User Guide

PERFORM-3D
Nonlinear Analysis and
Performance Assessment for
3D Structures

ISO PER053111M1 Version 5

Berkeley, California, USA June 2011
 Copyright
Copyright  Computers and Structures, Inc., 2006-2011
All rights reserved.

The CSi Logo is a registered trademark of Computers and Structures,

Inc. PERFORM is a registered trademark of Computers and Structures,
Inc. Windows is a registered trademark of Microsoft Corporation

The computer program PERFORM-3D and all associated

documentation are proprietary and copyrighted products. Worldwide
rights of ownership rest with Computers and Structures, Inc.
Unlicensed use of these programs or reproduction of documentation in
any form, without prior written authorization from Computers and
Structures, Inc., is explicitly prohibited.

No part of this publication may be reproduced or distributed in any

form or by any means, or stored in a database or retrieval system,
without the prior explicit written permission of the publisher.

Further information and copies of this documentation may be obtained

from:

Computers and Structures, Inc.

1995 University Avenue
Berkeley, California 94704 USA

Phone: (510) 649-2200

FAX: (510) 649-2299
e-mail: [email protected] (for general questions)
e-mail: [email protected] (for technical support questions)
web: www.csiberkeley.com

ii
 Disclaimer
CONSIDERABLE TIME, EFFORT AND EXPENSE HAVE GONE
INTO THE DEVELOPMENT, DOCUMENTATION AND TESTING
OF PERFORM-3D. IN USING THE PROGRAM, THE USER
ACCEPTS AND UNDERSTANDS THAT NO WARRANTY IS
EXPRESSED OR IMPLIED BY THE DEVELOPERS OR THE
DISTRIBUTORS ON THE ACCURACY OR THE RELIABILITY OF
THE PROGRAM. THE USER MUST EXPLICITLY UNDERSTAND
THE ASSUMPTIONS OF THE PROGRAM AND MUST
INDEPENDENTLY VERIFY THE RESULTS.
 Acknowledgments
Special recognition is extended to Dr. Graham H. Powell, Professor
Emeritus, University of California at Berkeley, who conceived and
developed the PERFORM series of programs. Professor Powell's
research, expertise and originality have resulted in many contributions
to our profession, particularly in the modeling of structures for practical
nonlinear analysis.

We extend thanks to the members of our profession for all of their

valuable feedback that has contributed to enhancement of the products
produced by CSI.

PERFORM-3D uses a node renumbering scheme developed by

Professor Marc Hoit (see Hoit, M.I., Stoker, D. and Consolazio, G.
"Neural Networks for Equation Renumbering" Journal of Computers
and Structures, Vol. 52, No. 5, 1994, pp. 1011-1021). We thank
Professor Hoit for allowing us to use his computer code.

ii
PERFORM-3D User Guide
Version 5

Contents

1 Overview 1-1
1.1 Screen Layout 1-1
1.2 Structures 1-2
1.3 Phases 1-3
1.4 Modeling Phase 1-3
1.5 Analysis Phase 1-5
1.6 View Direction and Perspective 1-8
1.7 Reports 1-9
1.7.1 Printed Reports 1-9
1.7.2 Saving Results to a File 1-9
1.7.3 ECHO File 1-9
1.7.4 Direct Access to Results Files 1-10
1.8 Reporting Errors 1-10
1.9 Some Things to Keep in Mind 1-10

2 Structures, Folders and Files 2-1

2.1 List of Folders 2-1
2.2 Saving and Opening a Structure 2-2
2.2.1 Saving a Structure 2-2
2.2.2 Opening a Structure 2-4
2.3 Location of ECHO file 2-5
2.4 Moving or Copying a Project or Structure 2-6
2.5 Other Folders 2-7
2.5.1 Records Folder 2-7
2.5.2 RecordsF Folder 2-7
2.5.3 Spectra Folder 2-8
2.5.4 User Folder 2-9
2.6 Do Folder Locations Matter? 2-10

i
3 Nodes 3-1
3.1 Methods for Specifying Nodes 3-1
3.2 Zooming 3-3
3.3 Closely Spaced Nodes 3-3
3.4 Supports 3-3
3.5 Masses 3-4
3.6 Slaving Constraints 3-5
3.6.1 General 3-5
3.6.2 Rigid Floor Constraint 3-5
3.6.3 Full Rigid Link Constraint 3-6
3.6.4 Eccentric Connection Constraint 3-6
3.6.5 Equal Displacement Constraint 3-8

