Modeling and Analysis of Structural Systems for

Efficient Design
Modeling the diaphragms (flexible, semi-rigid, rigid)

Week 5 Handout
2
 Week 5: Modeling Diaphragms and Shear Walls

◼ Diaphragms
◼ Basic concepts of floor and roof diaphragms
◼ Classification of diaphragms as flexible, rigid, or semirigid
◼ Determining gravity, wind, and seismic loads on diaphragms
◼ Modeling semirigid diaphragms using shell elements
◼ Influence of finite element mesh density on efficiency and accuracy of analysis

◼ Shear Walls
◼ Basic shear wall concepts
◼ Modeling shear walls using frame and/or shell elements
◼ Single planar walls
◼ Coupled walls
◼ More complex wall layouts (L, T, and box shapes)
◼ Influence of finite element mesh density on efficiency and accuracy of analysis

3
 Week 5: Learning Outcomes

◼ Upon completion of this week’s content you will be able to :

◼ Describe the basic configuration and behavior of diaphragms
◼ Correctly classify the diaphragm as flexible, rigid, or semirigid
◼ Appropriately and efficiently model diaphragms using shell elements
◼ Apply gravity, wind and seismic loads in an appropriate manner
◼ Appropriately and efficiently model shear walls using fame and/or shell elements

24
 Week 5: Significance on the Job

◼ Diaphragms and shear walls are a critical part of the load path and must be modeled and
detailed appropriately

◼ Modeling diaphragms and shear walls from a top-town full-system perspective creates the most
versatile analytical model without adding significantly to model development or analysis time

25
Next Scene: Basic Concepts of Floor and Roof
Diaphragms

6
7
Basic Concepts of Floor and Roof Diaphragms
9
 Diaphragm Concepts

◼ Diaphragms play an essential role in gravity and lateral load paths:

◼ Distribute gravity load to vertical elements
◼ Serve as a wind load transfer mechanism from exterior cladding to the lateral load resisting systems
◼ Transfer lateral inertial forces to vertical elements of the seismic force-resisting system
◼ Distribute forces among vertical lateral force elements
◼ Transfer forces from podium slabs and subterrain slabs to basement walls and supporting soil
◼ Provide bracing for gravity columns (leaner columns)

◼ Out-of-plane bending stiffness of the diaphragms is important to transfer gravity loads

◼ In-plane axial and shear stiffness (or flexibility) of diaphragms is critically important for resisting
lateral loads

◼ In the past, classifying the in-plane diaphragm stiffness as rigid or flexible had certain
advantages. Currently, diaphragms are commonly (and easily) modeled as semirigid with shell-
type finite elements.
10
 Actions Needed for Diaphragm Design

◼ Structural analysis results must provide the forces needed for the design and detailing of the
diaphragm. Typically, these forces are grouped into the diaphragm panel, diaphragm chords,
and diaphragm collectors (also called drag beams or drag struts).
◼ Structural analysis results must also include the design forces in the lateral load resisting system.
The force distribution can be highly dependent on the in-plane stiffness of the diaphragm relative
to the stiffness of the lateral load resisting system.
◼ Accurate calculation of the design forces requires modeling the diaphragm as semirigid.

Diaphragm Chord

Diaphragm Panel

Load Direction

Diaphragm Collector 11
 Reasons Why a Semirigid Diaphragm is
Preferred for Analysis

◼ For most building systems under wind or seismic loading, a diaphragm cannot be classified as
flexible by ASCE 7. Hence, a 3D model is required, and diaphragms must be explicitly modeled.

◼ Wind Loading: Diaphragms can be considered as rigid to determine which load cases are
required for design (See Section 27.3.5 of ASCE 7-16). However, Section
27.4.5 requires that the structural analysis consider the relative stiffness of diaphragms and the
vertical elements of the MWFRS.

◼ Seismic Loading: Section 12.3.1.2 of ASCE 7-16 states that a diaphragm can be considered as
rigid if it consists of a concrete slab or concrete filled metal deck with a span to depth ratio of 3 or
less in structures that have no horizontal irregularities. However, most structures have at least
one horizontal irregularity so the diaphragm cannot be classified as rigid, and Section 12.7.3
requires that the analysis shall account for the participation of the diaphragm stiffness.

12
Modeling a Diaphragm as Semirigid

◼ Modeling a diaphragm as semirigid can be accomplished several ways.

◼ The principal requirement is that the model allows in-plane deformations (axial, shear, and
bending)

◼ Use of a horizontal truss or strut and tie mode is possible but can be cumbersome. There are
some advantages for reinforced concrete diaphragms.

◼ Use of membrane elements is possible, but out-of-plane bending DOF will need to be restrained
if more than one element is utilized per rectangular bay.

◼ Use of plate bending elements is not applicable because in-plane deformation are not included.

◼ Use of shell elements is preferred. The number of DOF will generally increase significantly, but
solution times are not greatly affected. Where shell elements are used it may be necessary to
use a reduced bending stiffness to avoid unintended coupling of shear walls and other sources of
unintended stiffness.

13
Some Important Diaphragm Considerations (1)

Forces in the typical diaphragm

levels are highly dependent on the
relative in-plane stiffness of the
diaphragm and the stiffness of the
lateral load resisting system.

14
Important Diaphragm Considerations (2)
Forces in the typical diaphragm
levels are highly dependent on
the relative in-plane stiffness of
the diaphragm and the stiffness
of the lateral load resisting
system.

At transfer levels and other

levels closely adjacent to a
discontinuity, the influence of
the diaphragm stiffness is
even more critical.

15
Important Diaphragm Considerations (3)
Forces in the typical diaphragm
levels are highly dependent on
the relative in-plane stiffness of
the diaphragm and the stiffness
of the lateral load resisting
system.

At transfer levels and other

levels closely adjacent to the
discontinuity, the influence of
the diaphragm stiffness is
even more critical.

Forces in the podium level and

subgrade diaphragms and the
shears and moments in the
subgrade lateral systems
are most critically affected
by in-plane diaphragm stiffness.

16
 Diaphragm Loads due to Wind

◼ Wind loads develop as a result of pressures acting on the windward, leeward, and sidewalls of
the structure.
◼ Loads should be applied as distributed loads along the edge of the diaphragm. Element loads
and/or nodal loads may be used.

