© All Rights Reserved

0 views

Wall Buckling [PARRA, MOEHLE].pdf

© All Rights Reserved

- Effect of Ani Sot Ropy on the Deep Drawing of Mild Steel
- Byfield, Mike - Structural design from first principles (2018, CRC Press).pdf
- Elastic Buckling of Bionic Cylindrical Shells Based on Bamboo
- midterm
- Homework 3
- Beam Columns II
- 04.STAHL-5_EN
- CHP4
- Ultimate Limit State Design of Ship Structures(1)
- 2 Marks
- N. Ohno, D. Okumura and A. Okada- Buckling Behavior of Elastic Kelvin Open-Cell Foams Subjected to Uniaxial Compression
- ss
- 125760133 Buckling of Thin Metal Shells 287
- Failure Mode Maps in the Thin Film Scratch Adhesion
- FRAME IDEALISATION & ANALYSIS
- CE579 Half Course Summary
- Why is the Yield Strength Calculated on the Basis of 0.2% or 0
- design 3
- KKIHJ.pdf
- Warning Errors

You are on page 1of 11

Frontiers of Earthquake Engineering

July 21-25, 2014

10NCEE Anchorage, Alaska

CONCRETE WALLS

P. F. Parra1 and J. P. Moehle2

ABSTRACT

Structural walls are used worldwide to resist gravity and earthquake loads. In many countries,

structural walls commonly are constructed with a rectangular cross section, or a cross section

made up of interconnected rectangles, without an enlarged boundary element. In some countries,

design practice has resulted in walls that are more slender than those used in the past. For

example, in Chile and elsewhere it is not unusual to find rectangular wall edges having thickness

of 6 to 8 in. (150 to 200 mm), resulting in floor-to-floor slenderness ratios reaching hu/b = 16 or

greater. Such walls can be susceptible to overall wall buckling in which a portion of the walls

buckles out of the plane of the wall. Examples of this behavior were observed following the 2010

Chile earthquake.

Tendency to buckle is believed to depend primarily on the wall clear height to thickness

ratio hu/b and loading history. Two failure modes are hypothesized. One hypothesis is that tensile

yielding for loading in one direction softens the boundary for subsequent loading in the opposite

direction, leading to lateral instability of an otherwise intact wall. A second hypothesis is that the

wall crushes first, leaving an even smaller and irregular cross section. This crushed section may

become immediately unstable or, alternatively, subsequent tension and compression cycles may

lead to instability of the reduced cross section according to the first hypothesis, leading to a

secondary buckling failure.

A theory is presented for buckling of sections subjected to inelastic tension and

compression strain cycles. The theory is applied to tests of reinforced concrete prisms and one

Chilean building (Alto Huerto), which had a buckled wall following the 2010 Chile earthquake.

Based on this study, it is concluded that buckling most likely was a secondary failure that

occurred after initial crushing of the wall boundaries.

1

Graduate Student Researcher, Dept. of Civil and Environmental Engineering, University of California, Berkeley,

CA 94720

2

Professor, Dept. of Civil and Environmental Engineering, University of California, Berkeley, CA 94720

Parra PF, Moehle JP. Lateral buckling in reinforced concrete walls. Proceedings of the 10th National Conference in

Earthquake Engineering, Earthquake Engineering Research Institute, Anchorage, AK, 2014.

Lateral Buckling in Reinforced Concrete Walls

ABSTRACT

Structural walls are used worldwide to resist gravity and earthquake loads. In many countries,

structural walls commonly are constructed with a rectangular cross section, or a cross section

made up of interconnected rectangles, without an enlarged boundary element. In some countries,

design practice has resulted in walls that are more slender than those used in the past. For

example, in Chile and elsewhere it is not unusual to find rectangular wall edges having thickness

of 6 to 8 in. (150 to 200 mm), resulting in floor-to-floor slenderness ratios reaching hu/b = 16 or

greater. Such walls can be susceptible to overall wall buckling in which a portion of the walls

buckles out of the plane of the wall. Examples of this behavior were observed following the 2010

