridges (Part V: Seismic Design), Maruzen
Co.,Ltd., p. 222, 1990.2 (in Japanese)
30) National Research Institute for Earth Science and Disaster Resilience (NIED): Strong-motion
Seismograph Networks (K-NET, KiK-net):
[Link]
31) Japanese Geotechnical Society: Dynamic analysis of ground - From basic to applied theory (ge-
otechnical engineering and basic theory series), ISBN: 4886449514 [Japanese Import], 2007.2
(in Japanese)
[Link]
32) Yasui, Y.: Relationship between Building Damage and Strong Motion Records - Soil-Building
Interaction - AIJ, pp. 67–86, 1997.12 (in Japanese)
33) Tohdo, M. and Ishiyama, Y.: A Practical Evaluation Method of Seismic Load Considering Soil
Structure Interaction Effects, 13th World Conference on Earthquake Engineering, Paper No.
2641, 2004.8
34) Tohdo, M., Chiba, O. and Fukuzawa, R.: Impedance Functions and Effective Input Motions of
Embedded Rigid Foundations, 7th Japan Earthquake Engineering Symposium, pp. 1039-1044,
- C7-52 - Recommendations for Loads on Buildings

1986.12
35) Tajimi, H. and Izumikawa, M., Proceedings of the 6th Japan Earthquake Engineering Sympo-
sium, pp. 1745-1752, 1982 (in Japanese)
36) National Research Institute for Earth Science and Disaster Prevention: A Study on “National
Seismic Hazard Maps for Japan,” Document Supplement and Data Appendix, Technical Note of
the National Research Institute for Earth Science and Disaster Prevention No. 336, Appendix
2-1–2-59, 2009 (in Japanese)
37) Gumbel, E.J.: Statistics of Extremes, Columbia University Press, 1958
38) Kanda, J.: A New Extreme Distribution with Lower and Upper Limits for Earthquake Motions
and Wind Speeds, Proceedings of the 31st Japan National Congress for Applied Mechanics, pp.
351-360, 1981
39) Cornell, C.A.: Engineering Seismic Risk Analysis, Bulletin of the Seismological Society of
America, Vol. 58, pp. 1583-1606, 1968
40) Kanno, T., Narita, A., Morikawa, N., Fujiwara, H., and Fukushima, Y.: A new attenuation rela-
tion for strong ground motion in Japan based on recorded data, Bulletin of the Seismological
Society of America, Vol. 96, No. 3, pp. 879–897, 2006
41) The Earthquake Research Committee of the Headquarters for Earthquake Research Promotion:
Report: 'National Seismic Hazard Maps for Japan (2007)', 2007.4.18 (In Japanese)
42) Baker, J. W.: Conditional Mean Spectrum: Tool for ground motion selection, Journal of Struc-
tural Engineering, Vol. 137, No. 3, pp. 322-331, 2011
43) Cornell, C. A.: Calculating building seismic performance reliability: A basis for multi-level de-
sign norms, Proceedings of the 11th World Conference on Earthquake Engineering, 1996.6
44) Takada, T.: Probabilistic Expression of Seismic Performance and Seismic Loads, Proceedings of
Panel Discussion of AIJ Annual Conference, 1999 (in Japanese)
45) Annaka, T. and Yashiro, H.: A seismic source model with temporal dependence of large earth-
quake occurrence for probabilistic seismic hazard analysis in Japan, Proceedings of the 10th Ja-
pan Earthquake Engineering Symposium, pp. 489-494, 1998 (in Japanese)
46) Annaka, T. and Yashiro, H.: Uncertainties in a probabilistic model for seismic hazard analysis in
Japan, Second International Conference on Computer Simulation in Risk Analysis and Hazard
Mitigation, Risk Analysis II, pp. 369-378, 2000
47) Si, H. and Midorikawa, S.: New attenuation relationships for peak ground acceleration and ve-
locity considering effects of fault type and site condition, Journal of Structural and Construction
Engineering (Transactions of AIJ), No. 523, pp. 63-70, 1999.9 (in Japanese)
48) Building Guidance Division, Ministry of Housing: Technical Reference for structural standard
of buildings (2007 edition), 2007.8 (in Japanese)
49) BSSC: NEHRP Recommended Provisions for Seismic Regulations for New Buildings and Oth-
er Structures Edition, 1997
50) Veletsos, A.S. and Newmark, N.M.: Effect of Inelastic Behavior on the Response of Simple
CHAPTER 7 SEISMIC LOADS - C7-53 -

System to Earthquake Motion, Proceedings of the Second World Conference on Earthquake En-
gineering, 1960
51) Ishii, T. and Sato, T.: An evaluation methodology of design earthquake motions based on
site-related earthquakes, Journal of Structural and Construction Engineering (Transactions of
AIJ), Vol. 59, No. 462, pp. 31- 42, 1994 (in Japanese with English abstract)
52) Ishii, T. and Sato, T.: An Evaluation Methodology of Design Earthquake Motions based on
Site-specific Design Earthquakes, Proceedings of the Fifth International Conference on Seismic
Zonation, Vol. I, pp. 392- 399, 1995
53) National Astronomical Observatory of Japan: Chronological Scientific 2003, Vol. 76, Maruzen
Co., Ltd., 2002.11 (in Japanese)
54) Usami, T.: Materials for Comprehensive List of Destructive Earthquake in Japan (Revised and
Enlarged Edition 416–2001), University of Tokyo Press, 2003 (in Japanese)
55) Utsu, T.: Catalog of Large Earthquake in the Region of Japan from 1885 through 1980, Bulletin
of the Earthquake Research Institute, University of Tokyo, Vol. 57, Part 3, pp. 401-463, 1982 (in
Japanese)
56) Utsu, T.: Catalog of Large Earthquake in the Region of Japan from 1885 through 1980 (Correc-
tion and Supplement), Bulletin of the Earthquake Research Institute, University of Tokyo, Vol.
60, Part 4, pp. 639-642, 1985 (in Japanese)
57) The Japan Meteorological Agency: The Seismological and Volcanological Bulletin of Japan,
1961- (in Japanese)
58) The Japan Meteorological Agency: [Link]
59) Japan Meteorological Business Support Center: [Link]
60) The Research Group for Active Faults of Japan: Active Faults in Japan; Sheet Maps and Inven-
tories, Revised Edition, University of Tokyo Press, 1991 (in Japanese)
61) Matsuda, T.: Seismic Zoning Map of Japanese Islands, with Maximum Magnitudes Derived
from Active Fault Data, Bulletin of the Earthquake Research Institute, University of Tokyo, Vol.
65, pp. 289- 319, 1990 (in Japanese)
62) Matsuda, T.: Magnitude and Recurrence Interval of Earthquakes from a Fault, ZISIN; Journal of
the Seismological Society of Japan, Second Series, Vol. 28, pp. 269-283, 1975 (in Japanese with
English abstract)
63) Sato, R.: Parameter Handbook of Earthquake Faults in Japan, Kajima Publishing Co., Ltd., 1989
(in Japanese)
64) Kakimi, T., T. Matsuda, I Aida, and Y. Kinugasa: A Seismotectonic Province Map in and around
the Japanese Islands, ZISIN; Journal of the Seismological Society of Japan, Second Series, Vol.
55, pp. 389-406, 2003 (in Japanese with English abstract)
65) The Headquarters for Earthquake Research Promotion: [Link]
66) Association for the Development of Earthquake Prediction: FY1997 Report on a Review and
Applicative Studies of the Methods to Evaluate Strong Motions, 1998.3 (in Japanese)
- C7-54 - Recommendations for Loads on Buildings

67) Association for the Development of Earthquake Prediction: FY1998 Report on a Review and
Applicative Studies of the Methods to Evaluate Strong Motions, 1999.3 (in Japanese)
68) Architectural Institute of Japan: Earthquake Ground Motion - Phenomena and Theory -, 2005
69) Architectural Institute of Japan: Generation Guide for Seismic Input Motions Based on the Re-
cent Advancement of Ground Motions Studies, 2009
70) Architectural Institute of Japan: Earthquake Ground Motion and Strong Motion Prediction - Key
Items for Learning the Basics -, 2016
71) Architectural Institute of Japan: Seismic Loading – Toward Performance-Based Design, 2008.3
(in Japanese)
72) International Organization for Standardization: Quality management systems - Requirements,
ISO9001: 2008
73) Shiratori, M, Koshizuka, S., Yoshida, Y. and Nakamura, H.: Quality Assurance and Verification
& Validation of Engineering Simulation, Maruzen, 2013 (in Japanese)
74) Architectural Institute of Japan: Recommendation for the Design of Seismically Isolated Build-
ings, 4th Edition, 2013.10 (in Japanese)
75) Yoshida, N.: Earthquake Response Analysis of Ground, Kajima Publishing, 2010 (in Japanese)
76) Japan Society of Civil Engineers: Committee report of seismic performance of bridge subjected
to fault displacement, 2008.7 (in Japanese)
- C8-1 -

CHAPTER 8 THERMAL LOADS

Introduction
Expansion and bending deformations occur when a building or member undergoes a temperature
change. Strain is generated in the member by restraining this deformation. A temperature change that
causes a load effect on a building is called thermal load.
Thermal stress caused by a thermal load is called the self-constraining stress (secondary stress), unlike
the stress (primary stress) caused by an external force such as a fixed load, live load, seismic load, or
wind load. Self-constraining stresses are generated through balancing inside the building, and occur
depending on the rigidity of the structural frame or the nonstructural member of the building, as well as
the degree of external constraint. Although it is unlikely that the building will directly lead to a collapse
based solely on the influence of the thermal load, it is necessary to pay attention to the design in terms
of usability, such as the occurrence of cracks and harmful deformations depending on the scale and
building use.
In addition to long buildings and buildings with large spaces that are built in areas with large
temperature differences, where the expansion and contraction of the building members from temperature
change is large, it is also necessary to consider the influence of temperature load of chimneys, silos,
thermal storage tanks, refrigerated warehouses, power plants, and other structures.
In some cases, it may be necessary to consider the effects of temperature load during the construction
period of the building.
It is important to grasp the stress and deformation by conducting a thermal stress analysis through the
accurate calculation of the temperature, as well as the temperature change of the structural members
based on the temperature conditions, such as the outdoor air temperature surrounding the buildings, solar
radiation, earth temperature, indoor temperature, and heat source. The rigidity of the buildings and the
properties of the materials used should also be considered.
The thermal load dealt with in these recommendations assumes temperatures that occur in the normal
use of a building. Temperatures that occur during an emergency such as a fire, act of terrorism, or sudden
equipment breakdown can reach several hundred degrees Celsius or more, which can cause significant
degradation of the material properties such as Young's modulus and the material strength. We decided
not to include such temperatures during the design process in these recommendations because
performance requirements differing from the normal requirements assumed herein are required.
The main points of revision from the old recommendations (2004 edition) are as follows.
(1) For the annual maximum and minimum temperatures as the basis of the thermal load calculation, the
100-year recurrence temperature was calculated based on the Meteorological Agency data for a 52
year period, from 1961 to 2012.
(2) The annual maximum and minimum outdoor air temperatures corresponding to the recurrence period
can be calculated.
(3) Load combination factors for combining the thermal load with other loads are presented.
- C8-2 - Recommendations for Loads on Buildings

8.1 Scope of Application

(1) This chapter describes thermal loads for the design of buildings.
(2) When a building is subject to temperature change due to the site layout, the time, the size of the
building, the use of the building, or the service conditions, thermal loads need to be considered.
(3) When the thermal stress or deflection in structural members is reduced by having expansion joints
or some other effective measures, thermal loads may not be considered.

Thermal load is defined as a temperature that has an effect on a building, such as outdoor air
temperature, solar radiation, underground temperature, indoor air temperature, and the heat source
equipment inside the building. A change in temperature of the structural and non-structural member
causes thermal stress, and is defined as the thermal load effect.

8.1.1 Temperature change

Temperature change in a member is a combination of the average temperature change ∆𝑇𝑇𝑑𝑑 and the
temperature gradient ∆𝑇𝑇𝑔𝑔 /𝑡𝑡. (See Fig. 8.1.1) Each change causes a load effect, as shown in Fig. 8.1.2.
When the deformation is constraind, member forces is generated as shown in Fig. 8.1.2.

∆𝑇𝑇𝑑𝑑
1 ∆𝑇𝑇𝑔𝑔 /𝑡𝑡

average temperature change temperature gradient

∆𝑇𝑇𝑑𝑑 ∆𝑇𝑇𝑔𝑔 /𝑡𝑡

𝑇𝑇1 , 𝑇𝑇2 = ∆𝑇𝑇𝑑𝑑 ± ∆𝑇𝑇𝑔𝑔 /2
Figure. 8.1.1 Temperature change in a member

Figure. 8.1.2 Effect of thermal load in a member

CHAPTER 8 THERMAL LOADS - C8-3 -

8.1.2 Consideration of thermal load

Thermal load should be considered for the following building types: buildings constructed in an area
where there is a considerable change in outdoor air temperature, lengthy buildings, buildings with a
large interior space, buildings with a direct influence of solar radiation such as from a glass roof,
buildings or structures with a heat source such as a chimney, silos containing warm materials, heat
storage tanks, refrigerated warehouses, and electric power plants. When a building is divided into
smaller parts with expansion joints to reduce the movement in each part, or the temperature change in
the structural member is reduced by thermal insulation, the thermal load may be disregarded.

8.2 Method of Thermal Load Calculation

The basic value of the thermal load shall be a 100-year recurrence temperature change such as
outside air temperature, solar radiation, underground temperature, or equivalent value.

As shown in Fig. 8.2.1, the outdoor air temperature has repeated daily fluctuations in addition to
annual fluctuations, and the influence of such daily fluctuations of solar radiation is thereby included.
When setting the thermal load of the building, the maximum difference between the temperature
obtained by adding solar radiation to the highest outdoor air temperature during the summer and the
lowest outdoor air temperature in the winter is considered. Because the occurrence time of the minimum
outdoor air temperature during the winter is early in the morning, the influence of solar radiation can be
ignored. For this reason, the outdoor air temperature is set as a basic value of the annual maximum and
minimum outdoor air temperatures for a return period of 100 years. However, the temperature of the
structural members (load effect) from the outdoor air temperature and solar radiation is effected by the
type of structure (reinforced concrete construction, steel construction, etc.), specifications of the
finishing (color etc.), thermal inertia of the structure, and the heat transfer characteristics of the structure
and outdoor air temperature. It also depends on periodicity such as the annual and daily variations. This
is extremely complicated for various reasons. Therefore, although it is not easy to set the temperature of
a structural member immediately even if the thermal load is set as described above, the above-mentioned
temperature is applied as the basic value in these guidelines.

Figure 8.2.1 Fluctuation of outdoor air temperature and solar radiation1)

- C8-4 - Recommendations for Loads on Buildings

Outdoor air temperature and solar radiation have a certain periodicity. Comparing a case of setting
the temperature of a structural member as a steady state and ignoring the periodicity, and a case in which
the temperature of a structural member is set at an unsteady state, which evaluates the influence
appropriately that the temperature which fluctuates from time to time is treated as a time history in a
reinforced concrete building with large thermal inertia, the case of the steady state analysis gives an
excessively larger temperature of structural member than an unsteady state analysis. Originally, it was
necessary to set the temperature of a structural member based on a time history temperature response
analysis considering the unsteady effect under actual conditions, for example, but the thermal load of
the time history based on the probability of non-exceedance has not been provided for the present
situation. If necessary, it is conceivable to use directly observed records or temperature data in a periodic
steady state prepared for the capacity design of air conditioning equipment.
In addition, for solar radiation, because it influences the temperature increase of the structural member
as a daily fluctuation, the setting based on the probability of non-exceedance is expected, but it has
currently not shown sufficient numerical values. On the other hand, the normal solar radiation outside
the earth’s atmosphere shows a nearly constant value. Therefore, the value calculated theoretically when
taking the scattering and absorption attenuation within the earth’s atmosphere into consideration is taken
as the basic amount of solar radiation.
The solar radiation and outdoor air temperature are considered separately in these recommendations.
Originally, they are physically correlated and cannot be separated; however, it is currently difficult to
set the thermal load when considering the correlation between them, and we expect for the future
research. When using actual Meteorological Agency data, the correlation between them is clearly
satisfied.
In addition, the temperature distribution of a building or structural member is influenced by the heat
transfer coefficient and thermal conductivity of the building materials. For example, when a temperature
difference occurs between the inside and outside of the wall, its temperature distribution will be as in
shown Fig. 8.2.2. In this case, the temperature difference between the inner and outer surfaces of the
wall is lower than that between the inner and outer ambient temperature (T1 and T2) owing to the effects
of the heat transfer coefficient of the inner and outer surfaces, 𝛼𝛼1 and 𝛼𝛼2 , and the thermal
conductivity 𝜆𝜆 of the wall. The temperature distribution of the wall is influenced by the heat transfer
coefficient and thermal conductivity, and fluctuates repeatedly over time with the same pattern as the
variation of the outdoor air temperature. In the walls of chimneys, heat storage tanks, refrigerated
warehouses, and so on, the generation of a rapid temperature gradient is generally controlled by
providing an air layer through a liner or by using a heat insulating material. However, as in an example
in which the ground close to a refrigerated warehouse is frozen, if heat is constantly being supplied,
such effect is not utilized even if insulation is present, and thus proper consideration is necessary to set
the thermal load.
Please refer to the data for each temperature, shown below, that the designer sets for a basic
temperature change.
CHAPTER 8 THERMAL LOADS - C8-5 -

Inner ambient temperature T1

Temperature
Inner surface’s temperature

Interior Exterior
Time

Outer surface’s temperature

Wall Outer ambient temperature T2

𝛼𝛼1 , 𝛼𝛼2 : heat transfer coefficient of the inner and outer surfaces of the wall
𝜆𝜆: thermal conductivity of the wall

Figure 8.2.2 Temperature distribution of the wall

8.2.1 Outdoor air temperature

When calculating the thermal stress of a building, it is necessary to consider various influential factors
and a change in the time series, including the influence and azimuth of solar radiation. Therefore, the
basic outdoor air temperature data are organized, and high and low temperatures are indicated. To obtain
the thermal stress of a building, the thermal load surrounding the building is first set. The thermal load
is expressed as the temperature.

(1) One-hundred-year recurrence values of highest and lowest outdoor air temperatures
The 100-year-recurrence values of the highest and lowest outdoor air temperatures were calculated
using a fitting of the extreme value distribution on 52-year data from 1961 to 2012 obtained from the
Meteorological Office.
As a result of organizing the temperature data at 56 points in Japan, the 100-year-reccurence value of
the high outdoor air temperature is distributed within a range of about 32–41 °C, whereas that of a low
temperature is distributed within a wide range of -30.9–5.6 °C.
As a result of investigating the transition of high and low temperatures in Tokyo for 52 years (1961
to 2012), the high temperature increased by about 1 °C in 50 years, and the low temperature increased
by about 3.5 °C.
It seems that these increases were influenced through global warming and urban heat islanding.
Although these guidelines show the 100-year-recurrence value based on the current observation data
(1961 to 2012), these values may gradually increase.
Table 8.2.1 shows the 100-year-recurrence value of the highest and lowest outdoor air temperatures
and observation records for 12 major cities (Sapporo, Sendai, Maebashi, Tokyo, Nagano, Niigata,
Nagoya, Osaka, Hiroshima, Takamatsu, Fukuoka, and Naha).
- C8-6 - Recommendations for Loads on Buildings

Table 8.2.1 100-year recurrence value of highest and lowest outdoor air temperatures
and observation records for 12 major cities
The 100-year recurrence The 100-year recurrence Observation records
value of highest value of lowest (Start–2012)
temperature temperature
(1961–2012) (1961–2012)
highest lowest
city distribution ℃ distribution ℃ temperature temperature
℃ ℃
Sapporo Weibull 36.9 Weibull -20.9 36.2 -23.9
Sendai Weibull 38.0 Weibull -10.8 37.2 -11.7
Niigata Weibull 39.2 Gumbel -12.0 39.1 -13.0
Nagano Weibull 38.5 Weibull -14.7 38.7 -17.0
Maebashi Weibull 41.0 Weibull -9.8 40.0 -11.8
Nagoya Weibull 40.2 Weibull -7.9 39.9 -10.3
Tokyo Weibull 40.1 Weibull -5.8 39.5 -9.2
Hiroshima Weibull 39.3 Gumbel -9.2 38.7 -8.6
Osaka Weibull 39.5 Weibull -5.7 39.1 -5.5
Fukuoka Weibull 38.3 Weibull -5.6 37.7 -8.2
Takamatsu Weibull 39.0 Weibull -6.8 38.2 -7.7
Naha Gumbel 36.5 Gumbel 5.6 35.6 6.6

Thermal loads are influenced by the weather, which is a natural phenomenon, and vary with the
summer–winter cycle, fluctuating day and night as a single cycle. Therefore, attention is focused on the
daily average temperature, and the values for the annual high value of the daily average temperature and
the yearly low value for a 100-year recurrence are shown in Table 8.2.2.
CHAPTER 8 THERMAL LOADS - C8-7 -

Table 8.2.2 Values for annual high value of daily average temperature
for 100-year recurrence for 12 major cities
Annual high values of daily average
temperature Values of daily average temperature based
highest daily lowest daily average on observation records (1961–2012)
average temperature temperature
highest lowest average
city distribution ℃ distribution ℃ temperature temperature temperature
℃ ℃ ℃
Sapporo Weibull 31.0 Weibull -14.6 30.1 -14.1 8.6
Sendai Weibull 31.9 Weibull -6.5 31.2 -5.2 12.2
Niigata Weibull 32.8 Weibull -4.6 32.0 -3.9 13.6
Nagano Weibull 30.9 Weibull -8.5 30.7 -8.0 11.7
Maebashi Weibull 33.3 Weibull -4.3 32.6 -3.8 14.3
Nagoya Weibull 33.2 Weibull -3.9 32.7 -2.9 15.5
Tokyo Weibull 33.6 Weibull -1.1 33.1 -0.6 16.0
Hiroshima Weibull 32.7 Gumbel -6.3 32.7 -5.8 15.7
Osaka Weibull 32.9 Weibull -2.2 32.9 -2.1 16.6
Fukuoka Weibull 32.5 Weibull -3.2 32.4 -3.2 16.6
Takamatsu Weibull 33.4 Gumbel -4.0 32.3 -3.3 15.9
Naha Weibull 31.5 Weibull 8.7 31.1 9.1 22.8

(2) Weather data

Useful weather data for the statistical processing of meteorological data and for calculating the
temperature of a building are as follows.
1) Weather data of the Japan Meteorological Agency
Many meteorological elements are provided by the Japan Meteorological Agency.
Required weather data can be obtained from the Japan Meteorological Agency’s website.
2) Expanded AMeDAS Weather Data2)
Based on the Automated Meteorological Data Acquisition System (AMeDAS) data of the Japan
Meteorological Agency, “Expanded AMeDAS Weather Data” was developed by the Architectural
Institute of Japan for use in air conditioning load calculations and research and design.
It provides hourly weather data over a 15-year period, from 1981 to 1995, at 842 points in Japan.

8.2.2 Solar radiation

(1) Influence of solar radiation and sol-air temperature
When solar radiation is absorbed by the surface of a member, it turns into heat and increases the
temperature of the member. The increasing rate of the temperature of a member depends on the solar
radiation absorption rate of the member surface, that is, it depends on the solar radiation absorption rate.
However, if the amount of solar radiation incident on the outer surface is J [W/m2], the solar radiation
- C8-8 - Recommendations for Loads on Buildings

absorption rate is ‘a’, and the total heat transfer rate of the outer surface is αo [W/(m2·K)], and it is well
known that the effect of an outdoor air temperature increase is equivalent to aJ/αo. A value of 25
[W/(m2·K)] is usually used as the total heat transfer rate of the outer surface, αo. For example, when the
amount of solar radiation is 1,000 [W/m2] and the solar absorption is 0.8, the temperature increase
becomes 32 °C, and the magnitude of the influence of solar radiation will then be understood. The total
heat transfer rate of the outer surface varies strictly depending on the wind velocity and the long
wavelength radiation of the member surface, and thus a temperature increase can be 10 [W/(m2·K)] or
less when the wind velocity is low. Now, the value obtained by adding this temperature increase, aJ/αo,
to the outdoor air temperature, To, is called the sol-air temperature, TSAT.

𝑎𝑎
𝑇𝑇SAT = 𝑇𝑇o + 𝐽𝐽 (8.2.1)
𝛼𝛼o
When calculating the heat transfer of a member surface, using the sol-air temperature makes it easy
to handle the influence of outdoor air temperature and solar radiation collectively. When the amount of
heat flow toward the inside from the surface of the member can be ignored, the sol-air temperature
coincides with the surface temperature of the member. The surface temperature of an ideal thermal
insulating material (with complete thermal insulation) painted black (where the solar radiation
absorption rate is equal to 1) is the maximum surface temperature of the member that can occur.
Although the sol-air temperature can be directly measured in this manner, and the physical meaning is
clear, it depends on the solar radiation absorption rate of the member, and it would thus be appropriate
to indicate the amount of solar radiation as a basic value.

(2) Amount of solar radiation

However, the amount of normal surface solar radiation of outside the atmosphere is 1,367 W/m2, and
varies by about 3% depending on the season, but is nearly constant. Although the amount of solar
radiation cannot exceed this value unless it is condensed using a convex lens or concave mirror, adopting
this as the basic value will result in an overestimation. When solar radiation enters the atmosphere, it
damps owing to scattering by air molecules, clouds and aerosols, absorption by water vapor, and so on.
The amount of solar radiation in a sunny sky without clouds will be nearly the same for a particular
location, season, and time, and thus instead of finding the value of the solar radiation for 100 years of
recreation statistically, the theoretically calculated value is shown as a basic value. The minimum value
of solar radiation is equal to zero, that is, solar radiation is not taken into consideration.
With regard to the amount of water vapor and aerosol during the summer without clouds, it is
considerably small (potential water precipitation of 1.5 cm), where the muddiness coefficient in
Angstroms is 0.05 (RURAL aerosol). Using these values, the amount of solar radiation, which is
theoretically calculated using SMARTS23), is adopted as a basic value. The basic value of the daily
average solar radiation is shown in Fig. 8.2.3, and the basic value of the daily maximum solar radiation
is shown in Fig. 8.2.4. SMARTS2 is a typical solar radiation prediction model used in the engineering
CHAPTER 8 THERMAL LOADS - C8-9 -

field, and the value calculated by SMARTS2 is adopted for the “ASTM standard slope solar spectrum4).”
However, the amount of solar radiation according each direction is calculated using Perez's model5),
which takes into consideration the luminance distribution of the sky using the position of the sun on July
23. Because the amount of the solar radiation in each direction has a value that includes the amount of
reflective solar radiation (the reflectance is equal to 0.3) from objects on the ground and the ground
itself, it is necessary to be careful to use it when comparing it with the solar radiation amount for an air
conditioning design 6).