Principal Wind Load Direction

17
Pressures acting on building from wind tunnel test Pressures applied to model as line loads
ASCE 7 Requirements for Diaphragm Loading and Analysis under Wind
◼ Lateral loads are computed in accordance with Chapters 26 and 27 of ASCE 7 and applied to the structure
in accordance with Sections 27.3.5 and Figure 27.3.8.
◼ Torsional load cases 2 and 4 may be omitted in some cases (see Appendix D). Several of these exceptions
apply where the diaphragm is defined as flexible in accordance with Section 27.4.5. Note that only
diaphragms constructed from wood panels can be defined as flexible.
◼ In accordance with Section 27.4.5 the structural analysis must consider the relative stiffness of diaphragms
and the vertical elements of the MWFRS. This can only be accomplished by use of a semirigid diaphragm
model.

Required wind load

Cases on ASCE 7

18
 Diaphragm Loads due to Earthquakes

◼ Design of the diaphragm requires a separate loading than used for the design of the lateral
load resisting system.

◼ For the lateral load resisting system, the analysis is performed using either the Equivalent Lateral
Force method (ELF), the Modal Response Spectrum (MRS) method, or the linear Response
History (LRH) method. These procedures are described in Sections 12.8 and 12.9 of ASCE 7.

◼ ASCE 7 does not provide procedures for determining diaphragm design forces directly from MRS
or LRH analysis. Design of the diaphragm using forces obtained using MRS or LRH would not be
code compliant (NIST, 2011).

◼ For the design of the diaphragm the lateral forces are obtained using one of two procedures
available in Section 12.10. Note that these forces would be applied as static lateral loads
even if the lateral load resisting system is analyzed using MRS or LRH analysis.

19
ASCE 7-16 Requirements for Diaphragms under Seismic Loading
◼ Design and detailing of diaphragms, chords, and collectors is in accordance with Section 12.10.
These requirements are significantly more complex than that required by wind.

◼ Most structures will be designed based on Sections 12.10.1 and 12.10.2. Precast concrete
diaphragms in SDC C, D, E, and F must be designed using Section 12.10.3.

◼ The structural analysis used to determine forces within the diaphragm, chords, and collectors
is based on lateral loading that is different than that used to design the lateral load resisting
system.

◼ Design of the diaphragm, chords, and collectors may require the use of the redundancy factor r,
or the overstrength factor Wo.

20
Determining Lateral Seismic Loads for use in Designing Diaphragms

140

120

100

Height, ft
80

60

40

20

0
0.0 50.0 100.0 150.0 200.0 250.0 300.0
Diaphragm Force Fpx, kips
Lateral Force F Diaphragm Force Fpx Minimum Diaphragm Force

Forces computed for a 10 Story Building

SDS=1.0, SD1=0.4, story weight w = 1265.6 k

21
 Use of Response Spectrum or Response History for Determining
Diaphragm Forces
164.3 k 10 164 k 253 k 164.3 k

143.6 k 9 154 k 253 k 143.6 k

123.6 k 8 144 k 253 k 123.6 k

104.3 k 7 134 k 253 k 104.3 k

• A separate static loading
85.7 k 6 124 k 253 k 85.7 k is required for the
68.0 k 5 115 k 253 k 5 design of the diaphragm
253 k
at each level.
51.1 k 4 106 k 253 k 51.1 k • Diaphragms are designed
using static loads even if
35.5 k 3 97 k 253 k 35.5 k MSFRS is designed using
21.2 k 2 89 k 253 k 21.2 k MRS or LRH analysis.
• See Sabelli (2009) for an
8.8 k 1 81 k 253 k 8.8 k alternate approach.

ELF Lateral forces Computed lateral Minimum lateral Suggested lateral forces
for designing MSFRS diaphragm forces diaphragm forces for designing diaphragm
Eq. 12.8-12 determined from determined from at Level 5 (NIST 2011a)
Eq 12.10-1 Eq 12.10-2
22
Application of Diaphragm Loads in 3D Analysis

◼ Where the diaphragm is modeled as semirigid the loads should be distributed throughout the
diaphragm in accordance with the distribution of mass.
◼ This is true for the analysis used for designing the lateral load resisting systems (Section 12.8
and 12.9) and for designing the diaphragms (Section 12.10)

Seismic Load Direction

Diaphragm inertial forces applied to internal

and external nodes
23
Next Scene: Examples of Diaphragm Behavior
and Dependency of Behavior on Element Mesh Density

Adams (2020) 24
25
Basic Diaphragm Modeling Concepts:
Flexible, Semirigid, and Rigid
27
 Diaphragm Classifications

Fully Flexible:
Diaphragm essentially has no in-plane stiffness. Forces in
the lateral force resisting systems are based on tributary area.

Fully Rigid:
Diaphragm is of sufficient in-plane stiffness that forces in
lateral force resisting elements are based only on the stiffness
and location of these elements.

Semirigid:
A diaphragm that is not classified as either Flexible or Rigid.

In fact, ALL DIAPHRAGMS ARE SEMIRIGID

28
Is the Diaphragm Flexible, Semirigid, or Rigid?
12.3.1.3 Calculated Flexible Diaphragm Condition.
Diaphragms not satisfying the conditions of Sections 12.3.1.1 or 12.3.1.2 are permitted to be
idealized as flexible where the computed maximum in-plane deflection of the diaphragm under
lateral load is more than two times the average story drift of adjoining vertical elements of the
seismic force–resisting system of the associated story under equivalent tributary lateral load as
shown in Fig. 12.3-1. The loadings used for this calculation shall be those prescribed by Section
12.8.

Note: This test is required for determining the classification of the diaphragm's
rigidity under seismic loads. The same test could be used for wind loads. 29
Calculated Diaphragm Flexibility

Depth

Span

Note: Seismic loading is shown as a concentrated load in

ASCE 7. It is more reasonable to use a uniform load.

30
Example for Diaphragm Flexibility

Depth=

Span=

Note: ASCE 7 does not provide a clear definition for the diaphragm
span. The span shown is the most logical interpretation.
31
SAP 2000 Model

• Walls and slabs modeled with thin shell elements.