Chile earthquake. Tendency to buckle is believed to depend primarily on the wall clear height to

thickness ratio hu/b and loading history. Two failure modes are hypothesized. One hypothesis is

that tensile yielding for loading in one direction softens the boundary for subsequent loading in the

opposite direction, leading to lateral instability of an otherwise intact wall. A second hypothesis is

that the wall crushes first, leaving an even smaller and irregular cross section. This crushed section

may become immediately unstable or, alternatively, subsequent tension and compression cycles

may lead to instability of the reduced cross section according to the first hypothesis, leading to a

secondary buckling failure. A theory is presented for buckling of sections subjected to inelastic

tension and compression strain cycles. The theory is applied to tests of reinforced concrete prisms

and one Chilean building (Alto Huerto), which had a buckled wall following the 2010 Chile

earthquake. Based on this study, it is concluded that buckling most likely was a secondary failure

that occurred after initial crushing of the wall boundaries.

Introduction

Design practices prior to the 1990s favored rectangular walls with enlarged boundary elements,

contributing to stability of the flexural compression zone. More recently, prevailing practices in

many countries favor rectangular sections without enlarged boundaries. The more slender

flexural compression zones can be susceptible to inelastic lateral buckling.

When parts of a wall section are subjected to compressive strains, the possibility of

lateral instability arises. Although global wall buckling occurs when the wall boundary is in

compression, buckling can be strongly influenced by the magnitude of the tensile strain

experienced by the wall for prior loading in the opposite direction (Paulay and Priestley [1]; Chai

and Elayer [2]). This is because residual tensile strains in the previously yielded longitudinal

reinforcement leave the wall boundary with open cracks, resulting in reduced lateral stiffness.

Consider a multi-story wall as shown in Fig. 1. The foundation, floor diaphragms, and

roof diaphragm provide lateral support at every story level. Thus, the unsupported height of the

1

Graduate Student Researcher, Dept. of Civil and Environmental Engineering, University of California, Berkeley,

CA 94720

2

Professor, Dept. of Civil and Environmental Engineering, University of California, Berkeley, CA 94720

Parra PF, Moehle JP. Lateral buckling in reinforced concrete walls. Proceedings of the 10th National Conference in

Earthquake Engineering, Earthquake Engineering Research Institute, Anchorage, AK, 2014.

wall boundary can be taken equal to the story clear height, hu.

An effective length khu can be defined based on the rotational restraints at the different

floor levels. In the present analysis, which is concerned with very slender walls, it may be

reasonable to consider the wall to be fixed at top and bottom of the clear height. Accordingly, k

is taken equal to 0.5.

A typical wall boundary will be subjected to alternating tension and compression as a

building responds to an earthquake (Fig. 1). If the boundary yields in tension, a cracked section

is produced, with crack width dependent on the amplitude of the reinforcement tensile strain sm

during the tension excursion. In a previously yielded wall, crack closure under deformation

reversal may require yielding of the longitudinal reinforcement in compression. In a wall with

two curtains of reinforcement, any slight asymmetry in the reinforcement is likely to result in one

curtain yielding before the other, leading to out-of-plane curvature and a tendency to buckle out

of plane. In a wall with one curtain of reinforcement, out-of-plane curvature occurs even more

readily. Whether the wall remains stable depends on the amplitude of the prior tensile strain sm

and the slenderness ratio hu/b of the wall. As a design approximation, the critical slenderness

ratio can be related to the maximum prior tensile strain sm, as will be shown later.

Laboratory tests results for buckled columns [2] are used to evaluate a proposed buckling

theory. For walls in Alto Huerto building (Chile), different procedures are used to estimate the

strain demands sm. These values are compared with the limit given by the buckling theory. This

enables an assessment of whether the sections are likely to have buckled prior to concrete

crushing.

In the following theoretical development, some of the modeling concepts of Paulay and Priestley

[1] are followed. Consider the wall shown in Fig. 2. Wall lateral buckling is constrained by the

story clear height (Fig. 2a). We assume the wall has been flexed previously such that the

boundary yields in tension (Fig. 2b), with a unit length (measured in the horizontal direction) of

the boundary element developing tension force T, maximum tensile stress , and tensile strain

. Upon deformation reversal, just before the boundary element yields in compression, the

longitudinal reinforcement will have unloaded by strain = ⁄ and reloaded in

compression to − (ignoring the Bauschinger effect), such that the residual tensile strain is

approximately = ⁄ − . To simplify the model, the residual tensile strain is

approximated as ≈ − 0.005. Invariably, one curtain of reinforcement will yield before

the other, producing curvature φ as shown in Fig. 2d, and out-of-plane displacement as illustrated

in Fig. 2a and Fig. 2c. Whether the boundary remains stable depends on magnitude of the lateral

displacement relative to the wall thickness b, which relates to the maximum previous tensile

strain and the resulting curvature as illustrated in Fig. 2c.

