See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/269198353

Timber Tower Research: Concrete Jointed Timber Frame

Conference Paper · April 2014

DOI: 10.1061/9780784413357.113

CITATIONS READS
8 7,618

4 authors, including:

William F Baker Joshua Schultz

Skidmore Owings & Merrill Gonzaga University
90 PUBLICATIONS   1,205 CITATIONS    33 PUBLICATIONS   107 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Glass/Facade Design View project

Vibration/Dynamics of Systems View project

All content following this page was uploaded by Joshua Schultz on 22 August 2017.

The user has requested enhancement of the downloaded file.

 Timber Tower Research: Concrete Jointed Timber Frame

W.F. Baker1, D.R. Horos2, B.M. Johnson3, J.A. Schultz4

1
Partner, Skidmore, Owings & Merrill LLP, 224 S. Michigan Avenue, Chicago,
Illinois 60604, PH (312) 554-9090, email: william.baker@som.com
2
Director, Skidmore, Owings & Merrill LLP, 224 S. Michigan Avenue, Chicago,
Illinois 60604, PH (312) 554-9090, email: david.horos@som.com
3
Associate, Skidmore, Owings & Merrill LLP, 224 S. Michigan Avenue, Chicago,
Illinois 60604, PH (312) 554-9090, email: benton.johnson@som.com
4
Skidmore, Owings & Merrill LLP, 224 S. Michigan Avenue, Chicago, Illinois
60604, PH (312) 554-9090, email: joshua.schultz@som.com

ABSTRACT

Tall buildings pose a challenge to the sustainability movement because they

offer both positive and negative environmental impacts. Positive impacts include
reducing urban sprawl, promoting alternative transportation, and allowing efficient
energy use on a district scale. However, these benefits come with the cost of greater
carbon emissions associated with both material production and building construction
compared to a low-rise building. The goal of this research was to develop a structural
system for tall buildings using mass timber as the main structural material that
reduces the carbon dioxide emissions associated with the structure.

The structural system research was applied to a prototypical building based on

an existing concrete benchmark for comparison. The selected concrete benchmark
building is the Dewitt-Chestnut Apartments; a 120m tall, 42-story building in
Chicago designed by Skidmore Owings and Merrill and built in 1966. This building
was chosen as the benchmark because the geometry is a rectangular extrusion, the
lease depths are consistent with contemporary residential buildings, and the concrete
structural system is efficient in material usage providing a lower bound for
comparison with the prototypical building.

This paper discusses key design issues associated with tall mass-timber
buildings along with potential solutions. Specific challenges include low structural
weight and associated net uplift due to lateral loads, long term differential shortening,
floor vibrations, and fire performance. It is believed that the system proposed in the
research and discussed in the paper could mitigate many of these design issues. The
proposed system, the “Concrete Jointed Timber Frame”, relies primarily on solid
mass-timber for the main structural elements such as the floor panels, shear walls, and
columns. The main structural mass timber elements are connected by steel reinforcing
through cast-in-place concrete at the connection joints. This system plays to the
strengths of both materials and allows the designer to apply sound tall building
engineering fundamentals. The result is believed to be an efficient structure that could
compete with reinforced concrete and structural steel while reducing the associated
carbon emissions by 60 to 75%.
INTRODUCTION

The Timber Tower Research Project by Skidmore, Owings & Merrill, LLP
(SOM) was publically released in June of 2013 and is available for download at
SOM’s website. The goal of the research project was to develop a structural system
for tall buildings that uses mass timber as the main structural material and minimizes
the embodied carbon footprint of the building. The structural system research was
applied to a prototypical building based on an existing concrete benchmark for
comparison. The concrete benchmark building is the Dewitt-Chestnut Apartments; a
120m tall, 42-story building in Chicago designed by SOM and built in 1966. SOM’s
proposed system is the “Concrete Jointed Timber Frame”. This system relies
primarily on mass-timber for the main structural elements, with supplementary
reinforced concrete at the connecting joints. This system leverages the strengths of
both materials and allows engineers to apply sound tall building engineering
fundamentals. The result is believed to be an efficient structure that could compete
with reinforced concrete and structural steel systems while reducing the embodied
carbon footprint of the structure by 60 to 75%. A rendering of the structural system is
shown in Figure 1.

Figure 1: Rendering of Prototypical Building

The basis of the research project was rooted in sustainable urban development.
Recent population projections by the United Nations have estimated the current world
population of 7.0 billion people to increase to 11.0 billion people by the year 2050
[1]. More importantly, the number of people living in cities has been estimated to
double from 3.5 billion people to 7.0 billion people during the same period. Tall
buildings will probably be needed in order to house that many additional people in
growing cities. Tall buildings constructed to meet population demands need to be
done in sustainable ways to limit the environmental impacts.