350 1100
H H
1000

Amount of day maximum solar radiation2 [W/m2]

solar radiation2[]W/m2]

300
900 E,W
250 800
[W/m ]

日最大日射量 [W/m ]
NE,NW
E,W 700
200
日平均日射量

NE,NW 600
SE,SW
500
150
SE,SW
S 400
Amount of daily average

S
100 300
N
200
50 N
100
0 0
20 25 30 35 40 45 50 20 25 30 35 40 45 50
緯度 [°N] 緯度 [°N]
latitude [°N]
図 8.2.8 日平均日射量基本値（夏期）
latitude [°N]
Figure. 8.2.3 Basic value of daily average Figure. 8.2.4 Basic value of daily maximum
solar radiation (summer) solar radiation (summer)

(3) Nighttime radiation

When calculating the heat transfer on the wall surface using the total heat transfer rate, it is essential
to use the ambient temperature, which does not consider the influence of air temperature but rather the
influence of radiation. Therefore, when the outdoor air temperature (air temperature) is used instead of
the ambient temperature, an error occurs, and thus the nighttime radiation (effective radiation) may be
taken into consideration as a revision.
Long wavelength radiation incident on an outer surface has a portion from the atmosphere and a
portion from objects on the ground and the ground itself (considering a horizontal surface without
obstacles, long-wave radiation will be equal to the portion from the atmosphere because there are
exchanges of such radiation with everything visible from the outer surface). Long wavelength radiation
radiated downward from the upper atmosphere, air molecules, and clouds is called atmospheric radiation.
In addition, the average radiation temperature in the sky is lower than the outdoor air temperature near
the ground surface, and thus it is lower than the long wavelength radiation radiated upward from the
ground surface. The amount of horizontal surface radiation at nighttime, Jnight,H [W/m2], is the difference
between the atmospheric radiation amount, σTo4[W/m2], and the actual atmospheric radiation amount L
- C8-10 - Recommendations for Loads on Buildings

[W/m2], when the average radiation temperature of the sky is regarded as the outdoor temperature, To
[K]. Here, Jnight,H is defined as follows.
𝐽𝐽night,H = 𝜎𝜎𝑇𝑇4o − 𝐿𝐿 (8.2.2)
The value decreases when the water vapor pressure is higher and there are fewer clouds. However,
when the weather is fine during the winter, the value increases considerably, occasionally reaching 150
W/m2. Although this is called nighttime radiation, it must be noted that it exists not only during the
nighttime but also in the daytime, and furthermore its direction is opposite the solar radiation (orientation
towards the sky is positive direction).
The sol-air temperature TSAT [°C] is defined as follows when considering nighttime radiation.
𝑎𝑎 𝜀𝜀
𝑇𝑇𝑆𝑆𝑆𝑆𝑆𝑆 = 𝑇𝑇o + 𝐽𝐽 − 𝜑𝜑S 𝐽𝐽night,H (8.2.3)
𝛼𝛼o 𝛼𝛼o
Herein, the outside surface temperature is To [°C], the amount of solar radiation incident on the outer
surface is J [W/m2], the solar radiation absorption rate is a, the external surface long-wave emissivity is
ε, and the shape factor when viewing the sky from the outside surface is φs (1 in the horizontal plane
and 1/2 in the vertical plane). The ground surface temperature is considered to be the same as the outdoor
air temperature. Considering the nighttime radiation, the outdoor air temperature will be considerably
low, and thus it does not need to be considered during the summer. However, the sol-air temperature
can be several degrees lower than the outdoor air temperature in the horizontal plane at night in winter,
and thus ignoring the radiation cooling completely may be a problem.

8.2.3 Earth temperature

Underground structures are affected by underground soil temperature. The ground surface is subject
to periodic fluctuations of outside air temperature and solar radiation by day and year, and the cyclic
influence is transmitted to inside the ground.
The depth of influence of a daily change is considered to be about 0.5 m, and the depth of influence
of a yearly change is considered to be about 10 m.
In this section, we show the earth temperature based on a semi-infinite solid theoretical solution.

(1) Semi-infinite solid theoretical solution

When the ground surface temperature fluctuates at 𝜃𝜃 cos(𝜔𝜔𝜔𝜔), the earth temperature 𝜃𝜃𝑥𝑥 (𝑡𝑡) at an
arbitrary depth (x) and time (t) is expressed through the following equation.
𝜔𝜔 𝜔𝜔
−𝑥𝑥�
𝜃𝜃𝑥𝑥 (𝑡𝑡) = 𝜃𝜃 ⋅ 𝑒𝑒 2𝑎𝑎 cos �𝜔𝜔𝜔𝜔 − 𝑥𝑥� � + 𝜃𝜃𝑀𝑀 (8.2.4)
2𝑎𝑎
where
𝜃𝜃: amplitude of ground surface temperature (°C),
x: depth (m),
t : time (h),
𝜔𝜔: angular frequency (rad/h),
CHAPTER 8 THERMAL LOADS - C8-11 -

a: thermal diffusivity (m 2/h),

𝜃𝜃𝑀𝑀 : average temperature (°C).
Figure 8.2.5 shows an example of the earth temperature distribution. The amplitude of the temperature
decreases with an increase in depth, and the time lags with the surface and inside are observed.

Temperature（℃）
0 5 10 15 20 25 30 35 40 45
May. Nov.
0
Dec. Jun.

2 Oct.
Apr. Sep.
Mar. Jan. Jul. Aug.
4 Aug.
Feb.
Sep.
Oct.
6
Depth(m)

Nov.
Dec.
8 Jan.
Feb.
Mar.
10 Apr.
May.
Jun.
12
Jul.

Figure. 8.2.5 An example of earth temperature (amplitude 𝜃𝜃 = 18.91 °C,

average temperature 𝜃𝜃𝑀𝑀 =19.21 °C, a = 0.0013 m 2/h)

(2) Practical calculation equation of earth temperature

Based on the semi-infinite theoretical solution, the calculation formulas considering the practicality
are shown below.
The practical calculations in Eqs. (8.2.5) and (8.2.6) show the amplitude (the envelope of the semi-
infinite solid theoretical solution) at high and low temperatures of the underground temperature, which
were obtained using a = 0.0026 m2/h as the thermal diffusivity of the ground in Eq. (8.2.4).
Figure 8.2.6 shows the earth temperature distribution during winter and summer.
𝑡𝑡𝑠𝑠 + 𝑡𝑡𝑤𝑤 𝑡𝑡𝑠𝑠 − 𝑡𝑡𝑤𝑤 −0.4𝑥𝑥
High temperature (summer) 𝑇𝑇𝑠𝑠 (𝑥𝑥) = + 𝑒𝑒 (8.2.5)
2 2
𝑡𝑡𝑠𝑠 + 𝑡𝑡𝑤𝑤 𝑡𝑡𝑠𝑠 − 𝑡𝑡𝑤𝑤 −0.4𝑥𝑥
Low temperature (winter) 𝑇𝑇𝑤𝑤 (𝑥𝑥) = − 𝑒𝑒 (8.2.6)
2 2

where
𝑇𝑇𝑠𝑠 (𝑥𝑥): underground temperature (°C) at the depth of x(m) from the ground surface,
𝑡𝑡𝑠𝑠 : monthly average temperature of the daily highest temperature of the hottest month (°C),
𝑡𝑡𝑤𝑤 : monthly average temperature of the daily lowest temperature of the coldest month (°C).
- C8-12 - Recommendations for Loads on Buildings

Temperature（℃）
-5 0 5 10 15 20 25 30 35
0

Depth(m) 4

6
Winter
8
Summer
10

Figure. 8.2.6 An example of earth temperature distribution using practical equations

(ts = 30.1 °C, tw = -1.4 °C, a = 0.0026 m2/h)

8.2.4 Indoor temperatures

The indoor temperature should be determined by evaluating the scheduled air-conditioning
environment. If such condition is unknown, the values in Table 8.2.3 are available. In general, air -
conditioning temperatures are set at 22 °C in the winter, 24 °C during intermediate periods, and 26 °C
in the summer. Recently, however, for energy saving, such temperatures are changing to 20 °C in the
winter, 24 °C during intermediate periods, and 28 ℃ in the summer. The “Act on Maintenance of
Sanitation in Buildings,” which provides guidance for specific buildings (offices, shops, meeting halls,
etc. having a total floor area of 3,000 m2 or more, and schools with a total floor area of 8,000 m2 or
greater) with air-conditioning facilities, regulates the indoor temperature from 17 °C to 28 °C. This can
be considered to be an allowable value of indoor temperature during a period of air-conditioning. For
schools with a total floor area of less than 8,000 m2, the “Standard for School Environmental Sanitation,”
recommends 10 °C or more in the winter, and 30 °C or less in the summer, whereas the most
recommendable temperatures are 18 ~ 20 °C in the winter, and 25 ~ 28 °C in the summer. Table 8.2.4
defines the seasons in each region.
CHAPTER 8 THERMAL LOADS - C8-13 -

Table 8.2.3 Air-conditioned indoor temperature

summer season intermediate winter season
General bldg. (offices etc.) 26-28℃ 24℃ 19-22℃
General bldg. (allowable value)* 28℃ - 17℃
School (recommendation)** 28℃ - 18℃
School (allowable value)** 30℃ - 10℃
Industrial bldg. (factories etc.) 28℃ 24℃ 20℃
* Act on Maintenance of Sanitation in Buildings
** Standard for School Environmental Sanitation

Table 8.2.4 Definitions of seasons

summer season intermediate winter season
Nov., Dec., Jan.,
Hokkaido Jul., Aug., Sep. May, Jun., Oct.
Feb., Mar., Apr.
May, Jun., Jul.,
Kagoshima, Miyazaki, Naha Apr., Nov., Dec. Jan., Feb., Mar.
Aug., Sep., Oct.
Jun., Jul., Aug., Apr., May, Oct., Dec., Jan., Feb.,
Others
Sep. Nov. Mar.

As a “rare occurrence of a temperature change,” it is also important to consider the indoor temperature
under the conditions in which the building’s air-conditioning does not function as expected. Indoor
temperature during a period of air-conditioning for humans does not largely deviate from the values in
Table 8.2.3, but without air-conditioning, the indoor temperature greatly depends on the buildings
specifications. For example, the indoor temperature of a building with large openings, like a greenhouse,
can be several tens of degrees higher than the outdoor air temperature, but the indoor temperature of an
externally insulated reinforced concrete building with a small number of openings could be relatively
stable, namely, almost the same as the daily average outdoor air temperature. Without the thermal
environment simulation, it is difficult to estimate the indoor temperature when air-conditioning does not
function for a long period of time, particularly during summer or winter holidays. If there are no
simulation results, the only way to evaluate the indoor temperature is to use the daily maximum ambient
temperature as the highest value, and the daily minimum ambient temperature as the lowest value.
However, it should be noted that the indoor temperature will not be below the daily minimum ambient
temperature, but it can be above the daily maximum ambient temperature.
The values in Table 8.2.3 are the exact temperatures in habitable areas. The actual temperatures in a
space are not uniform, and because the temperature distributions are uneven in the vertical direction,
particularly in large spaces such as atriums, they cannot be ignored. The temperature of the heat sink at
the top of an atrium will depend on the finishing, but it is not rare for the temperature to reach 60 °C
during summer days. An evaluation of the vertical temperature distribution must be based on a numerical
- C8-14 - Recommendations for Loads on Buildings

simulation, but “The Society of Heating, Air-Conditioning and Sanitary Engineers of Japan” presented
a program called “DAIKUKAN,” which can help evaluate the results relatively easily.

8.2.5 Other temperatures

Other temperatures include temperatures related to nonstructural members that are affected more than
the structure itself, and temperatures from internal heat sources such as chimneys, silos, thermal storage
tanks, freezing warehouses, and nuclear power plants. These temperature conditions are determined
based on their respective uses, efficiency, and use restrictions.

(1) Nonstructural members

1) Roofing materials
Because roof materials are strongly influenced by solar radiation, and metal roofs are generally small
in terms of heat capacity and their temperature easily changes, it is assumed that their temperature
change throughout the year or day is greater than that of the other members. Therefore, it is necessary
to take appropriate measures against thermal deformation of the roofing material according to various
conditions such as the local conditions, size of the roof, fixation degree of the roof material, and the
frame.
It is necessary to pay attention to the fact that the temperature of the roof surface is higher than the
outside air temperature owing to the influence of solar radiation during summer days, and conversely it
is lower than the outside air temperature during nighttime radiation cooling in winter. For example,
when the temperature of a double-folded plate roof is measured during the summer7), the roof surface
temperature of the upper folded plate may reach a maximum of 70 °C or more, and the daylight
difference may reach a maximum of about 50 °C in certain cases. In addition, it is necessary to pay
attention to cases in which the relative displacement between the roof material and the frame becomes
large, as described below.
a) In the case of a metal roof, the temperature of the roofing material may increase by about 10 °C to
20 °C within a short period of time (about 20 to 30 min) after receiving sunshine after sunrise, and
thus the relative displacement of the roofing material and frame may temporarily become large in
certain cases.
b) In a double folded plate roof, which improves the heat insulating properties of a roof, the
temperature of the plate is liable to increase and fall owing to the influence of solar radiation, wind,
snow, and other elements. On the other hand, because the temperature of a downward folding plate
and frame is close to the room temperature owing to the effect of the heat insulating material, a
large temperature difference may occur between the upper folded plate and frame.
The load effect from a temperature change of the roofing material is restrained by the thermal
expansion and contraction of the roofing material owing to the temperature difference between the
roofing material and frame fixing the roofing material, as well as the relative displacement caused by
the difference in the linear expansion coefficient of the material used. As a result, there have been
CHAPTER 8 THERMAL LOADS - C8-15 -

reported inconveniences such as sounds from the buckling of the roof folding board and squeaks or
breakage caused by a slipping of the heat insulating fittings, called a hanging or saddle, supporting the
double folded board roofing. In addition, regarding the damage caused by Typhoon No. 16 of 2014, the
roof of an insulated double folded sheet made of steel was scattered over a very wide area8) because the
roof fixing bolt was damaged by fatigue owing to thermal expansion and contraction from solar radiation,
and it has been clarified that wind resistance was not achieved.
As a treatment against thermal deformation of a roofing material, the following method can be
considered.
a) Method for escaping the deformation of the roofing material (installing an expansion joint,
installing a roller support allowing the roofing material to be movable)
b) Method of restraining deformation (increase the strength of the base material)
c) A method using a material with a low coefficient of linear expansion for the roofing material
d) A method confirming that no malfunction will occur during the service period by appropriately
evaluating the load effect accompanying the temperature change, as well as the durability
performance (mainly fatigue resistance) of the roof against a temperature change
However, there are not many actual measurements of temperature fluctuation for roofing materials,
and the mechanism of the loading effect accompanying the temperature change is also unknown in many
parts. Therefore, in general, the temperature condition and its coping method are set based on the
experience of the designer.
The method for considering the average daily difference based on the number of days when setting
the loading effect accompanying a temperature change is convenient, and is considered to be the most
frequently used method at present. However, metal roofing materials are strongly influenced by solar
radiation, as well as wind and snow, and the temperature increases and decreases frequently even during
the day. As the following actual measurement shows, there is a risk of underestimating the actual
situation when using the method considering the load effect only from the difference in day and the
number of days. It is thought that it is essential to appropriately evaluate the increase and decrease in
temperature during the day as a load effect in order to reduce problems with the roof material.

(Appendix) Example Temperature Measurement of Double Folded Plate Roof9)

The object of the measurement is a heat-insulating double-folded plate roof (gal barium steel plate)
of a warehouse located in Koto Ward, Tokyo (steel frame, five floors above the ground, with an ALC
exterior). The height of the roof surface is 28 m above the ground. The roof shape is a gable type, and
the length of the folded plate is 27 m on one side. As shown in Fig. 8.2.7, the temperature measurement
position is located at two places, 5 (point A) and 11.7 m (point B) from the roof ridge. At point A, the
surface temperature and strain on the outer side of the upper folded plate located at ① to ③ are
measured, and at point B, the strain located at ① is measured. In this actual measurement example, the
measurements were carried out for 7 to 13 days during the spring, summer, autumn, and winter,
respectively, and the measurement results from summer are shown in Fig. 8.2.8. The main points of the
- C8-16 - Recommendations for Loads on Buildings

Folded Plate Temperature Outside Temperature

 Strain Sunshine houres
60 120


Temeratue (℃)、Strain (×10-6)

Sunshine Houres （min./1hour）

40 100


20 80






 0 60


-20 40



-40 20
 


-60 0
8/7 8/8 8/9 8/10 8/11 8/12 8/13 8/14 8/15 8/16 8/17 8/18
Measurement Date

 
Figure. 8.2.8 Temperature history (August 7～18)

Foldede Plate Temperature
Outside Temperature
60 Sunhine houres 120
② ③

Sunshine Houres （min./1hour）

① 

 August 11
50 100
Temperature (℃)


 

 40 80

30 60

20 40

a-a Section 10 20

0 0
Figure. 8.2.7 Measurement 0 4 8 12 16 20 24
Times of Day

Figure. 8.2.9 Temperature history (1 Day)

measurement results are summarized below.

a) The temperature of the folded plate rapidly changes with or without solar radiation, and the folding
plate temperature during the day without nighttime or solar radiation is approximately equal to the
outside air temperature. The daily range is almost constant except for in autumn.
b) For the summer measurement, the difference in strain on the day of measurement was about 50 ×
10-6, which is about 20% of the complete restraint.
c) Even during the day, as shown in the lower part of Fig. 8.2.9, the folded plate temperature repeatedly
increases and decreases in small increments.
d) Assuming a law of linear damage and counting the increase and decrease in temperature of the
folded plate using a rain flow method, it is considered that on August 11, a change in the folded
plate temperature of 11.2 ° C. was repeated 19 times in a single day. This had a loading effect of
about 10 fold higher compared with an evaluation by multiplying the daily temperature difference
by the number of days.
CHAPTER 8 THERMAL LOADS - C8-17 -

2) Curtain wall
As with roofing materials, curtain walls are strongly influenced by the outside air temperature and
solar radiation. “JASS14 Curtain Wall ” 10) construction work for construction work organizes the
effective temperature difference based on the annual temperature difference of members based on the
type of component and the color tone of the surface. The amount of thermal deformation of the curtain
wall can be calculated based on the linear expansion coefficient of the constituent member and various
conditions concerning the color and direction. On the other hand, a precast concrete curtain wall is
generally designed to support four points in the plane, two points of which are pin supported, and the
other two are pin roller supported. Therefore, deformation of a frame at the time of an earthquake and
deformation from a temperature change of the curtain wall itself are absorbed by the surrounding
fasteners, and thus the curtain wall itself is not affected. Regarding the out-of-plane deformation, besides
the wind load, warp deformation or other event from a temperature gradient occurring in the curtain wall
may be considered. However, even in such a case, it seems that it is common to ignore the load effect
accompanying the temperature change by choosing not to constrain the deformation of the curtain wall
by effectively functioning the movable mechanism of the fastener part.
Therefore, with respect to a curtain wall of a standard shape and size, it is considered unnecessary to
conduct a special study on the temperature load owing to the influence of outside air temperature or
solar radiation. However, with regard to the fastener portion and the joint portion (joint portion), care
must be taken such that the deformation of the member can be sufficiently absorbed. In addition, when
the size of the curtain wall is large and it is supported at three points (six points supported on the plane)
in order to cope with the wind load, it is necessary to determine that no harmful deformation occurs in
the curtain wall. Furthermore, even if the support point is movable, there has also been reported a
problem in which, as a result of excessive tightening of the mounting bolt at the time of construction, a
sound is generated from the bolt part owing to deformation of the curtain wall. Furthermore, even if the
supporting point is movable, as a result of excessive tightening of the mounting bolt at the time of
construction, it is reported that a sound is generated from the bolt owing to the deformation of the curtain
wall, and consideration of construction control is necessary.

3) Glass and sash

In a glass sheet used for an exterior, a breakage phenomenon called thermal cracking may occur owing
to the temperature difference inside the glass plate caused by the influence of solar radiation. In this
mechanism, deformation from thermal expansion accompanying a temperature increase of the glass
surface receiving solar radiation is caused by a constraint of the low temperature parts such as shaded
parts and parts below the sun line. That is, when the bending tensile stress generated in the peripheral
portion of the glass exceeds the edge strength of the glass, thermal cracking occurs. In particular, it is
known that heat ray absorbing glass and net-containing glass are prone to thermal cracking, and heat
cracking is likely to occur during sunny weather in the winter. The main factors affecting the occurrence
of thermal cracking are as follows, and appropriate consideration for these various factors is necessary.
- C8-18 - Recommendations for Loads on Buildings

a) Sunshine (shadow) conditions

b) Size and shape of the glass
c) Type of glass (heat absorbing glass, netted glass, etc.)
d) Glass construction condition (cutting surface precision, etc.)
e) Presence of curtains and blinds on the interior side
f) Types of sashes to use (sash color, heat capacity)
g) Installation status (clearance of glass and sash, etc.)
h) Sealing material (insulation performance between glass and sash)
For exterior sashes, joints are generally designed to follow the expansion and contraction deformation
associated with temperature changes, although the occurrences of mitral curvatures and buckling from
thermal expansion have also been reported. In many cases, problems often occur when sashes are
attached directly between long pillars of spans, under bouncing slabs, or under slabs.
As other troubleshooting examples, cases have been reported in which parts are scratched owing to
thermal expansion and contraction of the sash, and sounds are generated.
For example, in the case of an aluminum sash, the coefficient of linear expansion is about twice as
large as that of iron or concrete, and thus a loose (long) hole is generated for the joint clearance and
screw fastening part allowing it to follow the thermal expansion and contraction beforehand, and
measures such as inserting a sliding material into the bolt clamping part and load receiving part are taken.
Particularly in the case of an aluminum sash, because the coefficient of linear expansion is about twice
as large as that of iron and concrete, such considerations as making the clearance of the joint part loose
(a long hole) so that it can follow the thermal expansion and contraction, or inserting a sliding material
into the bolt tightening part and load receiving part, are taken. However, on the contrary, slipping may
be a source of sound, and it is thought that a sash has many metal parts that touch as one factor of sound
generation. In addition, problems may occur owing to thermal deformation of a sash caused by a
breakage of the waterproof seal, and determining the joint width when taking into consideration the
deformation of the sealing material is also essential.

(2) Building with internal heat source

1) Chimney
Because high-temperature flue gas flows inside a chimney, it is necessary to take appropriate
countermeasures according to the flue gas temperature for a structural framework under the influence
of thermal heat for long periods of time. Therefore, it is common to apply a lining inside the chimney,
the main purpose of which is as follows.
a) Act as a protective layer to prevent thermal exposure, corrosion, and deterioration of the structural
framework
b) Act as a heat insulation layer to prevent a reduction of exhaust gas temperature and improve the
diffusion effect.
Because the flue gas smoke temperature in the chimney greatly differs depending on the substance
CHAPTER 8 THERMAL LOADS - C8-19 -

burned, it is necessary to select the type and thickness of the lining after fully determining the flue gas
temperature.
The design standard for the temperature load of reinforced concrete chimneys and steel chimneys, is
stipulated in “Structural Design Recommendation for Chimneys”11) from the Architectural Institute of
Japan, and is as follows:
a) Reinforced concrete chimney
・Construct the lining so that the surface temperature of the cylinder does not exceed 100 °C.
・In principle, when the temperature difference between the inside and outside surfaces of the
cylinder exceeds 20 °C, the thermal stress is appropriately evaluated based on the thermal
characteristics and material properties of the material, and a sound cylinder is designed.
b) Steel chimney
・In principle, the lining must be constructed so that the surface temperature of the cylinder does
not exceed 165 °C.
・When painting is applied to the cylinder, it is necessary to consider the heat resistant temperature
of the paint.
・For a steel tower support and multi-leg type chimneys, they must be constructed to be safe against
thermal expansion of the barrel.
For reference, examples of exhaust gas temperatures of various chimneys currently in use are shown
in Table 8.2.5.
2) Silo
In the design of concrete silos that store high-temperature items, such as cement silos and clinker silos,
the temperature stress owing to the temperature difference of the inner and outer surfaces of the cylinder
together with the powder pressure from the stored materials becomes a serious problem. According to
actual temperature measurements of concrete silos, the temperature inside the silo fluctuates mainly by
around 100 °C in the case of a clinker silo, temporarily exceeding 120 °C, and in the inside of the
cylinder, it has been reported that the external surface temperature difference reaches 40 °C. The

Table 8.2.5 Various chimney exhaust gas temperatures

Exhaust gas temperature

By Industry Equipment name
(℃)
Heavy oil fired power 90～130
Coal fired power 90～100
Power company
LNG thermal power 90～110
Gas turbine 400～500
Ironmaking furnace, Heating
General company 50～300
furnace, Reactor etc.
Municipality Garbage incinerator 160～210
- C8-20 - Recommendations for Loads on Buildings

temperature condition of the silo is generally given as the temperature of the storage material, and there
are cases in which the temperature is 60 °C in the case of a cement silo, and 120 °C in the case of a
clinker silo.
In the design of concrete silos, it is necessary to consider the temperature load by the stored materials
as the design condition, as well as the hoop tension from the powder pressure.
3) Heat storage tank
Regenerative air conditioning systems include on-site construction types, unit types, and multi split
types. In the case of a building with a space (underground double slab space) partitioned by an
underground beam between the floor slab at the bottom floor and the pressure-resistant board, heat
insulation waterproofing is performed in this space, and space is often used as an on-site construction
type heat storage tank that uses cold/hot water as a heat storage medium. The temperature range of use
of a thermal storage tank is generally 5 °C for cold water and around 45 °C for hot water. On the other
hand, a tank may be used at a temperature below freezing as an ice thermal storage system, and within
a temperature range of around 80 °C for hot water storage. Figure 8.2.10 shows examples of temperature
changes in a warm water bath and an ice thermal storage tank during a 1-day period.
In the case of an on-site construction type heat storage tank utilizing an underground double slab
space, it is considered that the thermal influence on the body of the building from heat storage such as
cracks in the building concrete is usually alleviated by the heat insulating effect of a heat insulating
waterproof layer. However, when a cold water tank and hot water tank are adjacent to each other, or in
the case of a cold/hot water tank, an appropriate consideration is also required for the body of the
building. Particularly in areas where spring water is generated, spring water from outside of the tank,
such as groundwater, flows from the cracked part of the concrete and the joint part, which may damage
the heat insulating waterproof layer, and thus it is necessary to pay attention to such aspects.

50 15

48 10
温度（℃) (℃)
温度（℃) (℃)

Temperature
Temperature

46 5

44 0

42 -5

40 -10
0 4 8 12 16 20 24 0 4 8 12 16 20 24
時刻(時)
Times of Day 時刻(時)
Times of Day

Figure. 8.2.10 Example of temperature change during 1 day of thermal storage tank

To prevent cracks owing to changes in water temperature, it is important to keep the flow of water
constant so that the water temperature in the tank is made uniform, and no stagnation or short circuit
occurs. It is also important to provide communication pipe positions shifted up and down, and right and
left, for each tank, and to achieve a secure flow path of as long as possible.
CHAPTER 8 THERMAL LOADS - C8-21 -

4) Freezer
In a freezer, proper heat insulation and warmth not only contribute to securing and maintaining the
proper temperature inside the refrigerator, thereby contributing to energy conservation, they also have a
structural significance such as the prevention of frosting. Therefore, it is vital to examine the influence
of heat from the basic planning stage. The storage temperature zone of a commercial refrigerated
warehouse is classified as shown in Table 8.2.612).
The influence of thermal stress from the cold of the freezer on the structural frame and non-structural
member is important, and appropriate measures are required.
For example, in an examination of the temperature load for a 4-story wall-type RC construction
refrigeration warehouse, based on the measurement results of each floor part when cooling from room
temperature (20 ° C.) to -30 ° C, it has been pointed out that cracks occur on the first and second floors
from a deformation restraint by the foundation beam.
In addition, in the freezer, it is necessary to pay particular attention to “frosting” in parts in contact
with the surface and soil. As a typical example, if a freezer is installed on the floor in contact with the
ground surface, the cold of the freezer cools the soil through the frame, and the soil freezes and expands
with moisture, causing problems such as a swelling of the floor. To avoid frosting, a method for
providing an air layer between the freezer on the first floor and the ground surface with a high floor type
structure is generally reliable. However, simply raising the freezer to a higher floor may not be sufficient,
and it is important to ensure under-floor ventilation such that cold air is not stored under the floor. In
addition, when it is necessary to set a freezer on a solid first floor, it is necessary to increase the heat
insulating property of the floor and take measures against freezing, such as installing a vent pipe in the
ground and flowing hot air.