To estimate conditions for stability, we first approximate the effective length (height).

For a multi-story wall with length not less than the first-story clear height ℎ , it is reasonable

to assume that the flexural plastic hinge extends over the height of the first story. Assuming

fixity at top and bottom, with a simple harmonic buckled shape, the effective length in Fig. 2a is

ℎ = 0.5ℎ . Examining the effective length more closely (Fig. 2c), we can express the

maximum deflection as a fraction of wall thickness (that is, = ). The relation between

maximum deflection and the maximum curvature is:

ℎ (1)

= =

As a first approximation, the maximum curvature from Fig. 2d can be written as:

− 0.005 (2)

=

Equilibrium of forces and moments in the free-body diagram of Fig. 2d results in the

following two expressions:

=0→ = + (3)

=0→ = (4)

In Eq. 4, moments are taken about the centerline, such that moments of longitudinal

reinforcement compressive force resultants (assumed equal) cancel. Assuming longitudinal

reinforcement is stressed to and assuming the concrete compressive force is represented by

the usual rectangular stress block with depth and average stress 0.85 ′ , we can write

= (5)

= 0.85 ′ 1 − 2 (6)

(7)

1−2 −1 = =

0.85 ′ 0.85

in which = ⁄ ′ is the mechanical reinforcement ratio. This expression has real roots only

if the following is satisfied:

(8)

2 2 4

≤ 0.5 1 + − +

0.85 0.85 0.85

Substituting from Eq. 8 into Eq. 1, solving for ⁄ℎ , and defining width b as the

critical width results in:

1 − 0.005 (9)

=

ℎ

The main variables appearing in Eq. 9 are slenderness ratio ℎ ⁄ , maximum tensile

strain in longitudinal reinforcement, effective depth parameter for longitudinal

reinforcement, and . Parameter can be found from = , where it is noted that ≈ 0.8 for

thin walls with two curtains of reinforcement and 0.5 for walls with single layer of

reinforcement. From this, it is clear that walls with two curtains of longitudinal reinforcement are

inherently more stable than walls with a single curtain. Parameter relates to the mechanical

reinforcement ratio Eq. 8, which is an inconvenient parameter for preliminary design. For

practical construction, 0.4 ≤ ≤ 0.6.

Adopting values = 0.8 and = 0.5, Eq. 9 for walls with two curtains of

reinforcement becomes:

(10)

= 0.7 ∈ − 0.005

ℎ

If the cover is lost before buckling the wall (this occurs usually within a compressive

strain range of 0.003-0.005), it is more reasonable to use = 1 and equal to the confined

core width (out-to-out distance between vertical bars).

For typical slender wall geometries, the boundary can be approximated as fixed-fixed, in

which case k = 0.5. Thus, Eq. 10 becomes:

(11)

= 0.35 ∈ − 0.005

ℎ

Considering low-cycle fatigue, the maximum tensile strain normally accepted for

longitudinal reinforcement is approximately = 0.05. For = 0.05, Eq. 11 results in

hu/bcr = 13.

The preceding derivation is based on an idealized wall boundary subjected to uniform

compressive strain. Actual wall boundaries have strain gradient along the wall length, which

would tend to brace the edge of the wall. This suggests that the preceding results should be

conservative for actual wall boundaries.

Several tests have been done (Chai and Elayer [2]; Creagh et al. [3]; Acevedo et al. [4]) with

prismatic sections loaded under tension/compression cycles. The data from Chai and Elayer [2]

are especially relevant, as those tests gradually increased tensile and compressive strains until

overall prism buckling occurred. These data were based on tests of axial columns under large

strain amplitudes expected in the plastic hinge region of a ductile reinforced concrete wall. Fig. 3

compares results of Eq. 10 with test data from prismatic sections that buckled following tensile

strain excursions to . All test specimens had pin-ended boundary conditions (k = 1). The

results suggest that Eq. 10 is a reasonable approximation to describe behavior of uniformly

loaded prisms.