• Uncracked properties used for both walls and slabs
• Walls are assumed fixed at their base, in- and out-of-plane.
• Distributed load applied along the length of the slab in the Y-direction
32
Analysis Results for 4” Thick Slab

0.0425”

0.0295”
0.0216”

0.0136”

ADVE=(0.0136+0.0295)/2=0.0216 in.
MDD=(0.0425-0.0216)=0.0209 in.
MDD/ADVE=0.0209/0.0216=0.967<2.0 Diaphragm is NOT Flexible
33
Distribution of Forces in Walls for 4” Slab

The diaphragm is behaving like a “flexible” diaphragm

even though the Diaphragm is not “Flexible” !

34
Analysis Results for 2” Thick Slab

0.0595

0.0297

0.0134”

ADVE=(0.0134+0.0297)/2=0.0215 in.
MDD=(0.0595-0.0215)=0.0380 in.
MDD/ADVE=0.0380/0.0215=1.767<2.0 Diaphragm is NOT Flexible
35
Summary for SAP 2000 Analysis

▪ It is quite difficult (and unlikely) to classify a diaphragm as “Flexible”

using the analytical procedure of Section 12.3.1.3.

▪ For the example given, the diaphragm behaved like a “Flexible”

diaphragm in terms of distribution of forces to the walls, although it was
not classified as a “Flexible” diaphragm.

▪ All diaphragms are semirigid, and if possible, should be included as such

in the 3-Dimensional Structural Analysis

36
Classification of Diaphragms in IBC

Depth

Span

Note: Section 1604.4 of IBC states that the diaphragm can be considered as rigid for analysis if the deflection
MDD is less than 2(ADVE). This would mean for that almost all diaphragms can be analyzed as rigid. In many
cases, the use of a rigid diaphragm would not be suitable, and the IBC provision should be used with caution.

37
 Influence of Diaphragm Modeling on Efficiency
of Calculations and Computed Behavior
◼ Slides in this scene are based on a paper by Allen Adams Chief Structural Engineer at Bentley
Systems (developers of STAAD and Ram Structural System)

◼ Building is 4-Story structural steel with moment resisting frames as the lateral load resisting
elements

◼ Diaphragms are 3-1/4” light weight concrete over metal deck. They are modeled as semirigid
using shell elements

◼ For comparison purposes the diaphragm is also

molded as rigid

◼ Analyses performed:
◼ Varying the mesh density for shell elements
◼ Reanalysis with braced frames instead of moment frames
◼ Varying the mechanical properties of the shell elements
(not discussed here)
Ram Structural System Model 38
Influence of Mesh Density on Run Times

Each bay modeled with Each bay modeled with

30 by 30 mesh 2 by 2 mesh

1 ft Mesh, 900 per Bay 15 ft Mesh 4 per Bay

Run Times for ELF Analysis
◼ Run times for semi-rigid diaphragms with 4’, 8’ and 15’ elements are
reasonable compared to rigid diaphragm

◼ Run times for 1’ and 2’ elements may be problematic if there are

numerous analyses required that use different models. Run times are
not highly dependent on the number of load cases run for a given
model.
39
Results for Moment Frame Shears
for Loading in Y direction “Error” relative 2 3
Shear, k to 1’ mesh 1 1

▪ Frame shears for semirigid

diaphragm are insensitive to
mesh density.

▪ There is a significant
difference in frame shears
between semi-rigid and rigid
diaphragm results (differences
as high as 11.2% in upper level)

40
Results for Braced Frame Shears
for Loading in Y direction 2 3

▪ Frame shears for semi-rigid

diaphragm are insensitive to
mesh density.

▪ There is a very significant

difference in frame shears
between semi-rigid and rigid
diaphragm results (differences
as high as 58.9% in upper level)

41
Results for Braced Frame Story Drift
for Loading in Y direction 2 3

▪ Story drifts are more sensitive

to mesh density than is story
shear. However, differences
for semi-rigid diaphragms are
not excessive (generally
less than 6%).

▪ There is a very significant

difference in story drift at the
center of the model when the
rigid diaphragm is used. Note
that this is due to in-plane
deformations within the
diaphragm itself. 42
Results for Braced Frame in-plane Diaphragm Forces
for Loading in Y direction 2 3

▪ Significant differences in
Diaphragm stresses not diaphragm stresses for
available from rigid diaphragm lowest resolution model
model (2x2 mesh), particularly axial
stress due to chord action in the
bridging diaphragm.
▪ 4x4 mesh has results similar
to the 30x30 mesh.

43
Summary of 4-Story Building Example
◼ The sensitivity of meshing density is more significant for the braced frame building than for the
moment frame building. This is due to the the stiffness of the diaphragm relative to the frames
being lower for the braced frame building.

◼ The significance of the differences in results would be less if the building was rectangular in plan
(did not have the narrow connecting diaphragm).

◼ The highest resolution mesh with 900 elements per bay did not produce significantly better (or
different) results than the model with 16 elements per bay.

◼ Some of the the results using a rigid diaphragm were considerably different than obtained using
semrigid elements. The rigid diaphragm also had the disadvantage of not reporting diaphragm
forces.

◼ For this building, the run time for the analysis with 16 elements per bay (4x4) was only 4 seconds.
Hence, there is no consequential loss in analysis efficiency relative to the rigid diaphragm model
(1 second).

44
 Next Scene: Modeling the Podium and Basement
Diaphragms

1000

500

0
F

Shear Force (kips)

-500

-1000

-1500
H
-2000 Basement Wall 2
Basement Wall 1
-2500

-3000
0.0001 0.001 0.01 0.1 1 10 100 1000 10000 100000

Slab Stiffness Modifier

Slab 1

h Wall 1
Slab 2

h Wall 2

45
46
Modeling Diaphragms as Rigid
 1000

500

Shear Force (kips)

F
-500

-1000

-1500
H
-2000 Basement Wall 2
Basement Wall 1
-2500

-3000
0.0001 0.001 0.01 0.1 1 10 100 1000 10000 100000

Slab Stiffness Modifier

Slab 1

h Wall 1
Slab 2

h Wall 2

48
 Influence of Podium Slab and Shear Wall
Properties on System Response

◼ Analysis of a simple 20-story building with 2 basement levels

◼ Investigate the following analysis features on basement wall

shear and total roof drift: F

◼ Thickness of diaphragm slab

◼ Including shear deformation in shear walls H

◼ Analysis is a simple 2-D model using frame elements only. A

similar structure is analyzed in more detail in the next scene.