Table 8.2.6 Classification and storage temperature of freezing and refrigerating warehouse

By refrigerator Storage
class Temperature (℃)
C3 class +10 ～ -2
C2 class -2 to -10
C1 class -10 to -20
F1 class -20 to -30
F2 class -30 to -40
F3 class -40 to -50
F4 class -50℃ or less

Note that when the concrete temperature falls below the freezing point, depending on the temperature
and moisture content, its mechanical and thermal properties may be significantly different from those at
ordinary temperature, and thus care must be taken at the time of the design. The strength and thermal
properties of concrete at low temperature are described elsewhere13).
- C8-22 - Recommendations for Loads on Buildings

It is also necessary to pay attention to the gas pressure weld joint performance of steel bars at low
temperature and the weldability of the steel materials. Design and construction guidelines14) for gas
pressure welded joints when the structure temperature is expected to drop to -50 °C during use are issued
by the Japan Reinforcing Bar Joints Institute.
The properties of concrete and reinforcing bars at low temperature are summarized in the attached
document of the Japan Society of Civil Engineering, “Standard Specifications for Concrete Structures:
Design”15), which is a good design reference.

8.2.6 Reference temperature

To obtain the thermal stress of a building, the thermal load surrounding the building is first set. The
thermal load is expressed as a difference from a certain reference temperature. The reference temperature
is the temperature at which the load effect by the temperature of the structural member or the secondary
member becomes zero. Therefore, the reference temperature can also be said to be the temperature at
the time the building is constructed. For example, in the case of a steel frame construction, it can be
considered with reference to the outside air temperature at the time of the final bolt tightening. In the
case of a reinforced concrete construction, it can be thought of as the average outdoor air temperature
of the construction period in the construction area. However, because the construction period generally
varies from several months to several years, and the outside air temperature also changes, it is necessary
to set a certain degree of leeway when designing the reference temperature. When the construction time
is clear, for example, in the short term during summer or winter, the temperature during that period can
be considered as the reference temperature.
When conducting a temperature stress analysis, for example, when the foundation is a roller type and
there is no constraint from the outside, the structure generates stress only from the relative temperature
difference between the structural members. At this time, the reference temperature is irrelevant to the
stress and has no meaning. However, because the deformation occurs as it is, the annual average
temperature is taken as the reference temperature.
However, when restrained from the outside (ground), stress is generated from the temperature
difference. When in contact with the ground and receiving restraint, a reference temperature is required.
In this case, it is conceivable to set the ground temperature (underground temperature) at the time of
construction as the reference temperature. The underground temperature fluctuates around the annual
average temperature, and the temperature of the constant-earth temperature layer (depth of about 10 m)
almost agrees with the annual average temperature of the area. Considering that the underground
temperature of the restraint side (ground) also fluctuates when the construction time is long, and that
uncertainties in the degree of constraint of the building and ground are not clear, it is considered that the
average annual temperature of the construction site may be used. In the design standards of civil
engineering structures, the reference temperature is set at around the annual average temperature.
CHAPTER 8 THERMAL LOADS - C8-23 -

8.2.7 Load combination with other loads

When considering the thermal load, consider a combination with other loads appropriately according
to the situation assumed. The thermal load is constantly affecting the building as it changes throughout
the year, and can be regarded as a “constantly acting load” similar to dead and live loads. However, the
temperature load is a self-constraining stress that is generated depending on the rigidity of the structure
and constraint conditions. Therefore, even in combination with other loads, it is also necessary to
properly consider the influence of a rigidity reduction owing to cracking and plasticization.
When assuming the thermal load as the primary load, consider a combination with the constant load
(dead or live load). Evaluate the load factor of thermal load when assuming a snow load, wind load, and
seismic load as the primary loads. As a method of evaluation, “Turkstra's rule of thumb”16) (Method 1)
and a method for evaluating the probability distribution of the combined load effects are used17) (Method
2). Figures 8.2.11 and 8.2.12 are examples in which the combined load is evaluated according to Method
2. The points plotted in the figure represent the combined load effects of snow, wind, and thermal loads
based on the daily maximum weather data from 1961 to 2012 in three cities. The 100-year recurrence
load effect of each load and the combined load is indicated through the broken line.
At the point at which the combined load intersects with the primary load (the point indicated by ○
in the figure), the combined load is the basic value of the primary load (100-year recurrence) plus the
combined thermal load. The value obtained by dividing the combined thermal load by the basic value
of the thermal load (100-year recurrence) becomes the load coefficient of the thermal load. In Section
2, we compare the load factor of the temperature load evaluated using this method (Method 2) and
Method 1. In addition, for the sake of easier handling during the design, the value of the load coefficient
evaluated on the safety side is shown.
Sapporo Asahikawa Takada
Load effect of snow load Load effect of snow load Load effect of snow load
(100-year-reccurence) (100-year-reccurence) (100-year-reccurence)
Combined load effect Combined load effect Combined load effect
(100-year-reccurence) (100-year-reccurence) (100-year-reccurence)
Load effect of snow load(kNm)

Load effect of snow load(kNm)

Load effect of Load effect of Load effect of

thermal load thermal load thermal load
(100-year- (100-year- (100-year-
reccurence) reccurence) reccurence)

Load effect of thermal load (kNm) Load effect of thermal load (kNm) Load effect of thermal load (kNm)

Figure. 8.2.11 Distribution of the load effect of snow and thermal loads
- C8-24 - Recommendations for Loads on Buildings

Sapporo Tokyo Naha

Load effect of wind load Load effect of wind load Load effect of wind load

Load effect of wind load(kN)

(100-year-reccurence) (100-year-reccurence) (100-year-reccurence)
Load effect of wind load(kN)

Combined load effect Combined load effect

(100-year-reccurence) (100-year-reccurence) Combined load effect
(100-year-reccurence)

Load effect of Load effect of Load effect of

thermal load thermal load thermal load
(100-year- (100-year- (100-year-
reccurence) reccurence) reccurence)

Load effect of thermal load(kN) Load effect of thermal load(kN) Load effect of thermal load(kN)

Figure. 8.2.12 Distribution of the load effect of wind and thermal loads

References

1) Kushiyama, S., Obata, M.: Temperature for thermal stress analysis of reinforced concrete buildings,
Journal of Structural and Construction Engineering (Transactions of AIJ), No. 462, pp. 91–99, 1994.8
(in Japanese)
2) The Architectural Institute of Japan: Expanded AMeDAS Weather Data, 2000.1 (in Japanese),
Maruzen
3) Gueymard, C.: Parameterized transmittance model for direct beam and circumsolar spectral
irradiance, Solar Energy, 71-5, pp. 325-346. 2001
4) ASTM G 159, Standard Tables for Reference Solar Spectral Irradiance at Air Mass 1.5: Direct Normal
and Hemispherical for a 37° Tilted Surface, 1999
5) Perez, R., Ineichen, P., Seals, R., Michalsky, J. and Stewart, R.: Modeling daylight availability and
irradiance components from direct and global irradiance, Solar Energy, 44-5, pp. 271-289, 1990
6) The society of Heating, Air-conditioning and Sanitary Engineers of Japan, The Handbook of Heating,
Air-conditioning and Sanitary Engineers Vol.3 (12th Edition), pp. 25-27, 1997
7) Japan Metal Roofing Association: Standard of Steel Roofing SSR2007, 2007.11
8) Okuda, Y: Building damage caused by recent typhoon, The Kenchiku Gijutsu, No. 673, pp. 92-95,
2006.2
9) Tokinoya, H., Suzui, Y., Asai, H.: Effect of cyclic load on wind bearing capacity of metallic roof and
wall (part1) measuring steel roof temperature, Summaries of Technical Papers of Annual Meeting
Architectural Institute of Japan, C-1, pp. 921-922, 2005.9
10) Architectural Institute of Japan: Japanese Architectural Standard Specification JASS14 Curtain Wall,
1996
11) Architectural Institute of Japan: Structural Design Recommendation for Chimneys, 2007
12) Japan Association of Refrigerated Warehouses: Functions of refrigerated warehouse and logistics
service ([Link]
CHAPTER 8 THERMAL LOADS - C8-25 -

13) Japan Reinforcing Bar Joints Institute: Design and Construction Guidelines for Rebar Gas Pressure
Weld Joint for Low Temperature Structure, 1985
14) Japan Society of Civil Engineering: Standard Specifications for Concrete Structures, Design, 1996.
15) Japan Road Association: Specification for Road Bridge, 2013.7
16) Turkstra, C.J. (1970) Theory of structural design decision, Study No. 2. Solid Mechanics Division,
University of Waterloo, Waterloo, Ontario, Canada.
17) Nakashima, H., Takada, T., Ishikawa, T.: A study on the combination of snow
load and thermal load, AIJ Journal of Technology and Design, Vol. 21, No. 40, 931-936, Oct., 2015
(in Japanese)
- C9-1 -

CHAPTER 9 EARTH AND HYDRAULIC PRESSURES

Introduction

This chapter deals with earth and hydraulic pressures acting on the exterior basement walls of
buildings and retaining walls. These topics are represented consciously with the concept of the
basic load values that reflect the limit state design method, which have already been described
in the AIJ Recommendations for the Design of Building Foundations (2001) (hereafter referred
to as RDBF).
In general, earth pressure acting on an exterior basement wall is assumed to be earth
pressure at rest. At depths below the groundwater level, the influence of hydraulic pressure is
also considered, and if there is a surcharge on the ground surface, its influence is also taken into
account. In the design of an exterior basement wall, the serviceability limit states under earth
and hydraulic pressures that are acting continuously on the wall are taken into consideration.
The earth pressure may increase under the influence of the dynamic interaction between the soil
and wall during an earthquake. However, because experience has shown that significant damage
does not occur even during an earthquake if the walls are designed by taking into account the
serviceability limit states under the continuous earth and hydraulic pressures, checks on the
ultimate limit states under an earthquake loading may be omitted.
In general, earth pressure acting on a retaining wall is assumed to be active earth pressure,
and is determined by taking into account the influence of a surcharge on the ground surface
behind the wall and the influence of hydraulic pressure (if any exists). When designing a
retaining wall, the serviceability limit states under the continuous earth and hydraulic pressures
and the ultimate limit states under active earth pressure during an earthquake are examined. The
permanent earth pressure is calculated using Coulomb’s active earth pressure, and the earth
pressure during an earthquake is calculated using, as an example, Mononobe and Okabe’s
formula for active earth pressure during an earthquake.
The basic load values used in these guidelines correspond to a return period of 100 years or
to a 99% probability of non-exceedance. The basic values of hydraulic pressure dealt with in
this chapter correspond to a return period for the annual highest free water level of 100 years,
and such values are determined by appropriately taking into account the time-dependent
variations.
A continuous earth pressure varies slightly over time, but is strongly influenced by the
accuracy of the calculations and the uncertainty of the geotechnical parameters. It is therefore
necessary to determine the basic values corresponding to a 99% probability of non-exceedance,
when considering such uncertainty as a main factor. As an example, this chapter describes a
method for using the geotechnical parameter values at the site under consideration,
corresponding to a 99% probability of non-exceedance. In reality, however, a sufficient amount
- C9-2 - Recommendations for Loads on Buildings

of data is often unavailable. In such cases, the use of empirical values as the basic values is
permitted.

Symbols

The main symbols used in this chapter are as follows:

Capital letters
K0: coefficient of earth pressure at rest
KA: coefficient of active earth pressure
KEA: coefficient of active earth pressure during earthquake

Small letters
c: cohesion of soil (kN/m2)
c': cohesion of soil (kN/m2, expressed in terms of effective stress)
kh: design lateral seismic coefficient
h: depth of groundwater level depth (m)
p: earth pressure per unit area at depth z (kN/m2)
Δp: increase in earth pressure owing to surcharge on the ground surface behind
a retaining wall (kN/m2)
q: uniformly distributed load acting on the ground surface behind a retaining
wall (kN/m2)

Greek letters
α: inclination of ground surface behind a retaining wall (deg)
γt: wet unit weight of soil (kN/m3)
γ': submerged unit weight (kN/m3)
γw: unit weight of water (kN/m3)
θ: angle between the back of a retaining wall and the vertical plane (deg)
θk: resultant incidence angle of earthquake (＝tan−1kh, deg)
δ: friction angle of retaining wall (deg)
ϕ: internal friction angle of soil (deg)

9.1 Overview
(1) Earth and hydraulic pressures
In the design of an exterior basement or retaining wall, earth pressure reflecting the wet unit
weight of the soil is assumed to be the load at depths above the groundwater level, and the earth
and hydraulic pressures reflecting the submerged unit weight of the soil are assumed to be a
CHAPTER 9 EARTH AND HYDRAULIC PRESSURES - C9-3 -

load at depths below the groundwater level. In the case of a building with a basement, if the
bottom of the foundation is below the groundwater level, buoyancy is considered as a load
applied to the underside of the foundation. This chapter describes the calculation methods and
concepts of the earth pressure, hydraulic pressure, and buoyancy.
As shown in Figure 9.1.1, the amount of earth pressure varies with the relative displacement
between the structure and soil. The active earth pressure occurs when the soil behind the
structure reaches a plastic state as it pushes the wall forward. Passive earth pressure occurs
when the soil behind the structure reaches a plastic state while being pushed back by the wall.
Both active and passive earth pressures occur in a state of plastic equilibrium in which the
strength of the soil behind the wall is fully exerted.
Earth pressure at rest occurs when the wall in contact with the soil is at rest. Earth pressure
at rest is an intermediate state between the active earth pressure and passive earth pressure. In
general, as a retaining wall is displaced under the influence of soil behind the wall, the earth
pressure decreases as the displacement progresses, and the state of the earth pressure at rest
tends to approach the state of the active earth pressure.
Theories proposed to calculate the earth pressure in a plastic state include those by Coulomb
and Rankine. The former derives the earth pressure from the equilibrium conditions by
assuming a limit state in which the soil behind the wall fails and a wedge of soil slips as the
wall is moved. The latter derives the earth pressure by assuming a plastic state that occurs when
the entire soil block behind the wall fails. Coulomb’s earth pressure theory is a practical method
because it takes into consideration such parameters as the friction between the wall and soil,
the inclination of the wall, and the inclination of the ground surface behind the wall. It is
generally known that the earth pressure values derived by Coulomb’s method agree with those
derived by Rankine’s method when the parameters take certain values.

Figure 9.1.1 Relationship between displacement and earth pressure

- C9-4 - Recommendations for Loads on Buildings

(2) Earth pressure acting permanently on exterior basement walls

It is common practice to examine the serviceability limit state of exterior basement walls by
considering the continuous application of earth and hydraulic pressures. Because the
displacement of the exterior basement walls is slight under normal conditions, the continuous
earth pressure is calculated as the earth pressure at rest.
When designing exterior basement walls, the possibility of ultimate limit states in which a
continuous earth or hydraulic pressure predominately occurs cannot be precluded. Ultimate
limit states of this type may be neglected, however, because not much is known about the
variability of the coefficient of earth pressure at rest, and because no major damage has been
experienced in the past when the serviceability limit states were examined against continuously
applied earth and hydraulic pressures.

(3) Earth pressure acting continuously on retaining walls

It is generally considered that, although the earth pressure acting on ordinary retaining walls
varies depending on the soil behind the walls, a transition from earth pressure at rest to active
earth pressure is caused by a very small amount of displacement. With sandy soil, for example,
a transition to active earth pressure can be caused by a very small forward horizontal
displacement of the top of the retaining wall equal to about only 1/1,000 of the wall height.
Even when the earth pressure increases and approaches that at rest owing to seepage flow, such
as rainwater infiltration occurring over a long period of time, the retaining wall tends to be
inclined again, resulting in an active earth pressure. When designing ordinary retaining walls,
it is therefore acceptable to examine the serviceability limit states by using the active earth
pressure as a continuously occurring earth pressure. However, if very little displacement of the
retaining wall under consideration is expected, as in the case of a retaining wall in a dry area
whose top and bottom are connected to a building, it is better to use the earth pressure at rest
instead of the active earth pressure.
In the design of a retaining wall, the possibility of an ultimate limit state in which the
continuously acting earth pressure and the hydraulic pressure are the principal loads cannot be
precluded. For the same reason as in the design of exterior basement walls, however, an
examination of the ultimate limit states in such cases may be omitted. In the case of a retaining
wall, it is important to examine the ultimate limit states, taking into consideration the influence
of earthquakes described in item (4) below.

(4) Earth pressure during earthquake

The earth pressure acting on the exterior basement walls during an earthquake may become
greater than the earth pressure at rest depending on the differences in the vibration properties
of the surrounding ground and the building. No measurement results or damage verifying such
behavior have been reported, however, and it is thought that ignoring the increase in earth
CHAPTER 9 EARTH AND HYDRAULIC PRESSURES - C9-5 -

pressure during an earthquake is not likely to cause any major problems if the cross sections
conforming to the conventional allowable stress design criteria for sustained loading are used.
When designing exterior basement walls in non-liquefiable soils, it therefore is acceptable to
examine the serviceability limit states against continuously applied earth and hydraulic
pressures and omit an evaluation of the ultimate limit states that take into consideration
increases in earth pressure during an earthquake.
In soils that can liquefy during an earthquake, the earth and hydraulic pressures acting on a
wall can increase considerably. In such cases, it is necessary to examine the ultimate limit states
by assuming that a soil–water mixture with a unit weight of 18 to 20 kN/m3 applies pressure on
the wall. When liquefaction occurs, buoyancy acting on the underside of the foundation
increases considerably. It is therefore necessary to examine the safety against buoyancy when
a structure is floating in a soil–water mixture with a high specific gravity.
Cases of major earthquake-induced damage to retaining walls not accompanied by
liquefaction of the soil behind the walls have been reported. There is therefore a need for an
evaluation of the ultimate limit states during a major earthquake, in addition to an examination
of the serviceability limit states under continuous earth and hydraulic pressures. In the case of
a retaining wall, the stability of the entire ground behind the retaining wall, including the wall
itself, is often an important consideration. In such cases, it is necessary to conduct a stability
evaluation using an appropriate method, such as a circular slip method, in addition to an
examination of the stability under the active earth pressure during an earthquake.

(5) Basic earth and hydraulic pressure values and their uncertainty
The earth pressure is calculated by appropriately taking into account the influence of the
calculation accuracy and the uncertainty of the geotechnical parameters (unit weight and
strength parameters of the soil). Fluctuations of the groundwater level are also taken into
account as appropriate.
This guideline uses the values corresponding to a return period of 100 years or the values
corresponding to a probability of non-exceedance of 99% as the basic load values. A continuous
earth pressure does not change over time, and its values are greatly dependent on the degree of
uncertainty determined based on the accuracy of the calculation methods and the variability of
the geotechnical parameters. When considering the earth pressure, it is therefore necessary to
determine the basic values corresponding to a probability of non-exceedance of 99%, assuming
that these types of uncertainty are primary factors.
As an example of such a method, Section 9.6 describes a method for using the geotechnical
parameter values for the soil at the considered site corresponding to a probability of non-
exceedance of 99%. In reality, it is often not possible to obtain a sufficient amount of
geotechnical data. In such cases, values recommended in the Recommendations for the Design
of Building Foundations (backfill-related parameter values shown for reference in RDBF Table
- C9-6 - Recommendations for Loads on Buildings

3.4.2) are used, for example. These values are considerably conservative and have been
calculated so that they can be used in cases in which detailed surveys are not conducted. For
the time being, these values are therefore assumed to be comparable to the values corresponding
to a probability of non-exceedance of 99%.
In the calculation of earth pressure during an earthquake, values corresponding to a return
period of 100 years are used as basic values, taking into consideration not only the degree of
uncertainty resulting from the accuracy of the calculation methods applied, and the variability
of the geotechnical parameters, but also changes caused by earthquake motions. It should be
kept in mind, however, that as explained in Section 9.4, highly accurate estimations of
earthquake motions of various sizes are not necessarily possible at the present time.
For example, the Mononobe–Okabe equation2, 3) is used to calculate the earth pressure
acting on the retaining walls during an earthquake. It should be noted, however, that the design
horizontal seismic coefficient substituted in the Mononobe–Okabe equation does not
necessarily correspond to the value obtained by dividing the maximum horizontal acceleration
amax by the gravitational acceleration g. The reason for this is that the accuracy of the calculation
method applied to the earth pressure during a strong earthquake is not sufficiently high.
Although the Mononobe–Okabe equation is used in building designs, in cases involving a
strong earthquake, the design horizontal seismic coefficient is determined by referencing the
values back-calculated from data on the damage caused by the Hyogo-ken Nanbu Earthquake,
as an example.
In view of these circumstances, it has been decided to permit the method for using empirical
values for the design horizontal seismic coefficient in the Mononobe–Okabe equation, and
assume that the active earth pressure calculated from the Mononobe–Okabe equation has been
obtained using a value corresponding to a return period of 100 years.
It should be noted, however, that the concepts regarding the basic values of the earth
pressure recommended in these guidelines are provisional, and are intended for use prior to the
transition to the limit state design method. Depending on the accumulation of research findings
in the coming years, it will be necessary to create a more straightforward and clear-cut method
for calculating the basic values of the earth pressure.
For a method regarding such hydraulic pressure calculation, refer to Section 9.5.

9.2 Earth and Hydraulic Pressures Acting on Exterior Basement Walls

As shown in Figure 9.2.1, the earth pressure acting on the exterior basement walls is
assumed to be earth pressure at rest, and the hydraulic pressure is also taken into consideration
at depths below the groundwater level. If there is a surcharge on the ground surface, its effect
is also taken into account.

(1) Coefficient of earth pressure at rest, K0

CHAPTER 9 EARTH AND HYDRAULIC PRESSURES - C9-7 -

The use of the results of field measurements and laboratory tests on the collected samples has
been considered for a determination of the coefficient of earth pressure at rest, K0. Although the
amount of field measurement data obtained for the determination of the coefficient of earth
pressure at rest is limited, a substantial amount of data indicate a value of around 0.5.
Laboratory tests conducted to measure the coefficient of earth pressure at rest include
compression tests, as well as triaxial compression tests in which lateral displacement is
restricted. The coefficient of earth pressure at rest varies depending on such factors as the soil
type, strength, and stress history. However, it has been frequently reported that under a normal
consolidation, the measured values for both sand and clay show fairly good agreement with the
values given by the equation proposed by Jàky.4)

K 0 = 1 − sin φ ′ (9.2.1)

Figure 9.2.1 Earth pressure and hydraulic pressure acting on exterior basement wall

In this equation, if the internal friction angle ϕ' expressed in terms of effective stress is 30°, then
K0 is 0.5. Equation (9.2.1) was obtained by solving the stress at rest in the central vertical plane
of a two-dimensional triangular fill, and simplifying the equation obtained. At present, data that
clearly show the relationship with the actual state of exterior basement walls are lacking
therefore, although some theoretical bases do exist. In view of these circumstances, it is deemed
reasonable to use a value of approximately 0.5 for the coefficient of earth pressure at rest for
both sand and clay, except in cases in which highly reliable laboratory tests or field
measurement results are available.
The above discussion concerns the coefficient of earth pressure at rest when the ground
surface behind the wall is horizontal. For cases in which the ground surface behind the wall is
inclined, the following equation has been proposed5):
- C9-8 - Recommendations for Loads on Buildings

K 0α = K 0 (1 + sin α ) , (9.2.2)

where K0α is the coefficient of earth pressure at rest K0 for an inclined backfill, K0 is the
coefficient of earth pressure at rest for normally consolidated soil, and α is the inclination of the
backfill surface (deg).
Current knowledge regarding the nature of the coefficient of earth pressure at rest includes
the following. It is generally known from laboratory test results that the coefficient of earth
pressure at rest, K0u, for over-consolidated soil can be estimated using the following equation6):

K 0u = K 0 (OCR ) sin φ ′ , (9.2.3)

where K0u is the coefficient of earth pressure at rest for over-consolidated soil, K0 is the
coefficient of earth pressure at rest for normally consolidated soil, ϕ' is the internal friction
angle expressed in terms of effective stress (deg), and OCR is the over-consolidation ratio (the
value obtained by dividing the maximum effective stress experienced in the past by the present
effective vertical stress).
Ladd et al.7) report that the coefficient of earth pressure at rest for normally consolidated
undisturbed clay is strongly correlated with the plasticity index Ip. According to this report, K0
= 0.5 when Ip = 15, and K0 tends to converge to 0.8 as Ip increases. Kikuchi et al. measured the
coefficient of earth pressure at rest K0 for clay, and showed that it was about 0.5 regardless of
the value of Ip. These results indicate that the relationship between Ip and the coefficient of earth
pressure at rest varies considerably, and at present it is difficult to determine the general
tendency. In Kikuchi et al.’s study, the relationship between K0 and ϕ' is also explained using
Jàky's equation.8)
Hatanaka et al. conducted a study on the K0 of gravelly soil. According to this study, K0 is
strongly correlated with the in situ shear wave velocity VSF, and the following equation has
therefore been proposed9):

K0 = 0.0058VSF – 0.5 (VSF ≦ 300 m/s). (9.2.4)

(2) Case in which a load is applied to the ground surface

When a concentrated load is applied to the ground surface, the following equation can be used
to calculate the amount of increase in horizontal stress at the location of the exterior basement
wall (Fig. 9.2.2):

3Px 2 z
∆p 0 = , (9.2.5)
(
π r2 + z2 )
52

where P is the concentrated load (kN), r is the horizontal distance from the point of action of
the load to the location of interest (m), x is the shortest distance from the point of action of the
CHAPTER 9 EARTH AND HYDRAULIC PRESSURES - C9-9 -

load to the exterior basement wall, and z is the vertical distance from the point of action of the
load to the location of interest (m).

Figure 9.2.2 Horizontal stress components acting on exterior basement wall

Equation (9.2.5) calculates the horizontal stress acting on a displacement-free exterior

basement wall by assuming that the ground is a semi-infinite elastic body, determining the
horizontal stress in the ground when a concentrated load is applied to the ground surface
through use of the Boussinesq solution and multiplying the obtained value by 2 using the
principle of reflection. This calculation method is described in detail in RDBF.1)
If the load is distributed at a certain part of the ground surface, incremental increases in
earth pressure can be calculated by integrating Eq. (9.2.5). If a uniformly distributed load acts
on the ground surface, the earth pressure obtained by multiplying the amount of increase in
vertical stress by the coefficient of earth pressure is taken as the incremental earth pressure, Δp0.
If the increment of vertical earth pressure is assumed through an approximation to be equal to
the uniformly distributed load q acting on the ground surface, Eq. (9.2.6) can be obtained as

∆p0 = K 0 q , (9.2.6)

where q is a uniformly distributed load (kN/m2).