Comparison of Results from Buckling Model and Observed Damage in Chilean Building

Detailed studies have been carried out on two buildings in which out-of-plane buckling was

observed following the 2010 Chile earthquake. One of them is a 15-story building named Alto

Huerto, located in San Pedro de la Paz (near Concepción). This section includes only highlights

from analyses carried out on this building.

Alto Huerto was severely damaged after the Chile earthquake on February 27th, 2010. The

building was designed during 2007-2008 and constructed in 2009. It has fifteen stories and two

subterranean levels. The seismic force-resisting system is composed of reinforced concrete walls

of 7.87 in. (200 mm) typical thickness. The gravity force-resisting system comprises the

reinforced concrete walls plus some interior reinforced concrete columns. The typical story

height is 8.37 ft (2.55 m). There are some discontinuities in the vertical members in the first story

with respect to the upper stories. For example, walls K and Ñ (and other walls) step back from

the building perimeter by approximately 6.5 ft (2 m), resulting in reduced wall length in the first

story compared with typical stories above. The building sustained a variety of apparent damage

during the 2010 earthquake, with main damage characterized by wall crushing in the first story

or in subterranean levels. Some walls, and in particular the first story wall along line Ñ, showed

apparent out-of-plane buckling (Fig. 4). Fig. 5 shows the typical floor plan.

Figure 4. Alto Huerto overall view and characteristic wall damage, after DICTUC

[5].

Wall K

Wall Ñ

Instruments recorded the ground acceleration during the February 27th 2010 Chile

earthquake in a location close to Alto Huerto (Colegio San Pedro, Concepción). Fig. 6 presents

the linear displacement response spectra (2% damped) for the nearby site.

A linear, fixed-based model was developed using the software ETABS. Effective stiffness

is according to ASCE 41-06, Section 6.3.1.2, as summarized below:

• Walls 0.5EcIg (flexural); 0.4EcAcv (shear)

• Columns 0.3 EcIg (flexural); 0.4EcAcv (shear)

• Slabs 0.33 EcIg (flexural)

Using this model, the fundamental translational periods are calculated to be 0.57s in the

transverse (east-west) direction and 0.47s in the longitudinal (north-south) direction. According

to the linear response spectra in Fig. 6, in the transverse direction the spectral displacement is

approximately 5 in. (130 mm), which corresponds to roof displacement of approximately 6 in.

(150 mm), or roof drift ratio of approximately 6/1589 = 0.004. However, peak displacement

occurs for moderately longer period, resulting in spectral displacement of 16 in. (410 mm). The

corresponding roof drift ratio for this greater drift is 0.013. These latter values establish an

effective “upper bound” to the displacements based on the assumption that the recorded ground

motion represents the motion at the building site and the vibration period is moderately longer

than the calculated value.

Wall Ñ is the wall that showed the most obvious lateral buckling (Fig. 4b). Fig. 7a and 8b

present the wall elevation and the cross section in the first story. As shown, the wall has a 2-m

setback at the first level above grade, with a T-shaped cross section. A nonlinear, fixed-base

model for wall Ñ was implemented in the software PERFORM 3D (Fig. 7c). The model was

studied using input ground motion and nonlinear static analysis. Simplified models integrating

curvature over height, including a plastic-hinge model, were also investigated (Fig. 7d). All the

models produce the same conclusion. Specifically, for lateral loading that puts the flange in

tension, the stem will crush for roof drift ratio of approximately 0.005. According to the buckling

model presented earlier in this chapter, a prior peak tensile strain of ~ 0.03 is required to

soften the wall sufficiently to cause out-of-plane buckling. To reach a tensile strain of

~ 0.03, however, requires roof drift ratio around 0.014. Although this drift ratio is plausible

given the response spectrum for the site, it is approximately three times the drift ratio required to

cause wall crushing. Therefore, it seems much more likely that the wall crushed first for loading

that put the stem in compression, and the damaged section then buckled out of plane. If it is

assumed that spalling of cover concrete leaves an intact core, the buckling model indicates that

the reduced section would be prone to out-of-plane buckling at the roof drift ratio of 0.005.