Slab 1

h Wall 1
Slab 2

h Wall 2

49
System Analyzed: 20-Story Reinforced Concrete Shear Wall Building
Slab area contributing to spring
stiffness (next slide)

F H=240 ft
LW
h=15 ft
F
L=100 ft F
b=25 ft
H Lw=25 ft
tw=1.0 ft (12 in) F b
(2/3)H ts=0.5 ft (6 in)
EW=ES=5000 ksi
GW=GS=2000 ksi
F
F=180 k

Slab 1 F
h Wall 1
Slab 2 tW
h Wall 2

L
50
Podium Slab Influence on System Response (Frame Element Mathcad Model)

180 k
180k

EIW set to produce

H deflection of H/n
at roof

Slab 1 KS1
h Wall 1
Slab 2
h Wall 2
KS2

Spring stiffness = tsbES/0.5L

51
Podium Slab Influence on System Response (Shell Element Model)

180 k
180 k

H
5x110 mesh of
thin shell elements

Slab 1

h Wall 1
Slab 2 Grade Level

h Wall 2

Frame element and shell element models

produced very similar results (where shear
deformations are included in frame analysis). 52
 Basement Wall Shear Results (Shear Deformation in Wall)
1000 Range of probable
behavior

500
180 k
0 • Basement wall shears
Shear Force (kips)

are dependent on
-500 in-plane diaphragm
flexibility
-1000 • Probable wall shears
are as much as 3 times
-1500 shear just above grade
• Modeling slabs as rigid
-2000 Basement Wall 2 in plane produces
Basement Wall 1 unrealistic forces
-2500

-3000
0.0001 0.001 0.01 0.1 1 10 100 1000 10000 100000
Podium slabs Podium slabs
very flexible Slab Stiffness Modifier very rigid

53
Forces in Wall for Analysis (Shear Deformation in Wall)

180k 180

V=180k V=180k

Floor slab in Floor slab in

1617k 841k
compression compression

V=-1438k V=-661k

Floor slab in Floor slab in

1122k 11k compression
tension

V=-265k V=-650k

Diaphragm is Rigid Diaphragm is Semi Rigid

54
Lateral Displacement Results(Shear Deformation in Wall)
Range of probable
12 behavior

Droof=9.18” • Dslab is the lateral

10 Droof=10.8”
Dslab= 0.070” deflection at the edge
Dslab= 0.24”
Roof Displacement (inches)

of the wall at grade level

8 • Small amount of
deformation in
Droof=8.29” diaphragm has a
6 Dslab= 0.0” tremendous influence
on subgrade wall
shears
4
• Influence of diaphragm
deformation on roof
2 displacements is
small but not
negligible
0
0.0001 0.001 0.01 0.1 1 10 100 1000 10000 100000
Podium slabs Slab Stiffness Modifier Podium slabs
very flexible very rigid
55
 Basement Wall Shear Results (No Shear Deformation in Wall)
1000
More extreme differences
500 Range of probable
relative to analysis
behavior
180 k including shear deformations.
0
• Basement wall shears
Shear Force (kips)

-500 are very dependent on

in-plane diaphragm flexibility
-1000 • Probable wall shears are as
much as 6 times shear
-1500 just above grade
• Modeling slabs as rigid
-2000 in plane produces unrealistic
Basement Wall 2 forces
-2500 Basement Wall 1

-3000
0.0001 0.001 0.01 0.1 1 10 100 1000 10000 100000
Podium slabs Podium slabs
very flexible Slab Stiffness Modifier very rigid

56
Forces in Wall for Analysis (No Shear Deformations in Wall)

180k 180k

V=180k V=180k

Floor slab in Floor slab in

2649k 997k compression
compression

V=-2469k V=-817k

Floor slab in
Floor slab in 149k
3291k compression
tension

V=822k V=-1005 k

Diaphragm is Rigid Diaphragm is Semirigid

57
Lateral Displacement Results(No Shear Deformation in Wall)
Range of probable
12
behavior
Similar to analysis
including shear def.
10
Droof=10.8” Droof=8.8”
Roof Displacement (inches)

Dslab= 0.083” • Small amount of

Dslab= 0.23”
deformation in
8 diaphragm has a
tremendous influence
Droof=7.4” on subgrade wall
6
Dslab= 0.0” shears
• Influence of diaphragm
deformation on roof
4
displacements is
small but not
2 negligible

0
0.0001 0.001 0.01 0.1 1 10 100 1000 10000 100000
Podium slabs Slab Stiffness Modifier Podium slabs
very flexible very rigid
58
 Comparison of Wall Shears With and Without Shear Deformation

Shear Deformation in Walls No Shear Deformation in Walls

1000 1000

500 500

0 0
Shear Force (kips)

Shear Force (kips)

-500 -500

-1000 -1000

-1500 -1500

-2000 Basement Wall 2 -2000

Basement Wall 1 Basement Wall 2
-2500 -2500 Basement Wall 1

-3000 -3000
0.0001 0.001 0.01 0.1 1 10 100 1000 10000 100000 0.0001 0.001 0.01 0.1 1 10 100 1000 10000 100000

Slab Stiffness Modifier Slab Stiffness Modifier

59
 Comparison of Displacements with and without Shear Deformation

Shear Deformation in Walls No Shear Deformation in Walls

12 12

10 10
8.81in.