9.3 Continuous Earth Pressure Acting on Retaining Walls

As shown in Fig. 9.3.1, earth pressure acting continuously on a retaining wall is assumed to
be the active earth pressure. The resultant per unit width of this active earth pressure is
2
pA = 0.5 pAH0 , and the pressure is applied at the location of H0/3 from the bottom of the raining
wall at an angle of ‘θ + δ’.
Equation (9.4) is the Coulomb coefficient of the active earth pressure derived from the
rigid–plastic theory. When the ground surface is horizontal, and the back of the wall is vertical,
- C9-10 - Recommendations for Loads on Buildings

and thus the friction can be ignored, the Coulomb coefficient of the active earth pressure is
expressed as shown below and is equal to the Rankine coefficient of such pressure.

 φ
K A = tan 2  45° −  (9.3.1)
 2

The incremental earth pressure for a uniformly distributed load acting on the ground surface
can be calculated from the following equation:

∆p0 = K A q , (9.3.2)

where q is a uniformly distributed load (kN/m2), and KA is the coefficient of active earth
pressure.

Figure 9.3.1 Earth pressure acting continuously on retaining wall

The total earth pressure is calculated by calculating the incremental earth pressure from Eq.
(9.3.2), and adding the result to Eq. (9.3). Vertical stress transferred to the soil decreases with
depth depending on the angle of the wall friction. This is usually not taken into consideration,
and Eq. (9.3.2) is used regardless of depth.
If necessary, the influence of the hydraulic pressure is taken into consideration. The
calculation procedure is the same as the procedure used for Eq. (9.2) for exterior basement walls.
When designing the retaining walls, it is standard practice to provide drainage facilities. During
long periods of wet weather such as localized heavy rains and seasonal rains during tsuyu (the
rainy season in Japan), the backfill may become saturated, thereby failing because of high
hydraulic pressure. Therefore, if such cases are anticipated, it is necessary to conduct a study
using the unit weight of the soil in a saturated state and its strength parameters, and examine
the ultimate limit states under unusual water level conditions.
CHAPTER 9 EARTH AND HYDRAULIC PRESSURES - C9-11 -

9.4 Earth Pressure Acting on Retaining Walls During Earthquake

(1) Mononobe–Okabe formula for active earth pressure during earthquake
When calculating the earth pressure acting on a retaining wall during an earthquake, it is
common practice to use the Mononobe–Okabe formula (Eq. 9.6). Equation (9.6) was derived
by making the horizontal seismic coefficient kh during an earthquake acting on the back of the
retaining wall and considering the equilibrium of forces in a limit state. Although the values
given by the formula have shown good agreement with the results of shaking table tests
involving standing-wave excitation in certain cases, very little information is available
regarding the verification made using measurement data collected during an earthquake.
When calculating the earth pressure during an earthquake, it is desirable to determine the
value of the design horizontal seismic coefficient kh in Eq. (9.6) according to the value of the
maximum horizontal acceleration amax corresponding to a return period of 100 years, as
described in Chapter 7 (seismic load). It is necessary, however, to keep in mind that, if a wide
range of earthquake motions ranging from medium to small need to be covered, the calculation
accuracy of the earth pressure during an earthquake through the use of Eq. (9.6) is not
necessarily high.
In the examination of earth pressure during medium to small earthquakes, kh = amax/g (where
amax is the maximum horizontal acceleration of the soil block behind the wall, and g is the
gravitational acceleration) may be assumed. In the examination of the earth pressure during
large earthquakes, however, the calculation results obtained from Eq. (9.6) assuming kh = amax/g
will be excessively large. Because Eq. (9.6) is derived from the equilibrium of a rigid–plastic
body ignoring the deformation of the ground behind the wall, the equation does not reflect the
deformation of ground during a large earthquake or the softening of soil from the maximum
strength owing to the localization of the ground strains. Consequently, the determination of the
value of kh through the use of a large maximum horizontal acceleration during a large
earthquake will result in the assumption of an unrealistic slip line and hence an overestimation.
It is generally said that the range of acceleration in which kh may be calculated through the use
of the maximum horizontal acceleration is below 200 Gal or so.10, 11)
For these reasons, the RDBF recommends using kh values of around 0.2 for an examination
of the damage limit states during medium to small earthquakes with an acceleration of less than
about 200 Gal, and kh values of around 0.25 for an examination of ultimate limit states during
large earthquakes. The kh values for medium to small earthquake motions are empirical values
that have been used conventionally, and the kh values for large earthquake motions have been
determined mainly from the values14) back-calculated from the results of house retaining walls
and slope surveys conducted after the Hanshin–Awaji Earthquake. These guidelines follow the
same philosophy, and thus no effort is made to achieve correspondence between the design
- C9-12 - Recommendations for Loads on Buildings

horizontal seismic coefficient and the values corresponding to the 100-year return period of the
peak ground acceleration. Instead, empirical values recommended in RDBF are used for the
design horizontal seismic coefficient, and the active earth pressure at rest calculated from the
Mononobe–Okabe equation is regarded as the earth pressure obtained using the value
corresponding to a return period of 100 years.
It should be noted that the concept of the earth pressure during an earthquake described
above is a tentative approach used before a changeover to the limit state design method. It is
necessary to make a concrete method for determining the basic values of the earth pressure
more straightforward and clear-cut by accumulating more research results. As one way to do so,
a recently proposed modified Mononobe–Okabe method is introduced below.
The Mononobe–Okabe method assumes that the shear strength is isotropic and constant. In
reality, however, the internal friction angle of soil along a failed surface decreases from the peak
strength to the residual strength. Koseki et al. proposed a modified Mononobe–Okabe equation
that takes into consideration the effect of strain localization in the shear zone, and the effect of
the resultant strain softening.10, 11) With this method, the design horizontal seismic coefficient
kh is assumed to be equal to amax/g when the maximum horizontal acceleration amax of the
ground behind the wall is between about 200 and 700 Gal.
In Koseki et al.’s method, the peak strength ϕpeak and residual strength ϕres in the Mononobe–
Okabe equation are first evaluated using one method or another, for example, through
laboratory testing or according to empirical rules, and the conditions for the first active failure
are then determined. The design horizontal seismic coefficient assumed in this method is k =
kcr, and the inclination of the slip surface formed, ψ = ψcr, is calculated using the Mononobe–
Okabe method assuming ϕ = ϕpeak (Fig. 9.4.1). Then, from the equilibrium of forces acting on
the wedge whose bottom is the slip surface calculated earlier, the coefficient of active earth
pressure acting on the initial failure surface, K'EA, is calculated from the following equation:

Figure 9.4.1 Forces acting on a wedge of soil

CHAPTER 9 EARTH AND HYDRAULIC PRESSURES - C9-13 -

cos(ψ − φ )(1 + tan θ tanψ )(1 + tan θ tan α )(tan (ψ − φ ) + tan θ k )

′ =
K EA , (9.4.1)
cos(ψ − φ − θ − δ )(tanψ − tan α )

where K'EA is the coefficient of active earth pressure during an earthquake in the modified
Mononobe–Okabe equation, ψ is the inclination of the slip surface (deg), ϕ is the internal
friction angle of the soil (deg), θ is the angle between the back of the retaining wall and the
vertical plane (deg), δ is the friction angle of the wall (deg), α is the inclination of the ground
surface behind the wall, θk is the resultant incidence angle of an earthquake (tan−1kh, deg), and
kh is the design horizontal seismic coefficient.
In the above equation, the internal friction angle ϕ of the soil is lowered to ϕres. The
inclination of the failure surface, ψ, is fixed at the angle of the initial slip surface, ψcr. This
coefficient of active earth pressure, K'EA, is compared with the coefficient of active earth
pressure KEA calculated from Eq. (9.6) by assuming ϕ = ϕpeak. If the former is smaller than the
latter, it is assumed that the coefficient of active earth pressure K'EA acting on the initial slip
surface is still in effect. The horizontal seismic coefficient during an earthquake, kh,cr, used to
determine the first failure surface is determined arbitrarily. As of yet, there are no established
methods for determining this, and empirical values of about 0 to 0.2 are often used.
By taking into consideration decreases in the internal friction angle, the modified
Mononobe–Okabe equation gives a value of the coefficient of active earth pressure K'EA that is
greater than the value obtained by assuming ϕ = ϕpeak in Eq. (9.6), and is smaller than the value
obtained by assuming ϕ = ϕres in Eq. (9.6). The modified Mononobe–Okabe equation makes it
possible to calculate the active earth pressure corresponding to a large design horizontal seismic
coefficient (kh) to which the Mononobe–Okabe equation cannot be applied, and provides a more
realistic and smaller active failure region than the Mononobe–Okabe equation. As an example,
Fig. 9.4.2 shows the relationship between the design horizontal seismic coefficient kh and the
coefficient of active earth pressure K'EA. As shown, compared with the Mononobe–Okabe
equation, the modified Mononobe–Okabe equation provides more realistic values of active
earth pressure for large values of the design horizontal seismic coefficient kh.
- C9-14 - Recommendations for Loads on Buildings

Figure 9.4.2 Coefficient of active earth pressure K'EA in the modified Mononobe–Okabe
equation.

These guidelines only explain the evaluation method for the geotechnical parameters, ϕpeak
and ϕres, used in the above method partly because the method is not yet in wide use, partly
because the method involves relatively complex calculations, and partly because there are still
a number of unsolved problems such as the difficulty in determining the first failure surface. It
is generally known that in the Mononobe–Okabe equation, after a slip surface is formed from
the first active failure, the acceleration acting on the sliding block does not increase regardless
of the amount of further increase in the acceleration of the ground behind the wall.12,13)
According to this concept, there should be an upper limit to the active earth pressure during an
earthquake. Further study is therefore needed in this area.

Figure 9.4.3 Trial wedge method ignoring the cohesion in the Mononobe–Okabe equation
CHAPTER 9 EARTH AND HYDRAULIC PRESSURES - C9-15 -

(2) Trial wedge method

As shown in Figs. 9.4.3 and 9.4.4, the trial wedge method is used to determine the active earth
pressure during an earthquake from the equilibrium of forces acting on the soil mass behind the
wall by assuming the angle of active slip ψ during an earthquake. With this method, the active
earth pressure at the time the equilibrium condition is met is determined from a funicular
polygon by varying ψ. In the trial wedge method, the cohesion in the Mononobe–Okabe
equation can be taken into consideration. For the design horizontal seismic coefficient used for
calculation in the trial wedge method, a conceptual approach similar to the one used for the
design horizontal seismic coefficient in the Mononobe–Okabe equation is used.
The method described above is applicable to independent retaining walls whose
displacement is not constrained. In cases in which the bottom and top of a retaining wall are
connected to a building such that there is likely to be little wall displacement, as in the case of
a retaining wall in a dry area, the earth pressure occurring during an earthquake does not
necessarily become an active earth pressure. Instead, it may even become closer to the passive
side. As mentioned in Section 9.1, for such a retaining wall, it is appropriate to use the earth
pressure at rest for a continuously occurring earth pressure. When considering the earth pressure
during an earthquake, it is necessary to use the earth pressure calculated as the sum of the earth
pressure at rest and the increase in the earth pressure from the earthquake. Such a failure of a
retaining wall has yet to be reported, however, and sufficient knowledge is presently unavailable
regarding to what extent the earth pressure during an earthquake should be increased. This is
an area that requires further study. For design purposes, it will be necessary to conduct studies
on a case-by-case basis through an earthquake response analysis, for example, or to determine
the earth pressure by referring to the reported analysis results.15)

Figure 9.4.4 Trial wedge method taking into consideration the cohesion in the Mononobe–
Okabe equation
- C9-16 - Recommendations for Loads on Buildings

9.5 Design Groundwater Level

(1) Design groundwater level
Soil layers include high-permeability layers such as sandy soil layers, and low-permeability
layers such as clay layers, which often appear as alternating. Because perched water, free water,
and confined water each exists, it is necessary to judge which of these different types of water
will greatly influence a building.
Factors contributing to free water surface fluctuations include factors related to the natural
environment, such as rain, snowmelt, tide, and floods, and anthropogenic factors such as
pumping from wells, pumping regulations, and subway construction. Each city has its own
records of past groundwater level fluctuations. Government organizations such as the Ports and
Harbors Bureau, the National Land Agency (now restructured), and the Environment Agency
also have their own records.
If the hydraulic pressure is not a primary load, the design groundwater level is determined
by decreasing the value corresponding to a return period of 100 years, aiming at a target value
set at around the highest water level determined from estimated fluctuations over a period of 1
year. One-year fluctuations are not necessarily estimated from 1-year water level measurement
results. Instead, they are judged from the free water level observation records for the past
several years. If the hydraulic pressure is a primary load in the evaluation of the ultimate limit
states (examination of conditions during periods of abnormal water level), the design
groundwater level is determined by increasing the value corresponding to a return period of 100
years according to the estimated fluctuations. The water level fluctuations are estimated from
records of abnormal water levels.
Pressure acting at a point in water is consistent with a hydrostatic pressure distribution. In
the ground, however, impermeable layers or low-permeability layers of materials such as clay
and silt, and permeable layers or aquifers of sand and gravel, often form in alternation, and thus
pressures at particular points are not necessarily consistent with the overall hydrostatic pressure
distribution. However, because the measurement results indicate that hydraulic pressures at
depths below the free water level seldom exceed the hydrostatic pressure distribution, the design
groundwater level is determined based on the free water level. If the confined water level is
higher than the free water level because of the inclination of the geological formations, for
example, a detailed groundwater survey should be conducted.

(2) Buoyancy
In the floor slab of a basement, for example, because the hydraulic pressure in the groundwater
table is zero, buoyancy, which can be calculated by multiplying the water depth by the unit
weight of water, occurs. Although such buoyancy may cause the entire structure to be lifted up,
buoyancy that is stable over a long period of time can be expected to reduce the vertical loads.
For a conservative study on the uplift of a structure, a high groundwater level should be assumed.
CHAPTER 9 EARTH AND HYDRAULIC PRESSURES - C9-17 -

For a conservative expectation of a vertical load reduction, the groundwater level should be
assumed to be low.
If liquefaction occurs during an earthquake, buoyancy is caused by excess pore hydraulic
pressure. It is therefore necessary to pay careful attention in the design to the possibility of a
large amount of buoyancy. In such cases, a unit weight of muddy water of 18 to 20 kN/m3 is
used in place of γw in Eq. (9.1) to calculate the buoyancy.

9.6 Uncertainty of Earth Pressure and Geotechnical Parameters Used for Design
Purposes
(1) Coefficients of variation of geotechnical parameters
Parameters affecting the earth pressure include the following:

1) Geotechnical parameters
c: cohesion of soil (kN/m2)
ϕ: internal friction angle of soil (deg)
γ: unit weight of soil (kN/m3)
δ: wall friction angle (deg)
2) Geometric parameters
z: vertical depth from the upper end of the wall to the location at which the
earth pressure is to be calculated (m)
H: wall height (m)
α: inclination of the ground surface behind the wall (deg)
θ: angle between the back of the wall and the vertical plane (deg)
h: groundwater level (m)
3) Load-related parameters
q: surcharge on the ground surface behind the wall (kN/m2)
kh: design horizontal seismic coefficient

Factors contributing to the variability of the geotechnical parameters include the

inhomogeneity and anisotropy of the soil, the soil sampling methods, and the test methods. In
general, the coefficient of variation of cohesion, which is a parameter that is subject to
considerable variation, ranges between 0.2 and 0.4, and that of the internal friction is 0.1 to 0.2.
In contrast, the coefficient of variation of the unit weight is so small (0.02 to 0.08) that it may
be practically regarded as a determined value.16)
Using the Mononobe–Okabe equation, Matsuo et al.17) evaluated variations in the active
earth pressure of sand, silt, and clay using the Monte Carlo method. They examined the internal
friction angle (ϕ = 30, 35, 40°) for sand, the internal friction angle (ϕ = 10, 20, 30°) and cohesion
(c = 9.8, 14.7, 19.6 kN/m2) for silt, and the internal friction angle (ϕ = 5°) and cohesion (c =
- C9-18 - Recommendations for Loads on Buildings

14.7, 19.6 kN/m2) for clay. The coefficients of variation for the internal friction angle and
cohesion were Vϕ = 0.05, 0.10 and Vc = 0.2, 0.3, respectively, and a normal distribution was
assumed. The wall friction angle δ was assumed to be 2ϕ/3. Based on these results, it is generally
thought that a normal distribution may be assumed for the active earth pressure.
Variations in shape and dimensions include variations in wall height and angle. These
variations occur depending on the construction accuracy. Although their statistical properties
are still largely unknown, these variations are generally thought to be relatively small. The
surcharge on the ground surface behind the wall varies considerably depending on the
conditions around the structure.

(2) Values corresponding to 99% probability of non-exceedance

Site-specific geotechnical parameters are determined during the design stage on the basis
of a geotechnical survey and soil test results. When determining the characteristic values or
statistical properties, it is therefore necessary to take into consideration the uncertainty of the
sampling. Basically, the values corresponding to the 99% probability of non-exceedance for the
site under consideration are used. The values that are used, however, are limited to estimated
values, with consideration given to values that are abnormal from the viewpoint of engineering
and the number of datasets. Thus, because available information is limited mainly for economic
reasons, a lack of information is often compensated using empirical knowledge regarding the
structures of the same type, for example, or based on sites with similar geological conditions,
or engineering judgment that takes into consideration factors such as the region influencing the
limit states of the structure under consideration.
If a certain amount of data is available and a statistical method can be applied, characteristic
values can be estimated based on a sample distribution, as follows. The sample mean and
2
sample variance s of a sample xi (i=1, 2,..., n) can be calculated as

1 n
x= ∑ xi , (9.6.1a)
n i =1

∑ ( xi − x ) .
1 n
s2 = 2
(9.6.1b)
(n − 1) i =1

Because the values corresponding to the 99% probability of non-exceedance follow a t-

distribution with a degree of freedom n−1, characteristic values can be defined, considering a
confidence level 1−α, as follows18):

1
xk = x ± tα ; n −1 s 1 + . (9.6.2)
n
CHAPTER 9 EARTH AND HYDRAULIC PRESSURES - C9-19 -

In the above equation, the plus/minus (±) values are determined to be conservative with respect
to the limit states. The term ta;n−1 is a point with a confidence level of α percent. For example,
for a 99% confidence interval and n−1 = 20, t0.01;20 = 2.528. Table 9.6.1 shows representative
values of t0.01;n−1.

Table 9.6.1 representative values of t0.01;n−1.

5 10 15 20 25 30 60 120 ∞
3.365 2.764 2.602 2.528 2.485 2.457 2.390 2.358 2.326

As mentioned in Section 9.1, however, the amount of geotechnical data that are actually
available is not necessarily sufficient. In such cases, empirical values that have been used
conventionally, for example, the values recommended in RDBF (parameter values of the
Mononobe–Okabe equation shown as a reference in Table 3.4.2 in the RDBF) are used as the
basic values. Because these values are conservative for use in cases in which detailed surveys
are not conducted, they are regarded for the moment as values corresponding to the 99%
probability of non-exceedance.
In addition to the uncertainty of these parameters, the accuracy of the earth pressure formula
(uncertainty of the model) greatly influences the uncertainty of the earth pressure. Concerning
the accuracy of the earth pressure formula, it is necessary to compare the formula with in situ
measurement data, including accurate measurements of the geotechnical parameters. This is not
practiced widely, however, partly because an earth pressure measurement itself is difficult to
conduct. It is believed that a rational limit state design method can be established by
investigating the uncertainty of the construction methods in addition to accumulating such
measurement data.

References
1) Architectural Institute of Japan: Recommendation for design of building foundation
2) Mononobe, N. and Matsuo, H.: On determination of earth pressure during earthquake, Proceedings of
World Engneering Congress, Tokyo, Vol. 9, pp. 177-185, 1929.
3) Okabe, S.: General theory on earth pressure and seismic stability of retaining wall and dam, Journal of
Japan Society of Civil Engineers, Vol. 10, No. 6, pp. 1277-1323, 1924.
4) Jáky, J.: Pressure in soils, Proc. of 2nd ICSMFE, Vol. 1, pp. 103-107, 1948.
5) Danish Geotechnical Institute: Code of practice for foundation engineering, Danish Geotechnical Institute,
Copenhagen, Denmark, Bulletin No. 32, 1978.
6) Mesri, G. and Hayat, T. M.: The coefficient of earth pressure at rest, Canadian Geotechnical Journal, Vol.
30, pp. 647-666, 1993.
7) Ladd, C., Foot, R., Ishihara, K., Schlosser, F. and Poulus, H. G.: Stress deformation and strength
characteristics, State-of-the arts reports, Proc. of 9th ICSMFE, Vol. 4, pp. 421-494, 1977.
8) Kikuchi, Y., Tsuchida, T. and Nakashima, K.: Measurements of K0 values of marine clays by triaxial tests,
Technical Note of the Port and Harbour Research Institute, No. 577, p. 27, 1987.
9) Hatanaka, M., Uchida, A. and Taya, Y.: Estimating K0-value of in-situ gravelly soils, Soils and Foundations,
- C9-20 - Recommendations for Loads on Buildings

Vol. 39, No. 5, pp. 93-101, 1999.

10) Koseki, J., Tatsuoka, F., Munaf, Y., Takeyama, M. and Kojima, K.: A modified procedure to evaluate
active earth pressure at high seismic loads, Soils and Foundations, Special Issue on Geotechnical Aspects
of the January 17 1995 Hyogoken-Nambu Earthquake No. 2, pp. 209-216, 1998.9.
11) Koseki, J., Tatsuoka, F., Horii, K., Tateyama, M., Kojima, K. and Munaf, Y.: Procedure to evaluate active
earth pressure at high seismic loads levels for conventional and geosynthetic-reinforced soil retaining walls,
Proceedings of the 10th Japan Earthquake Engineering Symposium, pp. 1563-1568, 1998.
12) Kasuya, H., Okada, J., Hanafusa, T., and Enoki, M.: The experiment of earth pressure and motions of
retaining wall during earthquake, Proceedings of 38th Japan National Conference on Geotechnical
Engineering, pp. 1621-1622, 2003.7.
13) Okada, J., Kasuya, H., Hanafusa, T., and Enoki, M.: Theoretical study on the motion of retaining wall
during earthquake, Proceedings of 38th Japan National Conference on Geotechnical Engineering, pp.
1623-1624, 2003.7.
14) Commentary on the housing site disaster prevention manual＜Revision＞, supervised by the Private
Sector Building Land Development Office, Economic Affairs Bureau, Ministry of Construction, Gyousei
Corporation, pp.86-88, 1998.5 (in Japanese)
15) Aoki，M.: Evaluation of earth pressure and bearing capacity of foundations on slopes, The Kenchiku
Gijutsu, pp. 126-135, 2000.2 (in Japanese)
16) JGS: Reliability-based design for soil & foundation, Soil & Foundation Engineering Library 28,1985 (in
Japanese)
17) Matsuo, M.: Geotechnical engineering, idea and practice of reliability-based design, Gihodo Shuppan,
pp. 248-264, 1984 (in Japanese)
18) CEN: ISO2394, General Principles on Reliability for Structures, 1998
- C10-1 -

CHAPTER 10 TSUNAMI LOADS

Introduction
Severe tsunami damage to buildings and structures was induced through the Great East Japan
Earthquake in 2011. Ministerial Public Notice No. 1318, “Matter to establish the safety structural
methods against the supposed tsunamis in tsunami prone areas,” was set out by Ministry of Land,
Infrastructure, and Transport and Tourism in 2011. The AIJ added new guidelines for tsunami loads to
the revised edition of the “AIJ Recommendations for Loads on Buildings” in FY 2014.

These recommendations do not refer to variant proposals related to the calculation of tsunami loads
cyclopedically, but do refer to items that were recognized as a general structural design method against
tsunami loads in order to reasonably define tsunami loads on not only tsunami evacuation buildings but
also entire buildings and structures built in tsunami prone areas.

In the recommendations, the tsunami loads on buildings indicate the tsunami wave forces as fluid
forces and loads by waterborne debris. Tsunami wave forces were classified into the different states of
tsunamis on the buildings: at the surge front, behind the surge front, or in still water states. In principal,
the tsunami wave force of each state can be calculated from the inundation depth and flow velocity. The
calculation methods were categorized into one of three methods through the available input data: (A)
the time histories of the inundation depths and flow velocities, (B) the maximum inundation depth and
flow velocity, and (C) the maximum inundation depth.

Section 10.1 describes the scope of application of the recommendations, the estimation principle of
tsunami loads, three estimation methods, an evaluation of the design inundation depth, and design flow
velocity as the input data of the tsunami loads during calculation. In addition, it refers to the direct
evaluation of the time histories of tsunami wave forces on buildings from a simulation of flows around
the buildings based on the CFD described in this section. Sections 10.2 through 10.4 refer to the
calculation formulas of tsunami wave forces in both the horizontal and vertical directions in each state:
at the surge front, behind the surge front, or in a still water state. These calculation formulas are
categorized into a “drag-type evaluation” and a “hydrostatic-pressure type evaluation,” which are
utilized in the design guidelines on tsunami evacuation building designated from the Japanese cabinet
office in 2004 and the Ministerial Public Notice No. 1318 set out from the Ministry of Land,
Infrastructure, Transport, and Tourism in 2011. Section 10.5 refers to the reducing method of the tsunami
wave forces caused by the openings or open structures. Section 10.6 describes the calculation formulas
of the impact loads induced by waterborne debris, driftwoods, containers and other factors, as well as a
practical design approach against the waterborne debris impact problem. It also includes calculation
formulas of loads by the damming effects of waterborne debris. Section 10.7 describes matters to be
- C10-2 - Recommendations for Loads on Buildings

considered in the structural design of buildings against tsunami loads: residual damage of buildings
caused by earthquake ground motions before a tsunami, the liquefaction effects of the ground, and the
scouring effects of the ground around the buildings.

Symbols

Capital letters
A ：submerged horizontal projection area of building (m2)
AD ：projection area in flow (m2)
B ：width of building (m)
CD ：drag coefficient
CD1, CD2, CD3：coefficient for calculation of tsunami force
CDA, CDB, CDC：drag coefficient
D ：distance from shoreline (m)
FA ：vertical force (kN)
FB ：buoyancy (kN)
FD ：drag (kN)
FI ：impact load (kN)
FUL ：uplift (kN)
Fr ：Froude number
Tu ：tsunami load (kN)
Zs ：maximum scour depth (m)

Small letters
a ：inundation amplification coefficient
g ：gravity acceleration (m/s2)
hi ：inundation depth (m)
hmax：maximum inundation depth (m)
hf ：inundation depth at front face of building (m)
hr ：inundation depth at rear face of building (m)
vi ：flow velocity (m/s)
vmax：maximum flow velocity (m/s)
z ：height of the part of the building for calculation of tsunami pressure or tsunami force from the ground
(m)

Greek characters
ρ ：density of seawater (t/m3)
ρs ：density of seawater including grit (t/m3)

2
CHAPTER 10 TSUNAMI LOADS - C10-3 -

10.1 General

10.1.1 Scope of application

(1) Buildings and structures
This article can be applied to any buildings and structures on the ground in the supposed tsunami
inundation area. It does not depend on the construction type, such as timber, steel, or reinforced concrete.