Wall K is located immediately adjacent to wall Ñ (Fig. 5a). Therefore, it is reasonable to

conclude that the walls were subjected to nearly identical displacement histories. This wall

experienced minor failure in the boundary apparently due to compression. Fig. 8a and 9b present

the wall elevation and the cross section in the first story. As with wall Ñ, the wall has a setback

at the first level above grade. Unlike wall Ñ, wall K has a rectangular cross section. Simplified

models integrating curvature over height, including a plastic-hinge model, were used to study

likely strain demands in the first story, using the same approach as was used for wall Ñ.

According to these models, crushing of the wall boundary is expected for roof drift ratio of

approximately 0.006 (slightly larger than the value required for wall Ñ). Maximum tensile strain

is approximately 0.008 at this drift ratio. According to the buckling model presented earlier, a

prior peak tensile strain of ~ 0.03 is required to soften the wall sufficiently to cause out-of-

plane buckling after spalling. These combined results indicate that crushing of the wall boundary

would be expected to precede wall buckling. As noted, the wall sustained minor concrete

crushing, with no evidence of out-of-plane buckling.

The damage state for wall K (minor failure in the boundary apparently due to compression)

suggests that the wall did not undergo drifts significantly exceeding the drift at onset of crushing

(roof drift ratio ~ 0.006). Thus, it seems even less likely that wall Ñ could have reached lateral

drifts required for buckling to control the behavior (roof drift ratio ~ 0.014).

Conclusions

Wall boundaries can sustain overall buckling when subjected to earthquake loading. Tendency to

buckle depends primarily on the aspect ratio hu/b. Based on this study, it seems more likely that

buckling is a secondary failure that occurs after the wall crushes, leaving an even smaller and

laterally unstable cross section defined by the concrete core. Building codes should have a

slenderness ratio limit for the intended hinge zone of special structural walls. The UBC 97 [6]

limit of hu/b ≤ 16 is recommended for walls that maintain their concrete cover. The same limit

could be applied to walls for which cover concrete has spalled, substituting width bc of the

confined core for width b. The aforementioned limit applies to walls with two curtains of

distributed reinforcement. Walls with one curtain of distributed reinforcement are even more

vulnerable to lateral buckling, and should not be used within the intended plastic hinge zone of

special structural walls having aspect ratio hw/lw exceeding 2.

Acknowledgments

The research presented in this paper has been supported by the National Institute of Standards

and Technology through the ATC 94 project. Their support is gratefully acknowledged.

Opinions, findings, conclusions and recommendations in this paper are those of the writers and

do not necessarily represent those of the sponsor.

References

1. Paulay T, and Priestley MJN. Stability of Ductile Structural Walls. ACI Structural Journal 1993; 90 (4): 385-

392.

2. Chai YH, and Elayer DT. Lateral Stability of Reinforced Concrete Columns under Axial Reversed Cyclic

Tension and Compression. ACI Structural Journal 1999; 96 (5): 780-789.

3. Creagh A, Acevedo C, Moehle JP, Hassan W, Tanyeri AC. Seismic Performance of Concrete Special Boundary

Element, Network for Earthquake Engineering Simulation 2010, 18 pp.

4. Acevedo CE, Creagh A, Moehle JP, Hassan W, Tanyeri AC. Seismic Vulnerability of Non-special Boundary

Element of Shear Wall under Axial Force Reversals, Network for Earthquake Engineering Simulation 2010, 16

pp.

5. DICTUC. Inspección visual y levantamiento de daño edificio Alto Huerto Concepción, Report # 906575/10-

056-EE-01-R0, Santiago (Chile), 2010.