Roof Displacement (inches)

9.18 in.
Roof Displacement (inches)

8 8

6 6

4 4

2 2

0 0
0.0001 0.001 0.01 0.1 1 10 100 1000 10000 100000 0.0001 0.001 0.01 0.1 1 10 100 1000 10000 100000
Slab Stiffness Modifier Slab Stiffness Modifier

60
Example Summary
◼ 20-Story reinforced concrete shear wall analyzed using a simple model that varies in-plane
diaphragm stiffness and either includes or ignores shear deformations in the wall. The principal
finding are as follows:
1. It is essential to include in-plane deformations in the podium-level and basement slabs.
Using rigid diaphragms significantly overestimates basement level wall shear forces.
2. Including shear deformation in the shear wall also reduces the subgrade shears in the shear walls
3. Other modeling behaviors not explored here could also produce significant differences in
the behavior of the structure. F
• Including cracking (bending and shear) in the
subgrade levels of the wall. Here, the
diaphragm is becoming more rigid relative
to the wall.
• Including flexibility of the soil behind the basement wall. Here
the diaphragm is effectively becoming more flexible relative
to the wall
Added spring for
It would take several trial analyses to establish the appropriate soil flexibility
modeling details. Where there is considerable uncertainty a KS1
KSoil
bounding analysis should be used to capture the most critical Cracking in
effect. subgrade wall
KS2
61
 Next Scene: Modeling Diaphragms with Shell
Elements

5 6

Shell 1

a, r b,s c,t
r s t
3 4
a b c 3

Shell 2

Y
r s t
1 2

Z (up) X
62
 1000

500

Shear Force (kips)

F
-500

-1000

-1500
H
-2000 Basement Wall 2
Basement Wall 1
-2500

-3000
0.0001 0.001 0.01 0.1 1 10 100 1000 10000 100000

Slab Stiffness Modifier

Slab 1

h Wall 1
Slab 2

h Wall 2

63
Modeling Diaphragms as Semirigid using
Shell and Membrane Elements
 5 6

Shell 1

a, r b,s c,t
r s t
3 4
a b c 3

Shell 2

Y
r s t
1 2

Z (up) X

65
 Modeling Diaphragms with Shell Elements

◼ In almost all cases the diaphragm should be modeled with shell elements

◼ Some programs provide several different shell element types:

◼ Thin shell (recommended)
◼ Thick shell (includes through-thickness shear deformations)
◼ Multilayered shell (used primarily for nonlinear analysis)

◼ Materials may be isotropic (same in the two orthogonal directions) or orthotropic (different
properties in the two orthogonal directions). For a concrete slab or concrete slab over metal deck
the use of isotropic properties is recommended

66
Material and Section Properties Needed for Thin Shell Element
◼ Material properties include:
◼ Modulus of elasticity
◼ Poisson’s ratio
◼ Coefficient of thermal expansion
◼ Mass and weight density

◼ Section properties include:

◼ Element type (thin, thick, layered)
◼ Material property type
◼ Bending thickness
◼ Membrane thickness
◼ Property modifiers (optional)

Property modifiers are typically used to account for cracking, where typical values are in the range of 0.25
to 0.50. Modifiers can also be used to reduce bending stiffness to reduce unintended coupling provided
by the diaphragm. This is primarily an issue in seismic design.

67
Using Property Modifiers for Shell Elements (SAP 2000)

Note: Membrane and

bending thickness can
be set separately. Element
mass is based on
membrane thickness.

0.50
0.50 In-plane axial and shear. Value of 0.5 is
0.50 used to represent cracking.
0.01 Out of plane bending (m11 and m12)
0.01
and drilling torsion (m12). Value of 0.01 is
Caution: Using a low bending stiffness can cause used to eliminate unintended coupling
unrealistic high-period vertical vibration of the diaphragms. Out of plane (thru-thickness) shear
This can be a consideration in dynamic analysis and is
addressed in more detail in Weeks 7 and 12.

68
 Meshing the Diaphragm

◼ Rectangular elements are preferred. Some programs that perform automatic meshing may use
triangular or trapezoidal elements in some regions.

◼ The preferred element aspect ratio is between 0.5 and 2.0. Ideally the ratio should be as close
to 1.0 as possible.

◼ The mesh density to use depends on the application and on the information desired from the
analysis.
◼ If no stresses or forces are to be obtained from the diaphragm, a 2-by-2 or 3-by-3 mesh per rectangular
bay is likely acceptable.
◼ If stresses or forces are required a finer mesh is appropriate. In most cases a 5 by 5 mesh per
rectangular bay is sufficient but for odd geometries a larger number may be required. Trial and
error may be needed to determine the most appropriate mesh.

◼ The procedure used for meshing the diaphragm depends heavily on the program used. Also,
there may be numerous methods to create the same mesh using a given program.
69
 Meshing the Diaphragm

1. Establish grid lines 1-4 and A-E 1 2 3 4

2. Add nodes a grid line intersections
3. Draw frame elements between nodes as A Frame 1 Frame 2 Frame 3

required (only frame elements 1-10 are

Frame 10
Frame 7

Frame 8

Frame 9
identified here)
4. Draw full-bay shell elements in desired
Frame 4 Frame 5 Frame 6
locations (only elements 1 and 2 shown here). B
5. Mesh the shell elements into a 4-by-4
arrangement. Shell 1
6. If the frame elements are not automatically
meshed (into 4 segments each) this C
should be done manually.
7. The vertical deflections in the frame should Shell 2
be compatible with those in the shell element.
Y D
Note:
Steps 1-4 may be automatic.
E
Z (up) X 70
 Joining Elements with the Same Mesh

◼ Nodes 1-6 are defined as shown in previous 5 6

slide

◼ Nodes shown as “ ” are created when Shell 1

Shell 1 is meshed

◼ Nodes shown as “ “ are created when a, r b,s c,t

r s t
Shell 2 is meshed 3 4
a b c 3 4
◼ It is necessary to impose a boundary
constraint such that nodes 3, 4, a & r, b & s,
Shell 2
and c & t lie along the same line when the
elements is deformed. This may be
automatic in some programs. Y

1 2

Z (up) X 71
 Make Sure the Frame Elements Go Along
for the Ride

5 Shell 1
6
a, r b,s c,t

3 4

Correct: The frame element deforms

r s t consistently with the shells
3 4
a b c

r s
t
3 c 4
Y a b
r s t
Incorrect: The frame element deforms
1 2 inconsistently with the shells
Shell 2
Z (up) X 72
 Joining Elements with a Different Mesh