(2) Type of tsunami

This article can be applied to incoming and receding tsunamis (Fig. 10.1.1). Different evaluation
methods can be applied to the surge front and flow behind the surge front for incoming tsunamis. Loads
by the flow behind a surge front are created during receding tsunamis.

Loads Loads
Surge front Flow behind surge front Flow in the receding tsunami

Time Time
Incoming tsunami Receding tsunami

Figure 10.1.1 Idealized model of incoming and receding tsunamis

(3) Design inundation depth and design flow velocity

The design inundation depth and flow velocity should be evaluated based on the value associated with
a 100-year return period as well as other loads. However, a stochastic model or the return period scale
factors of these values have yet to be established.

(4) Density of seawater

The density of seawater is generally 1.03 (t/m3) for water with a salinity of 3.5 (%) at 15° Celsius, but
it increases if the seawater includes debris, soil, or sand. In the FEMA design guidelines1), the density
of seawater ρs is 1.2 (t/m3) when considering this effect.

(5) Froude number

The Froude number of applicable tsunamis is between zero to 2. The Froude number is a ratio of
inertia force to gravity force defined as Eq. (10.1.1), where v (m/s) is the flow velocity，g (m/s2) is
gravity acceleration, and h (m) is the inundation depth.

v (10.1.1)
Fr =
gh
- C10-4 - Recommendations for Loads on Buildings

10.1.2 Estimation principle

As shown in Fig. 10.1.2, tsunami loads indicate loads at the surge front, loads behind the surge front,
or loads in still water, which act on buildings not only in the lateral direction but also in the vertical
direction. They also include the loads of waterborne debris.

(1) Tsunami load in the design

Figure 10.1.3 shows the evaluation flow of the tsunami loads in the design. The evaluation methods
can be classified into three groups according to available information on the inundation depth and flow
velocity. Otherwise, design tsunami loads on a building can be calculated directly from the CFD without
evaluation formulas of tsunami loads.

(2) Tsunami forces in the front and back of a building

Tsunami forces on a building can be evaluated using the integration value of the wave pressures on
the front and back surface of the building.

(3) Combination of tsunami and other loads

In the recommendations, the seismic and wind loads can be ignored in the calculation of the tsunami
loads.

Load at surge front Bore force

Vertical force

Impulsive bore force

Tsunami loads Load behind surge front Drag

Vertical force

Load in still water Hydrostatic force

Buoyancy
Load by waterborne debris

Figure 10.1.2 Constitution of tsunami load

4
CHAPTER 10 TSUNAMI LOADS - C10-5 -

Figure 10.1.3 Evaluation flow of tsunami loads

10.1.3 Three estimation methods

Evaluation methods of design tsunami loads are classified into following three types according to the
available information on the inundation depth and flow velocity in the CFD. This design inundation
depth indicates a free field water depth at the site without a damming up by the building.
- C10-6 - Recommendations for Loads on Buildings

(1) The time history of the inundation depth and the flow velocity are available (A)
The design tsunami load can be evaluated based on the time history of the inundation depth and flow
velocity, if these values are derived from the CFD, which can simulate the geographical shape of the
city area. The calculation value should be verified through several cases or increased by the safety factor
for the design tsunami load. For the surge front, the hydrostatic-type triangle wave pressure distribution
is assumed in front of the building. For the flow behind the surge front, the hydrostatic-type triangle
wave pressure distribution of the design inundation depth is assumed in back of the building, and the
water height including a damming is assumed in front of the building (Figure 10.1.4). The integration
value of wave pressure in the front and back of the building coincides with the load of the flow behind
the surge front on the building. The wave pressure in back of the building cancels a portion of the wave
pressure in front of the building, and thus it can also be regarded as a trapezoid wave pressure distribution
in front of the building.

Building
Pressure in front
of the building

Pressure in back
Flow of the building
Pressure in front
canceled by that
in back

Figure 10.1.4 Hydrostatic-type wave pressure distribution in the front and back of the building for
tsunami loads behind the surge front

(2) Maximum inundation depth and the maximum flow velocity are available (B)
The design tsunami load can be evaluated based on the maximum inundation depth and the maximum
flow velocity, if the maximum flow velocity can be specified from CFD, which can simulate the local
fluctuations of the flow velocity. As with method (1), the wave pressure distribution can also be assumed.

(3) The maximum inundation depth is available (C)

The maximum inundation depth of a tsunami can be defined as a value on the tsunami hazard map.
The wave pressure shows a hydrostatic-type triangle distribution in front of the building assuming a
bore. It does not depend on the state of the flow (surge front or flow behind the surge front).

10.1.4 Design inundation depth and design flow velocity

(1) Tsunami hazard maps formulated by local governments
When it is difficult to evaluate the precise design inundation depths on the construction sites from
tsunami hazard maps, it is necessary to obtain the digital values of the depth by other countermeasures,
or carry out the following appropriate CFD.

6
CHAPTER 10 TSUNAMI LOADS - C10-7 -

(2) Computational fluid dynamics

When the design inundation depth and design flow velocity are evaluated through CFD, the
appropriate calculation should be carried out according to the guidebook on recommendations for
tsunami loads on buildings.

(3) Ground subsidence from crustal movement

Ground subsidence from crustal movement has to be reflected in the calculation of the design
inundation depths. Most tsunami hazard maps have already considered this phenomenon.

10.1.5 Estimation of tsunami wave forces using CFD

When the tsunami loads are evaluated through the sophisticated computational fluid dynamics,
appropriate calculations should be carried out according to the guidebook for recommendations of loads
on buildings.

10.2 Tsunami Loads at Surge Front

10.2.1 Wave force of tsunami bore

(1) When a time series of the inundation depths and the flow velocities are available (A)
When a time series of the inundation depth and the flow velocity are obtained through numerical
simulations or other methods, the tsunami force, F (kN), on a building can be estimated through the
following three calculation methods.

1) Equation (10.2.1) or (10.2.2) can be used.

𝐶𝐶D1
𝐹𝐹 = 2
𝜌𝜌𝜌𝜌(ℎi (𝑡𝑡)𝑣𝑣i (𝑡𝑡)2 )max (10.2.1)

5.4ℎimax ℎimax
𝐶𝐶D1 = 2.0 + 𝐷𝐷
, (0.01 < 𝐷𝐷
< 0.17)

𝐶𝐶D2 𝜕𝜕𝑣𝑣i (𝑡𝑡)

𝐹𝐹 = ( 𝜌𝜌𝜌𝜌ℎi (𝑡𝑡)𝑣𝑣i (𝑡𝑡)2 + 𝐶𝐶M 𝜌𝜌𝜌𝜌𝜌𝜌ℎi (𝑡𝑡) ) (10.2.2)
2 𝜕𝜕𝜕𝜕 max

𝐶𝐶D2 = 2.0,
ℎimax
𝐶𝐶M = 1.0, (0.01 <
𝐷𝐷
< 0.17)
where
CD1, CD2, CM: coefficients for estimation of tsunami forces,
B (m): width of the building,
W (m): depth in the cross section of the building,
hi(t) (m): time series of inundation depth of incident tsunami,
vi(t) (m/s): time series of flow velocity of incident tsunami,
- C10-8 - Recommendations for Loads on Buildings

𝜌𝜌 (t/m3): density of seawater,

( )max: maximum value,
D (m): distance from the shoreline.

Equation (10.2.1) reflects the effect of the inertial force term (the second term on the right side) of
Eq. (10.2.2) in the drag coefficient. Compared to the steady-state wave force, Eq. (10.2.1) may be more
convenient. Equations (10.2.1) and (10.2.2) are expressions calculated as average values from the results
of approximately 90 hydraulic experiments. Approximately 80% of the experimental cases were
evaluated on the safety side for 1.3 times (1.3 F) the calculated value.

2) It is also possible to calculate the Froude number, Fr, every moment from the inundation depth and
the flow velocity, and reflect the results in the water depth coefficient, a, of Eq. (10.2.7), and reduce the
water depth coefficient a to 1.5, if Fr <1 2).

3) Sakakiyama uses Fr at the time of the maximum inundation depth, and proposes the following water
depth coefficient, a, in expression (10.2.7), namely, for a three-dimensional structure with a small width,

𝑎𝑎 = 1 + 0.5𝐹𝐹𝐹𝐹 2 (0 < 𝐹𝐹𝐹𝐹 < 2), (10.2.3)

and for a wall-type structure,

𝑎𝑎 = 1 + 1.4𝐹𝐹𝐹𝐹 (0 < 𝐹𝐹𝐹𝐹 < 2). (10.2.4)

Both of Sakakiyama’s proposed formulas (10.2.3) and (10.2.4) calculate the average values, which are
relatively simple. The calculated values are in good agreement with the experimental values.

(2) When the maximum inundation depth and maximum flow velocity are available (B)
Equation (10.2.5) is available as the calculation formula of the tsunami wave force F (kN) when the
maximum inundation depth and the maximum flow velocity of the incident tsunami are obtained3).

𝐶𝐶D3 2
𝐹𝐹 = 2
𝜌𝜌𝜌𝜌ℎmax 𝑣𝑣max , (10.2.5)

ℎmax ℎmax
𝐶𝐶𝐷𝐷3 = 1.3 + 6.3 𝐷𝐷
, (0.01 < 𝐷𝐷
< 0.17)

where
CD3: coefficient for the estimation of tsunami force,
B (m): width of the building,
hmax (m): maximum inundation depth,
vmax (m/s): maximum flow velocity,
𝜌𝜌 (t/m3): density of sea water,

8
CHAPTER 10 TSUNAMI LOADS - C10-9 -

D (m): distance from the shoreline.

However, Eq. (10.2.5) is an expression for calculating the average value. Approximately 80% of the
experimental cases were evaluated on the safety side for 1.2 times (1.2 F) the calculated value.
It is difficult to judge whether the maximum inundation depth and the maximum flow velocity
obtained are at the time when the bore force is occurring. Particularly along a ria coast where the cove
is narrower and shallower than the bay mouth, the maximum inundation depth may occur at the time
when the flow velocity is almost zero. At that time, the application of formula (10.2.5) results in an
overestimation, and thus the formula should be applied with care.
Arimitsu et. al4) proposed the following Eq. (10.2.6) for the wave pressure distribution, using the
maximum inundation depth hfmax (m) at the front of the structure and the maximum flow velocity νmax
(m / s) within the vicinity of the structure without the variables of the incident wave.

𝜌𝜌g(ℎfmax − 𝑧𝑧) (𝑧𝑧 > inundation depth at the maximum velocity)

𝑝𝑝𝑚𝑚 = � 2
𝜌𝜌g(ℎfmax − 𝑧𝑧) + 𝜌𝜌𝑣𝑣max (𝑧𝑧 ≦ inundation depth at the maximum velocity)

(0 < Fr < 2) (10.2.6)

where
𝑝𝑝𝑚𝑚 (kN/m2): wave pressure,
hfmax (m): maximum inundation depth in front of the structure,
vmax (m/s): maximum flow velocity,
𝜌𝜌 (t/m3): density of seawater,
z (m): height from the ground.

Equation (10.2.6) uses the inundation depth at the maximum flow velocity in the estimation of the
wave pressure distribution, and thus strictly speaking, it cannot be evaluated solely based on the
maximum flow velocity and the maximum inundation depth, and it is unnecessary to also calculate the
incident wave.

(3) When the maximum inundation depth is available

When the maximum inundation depth is obtained, the tsunami force F (kN) is calculated through
formula (10.2.7) according to the Ministry of Land, Infrastructure, and Transport’s Public Notice No.
1318 from 20115).

𝑎𝑎 2 2
𝐹𝐹 = 2
𝜌𝜌𝜌𝜌gℎmax , (10.2.7)
where
a: water depth coefficient, see Table 10.2.1,
B (m): width of the building,
hmax (m): maximum inundation depth,
- C10-10 - Recommendations for Loads on Buildings

𝜌𝜌 (t/m3): density of seawater,

g (m/s2): gravity acceleration.

However, the possibility of an overestimation when the inundation depth is not obtained in the time
zone during which the bore force is acting requires close attention.

Table 10.2.1 Water depth coefficient, a

the building is located 500 m or the building is located within 500 other cases
more from the shoreline, and m from the shoreline, and other
other facilities are expected to facilities are expected to reduce
reduce the tsunami wave force the tsunami wave force
water depth
1.5 2 3
coefficient, a

10.2.2 Impulsive bore force

When a building is located close to the shoreline, there are some possibilities of occurrence of an
impulsive force depending on the topography and height of the tsunami (the height from the mean tide
level without a tsunami), and thus if it is necessary to consider the local destruction, it is advisable for
physical hydraulic experiments or numerical simulations using the VOF methodology6) to be conducted.

Figure 10.2.1 shows the state when an approximately 5 m tall tsunami breaks just in front of a building
and hits it. Under these conditions, the hydraulic experiments results indicate that the impulsive wave
pressure was estimated to be about 5-times of the hydrostatic pressure corresponding to the height of
the tsunami7).

However, in terms of the time series, a waveform looks like that shown in Fig. 10.2.2. Because the
duration time is short, it is considered that the influence differs depending on the building. Thus, it is
necessary to consider the response of the building if it is necessary to treat the impulsive wave pressure.
Hayashi et. al.8) proposed the use of formula (10.2.8) as an expression that takes into consideration the
dynamic response of the buildings.

𝑝𝑝m 1.83𝑈𝑈
𝜌𝜌g
= 𝐻𝐻 (10.2.8)
𝑇𝑇0 g
where
pm (kN/m2): maximum pressure at the impulsive bore force,
H (m): wave height from trough to crest of the wave,
g (m/s2): gravity acceleration,
𝜌𝜌 (t/m3): density of seawater,

10
CHAPTER 10 TSUNAMI LOADS - C10-11 -

U (m/s): wave celerity,

T0 (s): natural period of the structure.

Because the natural period in a medium-low storied building is about 0.5 to 1 s, formula (10.2.8) is
transformed into formula (10.2.9) to obtain the coefficient proportional to the wave height near the coast
revetment (sea wall). which shows that a hydrostatic force of about 1.5 to 3 times the tsunami height
may act on the structure.

𝑝𝑝m �1.83×�5.0g� 1.3

𝜌𝜌g𝐻𝐻
= 𝑇𝑇0 g
= (10.2.9)
𝑇𝑇0

Figure 10.2.1 The state in which about a 5 m tall tsunami breaks just in front of the building and hits it
(Kuji city, Iwate prefecture, March 11, 2011)

Figure 10.2.2 Time series of impulsive bore pressure and sustainable pressure in the case of physical
hydraulic experiments7)

10.2.3 Vertical force

The vertical force FA (kN) consists of the buoyancy FB (kN), (b) uplift FUL (kN), and (c) weight FW
(kN) of seawater flowing into the inside of a building, as represented through Eq. (10.2.10) and shown
in Fig. 10.2.39). The constitution of the vertical force differs depending on the conditions of the incident
tsunami, building, ground, and others.

FA = FB + FUL − FW (10.2.10)
where the vertical force acts upward when the value calculated from Eq. (10.2.10) is positive.
The wave pressure on the front of the building owing to the collision of a tsunami-inundated flow
consists of the hydrostatic and hydrodynamic pressures. The hydrostatic pressure is proportional to the
- C10-12 - Recommendations for Loads on Buildings

water depth from the surface of the inundating flow. In this section, the impulsive wave pressure is
classified as hydrodynamic pressure.
The buoyancy results from the hydrostatic pressure and functions are on the danger side in terms of
the movement and/or overturning of the building. The uplift results from the hydrodynamic pressure
(wave pressure minus hydrostatic pressure) on the front of the building and can occasionally function
on the safety side because it can have a negative value. The weight of influent seawater in the building
also functions on the safety side. Therefore, the assumption that the vertical force consists of buoyancy
and uplift with a positive value is equivalent to evaluating the vertical force on the danger side for the
movement and/or overturning of the building.

Figure 10.2.3 Main components of tsunami wave pressure and force on a building9)

(1) Buoyancy
The buoyancy FB (kN) on a building is basically calculated from Eq. (10.2.11), regarding a trapezoid
formed from the inundation depth hf (m) at the front, the inundation depth hr (m) at the back, and the
base width (depth direction) as the submerged portion of the building.

ρg (hf + hr )A
1 (10.2.11)
FB =
2
where
hf (m): inundation depth at the front of the building,
hr (m): inundation depth at the back of the building,
A (m2): horizontally projected area of the submerged portion of the building,
ρ (t/m3): density of seawater,
g (m/s2): gravitational acceleration.

However, the buoyancy should be appropriately evaluated when the horizontally projected area A (m2)
is a function of the height z (m) from the ground.
Buoyancy acts on the bottom of a building when inundating water intrudes under the building, i.e.,
when the building is on the ground with high permeability, the building has foundation piles on the

12
CHAPTER 10 TSUNAMI LOADS - C10-13 -

ground liquefied after an earthquake, and the building has ventilation holes under the first (ground) floor.
A method for evaluating the buoyancy caused by air trapped under the floor and/or ceiling has not been
proposed.

Inundation depths hf and hr needed to evaluate the buoyancy on a building are calculated as follows:
1) When the time series of the inundation depth and flow velocity are available (A)
The inundation depth obtained from an operation in which the tsunami wave force calculated from
Eq. (10.2.1) or Eq. (10.2.2) is converted into the hydrostatic force (the total hydrostatic pressure) is
adopted as the inundation depth hf (m) at the front of the building. A value of zero is adopted as the
inundation depth hr (m) at the back of the building.

2) When the maximum inundation depth and the maximum flow velocity are available (B)
The inundation depth obtained from an operation in which the tsunami wave force calculated from
Eq. (10.2.5) is converted into the hydrostatic force is adopted as the inundation depth hf (m) at the front
of the building. A value of zero is adopted as the inundation depth hr (m) at the back of the building.

3) When only the maximum inundation depth is available (C)

One-and-a-half times the maximum inundation depth of the incident tsunami (when a reduction of
the tsunami wave force can be expected by other facilities, and the building is located over 500 m from
the shoreline), 2 times the maximum inundation depth (when a reduction of the tsunami wave force can
be expected by other facilities, and the building is located within 500 m from the shoreline), and 3 times
the maximum inundation depth (conditions except the above) are adopted as the inundation depth hf (m)
at the front of the building. A value of zero is adopted as the inundation depth hr (m) at the back of the
building.

(2) Uplift
The uplift FUL (kN) on a building results from the hydrodynamic pressure on the front of the building,
and acts upward. As shown in Fig. 10.2.3, the pressure distribution of the uplift is assumed to be a right-
angle triangle with the maximum pressure at the front of the building, and zero at the back of the building,
whereas the maximum pressure is the same as the hydrodynamic pressure at the lowest end of the
building. The maximum pressure depends on the conditions of the incident tsunami, location of the
building, shape of the building, opening ratio of the building, the density of muddy seawater, and other
factors.
Both the buoyancy and uplift act on the bottom of a building on ground with a high permeability, a
building with foundation piles on liquefied ground, and a building with ventilation holes under the floor,
among other building types, and are two reasons for the movement and/or overturning of a building.
The details of uplift owing to the collision of a tsunami-inundated flow against a building are not well
known. A previous study10) reported that the hydrodynamic pressure at the lowest end of the front of the
- C10-14 - Recommendations for Loads on Buildings

building reaches 80% of the hydrostatic pressure at the same lowest end when the collision of a tsunami-
inundated flow occurs on horizontal land after dispersing into the sea area, and then flowing over a
vertical quay wall. Another study9) indicated that the hydrodynamic pressure at the lowest end reaches
40–50% of the hydrostatic pressure at the same lowest end in the case of the collision of a tsunami-
inundated flow on horizontal land after running up over a uniformly sloping beach.

The uplift on a building is calculated under the follows conditions:

1) When the time series of the inundation depth and flow velocity are available (A), and when the
maximum inundation depth and the maximum flow velocity are available (B)

Because the buoyancy on the building evaluated in subsection 10.2.3(1) includes the uplift, it is not
necessary to newly evaluate the uplift on the building.

2) When only the maximum inundation depth is available (C)

By assuming the wave pressure on the front of the building to be hydrostatic pressure, the uplift on
the building is set to zero.

(3) Weight of influent seawater

The weight FW (kN) of seawater flowing into the inside of a building acts in a vertical downward
direction and is calculated from Eq. (10.2.12). When the volume of influent seawater cannot be
appropriately evaluated, the weight is set to zero.

FW = ∑ ρgVi (10.2.12)
where
Vi (m3): volume of influent seawater in each floor,
ρ (t/m3): density of seawater,
g (m/s2): gravitational acceleration.

The volume of influent seawater depends on the incident tsunami conditions, opening ratio of the
building, locations of the openings, and other factors.

(4) Other factors

Lift FL (kN) on a building results from a flow along the front and bottom of the building, and acts
perpendicularly to the flow. The lift on the ceiling of a floor simply supported by columns acts downward,
and functions on the safety side for the movement and/or overturning of the building. However, the lift
coefficient for a non-submerged building is not well known.

14
CHAPTER 10 TSUNAMI LOADS - C10-15 -

10.3 Tsunami Loads behind Surge Front

10.3.1 Drag
Drag FD (kN) on a building acts in the flow direction, and is calculated from Eq. (10.3.1).

1
FD = ρCD v 2 hi B (10.3.1)
2
where
CD: drag coefficient (2.1 for a building having a square section and incident flow angle of 0° 11)),
v (m/s): velocity of flow behind the surge front,
hi (m): inundation depth of incident tsunami (refer to Fig. 10.2.3),
B (m): width of the building,
ρ (t/m3): density of the seawater.

In an incident tsunami inundation flow, which is not affected by buildings and other facilities, the
maximum surface slope initially appears. Next, the maximum flow velocity and momentum flux occur
at almost the same time, and finally, the maximum inundation depth appears.
Except for the duration of the collision of a surge front, it is known that the drag is dominant in the
horizontal flow-directional forces acting on the building. Although the drag consists of the hydrostatic
force, hydrodynamic force (total hydrodynamic pressure), and frictional force, the percentage of the
frictional force in the drag is small in the case of a bluff building forming a rear region with a lower flow
velocity than its ambient flow velocity. Therefore, the drag can be evaluated as the difference between
the horizontal forces on the front (hydrostatic force plus hydrodynamic force) and back (hydrostatic
force only) of the building, and is frequently called the “pressure drag” or “form drag.”
The velocity of an incoming tsunami inundation flow behind the surge front, which is not affected by
the building of interest, is adopted as the flow velocity, v. The “inundation depth of an incident tsunami
× the width of the building” is usually used in an evaluation of the projected flow-directional area AD of
the submerged portion of the building. Because the drag coefficient in this case is newly defined, there
is little accumulation of data on the coefficient. It is necessary to note that the coefficient is different
from the existing coefficient for a two- or three-dimensional submerged object. When intending to use
the existing coefficient, it is necessary to solve the evaluation problem of the projected flow-directional
area, AD.

The drag is calculated as follows.

1) When the time-series of the inundation depth and flow velocity are available (A)
As shown in Eq. (10.3.2), the drag FD (kN) on a building is evaluated as a case in which v2hi has the
largest value.

FD =
1
2
( )
ρCD v 2 hi max
B (10.3.2)
- C10-16 - Recommendations for Loads on Buildings

2) When the maximum inundation depth and the maximum flow velocity are available (B)
In Eq. (10.3.1), the maximum flow velocity is adopted as the flow velocity v, and the maximum
inundation depth is adopted as the inundation depth hi of the incident tsunami used to calculate the
projected flow-directional area AD = hiB of the submerged portion of the building.

3) When only the maximum inundation depth is available (C)

When the ground is horizontal, the flow velocity v at the maximum inundation depth adopted as the
incident inundation depth hi in Eq. (10.3.1) can be calculated using the empirical Eq. (10.3.3)12).

v = Fr gh , (10.3.3)
where v is the flow velocity (m/s), Fr is the Froude number, g is the gravitational acceleration (m/s2),
and h is the inundation depth (m). Although the Froude number of incident tsunami is around 0.7 to 2,
it is desirable to adopt 2 to remain on the safe side. The maximum inundation depth is also adopted as
the incident inundation depth hi to calculate the projected flow-directional area AD.
On the other hand, when the ground is uniformly sloping, the flow velocity can be calculated using
the nonlinear shallow water theory with no frictional resistance term13). Let us consider a tsunami that
propagates to the shoreline as a bore, and then runs up over a uniformly sloping beach with a slope angle
θ of a horizontal plane. The maximum inundation depth him (m) at an arbitrary position x (m) on land,
and the flow velocity v (m/s) at the same position and time, are expressed through Eqs. (10.3.4) through
(10.3.6) with time t (s) as a parameter.

him =
1
36 gt cos θ
2
(
2t 2 gR − gt 2 sin θ − 2 x)2
(10.3.4)

v=
1
3t
(
t 2 gR − 2 gt 2 sin θ + 2 x ) (10.3.5)

2x (10.3.6)
t=
g sin θ
where the x-axis with the origin at the shoreline goes landward along the sloping beach, 0≤x≤R/sinθ (m),
R is the run-up height (m), θ is the slope angle (°), and g is the gravitational acceleration (m/s2).
When him, its position x, and the slope angle θ are given, the run-up height R can be calculated from
Eqs. (10.3.4) and (10.3.6), and the flow velocity v can then be calculated from Eqs. (10.3.5) and (10.3.6)
and the calculated run-up height R. Because the inundation area is displayed in a tsunami hazard map,
it is possible to adopt the run-up height R as read from such a map. However, because there is a high
possibility that the run-up height R obtained through a numerical simulation taking into account the
frictional resistance force is lower than that calculated from Eqs. (10.3.4) through (10.3.6), which do not
take into account the frictional resistance force, use of the run-up height as read from the hazard map
through Eqs. (10.3.4) through (10.3.6) cannot be recommended.
With this method, the Froude number is √cosθ (<1) and depends only on the slope angle θ regardless
of position x on land and over time. In addition, the maximum inundation depth him is adopted as the

16
CHAPTER 10 TSUNAMI LOADS - C10-17 -

incident inundation depth hi to calculate the projected flow-directional area AD. It might be a way to
calculate the drag to deduct the hydrostatic force at the back of the building from the tsunami wave force
calculated from Eq. (10.2.7). In this case, the maximum inundation depth him is adopted as the inundation
depth at the back of the building.

10.3.2 Vertical force

Inundation depths of hf and hr, which are necessary to evaluate the buoyancy on a building, are
calculated as follows.
1) When the time-series of the inundation depth and flow velocity are available (A)
The largest inundation depth among the inundation depths obtained from an operation that the wave
forces by adding the hydrostatic force at the back of the building to the maximum tsunami wave forces
calculated from Eqs. (10.2.1), (10.2.2) and (10.3.2), are converted into the hydrostatic forces (refer to
Fig. 10.3.1) is adopted as the inundation depth hf (m) at the front of the building. The inundation depth
of the incident tsunami occurring at the same time as the largest inundation depth mentioned above is
adopted as the inundation depth hr (m) at the back of the building. However, hr = hf when there is an
escarpment immediately behind the building.