Tall buildings pose a challenge to sustainable city development because they

offer both positive and negative environmental impacts. Positive impacts include
reducing urban sprawl, promoting alternative transportation, and efficient energy use.
These benefits come at the cost of emitting more carbon dioxide to produce the
materials and to construct the building. These carbon emissions are referred to as the
embodied carbon footprint of a building. A tall building’s embodied carbon footprint
is significantly higher than that of low-rise buildings on a per square foot basis. This
is because the structure is usually responsible for the majority of the building’s
embodied carbon footprint, and tall buildings require far more structure to support
their height. The structural materials and systems chosen for a tall building can have a
significant impact on the overall embodied carbon footprint of the building.

Wood has been shown to be more sustainable than other materials because it
generally requires less energy to produce than structural steel and reinforced concrete
[5-7]. More importantly, wood is approximately 50% carbon by weight, a carbon sink
that is the natural result of photosynthesis. These sustainable aspects make wood an
attractive material in which to construct the sustainable cities of the future. Recently,
developments in mass-timber technology are overcoming strength and fire challenges
associated with wood structures. Mass timber products such as cross-laminated
timber (CLT) can be built up using small pieces of dimensional softwood lumber and
structural glue to achieve panels as large as 305mm thick and 12.2mm long. These
panels can be used as floors and shear walls with structural sizes necessary to support
a tall wooden building. Wooden members of this size have an equally important
characteristic; they behave like heavy timbers in a fire which form an insulating char
layer. The charring behavior is predictable and preserves a portion of the member’s
structural strength making performance based fire design of mass-timber structures
possible. Mass-timber has made wood a viable choice for multi-story buildings as
evidenced by projects in Europe, Australia, and many other proposals around the
globe. SOM’s proposed system for high-rise buildings the primarily uses mass-timber
such as CLT and glued laminated members to achieve the necessary structural
performance while minimizing the embodied carbon footprint of the structure.

PROTOTYPICAL BUILDING LAYOUT

The structure of the benchmark Dewitt-Chestnut Apartments consists of a

191mm thick reinforced concrete flat plate which spans between interior reinforced
concrete columns toward the back and inside of the apartment units. The flat plate
spans 6.71m from the interior columns to a reinforced concrete framed tube at the
exterior. The columns in the framed tube are typically spaced at 1.68m on center. The
intent of the research project was not to replicate the same structural system as the
reinforced concrete benchmark in mass-timber; rather it was to support the same
architectural program with a structural system suited to mass-timber. The structural
elements in the prototypical building are also arranged based on contemporary
interior planning strategies. A diagram of the typical floor is shown in Figure 2.

The primary goal of any structural system is to provide an efficient,

marketable and valuable building to the owner and occupants. A marketable building
must have adequate and flexible floor area to layout useful space for the occupants.
The most marketable building lay out is an open floor plan which allows a variety of
room layouts and maximum flexibility for future changes. A truly open floor layout
requires that the floor structure span the entire distance of the leasable area. This
distance in the Benchmark Chestnut Dewitt was 8.69m with a clear span of 8m. The
most advantageous mass timber system to span this distance is a flat CLT panel. This
floor system is redundant and can allow for varying floor penetration arrangements
while also minimizing floor to floor height and overall wind loads on the building.
The mass-timber floors are supported by mass-timber walls toward the center of the
floor-plate and timber columns spaced between 4.88m and 8.53m at the perimeter.
The shear wall core at the center of the floor plate is extended to the perimeter at four
locations to improve the lateral load resisting system while also eliminating a ‘corner
condition’ for the floor system. The apartment layouts are coordinated with the shear
walls in the typical lease areas to minimize their impact on flexible space.

Figure 2: Typical Floor Diagram

MASS-TIMBER AS A MATERIAL FOR HIGH-RISE BUIDINGS

Mass-timber is a lightweight, anisotropic material with lower stiffness and

strength characteristics compared to structural steel and reinforced concrete. Mass-
timber also experiences movements due to changes in moisture content as well as
long term creep. Structural systems which utilize mass-timber must consider each of
these characteristics in order to be successful.

The density of wood is typically 4.7kN/m3 for most species which is

approximately 20% of the density of concrete. The structural dead load of the
reinforced concrete benchmark building was 7.09kN/m2 compared to 2.97kN/m2 for
the prototypical building. This reduces the compressive demands on the vertical
structure including columns and bearing walls. However, wood’s lower density also
has the disadvantage of increasing the potential for net uplift due to lateral loads and
increased floor vibrations due to walking excitation.