◼ Nodes 1-6 are defined when developing 5 6

the model

◼ Nodes shown as “ ” are created when Shell 1

Shell 1 is meshed

◼ Nodes shown as “ “ are created when r s t

r s t
Shell 2 is meshed 3 4 a b
a b 3 4
◼ It is necessary to impose a boundary
constraint such that nodes 3, 4, a, b, r, s,
and t lie along the same line when the Shell 2
elements is deformed. This step may
not be automatic in all programs. Y
Note: Avoid using incompatible 1 2
meshes wherever possible (especially where
there are frame elements on the same boundary)
Z (up) X 73
Diaphragm Meshed Using SAP 2000
• Grid Lines A-H and 1-6 are formed
automatically when developing
the model.
• Beams (shown in blue) are
also formed automatically and span
between grid lines one element per bay
• Initially there is one shell element
generated for each bay (no meshing)
• The analyst requests each bay be meshed
as shown (6-by-4)
• Beams are automatically meshed with the
diaphragm (subdivided into 4 or 6 elements)
• Edge constraints are automatically formed
such that beam and shell deflections are
Plan view of one floor of 12-story building compatible along grid lines
• Meshing the openings takes extra manual
effort and is not described here

74
 Postprocessing After Analysis

◼ When the analysis is complete the following items may be required for design:
◼ Forces within the panel of the diaphragm
◼ Forces in the chords
◼ Forces in the collectors

◼ Many programs make the postprocessing easy by allowing the analyst to “cut” through the
diaphragm and report axial forces, shears, and moments at the cut location.

75
Next Scene: Basic Concepts for Modeling
Shear Walls

Fixed supports at base

76
 5 6

Shell 1

a, r b,s c,t
r s t
3 4
a b c 3

Shell 2

Y
r s t
1 2

Z (up) X

77
Modeling Flexible Diaphragms
Fixed supports at base

79
 Reinforced Concrete Shear Wall Modeling

◼ Emphasis on “slender” walls (i.e., walls with height to width ratio greater than 4)

◼ Modeling isolated wall systems, frame-wall systems, and coupled wall systems

◼ Modeling with frame elements, shell elements, or a combination of frame and shell elements

◼ Analytical models should be 3D for final design. Planar models are appropriate for preliminary
design.

◼ Planar models demonstrated in this Scene could be analyzed independently or as part of a 3D

system

80
 Factors Strongly Affecting Computed Behavior
of Structures Incorporating Walls

◼ Modeling the soil-foundation-structure interface (discussed in detail in Week 6)

◼ “Backstay” influence of in-plane diaphragm flexibility in grade-level and below grade diaphragms

◼ Unintended coupling due to out-of-plane diaphragm bending stiffness

◼ Shear deformations, especially in the wall and coupling beams

◼ Flexural and shear cracking in the shear wall

◼ Flexural and shear cracking in frame elements, especially in coupling beams

◼ Cracking in the diaphragms

Note: There is often considerable uncertainty in including the above effects in analysis.
All realistic ranges of behavior should be explored to ensure that the critical effects
and associated demands are captured.
81
 Modeling Individual Shear Walls Using Frame Elements:
Beams and columns contribute to lateral resistance

Structure Model • Columns

• Simple frame element
• End zones set to half beam depth
• End zone reduction factors as appropriate
• Beams
• Simple frame element
• Beam stiffness includes effective flange
width (due to diaphragm slab)
• End zones set to half column depth on
exterior and half of wall width on interior
• End zone reduction factors as appropriate
at exterior connections
• Special properties for grade level beam
to ensure proper backstay behavior
• Wall
• Simple frame element
• Properties should include influence of
boundary element (or wall flange) if
present
Fixed supports at base • Soil Spring
• Spring support or simple frame element
82
All frame elements include axial, bending, and shear deformations
 Modeling Individual Shear Walls Using Frame Elements:
Beams and columns do not contribute to lateral resistance

Structure Model
For cases where the beams and columns do not
contribute to the lateral resistance end releases
may be used.

Note that in some programs (e.g., SAP2000) the

end release is always located at the interior end of the
end zone (the face of the support)

End Releases
Fixed supports at base

83
Frame elements must be included to carry gravity loads that are stabilized by the lateral system
 Coupled Wall Systems Modeled with Frame
Elements

Structure Model

• Modeling approach is similar

to that shown previously

• Properties for coupling

beams and shear
walls should include effect of
flexural and shear cracking
where appropriate

Fixed supports at base

84
For all frame elements include axial, bending, and shear deformations
Cracked Section Properties (ACI 318)

Note: The tables provide properties related

to second order analysis, but are appropriate
for any analysis under ultimate loads.

No information is provided
in ACI 318 regarding cracking in shear for
the walls nor for the coupling beams.
See Scene 4 of Week 4 for guidance.

85
Effective Flange Width (ACI 318)

beff

Effective flange width applies only to walls modeled with

frame elements. Walls modeled with appropriately meshed
shell elements capture the “effective” width automatically.
86
 Example: 12-Story R/C System
Modeled Using Frame Elements

Frame Analyzed

87
Example Modeling Assumptions

◼ All lateral resistance is provided by Frame-Wall systems on Grids 3, 4, 5, and 6.

◼ Each frame can be analyzed separately for preliminary design because the diaphragm is
essentially rigid.
◼ Frames 1, 2, 7, and 8 resist gravity load only
◼ Columns are 28” by 28”
◼ Beams are 20” wide and 30” deep. Slab thickness is 4” at levels 2-R, and 6” thick at grade
◼ Wall thickness is 12 inches with 28” by 28” boundary elements
◼ Concrete is normal weight with 5000 psi compressive strength The same system will be modeled with
shell elements in the next scene.
88
Gross and Effective Section Properties
Gross Properties
Element Area (in2) Av (in2) I (in4)
Column 784 653* 51211
Typical “T” beam (bf=84 in) 856 600** 75666
Grade level “T” beam (bf=84 in) 984 600** 79869
Wall 4112 3427* 32,100,000
*Area/1.2 ** Web area

Analysis Properties (Using ACI 318 Table 6.6.3.1.1(a)