2) When the maximum inundation depth and the maximum flow velocity are available (B)
The larger inundation depth between those depths obtained from an operation in which the wave
forces, by adding the hydrostatic force at the back of the building to the maximum tsunami wave forces
calculated from Eq. (10.2.5), and under case (B) of the section 10.3.1, are converted to the hydrostatic
forces (refer to Fig. 10.3.1) is adopted as the inundation depth hf (m) at the front of the building. The
maximum inundation depth of the incident tsunami is adopted as the inundation depth hr (m) at the back
of the building. However, hr = hf when there is an escarpment immediately behind the building.

Figure 10.3.1 Inundation depth at the front of the building at the time when the wave force, by adding the
hydrostatic force at the back of the building to the tsunami wave force, is converted into the hydrostatic
force
- C10-18 - Recommendations for Loads on Buildings

3) When only the maximum inundation depth is available (C)

The larger inundation depth between a times the maximum inundation depth of an incident tsunami
(where a is 1.5 when a reduction of the tsunami wave force is expected by other facilities, and the
building is located over 500 m away from the shoreline, 2 when a reduction of the tsunami wave force
is expected by other facilities and the building is located within 500 m from the shoreline, and 3 in the
conditions other than the above), and the inundation depth obtained from an operation in which the wave
force, by adding the hydrostatic force at the back of the building to the tsunami wave force calculated
under case (C) of section 10.3.1 is converted into the hydrostatic force (refer to Fig. 10.3.1), is adopted
as the inundation depth hf (m) at the front of the building. The maximum inundation depth of the incident
tsunami is adopted as the inundation depth hr (m) at the back of the building. However, hr = hf when
there is an escarpment immediately behind the building.
When the flow velocity is negligibly small, the calculation procedure follows that indicated in
section 10.4.2.

10.4 Hydrostatic Loads in Still Water

10.4.1 Horizontal hydrostatic force

Horizontal hydrostatic force occurs when standing and slowly moving water encounters a building or
member. This force is caused by an imbalance of pressure owing to a differential water depth on opposite
sides of the building or member.

10.4.2 Buoyancy
The total buoyancy equals the weight of water displaced. The buoyancy must be resisted by the weight
of the building or member, and any opposing forces resisting floatation such as the weight of water
entering the building and the pile friction force.

10.5 Loads on Members with Openings and/or Open Structures

In general, buildings have openings (windows, doors, etc.) on their exterior and partitioning walls,
and some buildings are constructed as “open structures” supporting the building above simply using
columns. A tsunami load on such buildings shall be evaluated based on their area exposed to the wave
pressure, not on the total projected area of the inundated surface5).
When the load is calculated through the hydrodynamic force equation, and its pressure distribution is
not specified, the load on members with openings and/or an open structure shall be determined by
assuming a hydrostatic pressure distribution, as indicated in subsection 10.1.3.

10.5.1 Loads on members with openings

Past experimental hydraulic model research32) has indicated that the tsunami load on a building with

18
CHAPTER 10 TSUNAMI LOADS - C10-19 -

openings can be evaluated by multiplying the force on a building without openings and the closure ratio.
Herein, the closure ratio is defined as the ratio of the area of enclosure (i.e., area without openings) to
the total projected vertical plane area of the inundated pressure-exposed surface. The ratio shall not be
less than 70% when calculating the load5). It was also indicated through experimental research that the
larger of the closure ratio of an exterior pressure-exposed surface and that of the other downstream side
exterior surface shall be employed in calculating the load on members with openings14).
Breakaway members on the pressure-exposed exterior and partitioning walls, which are expected to
be easily broken away by the wave pressure, can be regarded as openings. Window glasses, doors, and
shutters, for example, can be regarded as openings unless they are made of high strength materials.
To employ the evaluation approach of the load mentioned above, the exterior and partitioning walls
shall be properly designed and arranged such that they may not hamper the wave passing through the
building.

10.5.2 Loads on open structures

An open structure is a building that has a floor (generally the first floor) supporting a structure above
simply with columns, where the wave flow can pass through smoothly. The tsunami load on the open
structure can be evaluated according to the closure ratio, as is applied for members with openings. The
ratio, however, can be less than 70%, unlike in the case for members with openings described in 10.5.1.
When stairs, elevators, and inner walls, which may hamper the passing wave, are located in the open
structure, the wave pressure on these components shall be considered in calculating the load on the
building. The wave pressure on all columns (not only the columns in the first front row exposed to the
wave) in an open structure shall also be considered, as discussed in the past test results of hydraulic
model experiments15).

10.5.3 Loads evaluated through hydraulic model experiments or CFD

The tsunami load on members with openings and/or an open structure can be determined based on
hydraulic model experiments or CFD, and not by the procedure mentioned above, when appropriately
considering the force on the exterior and the partitioning walls.

10.6 Loads from Waterborne Debris

10.6.1 Debris impact loads

The waterborne debris impact loads may be categorized into two types: one is a load causing local
failures of a building, and the other is a load that significantly affects the global building responses.
Failure of the axial-load carrying members of a building may lead to critical damage of the building
even if it is local. However, it is often difficult to determine the design impact loads owing to
uncertainties in the debris types, tsunami flow velocity, and other factors, and a building should be
carefully designed to eliminate its total or partial collapse following a local failure. A missing column
- C10-20 - Recommendations for Loads on Buildings

strategy, which has been employed in the Japanese Standards for Seismic Evaluation of Existing RC
Buildings16), can be a practical method to decrease the possibility of a progressive building collapse even
when the impact loads are not well defined.

10.6.2 Damming effect caused by waterborne debris

The damming effect caused by the accumulation of waterborne debris can be treated as the
hydrodynamic force considering the breadth of the debris dam against the front surface of the building.

10.7 Consideration of Resistant Tsunami Design of Buildings

10.7.1 Damage from earthquake ground motion

The load bearing capacity of buildings should be evaluated appropriately by accounting for damage
from earthquake ground motions.

10.7.2 Liquefaction of ground

Liquefaction of the ground from an earthquake shall be taken into consideration because it may
degrade the capacities of the buildings such as the tensile capacities of the piles for the overturning
resistance.

10.7.3 Scouring of ground

Local scouring shall be taken into consideration because it may be caused mainly along the corner of
the front resulting from the flow around the buildings.

References
1) FEMA：Guidelines for Design of Structures for Vertical Evacuation，FEMA P 646，2012
2) Institute of Industrial Science, The University of Tokyo: Interim Report of the Building Standards
Improvement Project No. 40 “A Study on Improvement of Building Standards in Tsunami Hazard Areas,”
Part.2, 2011 (in Japanese)
3) Achmad, F., Shigihara, Y., Fujima, K. and Mizutani, N.: A Study on Estimation of Tsunami Force
Acting on Structures, Journal of Japan Society of Civil Engineers, Ser. B2 (Coastal Engineering), Vol.65，
No.1, pp.321-325, 2009 (in Japanese)
4) Arimitsu, T., Ooe, K. and Kawasaki, K.: Proposal of Evaluation Method of Tsunami Wave Pressure
Using 2DDepth-integrated Flow Simulation, Journal of Japan Society of Civil Engineers, Ser. B2
(Coastal Engineering), Vol.68, No.2, pp. I_776-I_780, 2012 (in Japanese)
5) The Ministry of Land, Infrastructure, Transport and Tourism. “Establishment on Safety Building
construction method against assumed tsunami in the tsunami inundation estimation”, No.1318, 2011 (In
Japanese)
6) Arikawa, T., Yamada, F. and Akiyama, M.: A Study on Applicability of Tsunami Wave Forces in

20
CHAPTER 10 TSUNAMI LOADS - C10-21 -

CADMAS-SURF/3D，Proceedings of Coastal Engineering, JSCE, Vol.52, pp.46-50, 2005 (in Japanese)

7) Arikawa, T., Nakano, F., Ohtsubo, D., Shimosako, K. and Ishikawa, N.: Research on Distribution and
Deformation of Structures due to Surge Front Tsunami, Journal of Japan Society of Civil Engineers, Ser.
B2 (Coastal Engineering), Vol.54, pp.841-845，2007 (in Japanese)
8) Hayashi, T., Hattori, S. and Hayashi, K.: A Wave pressure and an Impulse of Breaking Waves，
Proceedings of Coastal Engineering, JSCE, Vol.5, pp.21-27, 1958 (in Japanese)
9) Matsutomi, H., Ohmukai, T. and Imai, K.: Fluid Force on a Large Structure Due to an Inundated Flow
Caused by a Tsunami: Annual journal of Hydraulic Engineering, JSCE, Vol.48, pp.559-564, 2004 (in
Japanese)
10) Asakura, R., Iwase, K., Ikeya T., Takao, M., Kaneko, T., Fujii, N. and Omori, M.: An Experimental
Study on Wave Force Acting on On-Shore Structures due to Overflowing Tsunamis, Proceedings of
Coastal Engineering, JSCE, Vol.47, pp.911-915, 2000 (in Japanese)
11) Takahashi, T., Nakagawa, H. and Kanou, S.: Risk Estimation against Washed away of Wooden
Houses by a Flooding, Disaster Prevention Research Institute Annuals. Vol.28, B-2, pp.455-470, 1985
(in Japanese)
12) Matsutomi H.: Problems in Recent Onshore and Land Tsunamis, Summer Seminar on Hydraulic
Engineering, JSCE, pp.B-3-1- 20, 2009 (in Japanese)
13) Peregrine, D. H., and S. M. Williams: Swash overtopping a truncated plane beach, Journal of Fluid
Mechanics，Vol.440，pp.391-399，2001
14) Building Performance Standardization Association, Institute of Industrial Science, the University of
Tokyo, Kajima Corporation：A Report of the Building Standards Improvement Project No. 49 “A Study
on Improvement of Building Standards for Tsunami Evacuation Buildings”, 2013 (in Japanese)
15) Matsutomi, H., Kettoku, G. and Shimazu, T. ： Front Inundation Depth, Sliding-Overturning
Condition and Decrease in Tsunami Horizontal Fluid Force of RC Building with Aperture，Proceedings
of JAEE Annual Conference，pp.41-42，2013 (in Japanese)
16) The Japan Building Disaster Prevention Association：Standard for Seismic Evaluation of Existing
Reinforced Concrete Building, 2001
- C11-1 -

CHAPTER 11 IMPULSIVE LOADS

Introduction
An impulsive load is a newly introduced topic in this edition. Japan is an earthquake prone nation
and lies directly within an active typhoon region. To prepare for such natural disasters, structural
building designs must emphasize earthquake and wind loads. In addition to countermeasures against
these types of natural phenomena, additional interest has been directed toward the impact and blast loads
from artificial phenomena1). The progressive collapse of the World Trade Center during the September
11 attack generated significant concern regarding impulsive loads within the architectural society. After
this event, the Architectural Institute of Japan immediately organized a special investigation committee
and published its “Special investigation committee’s report: Collapse of the WTC, Damage to the
Pentagon, and Lessons Learned” in 20032). Although a bomb-resistant design had once been considered
in the architectural field during World War II, little attention had been paid to such studies after the war.
Rapid urbanization and changes to the surrounding conditions in our society, however, have led to the
demand for an established design methodology that ensures the safety and security of buildings based
on a perspective appropriate to the current political climate3).
When considering recent domestic events, there have been several cases in which buildings have
been damaged through an impulsive load. In 2004, a military helicopter crashed into and damaged a
building on a college campus in Okinawa, and a gas explosion occurred in a museum in Chiba. In 2005,
express trains ran over some blocks and damaged a station building in Kochi, and derailed trains crashed
into an apartment building in Hyogo. In 2006, a road vehicle fell from a multilevel parking tower in
Yamaguchi. In 2007, natural gas exploded in an urban spa in Tokyo. In 2011, the Fukushima Daiichi
nuclear power plant exploded and released radioactive materials into the environment after the Great
East Japan Earthquake. In 2012, an explosion occurred at a chemical plant in Yamaguchi and damaged
residential houses in the surrounding area.
The impulsive load dealt with herein is limited to artificial phenomena occurring accidentally in
ordinary buildings. In Japan, an impulsive load has been treated in the field of civil engineering as a
design load for the construction of rockfall protective structures. The US and the UK have established
blast-resistant design guidelines addressing an external explosion. The Eurocode includes impact loads
caused by road vehicles, derailed trains, helicopters and forklift trucks, as well as a blast load caused by
an internal explosion. In the recommendations, an impulsive load acting on a building is determined
based on reference to the accumulated achievements4)-9) of the Subcommittee on Shock Resistant
Performance Evaluation at the Architectural Institute of Japan.
- C11-2- Recommendations for Loads on Buildings

Definition of symbols
The following is a list of symbols used in this chapter.

Capital letters
𝐴𝐴V (m2): Opening area
𝐸𝐸b (kJ/mol): Explosion energy
𝐹𝐹 (kN): Impact force
𝐹𝐹b (kN): Blast force
𝐼𝐼 (kPa∙s): Impulse of explosion
𝐾𝐾 (m/kg1/3): Scaled distance
𝑃𝑃 (kN∙s): Momentum
𝑄𝑄 (m/s): Moving speed of gas
𝑅𝑅 (m): Offset distance
𝑅𝑅: Scaled distance of peak pressure
𝑆𝑆L (m/s): Laminar combustion velocity
𝑉𝑉 (m3): Volume of explosive gas
𝑊𝑊 (kg): TNT equivalent mass of explosives

Small letters
𝑎𝑎0 : (m/s): Speed of sound
𝑒𝑒: Expansion ratio
𝑔𝑔 (m/s2): Gravitational acceleration
ℎ (m): Drop height
𝑙𝑙 (m): Length of impactor
𝑚𝑚 (ton): Mass of impactor
𝑣𝑣 (km/h): Impact velocity
𝑝𝑝 (kPa): Explosion pressure
𝑝𝑝0 (kPa): Atmospheric pressure
𝑝𝑝d (kPa): Maximum pressure
𝑝𝑝r (kPa): Peak reflected pressure
𝑝𝑝s (kPa): Peak incident pressure
𝑝𝑝stat (kPa): Release pressure at fragile part
𝑡𝑡 (s): Time

Greek letters
𝛼𝛼: Thermal diffusion coefficient of mixed gas
𝜀𝜀: Expansion ratio of mixed gas
𝜌𝜌: Density of mixed gas
CHAPTER 11 IMPULSIVE LOADS - C11-3 -

11.1 Scope and evaluation policy

11.1.1 Scope

(1) Target buildings

The buildings addressed in the recommendations cover offices, apartments, hotels, hospitals, schools,
and public and commercial buildings. The structural type is limited to multi-story frame structures
including reinforced concrete and steel buildings. The scope of the application of the recommendations
is not intended for the shock-resistant design of a non-typical building such as a fortress.

(2) Target events

Impact and explosion loads may be caused by natural phenomena such as earthquakes, heavy storms
or eruptions, and by artificial phenomena such as car accidents or gas explosions. The recommendations
cover only impact and explosion loads resulting from artificial phenomena. Events resulting from
natural phenomena are not considered. For example, the pounding between two neighboring buildings
caused by earthquake ground motion, or the impact of flying objects blown upward by high-velocity
winds are out of the recommendations. Artificial phenomena exclude intentional events such as war and
terrorist attack, as well as events related to nuclear power plants. That is, the recommendations cover
only impulsive loads on typical buildings resulting from accidental action. The impact and explosion
events that are within the scope of the recommendations are shown in Tables 11.1.1 and 11.1.2,
respectively.

(3) Target limit states

In general, an impulsive load acts as a remarkably large load on a portion of a building for a very
short time. We consider that an area where an impulsive load directly occurs follows the safety limit
states. The neighboring area where an impulsive load indirectly occurs follows the functionality limit
states and/or functional retainability limit states. It is necessary to consider any notable uncertainty in
an impulsive load and to prepare for cases of a large unexpected impulsive load in the design stage. We
must secure the building robustness to prevent a progressive collapse of the entire structural system
initiated by a local failure as an initial event.
- C11-4- Recommendations for Loads on Buildings

Table 11.1.1 Intended impact events

Impactor Actions
An impact into a building (columns, walls, and beams) adjacent to a road.
Road vehicle
Consider events in a parking lot, including impact and falling events.
An impact into buildings (columns, walls and beams) adjacent to a railway.
Derailed train
Consider an impact from a train that ran over the tail end of track
An impact into a building with a height of over 60 m.
Small aircraft
Consider a crash into the ground around an airport or under a flight route.
Helicopter A crash into a rooftop designed as a heliport during an emergency landing.
Forklift An impact of a forklift against a building and luggage frame during operation.

Table 11.1.2 Intended explosion events

Internal Ignition of leaked explosive gas (methane, propane, or city gas) in a closed or
explosion semi-closed space from static electricity, a spark, or a lighter.
Buildings near hazardous facilities including explosive material factories,
External chemical plants, or hydrogen stations.
explosion Suppose an explosive source, including explosive gas, water vapor, dust,
explosive materials, or high explosives.

11.1.2 Evaluation policy

(1) Evaluation policy of impact load

A significant amount of impact test data is available for road vehicles. Such data have been published
and shared in the US. In comparison, a few experiments have been conducted on other impactors.
Although impact experiments have been conducted on impactors other than road vehicles, such data
have yet to be opened to the public. For impact test data that are available, such as on road vehicles, we
use the maximum value observed in the time history of the impact force as a static design load. When
no reasonable impact test data are available, we determine the time history of the impact force based on
crash simulation results or by using a simplified energy method first, and then approximate the resulting
time history as a triangular wave with the equivalent impulse, and finally use the maximum value of the
triangular wave as a static design load.
The mechanical interaction between an impactor and a building governs the magnitude of the impact
force observed in such an impact event. Several factors affect the impact force, such as the weight of
the impactor, the impact velocity, the incident angle, the deformation properties of both the impactor
and the building, and the damping characteristics. The Eurocode10) provides an impact-related design
load for buildings as an equivalent static load. However, the basis for determining the design load is not
clear. To avoid ambiguity, we provide a static design load as the maximum value obtained from the time
CHAPTER 11 IMPULSIVE LOADS - C11-5 -

history of the impact force. The recommendations assume a hard impact in which all kinetic energy is
absorbed as internal energy caused by the deformation of the impactor. By assuming a hard impact, we
can assure a safer evaluation of design load. When preventive measures are used to reduce the
occurrence rate and magnitude of the impact force, the design load can be accordingly reduced.
An impact simulation analysis is one of the ways to obtain the time history of an impact force. The
finite element method (FEM) is commonly used in the simulation analysis and performed using
commercial software. The accuracy of the results depends on the appropriate model selection, parameter
settings, mesh sizes, and time steps. Significant attention should be paid to these details.

(2) Evaluation policy of blast load

For determining a design load associated with an explosion, there are many empirical formulas that
have been derived from tests on both internal and external explosions. For the recommendations, these
proposed empirical formulas were reviewed to obtain the time history of a blast force, and the resulting
maximum value is used as a static design load. Only a gas explosion inside a building is considered as
an internal explosion event. Additional impact effects of the debris generated from a blast, such as
concrete blocks, glass fragments and scattering, are ignored in the recommendations.

(3) Combination of impulsive and other loads

An impulsive load can be combined exclusively with dead and live loads. The basic value of an
impulsive load is determined as the force corresponding to an annual exceedance probability of 1% that
is equivalent to a return period of 100 years. Owing to insufficient data accumulation, the
recommendations do not consider a stochastic load combination. The simultaneous action of an
earthquake or wind load is ignored because of low probabilities of such load combinations. Additionally,
a combination with a snow load is ignored because the contribution of a snow load becomes on the safer
side.

11.2 Impact load

11.2.1 Road vehicle

(1) Buildings and road vehicles

Buildings with an expected risk of a road vehicle impact are described below.
1) Buildings located adjacent to a road
Determine the impact force by considering the items described below.
- Consider the impact of standard- and large-sized vehicles.
- Consider the driving speed appropriate for the location, such as cities and suburbs.
- Consider the types of vehicles appropriate for a main road or local street. Include large-sized vehicles
only for a main road.
- C11-6- Recommendations for Loads on Buildings

- Motorways are out of the scope of the recommendations because they have preventive measures for
falling and deviating events.
- Consider the types of roads, such as straight, curved, and road crossings.
- Consider the structural members of a building, such as the columns and walls. Include beams only for
large-sized vehicles.
- Mitigate the design load corresponding to any preventive measures installed on site for an impact event,
such as crash barriers, gaps, gullies, shrubberies, bollards, and parking blocks.
2) Buildings that permit vehicles to enter the site
Determine the impact force by considering the items described below.
- Consider the impact of standard- and large-sized vehicles.
- Consider erroneous operations by drivers and vehicles entering inside a building near any large parking
lots, such as a roadside restaurant, supermarket, or convenience store.
- Consider nonstructural members of the building, such as window glass used in a facade, front door, or
other object, in addition to structural members.
3) Buildings that are entirely or partially used as a parking lot
Determine the impact force by considering the items described below.
- Consider the structural members of the building, such as the columns and walls inside a parking garage.
- When the building has a parking garage on the 2nd floor or higher, or it is standing next to another
building that has a parking garage on the 2nd floor or higher, consider the impact of a road vehicle
falling from the building. Determine the impact height, velocity, and direction corresponding to the
details of the expected event.
- Ignore an impact from a road vehicle falling from an elevator-type parking garage because
driving/operating a car inside the building is not allowed.
- When providing some preventive measures to mitigate the impact occurrence rate and magnitude of
an impact force, the design load can be mitigated accordingly.
- As a reference, see “The guidelines for preventive measures and equipment for a falling road vehicle
incident in a parking garage (Guidelines to standard building structural technologies, Tokyo, 2007) 11)”.

(2) Approach to determine impact force for road vehicles

The impact force is determined by considering the items described below.
- Determine the impact force acting on a building when considering the surrounding conditions, the
expected weight, and the impact velocity of the road vehicle.
- Suppose a hard impact in which all kinetic energy is absorbed as internal energy through a deformation
of the impactor.
- For an expected road vehicle fall, the velocity of the fall is given by
𝑣𝑣 = �2𝑔𝑔ℎ, (11.2.1)
where v is the impact velocity, g is the gravitational acceleration, and h is the drop height.
- Determine the acting position and acting area of impact force corresponding to the height and width
CHAPTER 11 IMPULSIVE LOADS - C11-7 -

of the vehicle and its position relative to the building.

- Use conservative characteristic values for physical constants of both the impactor and the building.
Consider a strain-rate effect accordingly.
- For cargo trucks, calculate the impact force with the total weight of the vehicle body, which consists
of the vehicle weight, superimposed load, and occupant weight. Consider the effect of a fictitious force
from the cargo/occupant load.

(3) Design load used in Eurocode

For reference, the design load used in the Eurocode (an equivalent static load) is described below.
The Eurocode10) provides a road vehicle crash impact force for support members including columns
and walls of a building adjacent to a road, as shown in Table 11.2.1.

Table 11.2.1 Effective design load for a road vehicle impact on the support members of a building
adjacent to a road.
Material Force 𝐹𝐹dx (kN) Force 𝐹𝐹dy (kN)
Motorway, national road, main road 1000 500
Farm roads in rural areas 750 375
Urban roads 500 250
Patios, parking garages
Standard-sized vehicles 50 25
Large-sized vehicles (trucks) 150 75
* x = moving direction of the road vehicle, y = direction orthogonal to the moving direction
* A cargo truck indicates a vehicle with a load capacity more than 3.5 ton.

It is recommended to not assume that the load in the moving direction of the road vehicle 𝐹𝐹dx and
the load in the direction orthogonal to the moving direction 𝐹𝐹dy act simultaneously. Thus, 𝐹𝐹dx and 𝐹𝐹dy
should be described and determined separately. Impact forces acting on the support members are
illustrated in Fig. 11.2.1. The impact load is given as an equivalent distributed load, which is determined
by dividing the equivalent static load given in Table 11.2.1 by the acting area.
1) Case of standard-sized vehicle impact events
Assume that the position of the impact force (corresponding to ℎ in Fig. 11.2.1) is 0.5 m above the
road surface. The height of the acting surface is 0.25 m above the road surface (corresponding to 𝑎𝑎 in
Fig. 11.2.1), and the width is 1.5 m or equal to the width of the structural member if it is narrower than
1.5 m.
- C11-8- Recommendations for Loads on Buildings

2) Case of large-sized vehicle (truck) impact events

Assume that the position of the impact force (corresponding to ℎ in Fig. 11.2.1) is between 0.5 and
1.5 m above the road surface or equal to the height of the protection wall. The height of the acting
surface is 0.5 m above the road surface (corresponding to 𝑎𝑎 in Fig. 11.2.1), and the width is 1.5 m, or
equal to the width of the structural member if narrower than 1.5 m.

Structural members adjacent to the road such as columns and

Impact
𝑎𝑎 load
ℎ
ℎ: Height of impact load
𝑎𝑎: Range of acting position

Figure 11.2.1 Impact load acting on support members of the building adjacent to a road

Table 11.2.2 List of standard-sized vehicle impact test results12)-15)

Vehicle Impact Max. Peak Load
Momentum*1
Vehicle Weight velocity force time duration*2
𝑃𝑃 (kN∙s)
𝑚𝑚 (t) 𝑣𝑣 (km/h) 𝐹𝐹 (kN) 𝑡𝑡p (s) 𝑡𝑡end (s)
A 1.8 56.2 28 670 0.036 0.083
1.2 39.8 13 370 0.038 0.072
B 1.3 56.3 20 540 0.031 0.074
1.3 56.2 19 430 0.031 0.090
C 1.6 56 26 680 0.025 0.075
D 1.4 56.5 21 480 0.028 0.089
*1 The momentum is given by 𝑃𝑃 = 𝑚𝑚𝑚𝑚, where 𝑚𝑚 is the vehicle weight and 𝑣𝑣 is the impact velocity.
*2 The load duration is given by 𝑡𝑡end = 2𝑃𝑃/𝐹𝐹.
CHAPTER 11 IMPULSIVE LOADS - C11-9 -

Figure 11.2.2 Triangular wave approximation of the time history of impact force

Figure 11.2.3 Momentum 𝑃𝑃 (kN∙s) and maximum force 𝐹𝐹 (kN) in frontal impact test of road

(4) Calculation example of a design load for a standard-sized vehicle

1) Calculation example based on an impact test
The time history of the impact force obtained from a standard-sized vehicle impact test is shown in
Table 11.2.2. The resulting time history is approximated as a triangular wave, as shown in Fig. 11.2.2.
Herein, 𝐹𝐹 is the maximum force, 𝑡𝑡p is the peak time, and 𝑡𝑡end is the load duration. The maximum
force 𝐹𝐹 can be determined by giving momentum 𝑃𝑃 as the index of impact conditions. The relationship
between the momentum and maximum force is shown through Eq. (11.2.2) and in Fig. 11.2.3. The static
design load owing to a road vehicle impact is given as the maximum value obtained from Eq. (11.2.2).
Determine the acting position and the acting area of the impact force corresponding to the height and
width of the vehicle and its relative position to the building.