4 Frames 4-1
4.1 Plotting of Nodes and Elements 4-1
4.2 Defining a Frame 4-2
4.3 Duplicating a Frame 4-3

5 Component Properties 5-1

5.1 Elements and Components 5-1
5.2 Procedure for Specifying Components 5-2
5.2.1 Scope 5-2
5.2.2 Component Types and Names 5-3
5.2.3 Materials 5-4
5.2.4 Beam and Column Cross Sections 5-4
5.2.5 Fiber Cross Sections for Walls 5-5
5.2.6 Fiber Cross Sections for Frame Elements 5-7
5.2.7 Basic Structural Components 5-7
5.2.8 Plot Hysteresis Loops 5-9
5.2.9 Strength Sections 5-10
5.2.10 Frame Compound Components 5-10
5.2.11 Shear Wall Compound Components 5-11
5.2.12 General Wall Compound Components 5-12
5.2.13 Viscous Bar and BRB Compound Components 5-12
5.3 Managing Component Properties 5-13
5.3.1 General 5-13

ii
 5.3.2 Filter 5-13
5.3.3 Purge 5-13
5.3.4 Use Cross Section Strengths 5-14
5.3.5 Save As 5-14
5.3.6 Copy-Paste 5-14
5.4 Printed Reports 5-15
5.5 F-D Relationship 5-15
5.5.1 Actions and Deformations 5-15
5.5.2 Actual Components 5-15
5.5.3 PERFORM F-D Relationship 5-16
5.5.4 E-P-P and Bilinear Relationships 5-17
5.5.5 Additional Parallel Stiffness 5-17
5.6 Strength Loss 5-18
5.6.1 General 5-18
5.6.2 Total Strength Loss at X Point 5-19
5.6.3 Strength Loss Interaction 5-20
5.7 A Warning on Strength Loss 5-20
5.7.1 General 5-20
5.7.2 Components in Parallel 5-21
5.7.3 Components in Series 5-21
5.7.4 Effect on Analysis Method 5-22
5.7.5 Hinge Strength Loss in a Compound Component 5-23
5.8 Column Ductile Limit Depends on Axial Force 5-24
5.8.1 Effect to be Modeled 5-24
5.8.2 FEMA 356 Requirements 5-25
5.8.3 PERFORM Implementation of Variable
Ductile Limit 5-27
5.8.4 Procedure 5-28
5.9 Deformation and Strength Capacities 5-29
5.9.1 General 5-29
5.9.2 Demand and Capacity 5-29
5.9.3 Deformation Capacities 5-30
5.9.4 Drift Capacities 5-30
5.9.5 Strength Capacities 5-31
5.9.6 Overview of Limit States and Usage Ratios 5-31
5.9.7 Note on Maximum Demand-Capacity Ratio 5-32
5.10 Beam or Column Shear Strength Depends on
Hinge Rotation 5-33