6. UBC. Uniform Building Code. International Conference of Building Official, Whittier, California, 1997.

- Effect of Ani Sot Ropy on the Deep Drawing of Mild SteelUploaded byaraveeti
- Byfield, Mike - Structural design from first principles (2018, CRC Press).pdfUploaded byCharles Ajayi
- Elastic Buckling of Bionic Cylindrical Shells Based on BambooUploaded bymessman
- midtermUploaded byJessie Widder
- Homework 3Uploaded byKartika Purwitasari
- Beam Columns IIUploaded byMihajloDjurdjevic
- 04.STAHL-5_ENUploaded byEmax01
- CHP4Uploaded byNajmi Ishak
- Ultimate Limit State Design of Ship Structures(1)Uploaded byProrok Nemanja
- 2 MarksUploaded bychandru1chandru
- N. Ohno, D. Okumura and A. Okada- Buckling Behavior of Elastic Kelvin Open-Cell Foams Subjected to Uniaxial CompressionUploaded byFlaoeram
- ssUploaded byThulasi Raman Kowsigan
- 125760133 Buckling of Thin Metal Shells 287Uploaded bypawkom
- Failure Mode Maps in the Thin Film Scratch AdhesionUploaded byÇiğdem Albayrak
- FRAME IDEALISATION & ANALYSISUploaded byAnonymous iS33V5
- CE579 Half Course SummaryUploaded bySwarna Rk
- Why is the Yield Strength Calculated on the Basis of 0.2% or 0Uploaded bycivilsalman
- design 3Uploaded byBoopathi Yoganathan
- KKIHJ.pdfUploaded byVisu Viswanath
- Warning ErrorsUploaded byRakesh KR
- 109CE0454.pdfUploaded bysom_bs79
- Buckling of Thin PlatesUploaded byOSCARDELTA
- analisis of suspension arm.pdfUploaded bySantiago Urgiles
- Buckling Analysis of Composite PanelsUploaded byRAHMON RAUF OYEDOKUN
- Numerical Analysis of Rail InteractionUploaded bypandeydb
- AISC 360Uploaded byShervin2010
- conic p ext.pdfUploaded byMaria Tns
- Vigas celularesUploaded byjose ronaldo
- Seismic Design and Behavior of Concentrically BracUploaded byarjun
- Group pileUploaded byChien Chieu

- SAS InstituteUploaded byapi-3716851
- lsmw idocUploaded bykocherlakotapavan
- Intro data miningUploaded byapi-19730613
- Annual Report 09 FerozeUploaded byNabeel Ahmed Khan
- Anemia in HIV Handout 97-2003Uploaded byK. Vijay Ananth
- Battle of Trench and Fate of Banu QuraydhaUploaded byVero
- Getting Started with Python Data Analysis - Sample ChapterUploaded byPackt Publishing
- Winter 2013 Issue: Zeta DataUploaded byAleta Sullivan
- sboa017Uploaded byapi-3725139
- BFS DFS ReferenceUploaded byThomas Ramos
- Grandparents TriviaUploaded byLea Colbongan Papay
- Cult of Freyr and FreyjaUploaded byDaniel Cody
- 5e Crafting Potions and PoisonsUploaded byH Vascovich
- COLLEGE PHYSICS I-jc problems.docxUploaded byHelen Grace Llemos Cabalag
- 4125Uploaded byMalcolm Christopher
- Introduction to Socio-Anthropology _rlUploaded byThea Mallari
- Heart FailureUploaded bycoollist33
- Communities_Rural and UrbanUploaded bydevendrachoudhary_upsc
- Ecological Services and Ecological Release ...Uploaded byThe Wecskaop Project
- Links MuitosUploaded byJosé Carlos Alves
- Feyerabend the Role of Reason in ScienceUploaded byrenato
- Statistical Methods MCQ'SUploaded byGuruKPO
- ier-march-2010Uploaded byNikhil Choudhary
- Future Tense - Future Perfect TenseUploaded byDonat Nabahungu
- PAYING FOR INFORMATION A Report of the Ethics Advisory Committee of The Canadian Association of JournalistsUploaded byjsource2007
- 5_DOWNLOAD Lactobacillus Acidophilus Rosell 52Uploaded byeverstrong
- sao maths plan std 2 and 3Uploaded byapi-327938582
- Page 10Uploaded byLettrariH
- Timesheet Multiple Employee Weekly TimeIn TimeOut TemplateUploaded bychanduravilla
- Drawing GameUploaded byMayur Malpani

## Much more than documents.

Discover everything Scribd has to offer, including books and audiobooks from major publishers.

Cancel anytime.