𝐹𝐹(kN) = 23.11𝑃𝑃(kN∙s) +37.45 (11.2.2)

- C11-10- Recommendations for Loads on Buildings

2) Calculation example based on crash simulation

Figure 11.2.4 Road vehicle FEM model

(a) Simulation result (b) Simplified waveform

Figure 11.2.5 Time history of impact load for each simulated impact velocity for road vehicle

The finite element model of the vehicle is used to obtain its impact force. A crash simulation was
conducted using the finite element model provided by the National Crash Analysis Center (NCAC)16)
and crash simulation software, LS-DYNA17).
To simulate the behavior of a hard impact, a frontal crash into a rigid wall is applied for four impact
velocities, namely, 20, 40, 60, and 80 km/h. Figure 11.2.4 shows the vehicle model used. Simplified
waveforms for the respective velocities obtained from the simulation are shown in Fig.11.2.5 and their
values are listed in Table 11.2.4. The simplified waveform is assumed to be a triangular wave, and the
maximum force value is determined in accordance with the equivalent impulse from the simulation
results. The static design load for a road vehicle is given as the maximum force of the triangular wave.
The acting position and acting area of the impact force are determined corresponding to the height and
width of the vehicle and its relative position to the building.

Table 11.2.4 Crash simulation results: maximum load, peak time, and load duration of impact force
for road vehicle
Impact force Maximum force *1 Peak time*2 Load duration*1
𝑣𝑣 (km / h) 𝐹𝐹 (kN) 𝑡𝑡p (s) 𝑡𝑡 (s)
20 275 0.045 0.085
40 469 0.056 0.090
60 700 0.043 0.092
80 921 0.031 0.092
*1 The load duration and maximum force are determined in accordance with the equivalent
impulse of the simulation result.
*2 The peak time is the time at which the maximum force occurs during the simulation.
CHAPTER 11 IMPULSIVE LOADS - C11-11 -

(5) Example calculation of design load for a large-sized vehicle

Figure 11.2.6 Finite element model of a truck

(a) Simulation result (b) Simplified waveform

Figure 11.2.7 Time history of impact force for each simulated impact velocity for trucks

To represent a cargo truck, we use a Ford truck model16). The vehicle model is shown in Fig.11.2.6.
In the same way as for a standard-sized vehicle, a hard impact event is supposed (frontal impact into a
rigid wall17)) for three impact velocities, namely, 20, 40, and 60 km/h. The time history of the impact
force for the respective velocities obtained from the simulation is shown in Fig.11.2.7, and the values
are listed in Table 11.2.5. The simplified waveform is assumed to be a triangular wave, and the maximum
force is determined in accordance with the equivalent impulse of the simulation result. The static design
load for a large-sized vehicle is given as the maximum force of the triangular wave. The acting position
and acting area of the impact force corresponding to the height and width of vehicle are determined,
along with the position relative to the building
.
Table 11.2.5 Simplified waveform for each simulated impact velocity for trucks
Impact force Maximum force*1 Peak time*2 Load duration*1
𝑣𝑣 (km / h) 𝐹𝐹 (kN) 𝑡𝑡p (s) 𝑡𝑡 (s)
20 1240 0.02 0.085
40 1690 0.015 0.12
60 2030 0.01 0.14
*1 The load duration and maximum force are determined in accordance with the
equivalent impulse of the simulation result.
*2 The peak time is the time at which the maximum force occurs during the
simulation.
- C11-12- Recommendations for Loads on Buildings

(6) Mitigation of impact force

The design load is determined for an impactor directly impacting a building or its structural members.
However, the impact force can be mitigated using preventive measures to absorb the impact energy and
avoid a crash. An impact simulation of a frontal crash of a standard-sized vehicle into crash barriers is
performed17) to evaluate the buffer effect of structural members when using preventive measures. A case
in which a standard-sized vehicle crashes head into the center of a crash barrier at an impact velocity of
40 km/h is assumed.
The simulation model and time history of the impact force are illustrated in Fig. 11.2.8. The maximum
magnitude of the impact force is mitigated by approximately 45% compared with the case in which no
barriers are used. The impact energy increases in proportion to the mass of the impactor and the impact
velocity. The increased impact energy results in a plastic deformation of the barrier, and the energy
absorption capability is accordingly increased. However, an overly deformed barrier has less energy
absorption capability owing to a decrease in resistance and a fracture of the barrier itself. The change in
energy absorption capability should be properly evaluated depending on the circumstances.
.

(a) Simulation (b) Impact

Figure 11.2.8 Time history of impact force for an impact into a crash

Table 11.2.6 Example of preventive measures installed in building subjected to a road vehicle impact
event
Purpose Measures
Impact prevention Bollards, parking blocks, guard, barrier, fence, gate,
Bench, hedge (trees, stones, rocks)
Mitigation of the impact velocity Gullies, gaps, steps, up-hill grade
Speed limit Crank turns, slalom turns, humped road
Traffic congestion reduction Cul-de-sac (dead end street), looped street
Fall prevention Fence, car stop
CHAPTER 11 IMPULSIVE LOADS - C11-13 -

Examples of preventive and mitigation measures for buildings subjected to a road vehicle impact
event are shown in Table 11.2.6. Possible preventive and mitigation measures against a road vehicle
impact include the installation of obstacles or grooves to reduce the impact velocity as a hardware
solution, and notifications and warnings to reduce driving speed or traffic volume as a software solution.

11.2.2 Derailed train

(1) Buildings and derailed trains

Buildings with a risk of a derailed train impact are expected, and derailed trains as impactors are
described below.
1) Buildings located adjacent to railway track
Determine the impact force by considering the items described below.
- Determine the weight, velocity, and impact direction of the train corresponding to the type of train
(local, rapid, or express).
- Consider the shape of the railway track (straight, slow curve, or sharp curve).
- Consider the structural members of the building, such as columns and walls.
If preventive measures for an impact are installed and their effectiveness is confirmed through
experiments and/or simulations, the design load can be mitigated accordingly.
2) Buildings with train tracks inside
Determine the impact force by considering the items described below.
- For buildings that enclose the train tracks inside, such as station buildings and train sheds, consider
the impact on the interior of buildings owing to erroneous operations.
- Consider nonstructural members of the building, such as window glass for a facade, front door, or
other objects in addition to structural members.
- Mitigate the design load corresponding to any preventive measures installed for an impact on site, such
as train stops, gaps, gullies, or shrubberies.

(2) Approach to determine the impact force for derailed trains

The impact force is determined by considering the items described below.
- Determine the impact force corresponding to the building, the surrounding conditions, and the weight
and impact velocity of the derailed train.
- Suppose a hard impact in which all kinetic energy is absorbed as internal energy caused by a
deformation of the impactor.
- Consider the types of train, such as local and express trains in a conventional railway. Shinkansen
(bullet train) is out of the scope of this research.
- Consider only the impact of the first carriage. The impact of more than one carriage can be considered
individually, or as a continuous or integrated load of the impact force of the first carriage.
- The acting area and acting position of an impact load are determined corresponding to the height and
- C11-14- Recommendations for Loads on Buildings

width of the carriage and its position relative to the building.

- Conservative characteristic values of physical parameters are used for both the impactor and building.
Consider a strain-rate effect accordingly.

(3) Calculation example of a design load for a derailed train

Use the finite element model of a train created by Hisamori, Tachibana et al.18). The carriage is made
from stainless material. The total length is 20 m, and the weight is 26.3 t. The carriage model is shown
in Fig.11.2.9. In the same way as the approach for road vehicles, suppose a hard impact event (frontal
impact into a rigid wall) for four impact velocities of 20, 40, 60, and 80 km/h.
The time history of the impact force for the respective velocities obtained from the simulation is
shown in Fig. 11.2.10. The major deformation is found at the front part of the carriage and is caused by
buckling. The result indicates that the deformation of the carriage develops down to the front wheels,
which are stiff. The maximum force in the time history occurs when the wheels experience an impact.
The values for the simplified waveform are shown in Table 11.2.7. The waveform is simplified as a
triangular wave. Occasionally a triangular approximation of the time history becomes difficult, for
example, at impact velocities of 60 and 80 km/h it is not easy to approximate as triangular waves. In
such cases, the triangular wave is determined using the maximum value and slope around the maximum
force, as described below:
① The time histories of the impact force for impact velocities of 60 and 80 km/h show the maximum
force at the time of the second peak corresponding to a wheel impact.

Figure 11.2.9 FEM model of derailed train

(a) Simulation results (b) Simplified waveforms

Figure 11.2.10 Time history of impact force for each simulated impact velocity of derailed trains
CHAPTER 11 IMPULSIVE LOADS - C11-15 -

Table 11.2.7 Simplified waveform of the time history of the impact force for each simulated impact
velocity for derailed trains
Impact force Maximum force*1 Peak time*2 Load duration*1
𝑣𝑣 (km / h) 𝐹𝐹 (kN) 𝑡𝑡p (s) 𝑡𝑡 (s)
20 3500 0.05 0.12
40 5000 0.02 0.17
60 10500 0.015 0.10
80 28100 0.011 0.07
*1 The load duration and maximum force are determined in accordance with the equivalent
impulse of the simulation results.
*2 The peak time is the time at which the maximum force occurs during the simulation. For
velocities of 60 and 80 km/h, determine the peak time to reproduce the slope around the
maximum force. The peak time is determined as the value calculated by dividing the maximum
force by the slope around the maximum force.

② The maximum response of the building is caused by this maximum force. Therefore, this value
is used as the maximum force of a triangular wave.
③ Determine the peak time to reproduce the slope around the maximum force. The peak time is
determined as the value calculated by dividing the maximum force by the slope around the
maximum force.
④ By considering the maximum force and the slope around the maximum force, the maximum
response of the building and the dynamic effect are expected to be reproduced.
The impulse of a triangular wave is determined in accordance with the equivalent impulse of the s
imulation time history. The static design load for a derailed train is given as the maximum force of a
triangular wave. The acting position and acting area of an impact force corresponding to the height and
width of the carriage and its position relative to the building are determined.

11.2.3 Small aircraft

(1) Buildings and small aircrafts

Buildings with a risk of a small aircraft impact are expected, and small aircrafts as impactors are
described below.
- Consider buildings 60 m in height or taller that stand along a flight route or near an airport that accepts
small aircrafts.
- Determine the size and weight of a small aircraft and the impact direction corresponding to the flight
conditions.
- Determine the impact velocity, the height of the impact point, and the impact area corresponding to
the flight route of the small aircraft.
- Consider the structural members of the building, such as the columns, walls, and beams.
- C11-16- Recommendations for Loads on Buildings

(2) Approach used to determine impact force for a small aircraft

The impact force for a small aircraft is determined by considering the items described below.
- Determine the impact force corresponding to the weight and impact velocity of small aircraft and the
surrounding condition of the building.
- Suppose a hard impact in which all kinetic energy is absorbed as internal energy caused by the
deformation of a small aircraft.
- Determine the acting area of impact force by considering the height and width of a small aircraft and
its relative position to the building.
- Use conservative characteristic values for physical constants for small aircraft and structural and
nonstructural members of the building. Consider a strain-rate effect accordingly.

(3) Calculation example of a design load for a small aircraft

According to the results of an aircraft impact test, it was shown that the fracture strength has a
relatively small effect on the impulsive load, and the inertia force of the mass governs the impulsive
load instead. It is known that the shape of the time history of an impact force resembles the mass
distribution of an aircraft body20). We suppose that small aircrafts have a similar tendency and assume
that the mass distribution of a small aircraft is approximated as a triangular shape. Then, the shape of
the time history of the impact force can be approximated as a triangular wave. We also assume that a
small aircraft does not slow down before or at the time of impact, the load duration is then equal to the
value of the total length of the small aircraft divided by the impact velocity. Furthermore, by assuming
that the momentum of the small aircraft (the weight of the body multiplied by the impact velocity) is
transformed into the impulse at the time of impact, the time history of the impact force can be
estimated21). The maximum force is determined based on the equivalent impulse and the load duration.
The impact force is calculated based on the expected specifications of a small aircraft including the
weight, length and impact velocity, and by using the procedures described above, Examples are shown
in Fig11.2.11 and Table 11.2.8. The static design load for a small aircraft is given as the maximum force
of a triangular wave. The acting position and acting area of the impact force should be determined by
considering the height and width of the aircraft and its position relative to the building.

Figure 11.2.11 Impact force of a small aircraft

CHAPTER 11 IMPULSIVE LOADS - C11-17 -

Table 11.2.8 Calculation example of the impact force of a small aircraft

Load duration Maximum force
Weight Length Impact velocity
𝑡𝑡 = 𝑙𝑙/𝑣𝑣 𝐹𝐹 = 2𝑚𝑚𝑚𝑚/𝑡𝑡
𝑚𝑚 (t) 𝑙𝑙 (m) 𝑣𝑣 (km / h)
𝑡𝑡 (s) 𝐹𝐹 (kN)
180 0.17 588
1 8.5
200 0.153 726
180 0.17 1176
2 8.5
200 0.153 1452

11.2.4 Helicopter

(1) Buildings and helicopters

Buildings with a risk of a helicopter crash and helicopters as impactors are described as follows:
- Consider buildings that have heliports on their rooftop.
- Determine the size and weight of a helicopter corresponding to the flight condition.
- Determine the drop height and impact velocity corresponding to the flight condition.
- Consider a crash on the rooftop, and as appropriate, consider the impact on the columns, walls, and
beams of the upper floors owing to erroneous operations.
- Mitigate the design load corresponding to any preventive measures installed at the site for a helicopter
impact event: fences, buffers and as such.

(2) Approach to determine the impact force for a helicopter

Determine the impact force by considering the items described below.
- Determine the impact force corresponding to the conditions of the heliport, the weight of the helicopter,
and the impact velocity.
- Suppose a hard impact in which all kinetic energy is absorbed by the internal energy caused by the
deformation of a helicopter.
- Suppose a case in which a helicopter crashes on the rooftop from midair.
- Determine the acting area of the impact force by considering the height, width, and length of the
helicopter and its position relative to the building.
- Use conservative characteristic values of the physical constants for a helicopter and the structural and
nonstructural members of the building. Consider the strain-rate effect accordingly.

(3) Calculation example of a design load for a helicopter crash

The impact force is calculated under the assumption that the impulse that is the time integration of
the impact force at the time of impact is equal to the momentum that is the weight multiplied by the
- C11-18- Recommendations for Loads on Buildings

impact velocity. In addition, we assume that the helicopter does not slow down before or at the time of
impact, the load duration is equal to the value of the total length of the helicopter divided by the impact
velocity. Determine the maximum force based on the equivalent impulse and load duration. To represent
the considered emergency-landing event in which the helicopter drops from its bottom first, the height
of the cabin is used as the length of the body.
The impact force is calculated using the expected specifications of a helicopter, such as the weight
and height of the cabin and the impact velocity. Examples are shown in Fig.11.2.12 and Table 11.2.9.
The static design load for a helicopter is given as the maximum force of a triangular wave. Determine
the acting position and acting area of impact force corresponding to the height and width of the
helicopter and its position relative to the building.

Figure 11.2.12 Impact force for a helicopter

Table 11.2.9 Calculation example of impact force for a helicopter crash

Cabin Drop Load duration Maximum force
Weight Impact velocity
height height 𝑡𝑡 = 𝐻𝐻/𝑣𝑣 𝐹𝐹 = 2𝑚𝑚𝑚𝑚/𝑡𝑡
𝑚𝑚 (t) 𝑣𝑣 (km/h)
𝐻𝐻 (m) ℎ (m) 𝑡𝑡 (s) F (kN)
5 35.6 0.152 261
2
10 50.4 0.107 523
1.5
5 35.6 0.152 523
4
10 50.4 0.107 1045

11.2.5 Forklift

(1) Buildings and forklifts

Buildings with a risk of a forklift impact are expected, and forklifts as impactors are described.
- For buildings that have working areas for forklifts, consider the impact inside and outside of the
building owing to erroneous operations.
CHAPTER 11 IMPULSIVE LOADS - C11-19 -

- Consider nonstructural members of the building, such as window glass for a facade, front door, or
other objects in addition to the structural members.
- Mitigate the design load corresponding to any preventive measures for a forklift impact installed on-
site, such as fences or gullies.

(2) Approach used to determine the impact force of a forklift

The impact force is determined for a forklift by considering the items described below.
- Determine the impact force corresponding to the surrounding condition of the building, the work
environment of the forklift, the total weight of the forklift (the sum of the weight of the forklift and its
superimposed load), and the impact velocity.
- Suppose a hard impact in which all kinetic energy is absorbed as internal energy caused by the
deformation of a forklift.
- Determine the acting area and acting position of the impact force by considering the height, width, and
length of the forklift and its position relative to the building.
- Use conservative characteristic values for the physical constants of the forklift and structural and
nonstructural members of the buildings. Consider the strain-rate effect accordingly.

Figure 11.2.13 Impact force of a forklift

(3) Calculation example of the design load of a forklift

By assuming that the impulse (time integration of the impact force) at the time of impact is equal to
the momentum (the total weight of the forklift multiplied by the impact velocity), the time history of the
impact force is estimated. In addition, we assume that the forklift does not slow down before or at the
time of impact, the load duration is then equal to the value of the total length of the forklift divided by
the impact velocity. Determine the maximum force based on the equivalent impulse and load duration.
The impact force is calculated using the expected specifications of a forklift, such as the total weight
and length and its impact velocity. Examples are shown in Fig. 11.2.13 and Table 11.2.10. The static
design load for a forklift is given as the maximum force of a triangular wave. The acting position and
acting area of the impact force should be determined by considering the height and width of the forklift
and its position relative to the building
- C11-20- Recommendations for Loads on Buildings

Table 11.2.10 Estimation examples of the impact force of a forklift

Total weight of Load duration Maximum force
Length Impact velocity
a forklift 𝑡𝑡 = 𝑙𝑙/𝑣𝑣 𝐹𝐹 = 2𝑚𝑚𝑚𝑚/𝑡𝑡
𝑙𝑙 (m) 𝑣𝑣 (km / h)
𝑚𝑚 (t) 𝑡𝑡 (s) 𝐹𝐹 (kN)
10 1.08 15
3 3
15 0.72 35
10 1.73 29
9 4.8
15 1.15 65
10 2.63 59
28 7.3
20 1.31 237

11.3 Blast load

11.3.1 Internal explosion

(1) Explosion source and estimated explosion scenario

More than one explosions caused by natural gas have damaged buildings in recent years and killed
and injured people within the buildings. Such events occurred in the Kujukuri sardine museum and in
an urban spa called “Shiespa” in 2004 and 2007, respectively, both of which received significant
attention. The cause of both incidents is thought to be natural methane from the southern Kanto gas field.
The methane filled up a closed space, caught fire and exploded. An impulsive load is also caused by the
internal combustion of explosive gas. Possible explosive sources include leaking gas, city gas, and
petroleum fumes. Although vaporized gasoline is a rather familiar explosive gas in open air, we can
disregard it in space such as an underground parking garage because the gasoline leaking from the tank
of a vehicle may not be sufficiently vaporized to cause an internal explosion.
The conditions required for such an explosion are the presence of an explosive gas whose
concentration in air (mixed gas) is sufficient for combustion, as well as the presence of a fire source
providing sufficient energy to cause the mixed gas to explode. A “sufficient” concentration of a gas
depends on the type of gas present. When using the incoming velocity of an explosive gas into a closed
(or semi-closed) space, 𝑄𝑄g , and the venting velocity from the space, 𝑄𝑄a , (both of which are volumetric
flow rates), the average gas concentration in a static state is 100𝑄𝑄g /�𝑄𝑄a + 𝑄𝑄g �%. This indicates that,
when using a large value for 𝑄𝑄a , the internal gas concentration can be kept lower than the lower bound
of the combustion. Thus, installing a sufficient venting system is important to prevent an internal
explosion (Fig.11.3.1).
CHAPTER 11 IMPULSIVE LOADS - C11-21 -

A gas alarm system is an efficient measure for preventing a gas explosion. Other possible preventive
measures for a gas explosion include an odor added that can help detect an increased concentration of
city or petroleum gases, which are originally odor-free.
The following describe the scenario for an internal explosion. Assume circumstances in which an
explosive gas flows into a space, and when none of the preventive measures described above worked as
intended. The gas concentration increases and exceeds the lower bound of combustion. Finally, the gas
catches fire from a particular source. During a gas explosion (deflagration), the internal pressure
increases uniformly, and a uniform load is applied to the structural members including the columns,
beams, walls, floors, and ceilings corresponding to their surface area (the force applied is the pressure
multiplied by the surface area). Therefore, in a typical internal gas explosion, the weakest parts, e.g.,
windows or doors, are destroyed first. We call such parts “fragile.”

(2) Outline of an explosion event

An illustration of a typical internal explosion process of an explosive gas is shown in Fig. 11.3.2.
Figure 11.3.3 shows the experimental model (0.6 m on one side with a fragile part of 0.2 m in diameter).
During an internal explosion event, the internal pressure increases first (middle of Fig. 11.3.2, and
Fig.11.3.3, A). Next, the fragile part is blown away and forms an open area (Fig. 11.3.2 bottom). The
internal pressure is then governed based on the conflict between the increase in pressure caused by a
combustion and the decrease in pressure caused by the outflow of air through the opening. This conflict
forms the first peak (𝑃𝑃1 ) (Fig. 11.3.3, B). Following the form of the opening, the internal flow transits
into a turbulent flow, thereby promoting a combustion event (Fig.11.3.3, C). When the combustion is
promoted drastically, the internal pressure increases and forms a second peak (𝑃𝑃2 ). However, the exact
conditions of 𝑃𝑃2 are unclear at this moment and are contributed to by several factors including the gas-
air ratio, the surface area of the fragile part, and release pressure. Figure 11.3.3 shows a case in which
- C11-22- Recommendations for Loads on Buildings

𝑃𝑃2 is not formed. As shown in the figure, after period 𝑃𝑃2 is formed, a peak marked as 𝑃𝑃2 ′ with a
fluctuation component is formed (Fig. 11.3.3, D). Such phenomenon occurs when the (internal) space
has a nearly cubic shape and a solid wall surface such as metal. It is thought that the interaction of
acoustic wave reflections on a wall surface and a combustion flame result in this phenomenon. It is not
realistic to consider the form of 𝑃𝑃2 ′ because the actual space is typically not cubic, and the material of
the wall can absorb some of the acoustic waves, or objects may prevent their propagation.
Consequently, to design a building in consideration of the risk of an internal explosion, it is important
to include the destruction of fragile parts, which result in the form of an opening, thereby reducing the
maximum pressure. For a special case in which the space has a cubic shape and the wall surface is
constructed of flat metal, we should prepare for some acoustic wave absorption to prevent a form of 𝑃𝑃2 ′.
Even for cases in which fragile parts are not applicable to a space owing to certain restrictions by the
building itself, or other possible reasons, it is not realistic to design structural members to bear the
[NOTE]
maximum pressure, from 0.7 to 0.8 MPa (gauge pressure, see ), that the mixed gas and air may
produce. In this case, the internal pressure continues to increase until the weakest structural member is
blown away and an opening area is formed. To deal with such cases, measures are examined to prevent
an explosion from occurring. A continuous venting system is one possible measure. In addition, consider
the possible inflow of an explosive gas to the space from outside.

[NOTE]:The standard (zero) value for the pressure is atmospheric pressure. The value of the pressure differs
from the absolute pressure (void-standard) based on the value of the atmospheric pressure.
CHAPTER 11 IMPULSIVE LOADS - C11-23 -

(3) Estimated formulas of explosion pressure

It is important to estimate the time history of the internal pressure to reproduce the increasing internal
pressure. To obtain the appropriate temporal change of the internal pressure, compare the estimated
formulas presented based on typical test results22)-24) and excerpted from different papers25). For the
pressure history, compare the increasing pressure history measured by Dahoe et al. with 20 L spherical
containers. For the maximum pressure, compare the experimental results given by Chao et al. and
Cooper et al. Figs. 11.3.4, 11.3.5, and Fig.11.3.6 show comparisons of the experimental results presented
by Dahoe et al., Cooper [Link]., and Chao et al., respectively. Considering that an overestimation of the
maximum pressure leads to an excessively conservative design, the estimated formula given by
Eurocode10) is judged to be the most reasonable expression. From Fig. 11.3.5, we can observe that most
of the existing formulas overestimate the maximum pressure. Figs. 11.3.5 and 11.3.6 show the
experimental results for a case in which the releasing pressure at the opening is low. These preconditions
can represent the actual internal explosion scenario, including the releasing pressure from glazing.
Because the predefined internal volume for the experiments conducted by Cooper et al. and Chao are
2.41 m3 and 63.5 m3, respectively, both of which can be recognized to represent the internal volume of
typical buildings that are targeted in the recommendations. These figures demonstrate the reasonability
of the Eurocode expression regarding the internal explosion in a typical building. For the pressure history,
the estimated formula given by Harris26) is judged to be the most reasonable. As a result, the time history
is obtained from the expression proposed by Harris, namely, Eq. (11.3.1), and the maximum pressure is
provided in Eurocode, Eq. (11.3.2). The maximum value in the time history is selected as a static design
load.
4𝜋𝜋 (𝑆𝑆L 𝑡𝑡)3
𝑝𝑝 = 𝑝𝑝0 exp �𝑒𝑒 2 (𝑒𝑒 − 1) � (11.3.1)
3 𝑉𝑉
- C11-24- Recommendations for Loads on Buildings

𝑝𝑝d = 3 + 𝑝𝑝stat
Max � 𝑝𝑝stat 0.04 � (11.3.2)
𝑝𝑝d = 3 + + 2
2 (𝐴𝐴V ⁄𝑉𝑉 )

𝑉𝑉 < 1,000m3, 0.05 < 𝐴𝐴V ⁄𝑉𝑉 (1/m) < 0.15, 𝑝𝑝d ≤ 50kN/m2

Equation (11.3.1) is a reasonable way to represent the time history until the maximum pressure is
achieved. However, it is not reasonable to represent the behavior after a fragile part has failed. As
previously mentioned, the behavior of the internal pressure following the fracture of a fragile part is
complicated: the combustion increases the pressure, and the outflow decreases the pressure. The
experiments conducted by Cooper et al. demonstrate that the pressure requires about 20 to 30 ms to
decrease to the atmospheric pressure. To allow a conservative verification (let the pressure apply longer),
suppose an explosion pressure history in which it takes about 50 ms to decrease to an atmospheric
pressure after it has achieved the maximum pressure. We call this explosion pressure history a 𝑝𝑝-𝑡𝑡
(pressure vs. time) curve22).
In the expression Eurocode employing Eq. (11.3.2), 𝑃𝑃d (kPa) and 𝑃𝑃stat (kPa) are the maximum
pressure and releasing pressure of a fragile part, respectively, and both use the gauge pressure. Here, 𝐴𝐴V
represents the area of the opening. This expression is effective if 𝑃𝑃d is less than or equal to 50 kN/m2.
The evaluation expression for the maximum pressure estimates the larger between the two pressure
peaks, 𝑝𝑝1 and 𝑝𝑝2 . Meanwhile, Eq. (11.3.1) can represent the behavior until achieving 𝑝𝑝1 . It is not easy
to determine systematically when and how 𝑝𝑝2 occurs, and currently there is no reasonable model
available to represent the pressure history from 𝑝𝑝1 to 𝑝𝑝2 . Therefore, Eq. (11.3.1) is used to represent
the behavior until achieving the maximum pressure calculated from Eq. (11.3.2). As stated before, Eq.
(11.3.1) can represent a slower pressure increase than the actual behavior, and therefore, the
conservativeness of the evaluation can be retained.
CHAPTER 11 IMPULSIVE LOADS - C11-25 -
- C11-26- Recommendations for Loads on Buildings

(4) Example of the time history of internal pressure

Figure 11.3.7 shows an example of a 𝑝𝑝-𝑡𝑡 curve obtained under the conditions described below. This
example is the case of a gas explosion event in a three-story office building used in the “Example of
structural calculation 2 (Architectural Institute of Japan: Standards and practical guide for structural
calculation of reinforced concrete buildings, 1991).” The assumed conditions are as follows:
- The gas explosion occurs in a room of W 7.0 m × D 5.5 m × H 3.6 m = 137 m3 in dimensions.
- The room has a glazing of 3 mm in thickness: W 4.0 m × H 2.0 m = 8 m2, 𝑝𝑝𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 ≈ 2 kPa.
- The explosion is caused by mixed methane gas.
- The concentration of the gas in the room is uniformly 10%.
First, calculate the pressure history using Eq. (11.3.1). The initial pressure 𝑝𝑝0 is the atmospheric
pressure of 101.3 kPa. A value of 7.4 is used for the expansion ratio 𝑒𝑒, and 0.37 m/s is applied as the
combustion velocity 𝑆𝑆L , respectively, from Table 11.3.1. Here, 𝑉𝑉 represents the volume of the room.