iii
 5.10.1 General 5-33
5.10.2 PERFORM Implementation 5-33
5.10.3 Procedure for Strength Section Properties 5-34
5.10.4 Procedure for Frame Compound Component 5-35
5.11 Cyclic Degradation 5-36
5.11.1 Backbone Relationship 5-36
5.11.2 Hysteresis Loops 5-37
5.11.3 Dynamic Analysis 5-38
5.11.4 Procedure for Degradation Factors 5-38
5.11.5 Push-Over Analysis 5-39
5.12 Control of Hysteresis Loop Shape 5-40
5.12.1 Alternative Loops 5-40
5.12.2 E-P-P Relationship 5-41
5.12.3 Trilinear Relationship After U Point 5-42
5.12.4 Procedure for Unloading Stiffness Factor 5-43
5.13 Cross Sections 5-43
5.13.1 The Role of Cross Sections 5-43
5.13.2 Beam and Column Sections 5-44
5.13.3 Fiber Cross Sections for Beams and Columns 5-45
5.13.4 Fiber Cross Sections for Walls 5-45
5.13.5 Other Sections 5-46
5.14 Using Cross Section Dimensions 5-46
5.14.1 General 5-46
5.14.2 Length of Uniform X-Section Segment 5-46
5.14.3 Tributary Length for a Curvature Hinge 5-47
5.15 "Auto" Components 5-48
5.15.1 Auto End Zones for Beams and Columns 5-48
5.15.2 Auto Panel Zones in Steel Frames 5-48
5.15.3 Panel Zone Limit States 5-50
5.16 Upper and Lower Bounds 5-50
5.16.1 Overview 5-50
5.16.2 Procedure for Specifying Upper/Lower Bounds 5-51
5.16.3 Procedure for Using Upper/Lower Bounds 5-52

6 Elements 6-1
6.1 Element Types 6-1
6.1.1 General 6-1
6.1.2 Simple Bar Elements 6-1

iv
 6.1.3 Beam Elements 6-1
6.1.4 Column Elements 6-2
6.1.5 Brace/Other Frame Elements 6-3
6.1.6 Shear Wall Elements 6-4
6.1.7 General Wall Elements 6-5
6.1.8 Infill Panel Elements 6-5
6.1.9 Connection Panel Zone Elements 6-5
6.1.10 BRB Elements 6-5
6.1.11 Viscous Bar Elements 6-6
6.1.12 Seismic Isolator Elements 6-6
6.1.13 Slab/Shell Elements 6-6
6.1.14 Support Spring Elements 6-7
6.1.15 Deformation Gage Elements 6-7
6.2 Specifying Elements 6-8
6.2.1 Element Groups 6-8
6.2.2 Steps 6-9
6.2.3 Start a New Element Group 6-9
6.2.4 Add Elements 6-10
6.2.5 Parallel Elements 6-10
6.2.6 Delete Elements 6-10
6.3 Element Orientations 6-11
6.4 Element Properties 6-12
6.5 Moving Elements Between Groups 6-13
6.6 P-, P- and Large Displacement Effects 6-13
6.6.1 General 6-13
6.6.2 P- vs. True Large Displacements 6-14
6.6.3 P- Effect 6-17
6.6.4 Do You Need To Consider P- Effects? 6-19
6.6.5 Axial Shortening Due to Bending 6-19
6.6.6 PERFORM Options 6-20

7 Load Patterns 7-1

7.1 Pattern Types and Restrictions 7-1
7.2 Nodal Load Patterns 7-2
7.3 Element Load Patterns 7-2
7.4 Self Weight Load Patterns 7-4

v
8 Drifts and Deflections 8-1
8.1 Drifts 8-1
8.1.1 Simple Drift 8-1
8.1.2 Distortion Drift 8-1
8.1.3 Procedure for Specifying a Drift 8-2
8.1.4 Reference Drift 8-2
8.2 Deflections 8-3

9 Structure Sections 9-1

9.1 Concept 9-1
9.2 Specifying Element Cuts 9-3
9.3 Shear Strength for a Wall Section 9-4
9.4 Section Groups and Strength Limit States 9-5
9.5 Moment, Shear and D/C Ratio Diagrams 9-6
9.5.1 Procedure 9-6
9.5.2 Save D/C Values to a File 9-6
9.5.3 Signs and Sign Reversal 9-7

10 Limit States and Usage Ratios 10-1

10.1 Limit State Types 10-1
10.2 Deformation Limit States 10-2
10.3 Strength Limit States 10-4
10.4 Drift Limit States 10-4
10.5 Deflection Limit States 10-5
10.6 Structure Section Limit States 10-5
10.7 Limit State Groups 10-5