Wood is anisotropic because of its grain structure. Consequently, wood is

approximately three times stronger when load is applied parallel to the grain
compared to perpendicular to the grain. Wood is also stronger in compression than
tension. The allowable tension parallel to grain is approximately half that in
compression and tension perpendicular to grain is not permitted due to the risk of
splitting failures. Wood and mass-timber can be most competitive with structural steel
and reinforced concrete when the material is used primarily to resist compression
parallel to grain. The design of connections in mass-timber buildings require care to
avoid tension stress perpendicular to grain.

The strength and stiffness of wood related to the support of a high-rise

building is demonstrated in Figures 3 and 4. These comparisons show the required
concrete, steel, and timber column sizes to support a 5338kN load (typical load
toward the base of the prototypical building). Figure 3 shows the required column
size to satisfy strength with the load applied parallel to grain in the case of timber.
This comparison shows that a reasonably sized timber column can support the load of
a high-rise building. Figure 4 shows the resulting column shortening due to 5,3338kN
over a 3.05m length. The timber column experiences more movement primarily due
to the lower modulus of elasticity of the material which is typically 30% of the elastic
modulus for normal weight concrete. Thus, a mass-timber system will typically need
to have a larger cross-sectional area compared to a conventional structural steel or
reinforced concrete system in order to achieve same deformations.
 Figure 3: Column Strength Comparison

Figure 4: Column Stiffness Comparison

Wood experiences movements due to changes in moisture content and long

term creep. The moisture content of wood is typically 30% when harvested. The
moisture content of ordinary lumber is typically less than 20% when installed. Mass-
timber is typically dried to 12% in order to be processed and 12% is near the
equilibrium moisture content for most North American climates. This minimizes
potential shrinkage movement common with ordinary wood construction. Mass-
timber panels such as CLT have an additional benefit of material stability due to the
alternating orientations of wood members. Wood members experience very little
movement parallel to grain and the alternating pattern restrains shrinkage movement
from layer to layer. However, creep needs to be considered in the design of the floors
to avoid excessive deflections as well as the vertical structure to avoid long term
differential shortening. For example, the compression forces transferred through
bearing on the concrete joints from floor to floor result in significant shortening of the
vertical elements. The long term differential shortening between the shear wall core
and perimeter columns would be approximately 76mm based on a long term creep
modifier of 1.5. Additional research on methods to compensate for these movements
is necessary and was recommended in SOM’s report.

GRAVITY LOAD RESISTING SYSTEM

Since mass-timber floor systems have relatively low weight and stiffness
characteristics, floor deflections and vibrations due to walking excitation are primary
design concerns. The floors of the prototypical building were designed to meet an
instantaneous deflection of L/360 and a total including long term deflection of L/240.
A long term creep deflection multiplier of 1.5 times the sustained load deflection was
used based on NDS guidelines [3]. The floors were also designed to satisfy vibrations
according to AISC Design Guide 11 using the velocity based methodology which was
found to be more useful for flat slab type floors [4].

The floor panels were initially designed to span with simple supports between
the shear walls at the center of the floor plate and spandrel beam at the perimeter. The
required panel thickness for the 8m span was determined to be 343mm thick and was
controlled by vibrations. The materials quantities associated with that thickness was
thought to be too great compared to the material required for the concrete benchmark
to be economically viable. Therefore, alternative methods were investigated to reduce
the material used.

Evaluation of the criteria shows that floor stiffness is the most effective way
to control vibrations. The floor stiffening effect of end rotation restraint (fixed end
condition) was quickly realized as an efficient way to reduce vibrations. It was
determined that a 203mm thick mass-timber floor panel satisfies vibration criteria if
end restraint was provided. This requires moment connections at the end connections
of the mass timber floor panels to the vertical elements such as mass-timber shear
walls and structurally glued laminated perimeter columns. Several connection
schemes were investigated to provide the required moment connections and rotational
restraint from the vertical elements. Steel reinforcing epoxy connected to the mass
timber and cast in reinforced concrete joints was determined to be the most
reasonable solution due to the ability of reinforced concrete to resist complex load
paths such as torsion. Details of the connections are shown in Figure 5. The
reinforced concrete joints are able to resist floor to floor compression, shear, bending
moments, and torsion due to gravity loads while the timber is only stressed in simple
compression, shear or flexure. This approach results in more efficient floor panels and
contributed to the overall efficiency of the composite timber system.
 The columns and walls were designed for combined compression and bending
due to the restraint of the floors. The bending moments from the floors are transferred
to the vertical elements through epoxy connected reinforcing similar to the floor
panels.