Figure 11.3.7 Calculation example of an explosion pressure history (𝑝𝑝-𝑡𝑡 curve) for internal
explosions and 𝐹𝐹b -𝑡𝑡 curve

Table 11.3.1 List of combustion velocity, expansion ratio, etc.26)-27)

Combustion limit Maximum Maximum
combustion laminar
Lower Upper Expansion ratio
Explosive gas velocity combustion
limit limit
concentration velocity
(%) (%) (%) 𝑆𝑆L (m/s) 𝑒𝑒 (-)
Methane CH4 5 15 9.98 0.37 7.4
Ethane C2H6 3 12.5 6.28 0.401 7.5
Propane C3H8 2.1 9.5 4.56 0.43 7.6
Butane C4H10 1.6 8.4 3.52 0.379 7.5
CHAPTER 11 IMPULSIVE LOADS - C11-27 -

Figure 11.3.8 Area dependency of the pressure that breaks the window glass26)

These conditions give the time history of increasing pressure. Next, the maximum pressure is
calculated using Eq. (11.3.2). Figure 11.3.8 shows the dependency of the releasing pressure on the area
and thickness of the glazing. Note that the abscissa is the inverse of the area of glazing. The assumptions
of a “glazing of 3 mm in thickness” and an “area of 8 m2” indicate that the releasing pressure of the
fragile part of the glazing, 𝑝𝑝stat , is about 2 kPa. The area of the fragile part, 𝐴𝐴V , can be considered as
the area of glazing, 8 m2. Here, 𝑉𝑉 is 137 m3. The calculation then gives 𝑝𝑝d =15.7 kPa. Now, we employ
the pressure history calculated through expression (11.3.1) until point 𝑝𝑝= 𝑝𝑝d . By assuming that the
pressure takes 50 ms to decrease until achieving atmospheric pressure, a 𝑝𝑝-𝑡𝑡 curve can be obtained. The
area of each member is multiplied by the pressure acting uniformly on the wall of the room to obtain
force 𝐹𝐹 and the history of the explosion force (𝐹𝐹b -𝑡𝑡 curve). In Fig. 11.3.7, the solid line represents the
pressure, and the dotted lines are the forces acting on the surfaces of the ceiling, floor, wall (7.0 m × 3.6
m), and window, respectively.
Give the maximum values of the 𝑝𝑝-𝑡𝑡 curve and 𝐹𝐹b -𝑡𝑡 curve as static design loads.

11.3.2 External explosion

(1) Explosion source and estimated explosion scenario

To evaluate the effect of an external explosion on the neighboring buildings, several examples of the
explosion sources are shown below.
- Fireworks factory
- Chemical plant
- Liquefied petroleum gas storage facility
- Hydrogen station
When the wind resistance design of a building covers the damage from an explosion event expected
for the building, the verification described here is not required. When any countermeasures or preventive
measures for external explosion events are installed, and the effectiveness of these measures is
- C11-28- Recommendations for Loads on Buildings

confirmed through experiments and/or simulations, the design load can be mitigated accordingly.
As in an internal explosion, the explosive force is determined based on the explosion pressure history
curve (𝑝𝑝-𝑡𝑡 curve). If the 𝑝𝑝-𝑡𝑡 curve is given, the pressure 𝑝𝑝 is integrated with time to obtain the impulse.
The time history of the explosion force applied to each member is given by multiplying the explosion
pressure 𝑝𝑝 by the acting area. Note that one-dimensional structural members such as columns and
beams have a relatively small area, being subject to an impulsive load, and thus there is little possibility
that the area will be damaged. Therefore, only two-dimensional members such as walls, floors, and roofs
are considered as subjected members.
The phenomenon of an external explosion varies in accordance with explosion scenario. In other
words, even when the material is identical, if the scenario differs, then the phenomenon differs as well.
The following are approaches used to derive the design load under the expected scenarios. For all
scenarios, the pressure used is the gauge pressure [NOTE].

(2) Buildings with an expected risk of an external explosion

1) Buildings neighboring a fireworks factory
Fireworks factories, which store large amounts of explosive materials, are usually located sufficiently
far from any ordinary buildings. Thus, an explosion occurring at a fireworks factory can be treated as a
point source explosion. The blast wave caused by a detonation has been widely studied, and it was
confirmed that the TNT equivalent describing the released energy in an external explosion event can be
used. Although black powder rarely explodes to detonation, the TNT equivalent can be applied for the
case of high explosives. The measure of scaled distance measure is then used to estimate the peak
pressure and impulse for the target building. The scaled distance (also known as the 𝐾𝐾 value) is given
by
𝑅𝑅
𝐾𝐾 = 𝑊𝑊 1⁄3 . (11.3.3)

The unit is m/kg1/3. Herein, 𝑅𝑅 is the distance to the point to be considered and 𝑊𝑊 is the TNT equivalent
mass of the explosive materials. The relationship between the value of 𝐾𝐾 and the peak pressure or
impulse can be represented through various formulas. One of the formulas is the relationship called
MITI8728), which is determined based on the experiments conducted by the former Agency of Industrial
Science and Technology in the 1980s. The relationship between scaled distance and blast pressure for
the MITI87 is shown in Fig. 11.3.9.
As an example of damage to human bodies, the US Department of Defense indicates in DoD6055.9-
STD29) that the peak pressure at which a drum membrane may incur damage with 1% probability is 20.7
kPa. The combination of distance and mass of a black powder at which the indicated pressure may occur
is shown in Table 11.3.3. Figure 11.3.9 indicates that the peak pressure is 21 kPa when the value of 𝐾𝐾
is 6.8. For example, the TNT equivalent to 1,000 kg corresponds to 68 m based on the scaled distance,

[NOTE]
: The standard (zero) value for the pressure is the atmospheric pressure. The value for the pressure is different from the absolute
pressure (void-standard), based on the value of the atmospheric pressure.
CHAPTER 11 IMPULSIVE LOADS - C11-29 -

as shown in Eq. (11.3.3). The TNT equivalent for a case in which the distance is double or half the
original can be similarly estimated. For the TNT conversion factor of black powder, although it depends
on the composition, a content percentage of 50% is assumed. The combination of the minimum mass of
powder and the distance required to damage the drum membrane can be estimated as below.
The scaled distance, shown in Eq. (11.3.3), indicates that reducing the mass of a high explosive to 1/8
has the same result as doubling the distance. Table 11.3.3 indicates this as well. Similarly, the impulse
is given by calculating the scaled distance based on the mass of black powder and the distance, and by
referring to Fig.11.3.9 (b).

Table 11.3.3 Mass of black powder and the distance, for 𝐾𝐾= 6.8
Distance (m) 136 68 34
TNT equivalent mass (kg) 8,000 1,000 125
Black powder mass (kg) 16,000 2,000 250

Pressure (bar) Impulse (bar ms/kg1/3)

10 10
MITI 87 MITI 87
data data

1 1

0.1 0.1

0.01 0.01
1 10 100 1 10 100
Scaled distance (m/kg1/3) Scaled distance (m/kg1/3)
(a) Blast pressure (b) Scaled impulse
Figure 11.3.9 MITI87: Blast pressure and scaled impulse for 𝐾𝐾 value
- C11-30- Recommendations for Loads on Buildings

The load duration of the 𝑝𝑝-𝑡𝑡 curve for an external explosion is sufficiently short compared to the
natural period of each structural member. Therefore, important parameters are the peak pressure and
impulse. On the other hand, the details of the shape of pressure history do not affect the response
calculation of the structural members. For simplicity, the time history of blast pressure is approximated
as a triangular wave represented by the peak pressure and the load duration. The scaled distance can
give the peak pressure and impulse. The load duration is then calculated by dividing two-times the
impulse by the peak pressure because the impulse is the area of a triangular wave. The 𝑝𝑝-𝑡𝑡 curve can
be determined as shown in Fig. 11.3.10. To consider an explosion-resistant design for buildings
neighboring a firewo rks factory, give the maximum value in the 𝑝𝑝-𝑡𝑡 curve as a static design load.

2) Buildings neighboring a chemical plant

Chemical plants are, as in the case of fireworks factories, usually located far enough away from any
ordinary buildings. Thus, we can treat the explosion occurring at a chemical plant as a point source
explosion and evaluate the effect of the explosion to buildings based on the relationship between the
energy and distance.
The cause of most of explosion accidents at chemical plants is a fluid explosive event, in which the
fluid is leaked, vaporized, and mixed with air to form an explosive mixture. The detonation of gas in an
open space remains a rare event, and we can assume it is a deflagration event. A deflagration differs
from a detonation to the extent of the estimated magnitude of the load. If Eq. (11.3.3) is used for the
detonation, the estimated load becomes approximately 1,000 times larger than the actual load for the
deflagration. Thus, we apply other scaled distance equations proposed by Dobashi et al.30) that are
applicable to a deflagration event.
In this equation, the scaled distance of the peak pressure is calculated by
𝑅𝑅
𝑅𝑅r1 = (𝐸𝐸⁄𝑝𝑝 ) 1⁄3 (𝑆𝑆 ⁄𝑎𝑎 )8⁄3 , (11.3.4)
0 L 0

and the scaled distance for the impulse is

CHAPTER 11 IMPULSIVE LOADS - C11-31 -

𝑅𝑅
𝑅𝑅r2 = (𝐸𝐸⁄𝑝𝑝 ) 1⁄3 (𝑆𝑆 ⁄𝑎𝑎 )4⁄3 , (11.3.5)
0 L 0

where 𝐸𝐸 is the explosion energy, 𝑝𝑝0 is the atmospheric pressure, 𝑆𝑆L is the laminar combustion velocity,
and 𝑎𝑎0 is the velocity of sound. The values for 𝐸𝐸 and 𝑆𝑆L for typical explosive gases are shown in
Table 11.3.431). Because all parameters are known for some common explosive gases, these expressions
can give the scaled distance without experiments.

Table 11.3.4 Combustion energy and laminar combustion velocity for typical explosive gases
Material 𝐸𝐸 (kJ/mol) 𝑆𝑆L (m/s)
Methane 888 0.362
Propane 2,215 0.463
Hydrogen 284 2.20
𝑇𝑇0 = 298 𝐾𝐾 ，𝑝𝑝0 = 1 atm = 101.3 kPa

The expressions proposed by Dobashi et al. are applied to the existing results of the deflagration
experiments. Double-logarithmical plots are shown in Fig.11.3.11. Note that the scaled pressure is the
dimensionless value of the peak pressure divided by the atmospheric pressure. Both the peak pressure
and the impulse are plotted along nearly the same lines regardless of the type of gas. The expressions
can explain the experimental results in a unified manner. For a conservative design, it is recommended
to use slightly larger values than the actual peak pressure and impulse of all gases. For example, referring
to Fig. 11.3.11, the expression used to approximate the relationship between the maximum pressure and
scaled distance is Eq. (11.3.6), and the expression used to approximate the relationship between impulse
and scaled distance is Eq. (11.3.7).
𝑃𝑃S −0.9
= 12724 × 𝑅𝑅r1 (11.3.6)
𝑃𝑃0

−0.954
𝐼𝐼m = 30325 × 𝑅𝑅r2 (11.3.7)
- C11-32- Recommendations for Loads on Buildings

As an example, the scaled distance is calculated for the case in which the explosion source is a mixture
(125 m3) of 10% methane and air. The calculated (actual) conservative peak pressures are given in Table
11.3.5. To consider the explosion-resistant design for buildings neighboring a chemical plant, the
maximum value in the 𝑝𝑝-𝑡𝑡 curve is given as a static design load.

Table 11.3.5 Peak pressure for deflagration with 10% methane and air mixture
Distance (m) 136 68 34
𝑉𝑉 : Volume of explosive gas (m3) 125 125 125
Peak pressure (kPa) 0.09 0.18 0.33

3) Buildings neighboring a liquefied petroleum gas storage facility

Liquefied petroleum gas storage facilities located near buildings are not uncommon. When a liquefied
gas is leaked and vaporized, it is believed that an explosive mixture will be distributed on the surface of
the building. In such cases, the pressures caused from an explosion act directly on the exterior wall of
the building at the same time. The same 𝑝𝑝-𝑡𝑡 curve is applied directly on the wall surface because of the
absence of decay owing to the distance. Note that the pressure is not maintained but rapidly decays after
achieving the peak pressure because it is not confined.
Although it is not easy to obtain a p-t curve for a deflagration owing to the instability of the
phenomenon, Dobashi et al. proposed an estimated formula derived from the governing equation for
acoustics30).
21𝜌𝜌𝜌𝜌 𝜀𝜀−1 3 5⁄2
𝑃𝑃(𝑡𝑡) = 4𝑙𝑙 𝜀𝜀
𝐴𝐴 𝑡𝑡 �𝐴𝐴 = 𝑐𝑐g 𝜀𝜀 2 𝑆𝑆L 2⁄√𝛼𝛼 �, (11.3.8)

where 𝜌𝜌 is the density of the mixture, 𝜀𝜀 is the expansion ratio, 𝑙𝑙 is the distance from the ignition point,
𝑡𝑡 is the time, 𝑐𝑐g is the model coefficient, and 𝛼𝛼 is the thermal diffusion coefficient. Because all
parameters in the expression are known for common explosive gases, we can obtain the 𝑝𝑝-𝑡𝑡 curve
without conducting experiments. Figure 11.3.12 shows an example of a comparison between
experimental and calculated results based on Eq. (11.3.8). To consider an explosion-resistant design for
buildings neighboring a liquefied petroleum gas storage facility, the maximum value in the 𝑝𝑝-𝑡𝑡 curve
is given as a static design load.

Figure 11.3.12 Example of 𝑝𝑝-𝑡𝑡 curve for deflagration30)

CHAPTER 11 IMPULSIVE LOADS - C11-33 -

4) Buildings neighboring a hydrogen station

The number of hydrogen stations is expected to be increased along with the popularization of vehicles
powered by fuel cells. Studies on safety at hydrogen stations including the prevention of leaks, explosion
protection, and reduction measures have progressing recently. Herein, we consider the possible
explosion of a mixture of leaked hydrogen and air.
Because hydrogen rarely causes deflagration to detonation in an open space, we treat the explosion
as an external deflagration of the mixture. Such phenomenon is considered in the same way as buildings
neighboring a chemical plant. The explosive energy varies with the type of gas, even with the same
distance from the explosion and the same volume of explosive involved. As an example, consider an
explosion with 30% hydrogen, as shown in Table 11.3.6. The distance and volume are the same as the
case of 10% methane. To consider the explosion-resistant design for buildings neighboring a hydrogen
station, give the maximum value in the 𝑝𝑝-𝑡𝑡 curve as a static design load.

Table 11.3.6 Peak pressure of the deflagration with 30% hydrogen and air mixture
Distance (m) 136 68 34
𝑉𝑉: Volume of explosive gas(m3) 125 125 125
Peak pressure (kPa) 7.1 13 25

(3) Reflection pressure with incident angle

In the event of an explosion facing normal to a building, the peak pressure will be greater than the
incident pressure in the free field because of the effect of the reflected pressure. When the pressure wave
acts directly on the wall surface from the front, the incident angle 𝛼𝛼 is defined as 0°. T incident angle
𝛼𝛼 is then 90° when the propagating direction of the pressure wave is parallel to the wall surface and the
peak pressure is equal to that in the free field.
The calculated peak pressure (incident pressure) in the free field, 𝑝𝑝s , combined with the Rankine-
Hugoniot expression gives the peak reflected pressure, 𝑝𝑝r , for an incident angle of 0°, which is given as
7𝑝𝑝0 +4𝑝𝑝s
𝑝𝑝r = 2𝑝𝑝s 7𝑝𝑝0 +𝑝𝑝s
, (11.3.9)

where 𝑝𝑝0 represents the atmospheric pressure. From Eq. (11.3.9), if the incident pressure is sufficiently
smaller than the atmospheric pressure, then the reflected pressure is approximately twice the incident
pressure.
The peak reflected pressure and the reflected impulse are then calculated. For the incident angle 𝜃𝜃,
the peak pressure 𝑝𝑝load is calculated with 𝑝𝑝s given by the scaled distance and 𝑝𝑝r given using Eq.
(11.3.9). For the impulse 𝐼𝐼load , 𝐼𝐼s is defined corresponding to Fig. 11.3.9, and 𝐼𝐼r is assumed to be 𝐼𝐼r
= 𝐼𝐼s because the available data are insufficient. The ConWep formula is then applied:
𝑝𝑝load = 𝑝𝑝r × cos2 𝜃𝜃 + 𝑝𝑝s × (1 + cos 2 𝜃𝜃 − 2 cos 𝜃𝜃), (11.3.10)
𝐼𝐼load = 𝐼𝐼r × cos2 𝜃𝜃 + 𝐼𝐼s × (1 + cos 𝜃𝜃 − 2 cos2 𝜃𝜃). (11.3.11)
The ConWep formula is a numerical package used to obtain the 𝑝𝑝-𝑡𝑡 curve based on the Technical
- C11-34- Recommendations for Loads on Buildings

Manual TM5-1300 of the US Department of Army, Navy, and Air Force 32) and the relationship between
the scaled distance and incident angle at the wall surface. This formula is widely used and it has been
embedded in many dynamic structural simulation codes.
The peak pressure applied to the rear wall of a building is mitigated by diffraction. Unlike a case in
which the front wall faces the source of an explosion, the size effect of the structural members makes it
difficult to uniformly define and accurately obtain the relation with the incident angle. Although it is
necessary to use computational fluid dynamics (CFD), which can consider the diffraction, it is not
realistic to use CFD because of the increasing cost from a design perspective. For simplicity, TM5-1300
extends the sidewall of the building to treat the rear wall and uses the same value as the one for the side
wall to evaluate the pressure on the rear wall conservatively.

(4) Design load for external explosion

As indicated above, we can obtain the 𝑝𝑝-𝑡𝑡 curves for each member of a building. The approach used
to obtain the design load distributed over the entire building, as shown in Fig. 11.3.13, is described here.
First, the distance and incident angle to the explosion source from the center of each wall surface of the
building, and the incident angle based on the relative location of the building and explosion source, are
calculated. The rear wall is considered a horizontal extension of the roof against the explosion source.
If the wall surface is large, the wall surface is divided into several segments. Then, the peak pressure
and impulse for each segment corresponding to the explosion source, distance, and incident angle are
calculated.
Note that the arrival time is different because the propagation speed of the pressure wave generated
by an explosion is the speed of sound. To treat it as a load distributed throughout the entire building, the
arrival time is determined by dividing the distance from the explosion source by the speed of sound in
the air. The time history of the distributed loads is shown in Fig. 11.3.13. The 𝑝𝑝-𝑡𝑡 curve for each wall
and roof is illustrated with its peak pressure and load duration.
As shown in the figure, the explosion source is 30% H2-Air 27 m3 and is located 20 m away from the
horizontal center of the building. Wall A is 20 m away from the explosion source. The peak pressure
when no building is present is 3.9 kPa, and the impulse is 25.8 Pa∙s, as determined through Eqs. (11.3.4)
and (11.3.5), and referring to Fig. 11.3.11. The incident angle is approximately 0°, and Eq. (11.3.9)
provides a peak reflected pressure of 7.8 kPa. The load duration of the triangular wave is then determined
to be 6.6 ms. The arrival time is 59 ms, as determined by dividing the distance by the speed of sound in
the air, i.e., 340 m/s.
Wall B is 21.2 m away from the explosion source, and the incident angle is 22°. The peak reflected
pressure and reflected impulse for an incident angle of 0° are 7.3 kPa and 24.1 Pa∙s, respectively. The
peak reflected pressure for an incident angle of 22° is 6.3 kPa, determined through Eqs. (11.3.10) and
CHAPTER 11 IMPULSIVE LOADS - C11-35 -

Figure 11.3.13 Example of the pressure distributed over the whole building against external
explosion

(11.3.11). The load duration is 7.6 ms and the arrival time is 63 ms, which are determined similarly as
wall A.
Wall C is on the roof and is 23.5 m away from the explosion source. The cutoff of the incident angle
is 90°. The peak reflected pressure and the reflected impulse are 3.4 kPa and 22.2 Pa∙s, respectively. The
load duration is 13.2 ms, and the arrival time is 69 ms.
Wall D is on the back side of the building and is treated as the horizontal extension of the roof. The
peak reflected pressure and the reflected impulse are 2.9 kPa and 18.8 Pa∙s, respectively. The load
duration is 13.1 ms, and the arrival time is 82 ms.
The calculations described above are repeated for all walls to obtain the 𝑝𝑝 -𝑡𝑡 curve of the blast
pressure acting on the entire building. To consider an external blast force, the maximum value along the
𝑝𝑝-𝑡𝑡 curve is given as a static design load.
- C11-36- Recommendations for Loads on Buildings

References

1) The Nikkei Architecture: Feature pages, The third force, “Impact Force,” the Nikkei Architecture,
No. 957, pp. 8-10. 2011 (in Jpanese)
2) The Architectural Institute of Japan, the special investigating committee for the collapse of the
WTC: The special investigating committee’s report, the collapse of the World Trade Center and the
damage of the Pentagon: Materials for the special investigating committee for the collapse of the
WTC panel discussion, 2003 (in Japanese)
3) Hamamoto, T.: Toward design guideline of buildings against impact and explosion loads, Academic
journal of Japan Society of the Defense Facility Engineers: JSDFE, Vol. 10, pp. 28-37, 2010 (in
Japanese)
4) Architectural Institute of Japan: Evaluation of the collision and blast resistance and the preventive
measures for important structures, Panel discussion of annual AIJ convention, 2006 (in Japanese)
5) Architectural Institute of Japan: The idea for damage mitigation on buildings against impact and
explosion actions, Panel discussion of annual AIJ convention, 2009 (in Japanese)
6) Architectural Institute of Japan, Tokyo Institute of Technology: Workshop for shock resistance
evaluation of buildings, Kanagawa, 2010 (in Japanese)
7) Architectural Institute of Japan and Japan Society of Civil Engineers: Symposium for shock
resistance design for structures, Tokyo, 2010 (in Japanese)
8) Architectural Institute of Japan: The idea for the shock resistance design for buildings, Panel
discussion of annual AIJ convention, 2012 (in Japanese)
9) Hamamoto, T.: Design Consideration in Tentative AIJ Guideline; Ohta, T. et al.: Design Loads in
Tentative AIJ Guideline; Nakamura, N. et al.: Dynamic Response in Tentative AIJ Guideline;
Mukai, Y. et al.: Member Design in Tentative AIJ Guideline; Nishida, A. et al.: Design Examples
in Tentative AIJ Guideline, Proceedings of the 9th International Conference on Shock & Impact
Loads on Structures, pp. 115-162, 2011
10) Eurocode 1- Actions on Structures – Part 1-7: General actions – Accidental actions, BS EN 1991-
1-7, British Standards, 2006
11) The guideline for preventive measure and equipment for road vehicle falling incident at a car park,
Guide to the standard building structural technologies, pp. 276-277, 2007 (in Japanese)
12) FHWA/NHTSA National Crash Analysis Center: Finite element model of Ford Taurus report, 2008
13) National Crash Analysis Center: 2010 Toyota Yaris FE model report, 2010
14) FHWA/NHTSA National Crash Analysis Center: Finite element model of Toyota Rav 4 report,
2008
15) FHWA/NHTSA National Crash Analysis Center: Finite element model of Dodge Neon report, 2006
16) National Crash Analysis Center: [Link]
17) JSOL: LS-DYNA ver. 970 User’s Manual, 2003
18) Hisamori, T and Tachibana, E.: “Performance evaluation of RC protective barrier under impact
CHAPTER 11 IMPULSIVE LOADS - C11-37 -

load caused by train collision,” Proceedings of International Symposium on Disaster Simulation &
Structural Safety in the Next Generation, pp.115-120, 2011
19) Civil Aeronautics Act, Act No. 231 of July 15, 1952. Amendment Act No. 54 of May 25, 2011.
Ministry of Land, Infrastructure, Transport and Tourism
20) Barr, P.: “Guideline for the design and assessment of concrete structure subjected to impact,” SRD
R439 Issue 3, AEA Technology, 1990
21) Kasai, Y.: “Impact load subjected to aircraft crash onto buildings,” Proceedings of 6th International
Symposium on Structures under Impulsive Loading, Japan Society of Civil Engineers, 2002
22) Dahoe A. E., de Goey, L. P. H.: On the determination of the laminar burning velocity from closed
vessel gas explosion, Journal of Loss Prevention in the Process Industries, Vol. 16, pp. 457-478,
2003.
23) Ohta, T. et al.: “Design Loads in Tentative AIJ Guideline,” Proceedings of 9th International
Conference on Shock and Impact Loads on Structures, 2011
24) Chao, J., Bauwens, C.R., Dorofeev, S.B.: “An analysis of peak over pressures in vented gaseous
explosions,” Proceedings of the Combustion Institute, Vol. 33, No. 2, pp. 2367-2374, 2011
25) Architectural Institute of Japan: Introduction to Shock-Resistant Design of Buildings. Architectural
Institute of Japan, 2014 (in Japanese)
26) Harris, R. J.: The Investigation and Control of Gas Explosions in Buildings and Heating Plant,
E&FN Spon Ltd, 1983
27) Hirano. T.: Preventive technologies for gas explosions, Kaibundo, 1983 (in Japanese)
28) Yoshida, M. et al.: “Safety of explosives (11) propagation of the blast wave and its hazardous
effects,” Explosion, Vol. 17 No. 1, pp. 2-5, 2007 (in Japanese)
29) Department of Defense, US: DoD Ammunition and Explosives Safety Standards, DoD6055.9-STD,
2004
30) Dobashi, R. et al.: “Study on the effect of blast pressure caused by gas explosions,” Proceedings
for Japan Society for Safety Engineering Symposium, pp. 194-197, 2009 (in Japanese)
31) Glassman, I.: Combustion, 3rd edition, Academic press, 1996
32) US Department of Defense: US Department of the Army, Navy and Air Force Technical Manual
(TM5-1300). Structures to resist the effects of accidental explosions [M]., Washington, DC, 1990.

Common questions