11 Inactive Elements 11-1

11.1 Inactive Elements 11-1
11.2 Procedure 11-1

12 Gravity Load Cases 12-1

12.1 Gravity Loads 12-1

vi
 12.2 Procedure 12-2
12.2.1 Steps 12-2
12.2.2 Number of Load Steps 12-2
12.2.3 Maximum Number of Events 12-3
12.2.4 Initial Step to First Event 12-3
12.2.5 Analysis Limit State 12-4

13 Static Push-Over Load Cases 13-1

13.1 Push-Over Loads 13-1
13.1.1 Load Distribution and Magnitude 13-1
13.1.2 Choice of Load Distribution 13-1
13.1.3 Load Distribution Based on Mode Shapes 13-2
13.1.4 Linear Analysis Option 13-3
13.2 Procedure 13-3
13.2.1 Steps 13-3
13.2.2 Reference Drift 13-5
13.2.3 Number of Steps 13-5
13.2.4 Maximum Number of Events 13-6
13.2.5 Initial Step to First Event 13-7
13.2.6 Analysis Limit State 13-7
13.3 Loads Based on Mode Shapes 13-8
13.3.1 Symmetrical Structure 13-8
13.3.2 Unsymmetrical Structure or Diagonal Push-Over 13-9
13.3.3 Other Key Points 13-10

14 Dynamic Earthquake Load Cases 14-1

14.1 Earthquake Loads 14-1
14.2 Procedure 14-1
14.2.1 Steps 14-1
14.2.2 Total Time 14-2
14.2.3 Time Step 14-3
14.2.4 Maximum Number of Events 14-4
14.2.5 Save Interval 14-4
14.2.6 Analysis Limit State 14-4

15 Earthquake Records 15-1

15.1 Organization 15-1
15.1.1 Overall Steps 15-1
15.1.2 Organization of Earthquake Records 15-1

vii
 15.1.3 Copying Earthquake Records 15-3
15.2 Setting Up a New Earthquake Record 15-3
15.2.1 Text File Format 15-3
15.2.2 Adding A New Record 15-4
15.3 Reviewing Existing Records 15-5

16 Dynamic Force Load Cases and Analysis 16-1

16.1 Purpose and Overall Procedure 16-1
16.1.1 Purpose 16-1
16.1.2 Overall Procedure 16-1
16.2 Organization of Dynamic Force Records 16-2
16.2.1 RecordsF Folder 16-2
16.2.2 Steps for Setting Up Dynamic Force Records 16-3
16.2.3 Copying Dynamic Force Records 16-3
16.3 Setting Up a New Dynamic Force Record 16-4
16.3.1 Text File Format 16-4
16.3.2 Adding a New Record 16-4
16.4 Reviewing Existing Records 16-5
16.5 Dynamic Force Load Cases 16-6
16.5.1 Load Cases 16-6
16.6 Multi-Support Earthquake Analysis 16-7
16.6.1 Purpose 16-7
16.6.2 Analysis Method 16-7
16.6.3 Imposed Dynamic Support Displacements 16-8
16.6.4 Ground Displacements 16-9
16.6.5 Ground Displacement Records 16-10
16.6.6 Alpha-M Damping 16-12
16.6.7 Results Interpretation 16-12

17 Analysis Series 17-1

17.1 Analysis Series Concept 17-1
17.1.1 Concept 17-1
17.1.2 Procedure 17-1
17.2 Basic Parameters and Masses 17-3
17.2.1 Event Overshoot Factor 17-4
17.3 Damping 17-6

viii
 17.3.1 General 17-6
17.3.2 Procedure for Modal Damping 17-7
17.3.3 Procedure for Rayleigh Damping 17-7
17.4 Procedure for Using Upper/Lower Bounds 17-7
17.5 Quick'n'Dirty 17-10