Element Area (in2) Av (in2) I (in4)
Column 784 653x0.5 51211x0.70
Typical “T” beam 856 600x0.5 75666x0.35
Grade level “T” beam 912 600x0.5 79869x0.35
Wall 4112 3427x0.5 32,100,000x0.35
E=4065 ksi G=1625 ksi
89
Development of Mathematical Model
Wind loads
Austin, TX
Exp. C
V=110 mph.
Frame carries
¼ of load on
Interior rigid end=120”
212’ width Reduction factor = 0.0

Exterior rigid end=14”

Reduction factor = 0.5

Soil spring
Assumed as rigid
Z
Y
Note: Beams and columns
Columns and wall assumed
X fixed at base
in this frame are part of
the lateral load resisting system
Total lateral force = 335k
Total Overturning Moment (at grade level)= 363393 in-k 90
Checking Equilibrium on the Computed Response

Values Check

Model showing axial forces

and section cut location

91
Computed Lateral Drift
180

160

140
Height Above Grade, ft

120

100

80
H/500
60

Deflected Shape
40

Cracked Sections
20
As expected, cracking has a significant
Uncracked Sections
0
effect on the computed displacements
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00
Lateral Displacement, inches

92
Computed Shears in Beams
Shear in Beam
◼ For both cracked and uncracked
Level Cracked (k) Uncracked (k)
sections the shears in the stories,
R 10.31 9.38 working from the top, increase until
12 11.76 11.49 mid-height of the structure and then

Increase
11 12.11 11.83 decrease. This is due to the beams
10 12.74 12.42 acting as an outrigger.
9 13.35 13.02
◼ The shears are affected by cracking
8 13.88 13.54
but not to a large extent
7 14.22 13.87
6 14.27 13.95
Decrease

5 13.94 13.65
4 13.11 12.88
3 11.65 11.51
2 9.41 9.35
G 4.85 5.20

93
Moments in Shear Wall ◼ For both cracked and uncracked
sections the moments in the upper
Moment in Wall at Bottom of Story levels of the wall are negative and
Story Cracked (in-k) Uncracked (in-k) are positive in the lower stories
12 -12772 -11516
◼ The change in sign is due to the
11 -21237 -17403
outriggering provided by the frame
10 -27492 -18872
elements
9 -18248 -16119
8 -11808 -9248 ◼ The moments are affected by
7 -1280 1735 cracking but not to a large extent
6 13439 16924
Structure Analyzed
5 32561 36505
4 56401 60757 Maximum outrigger
influence
3 85202 89975
Minimal outrigger
2 120385 125106 influence
1 179520 185224
B 176132 204281

Moment at Base of Wall 94

 Next Scene: Modeling Walls with Shell
Elements

Original node Shell elements Shell elements Common nodes Foundation nodal
layout added meshed constrained Restraints added

95
Fixed supports at base

96
Analysis of a Structure with Different
Diaphragms Models and Comparison of Results
Original node Shell elements Shell elements Common nodes Foundation nodal
layout added meshed constrained Restraints added

98
 Modeling Walls with Shell Elements

◼ Use of “thin” shell elements is recommended

◼ Rectangular elements are generally preferred (instead of triangular elements)

◼ Keep the aspect ratio of elements within the range of 0.5 to 2.0 with values closer to 1.0
preferred.

◼ Use property modifiers to model cracking, where needed. The modifiers would generally be
applied independently to in-plane axial and shear stiffness.

99
 Modeling Individual Shear Walls Thin Shell Elements:
Beams and columns contribute to lateral resistance

Structure Model • Columns

• Simple frame element
• End zones set to half beam
depth
• End zone reduction factors as appropriate

• Beams
• Simple frame element
• End zones set to half column depth on exterior
and half of wall width on interior
• End zone reduction factors as appropriate at
exterior connections
• Special properties for grade level beam
to ensure proper backstay behavior

• Wall
• 4-node thin shell elements
• Boundary elements or flanges must be
included separately using frame or shell
elements as appropriate

• Soil Spring
Fixed supports at base • Spring support or simple frame element

All frame elements include axial, bending, and shear deformations 100
 Creating the Shell Element Mesh

Original node Shell elements Shell elements Common nodes Foundation nodal
layout added meshed constrained restraints added

All programs have a variety

of procedures for developing
the mesh.

The procedure may be virtually

Automatic in some programs.

Note:
Beams and columns
on panel boundaries
(where they exist)
should be meshed
consistently
with the panel

101
Note: Procedure may depend on specific software package used.
 Modeling Link Beams in Coupled Walls

“x” are Element edge constraints

x
x
x

Detail of wall and

link beam meshing
Link beams modeled with shell elements.
Preferred when beam span to depth of 4 or less

Link beams modeled with frame elements

Preferred when beam span to depth > 4
102
 Wall with Boundary Elements
Beams in wall panel
may or may not be present
in actual structure Column area adjusted
For overlap with wall

x x

x x Column meshed and

constrained to shells
x x

Fixed supports at base

103
 Example: 12-Story R/C System
Modeled Using Frame Elements

Frame Analyzed

104
Gross and Effective Section Properties
Gross Properties
Element Area (in2) Av (in2) I (in4)
Exterior Column 784 653* 51211
Wall Boundary Element Column 616 - -
Typical “T” beam (bf=84 in) 856 600** 75666
Grade level “T” beam (bf=84 in) 984 600** 79869
Wall thickness=12” *Area/1.2 ** Web area

Analysis Properties (Using ACI 318 Table 6.6.3.1.1(a)

Element Area (in2) Av (in2) I (in4)
Exterior Column 784 653x0.5 51211x0.7
Wall Boundary Element Column 616x0.35 - -
Typical “T” beam 856 600x0.5 75666x0.35
Grade level “T” beam 912 600x0.5 79869x0.35
Wall thickness=12”, shear modifier=0.5, bending modification=0.35

f’c=5 ksi E=4065 ksi. G=1625 ksi 105

Development of Mathematical Model
Analysis to be run with and
without beams in wall region

Each wall panel

is a 5-by-5 of mesh of
thin shell elements with
t=12 inches

Boundary elements modeled

as columns with area
reduction to account for
wall-column overlap

Beam Rigid end=14”

Column rigid end 15”
Reduction factor = 0.5

Soil spring
Assumed as rigid

Columns and wall assumed

fixed at base

106
Section Cut for Obtaining Overall Model Equilibrium

Applied shear = 335 k (vs 333.6k computed))