Figure 5: Typical Structural Connections

LATERAL LOAD RESISTING SYSTEM

The lateral load resisting system for the prototypical building required greater
geometric stiffness than a conventional reinforced concrete building to overcome the
relatively low material stiffness of mass-timber. The lateral load resisting system
simultaneously needs to be efficient while maximizing the radius of gyration to
minimize potential for net uplift due to the relatively light composite timber system.
This resulted in the addition of the four shear walls located in the lease span of the
typical floor area. These walls add both the necessary stiffness while improving the
system efficiency and resistance to net uplift due to lateral loads.

The potential stiffness of the efficient mass-timber shear wall layout cannot be
realized without adequate link beams. The link beams need to be strong enough to
resist the coupling shear forces due to differential vertical movements between
individual wall panels. The ultimate coupling shear determined for the prototypical
building was in the range of 222kN to 445kN depending on the location. Possible
designs for these link beams were evaluated using reinforced concrete, structural
steel, and timber as shown in Figure 6. The maximum available shear capacity of a
structural select timber link beam was only 178kN which is inadequate. Another
concern with the choice of a timber link beam is the development of the shear and
moment at the connection to the shear wall. This connection would be complicated by
the horizontal grain orientation at the link beam connected to the vertical grain
orientation of the wall and would require embedded steel plates in order to transfer
the load. Thus, the load demands required the link beam to be designed as either
reinforced concrete or structural steel. Reinforced concrete was the logical choice
since the material had already been introduced to restrain the end rotation of the floor
panels. The reinforced concrete link beam is adequate to resist the required forces and
simplifies the end connection to the timber shear wall through a combination of
bearing and epoxy reinforcing connections.

Figure 6: Maximum Available Shear of Various Link Beams

Net uplift due to wind overturning moments was a primary concern due to the
lightness of the composite timber system. Potential for net uplift was reduced for the
prototypical building with the following strategies:

 Flat panel floor system minimizes floor-to-floor height and therefore wind
loads
 Floor spans from core to perimeter without additional intermediate
columns, maximizing the dead load on the core
 Vertical walls toward the middle of the floor plate are avoided
 The concrete spandrels act as moment frames between the ends of the
outrigger walls and adjacent columns, transferring dead load from the
columns to the core
 The concrete joints provide ballast, accounting for 55% of the structural
dead weight on a typical floor

The combination of these strategies could not completely eliminate net uplift which
was determined to occur from the foundations to Level 6. The maximum ultimate
uplift was determined to be approximately 4450kN and occur at the four pilaster
columns which connect directly to the ends of the outrigger walls. The prototypical
building was designed to resist this net uplift with structural steel plates embedded
within the four pilaster columns. The structural steel plates are connected from floor-
to-floor with bolted connections at the location of the reinforced concrete joints.

The horizontal tip deflection at the top of the prototypical building due to
wind load was determined to be H/600 which is stiffer than the target of H/500.
Strength checks of the shear wall elements were typically governed by gravity load
combinations and not wind load combinations. These design checks reveal that
managing net uplift was the driving factor in the design of the lateral load resisting
system.
CONSIDERATIONS RELATED TO FIRE

Mass-timber behaves similarly to heavy timber members in a fire, in that they

both form a char layer. The char layer insulates the core of the member, keeping it
cool and preserving a portion of the overall structural member strength. The
formation and growth of the char layer thickness over time is a predictable behavior
which can therefore be engineered. Structural members can then be proportioned to
maintain stability in the reduced strength state under service loads. This performance
based design approach is permitted in European standards but not by current U.S.
codes. Fire engineers will need to develop the appropriate code provisions and fire
testing necessary to verify the behavior of this and other systems before they can be
implemented in high-rise construction in the United States. Additionally, the design
community must continue to work creatively with forward thinking municipalities
and code officials using the latest in fire engineering and performance based design to
make timber buildings a viable alternative for more sustainable tall buildings.

NON-STRUCTURAL SYSTEMS

Moisture control is a primary design consideration for mass-timber buildings.

Protection from moisture will help ensure a durable building with a long service life.
The mass-timber is protected from moisture in the prototypical building by providing
shop applied damproofing at all timber/concrete interfaces. This limits moisture
exposure from water in the concrete prior to curing and potential moisture migration
through the concrete in service. The mass-timber is also protected from moisture with
a 51mm thick lightweight gypsum concrete screed on top of the floor panels. This
topping protects the timber from leaking plumping in services or infrequent events
such as a pipe failure or sprinkler activation.