18 "Elastic" Viscous Damping 18-1

18.1 Modal Damping 18-1
18.1.1 Damping Matrix 18-1
18.1.2 Physical Interpretation 18-2
18.2 Rayleigh Damping 18-3
18.2.1 Rayleigh Damping Model 18-3
18.2.2 Implied Modal Damping Ratio for
Linear Analysis 18-4
18.2.3 Implied Damping for Nonlinear Analysis 18-6
18.2.4 PERFORM Implementation 18-7
18.2.5 Procedure for Specifying Rayleigh Damping 18-8
18.2.6 Special Assumption for Concrete Fibers 18-8
18.2.7 Alternative Procedure for Base Isolation 18-9
18.3 A Case Where Modal vs. Rayleigh Damping May Be
Substantially Different : Coupled Shear Walls 18-11
18.4 Combined Modal and Rayleigh Damping 18-12

19 General Load Sequence 19-1

1.1 Standard and General Load Sequences 19-1
1.2 Points to Note for General Sequence 19-2
19.1.1 General Push-Over and Target
Displacement Plots 19-2
19.1.2 Unload Push-Over Load Case Type 19-2
19.1.3 Analysis Limit State 19-3
19.1.4 Controlled Drifts 19-3
19.1.5 Removing Gravity Loads 19-4
19.1.6 Return To Static State After Dynamic Analysis 19-4
19.1.7 Graphs That Cover Several Analyses 19-5
19.1.8 Inactive Elements 19-5

ix
20 Running Analyses 20-1
20.1 Steps for Running Analyses 20-1
20.2 Adding and Deleting Analyses 20-2

21 Mode Shapes 21-1

21.1 Plotting Mode Shapes 21-1
21.2 Using Mode Shapes for Push-Over Analysis 21-2
21.3 Vertical Masses 21-2

22 Response Spectrum Analysis 22-1

22.1 Response Spectrum Load Cases 22-1
22.1.1 Spectra 22-1
22.1.2 Response Spectrum Load Cases 22-2
22.2 Response Spectrum Analyses 22-3
22.2.1 Run Analysis 22-3
22.2.2 Analysis Results 22-4
22.2.3 Lateral Load Patterns from Analysis Results 22-4

23 Energy Balance 23-1

23.1 Types of Energy 23-1
23.1.1 Dynamic Analysis 23-1
23.1.2 Static Analysis 23-2
23.1.3 Calculation of Inelastic and Strain Energies 23-2
23.2 Energy Plots 23-5
23.2.1 Complete Structure 23-5
23.2.2 Inelastic Energies in Element Groups 23-5
23.2.3 -K Energies in Element Groups 23-5
23.2.4 Energy Error 23-6
23.3 Approximate Damping Ratio 23-6
23.3.1 Overview 23-6
23.3.2 Theory 23-6
23.3.3 Procedure 23-8

24 Deflected Shape Plots 24-1

24.1 Deflected Shape 24-1
24.2 Procedure 24-2

x
 24.2.1 Deflected Shape Only 24-2
24.2.2 Deflected Shape with Usage Ratio Coloring 24-3
24.2.3 Some Points on Components vs. Entire Element 24-4
24.2.4 Other Aspects 24-4
24.2.5 Save Usage Ratios to a File 24-5

25 Time History Plots 25-1

25.1 Plot History for a Single Node 25-1
25.1.1 Options 25-1
25.1.2 Procedure 25-2
25.1.3 Save Plotted History to a File 25-2
25.2 Save Histories for Multiple Nodes 25-2
25.3 Plot History for a Single Element 25-3
25.3.1 Procedure 25-3
25.3.2 Save Plotted History to a File 25-4
25.4 Save Histories for Multiple Elements 25-5
25.5 Drift or Deflection Histories 25-6
25.6 Structure Section Force Histories 25-6
25.7 Plot Section Results for Multiple Load Cases 25-7
25.7.1 Purpose 25-7
25.7.2 Procedure 25-7