Applied OTM = 365393 in-k (vs 360084 in-k applied to model)

Equilibrium satisfied within 1%

107
Lateral Displacement Results for Uncracked System
180 ◼ The lateral displacements for the shell
element model without beams in
160 the wall panel are significantly greater
than the displacements for the frame model
140

◼ The lateral displacements for the shell

Height Above Grade, ft

120
element model with beams in
100 the wall panel are close to, but slightly
greater than the displacements for the
80 frame model

60
◼ The differences in the results for the frame
Shell w/o Beams model and the shell model without beams in
40
Shell with Beams the wall panel is due to excess flexibility at
20 Frame Model
the beam connecting point in the shell
model and excess stiffness at the same
0 point in the frame model
0.00 0.50 1.00 1.50 2.00 2.50
Lateral Displacement, inches
◼ The results shown in the next few slides are
for the shell model with beams in the wall
Note: the lateral displacements for the shell model and the frame model panel. It is assumed that these beams are
are virtually identical when only the wall is modeled present in the actual structure.

108
Computed Lateral Drift
180

160

140
Height Above Grade, ft

120

100

80 Deflected Shape
H/500
60
As expected, cracking has a significant
40 effect on the computed displacements.
Cracked Sections
20 Values are similar to those obtained using
Uncracked Sections frame elements for wall.
0
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00
Lateral Displacement, inches

109
Computed Shears in Beams
Shear in Beam
◼ For both cracked and uncracked
Level Cracked (k) Uncracked (k)
sections the shears in the stories,
R 10.24 9.13 working from the top, increase until
12 11.72 11.28 mid-height of the structure and then

Increase
11 12.02 11.74 decrease. This is due to the beams
10 12.62 12.10 acting as an outrigger.
9 13.19 12.62
◼ The shears are affected by cracking
8 13.67 13.05
but not to a large extent
7 13.95 13.32
6 13.97 13.33 ◼ Values are similar to those obtained
Decrease

5 13.60 13.00 using frame elements for the wall

4 12.74 12.23
3 11.29 10.92
2 9.00 8.86
G 4.55 4.85

110
Section Cut for Obtaining Wall Moment (Uncracked System)

Note that section cut does not include exterior columns

But does include wall boundary columns
111
Moments in Shear Wall ◼ For both cracked and uncracked
sections the moments in the upper
Moment in Wall at Bottom of Story levels of the wall are negative, and
Story Cracked (in-k) Uncracked (in-k) are positive in the lower stories.
12 -12401 -10802
◼ The change in sign is due to the
11 -18707 -17024
outriggering provided by the frame
10 -20833 -18247
elements
9 -18525 -15646
8 -12669 -8434 ◼ The moments are affected by
7 -1683 2730 cracking but not to a large extent
6 13338 18119
Structure Analyzed
5 32531 37165
4 55402 63639 Maximum outrigger
influence
3 84920 92864
Minimal outrigger
2 119179 127582 influence
1 179134 185878
B 161928 204356

Moment at Base of Wall 112

 Summary of Analysis using Frame and Shell
Elements for Shear Wall

◼ In terms of displacements and internal forces the results for analysis using frame elements
and using shell elements for the wall are similar

◼ For the analysis using shell elements there is some uncertainty regarding the analysis where
beam elements are not extended through the wall. Analysis without these elements seemed to
produce considerably more displacement than analysis without the beams extending into the wall.

◼ Where analysis with cracking is performed it is not possible to independently set cracking
factors for axial forces and bending moments where shell elements are used.

113
Next Scene: Overview of 3D Model

114
Original node Shell elements Shell elements Common nodes Foundation nodal
layout added meshed constrained Restraints added

115
Important considerations when rigid
diaphragms are modeled
117
 3D Model of Reinforced Concrete Structure

◼ The same structure modeled using 2D

models will be modeled in 3D in Week 9 of
the course.

◼ Shear walls will be modeled using shell

elements

◼ Diaphragms will be modeled as rigid, and

by use of shell elements

◼ The 5-bay frames in Grid Lines 1 and 8

will be designed as moment resisting
frames (instead of gravity only frames).

118
Three-Dimensional Model with Semirigid Diaphragms

119
Diaphragm Modeling

Roof Diaphragm Typical Floor Diaphragm with Openings

120
Shear Wall Modeling

Shear Wall with 4 by 4 mesh

And boundary elements

121
Next Scene: Summary of Week 5

122
123
Conclusion
Plan view of one floor of 12-story building

125
 Recap of Week 5

◼ Diaphragm Behavior

◼ Significance of the backstay effect in grade and subgrade diaphragms

◼ Influence of mesh density on diaphragm analysis time and results

◼ Modeling diaphragms using finite elements

◼ Shear wall behavior

◼ Modeling shear walls using frame elements

◼ Modeling shear walls using finite elements

◼ Example 12-story structure analyzed using SAP 2000

126
 Week 5 References

1. Sabelli, R., Pottebaum, W, and Dean, B. (2009). “Diaphragms for Seismic Design”, Structural Engineer,
(January Issue).
2. NIST (2011a). Seismic Design of Steel Deck and Concrete-filled Diaphragms, [NIST GCR 11-917-10],
National Institute for Science and Technology, Gaithersburg, MD.
3. NIST (2011b). Seismic Design of Cast-in-Place Concrete Special Structural Walls and Coupling Beams
[NIST GCR 11-917-11REV-1], National Institute for Science and Technology, Gaithersburg, MD.
4. NIST (2016). Seismic Design of Cast-in-Place Concrete Diaphragms, Chords, and Collectors
[NIST GCR 16-917-142], National Institute for Science and Technology, Gaithersburg, MD.
5. Adams, A. (2020). “Study on the Sensitivity of Mesh Size and Properties in the Analysis of Semirigid
Diaphragms”, Proceedings of the 202 SEAOC Convention.

127
 What’s Coming in Week 6?

◼ What is Soil-Structure Interaction?

◼ Effects of Soil-Structure Interaction on Building Structures
◼ Inertial Soil-Structure Interaction
◼ Kinematic Soil Structure Interaction
◼ Case Study of Structure with SSI

128
Plan view of one floor of 12-story building

129