The lightweight gypsum concrete screed was also useful in controlling

acoustic performance. The acoustic performance desired for the prototypical building
was a sound transmission class (STC) and impact insulation class (IIC) greater than
55. The buildup of the gypsum concrete over an underlayment mat approximately
doubled the acoustic rating of the bare CLT floor panel and allowed the acoustic
performance to be met.

Attachment of non-structural systems such as the exterior wall and elevators

was also considered in the prototypical building and shown in Figure 7. The
reinforced concrete joints again proved useful in the support of these elements. The
exterior cladding can be mounted to the reinforced concrete spandrel beam with
similar attachments used in a reinforced concrete building slab edge or a steel
building with composite metal deck slabs. The reinforced concrete spandrel also
provides a stiff element at the edge of the floor plate which can be designed to
minimize cladding joint deformations caused by differential floor to floor live loads.
Elevator guide rails and divider beams can also utilize the reinforced concrete joints
located on the inside face of the shear walls inside the elevator shafts. The guide rails
and divider beams can be connected to these concrete joints similar to connecting to a
typical concrete beam in an all reinforced concrete building. The ability to minimize
the impact on other trades will prove useful in the implementation of this system.

Figure 7: Non-Structural Systems Integration

EMBODIED CARBON FOOTPRINT COMPARISON

The cradle to gate embodied carbon footprints of the benchmark and

prototypical building are compared in Figure 8. Two scenarios are considered. The
first scenario assumes standard material production. The second scenario considers
60% cement replacement for both structures and air drying of the timber elements.
Embodied carbon footprints are calculated based on references [5-9]. Both scenarios
consider a construction carbon cost of 2.97kN/m2. This preliminary study suggests
that the embodied carbon footprint of the prototypical building will be 60 to 75% less
than the benchmark building.
 Figure 8: Embodied Carbon Footprint Comparison

CONCLUSIONS

SOM believes that the proposed “Concrete Jointed Timber Frame” system is
technically feasible from the standpoint of structural engineering, architecture,
interior layouts, and building services. Additional research and physical testing is
necessary to verify the actual performance of the structural system. SOM has also
developed the system with consideration to constructability, cost, and fire protection.
Reviews from experts in these fields and physical testing related to fire is also
required before this system can be fully implemented in the market. Once
implemented, this system could be competitive with conventional construction
technologies due to the flexible interior spaces provided, comparable material
consumption, and drastically reduced embodied carbon footprint.

REFERENCES

[1] United Nations, Department of Economic and Social Affairs, Population Division
(2011). World Population Prospects: The 2010 Revision, Volume I and 2:
Comprehensive Tables. ST/ESA/SER.A/313.

[2] Karacabeyli, E and Douglas, B., (2013). Cross Laminated Timber Handbook. SP-
529E. U.S. Department of Agriculture, Forest Products Laboratory, Binational
Softwood Lumber Council.

[3] American Wood Council (AWC). (2012). National Design Specification for
Wood Construction with Commentary. Leesburg, VA.

[4] Murray, T.M., Allen, D.E. and Ungar E.E. (1997). “Floor Vibrations Due to
Human Activity.” AISC Steel Design Guide #11, American Institute of Steel
Construction, Chicago, IL.
 [5] Puettmann, M.E., and Wilson, J.B., (2005). “Life-Cycle Analysis of Wood
Products: Cradle-To-Gate LCI of Residential Wood Building Materials.” Wood
and Fiber Science. Vol. 37 Corrim Special Issue. pp18-29.

[6] Puettmann, M.E., Bergman, R., Hubbard, S., Johnson, L., Lippke, B., Oneil, E.,
and Wagner, F. (2010). “Cradle-to-Gate Life-Cycle Inventory of US Wood
Products Production: CORRIM Phase I and Phase II Products,” Wood and Fiber
Science, 42(CORRIM Special Issue), 2010, pp. 15–28.

[7] Puettmann, M.E., Oneil, E., Milota, M.R. and Johnson. L.N. (2013). “Cradle to
Gate Life Cycle Assessment of Glue-Laminated Timbers Production from the
Pacific Northwest.” CORRIM Report Update. 25pp.

[8] Hammond, G. P. and Jones, C. I. (2008). “Embodied energy and carbon in

construction materials.” Proceedings of the Institution of Civil Engineers -
Energy, 161 (2). pp. 87-98. http://dx.doi.org/10.1680/ener.2008.161.2.87

[9] British Standards Institution (BSI). (2011). “BS PAS 2050:2011” London, UK.

View publication stats