26 Hysteresis Loop Plots 26-1

26.1 Purpose 26-1
26.2 Procedure 26-1
26.3 Cut-Off Corners 26-3
26.4 Save Loops to a File 26-3

27 Moment and Shear Diagrams 27-1

27.1 Detailed Results for a Single Element 27-1
27.1.1 Element Selection 27-1
27.1.2 Element Loads Option 27-2
27.1.3 Moment and Shear Envelopes Option 27-2
27.1.4 Moment and Shear Histories Option 27-2
27.1.5 Deflected Shapes Option 27-3

xi
 27.2 Moment and Shear Diagrams for a Line of Elements 27-4
27.3 Moment and Shear Diagrams Based on Structure
Sections 27-5

28 General Push-Over Plots 28-1

28.1 Overview 28-1
28.1.1 Available Methods 28-1
28.1.2 Differences Among Methods 28-1
28.1.3 Some Key Points 28-6
28.1.4 Main Steps 28-8
28.2 Procedure for Defining Response Spectra 28-10
28.2.1 Spectrum Types 28-10
28.2.2 Add or Edit a Spectrum 28-10
28.2.3 Copy Spectra from a Different Structure 28-11
28.3 Procedure for Capacity and Demand Curves 28-11
28.3.1 Choose Task 28-11
28.3.2 Plot Capacity Curve 28-11
28.3.3 Sa vs. Sd Option 28-12
28.3.4 Spectral Acceleration Capacity 28-12
28.3.5 Define Trial Points 28-14
28.3.6 Period Calculation 28-15
28.3.7 Choose Push-Over Method and Response
Spectrum 28-20
28.3.8 Plot Demand Curve 28-22
28.3.9 Capacity Curve Scaling for Sensitivity Studies 28-23
28.4 Performance Evaluation 28-23
28.5 Consistency 28-24
28.5.1 General 28-24
28.5.2 Calculation of Spectral Displacement 28-24
28.5.3 Suggested Methods 28-26
28.5.4 A Reason for Using Sa vs. Sd Option 28-26
28.6 Energy Dissipation and Damping 28-27
28.6.1 General 28-27
28.6.2 Stiffness and Degradation and Energy Ratio 28-27
28.6.3 Equivalent Damping Ratio 28-28
28.6.4 Degradation in Push-Over Methods 28-29
28.6.5 PERFORM Method for Energy Ratio 28-30
28.6.6 Extra Damping from Fluid Dampers 28-30

xii
 28.6.7 Strength Loss and Degradation 28-32
28.7 Theory for Relationship Between Sa and H 28-32
28.7.1 General 28-32
28.7.2 Special Case With Modal Load and
Displaced Shape 28-33
28.7.3 General Case Where Load Matches
Displaced Shape 28-36
28.7.4 Case Where Load Does Not Match
Displaced Shape 28-36

29 Target Displacement Plot 29-1

29.1 Target Displacement Method 29-1
29.1.1 Concept 29-1
29.1.2 Target Drift 29-2
29.1.3 Spectral Drift 29-3
29.1.4 Factor C0 29-3
29.1.5 Factor C1 29-3
29.1.6 Factor C2 29-4
29.1.7 Factor C3 29-4
29.2 PERFORM Implementation 29-4

30 Usage Ratio Graphs 30-1

30.1 Usage Ratios 30-1
30.2 Procedure 30-2

31 Load Case Combinations and Envelopes. 31-1

31.1 Overview 31-1
31.1.1 Main Steps 31-1
31.1.2 Load Case Combinations 31-2
31.1.3 Combination Methods 31-3
31.2 Procedure 31-4
31.2.1 Load Combinations 31-4
31.2.2 Usage Ratios for Structure 31-5
31.2.3 Usage Ratios for Elements 31-6

32 Importing/Exporting Structure Data 32-1

32.1 Importing Structure Data 32-1
32.1.1 Text Files 32-1

xiii
 32.1.2 File for Elements (with Optional Nodes) 32-2
32.1.3 File for Nodes 32-3
32.1.4 File for Nodal Masses 32-4
32.1.5 File for Nodal Loads 32-5
32.2 Note on Element Import 32-6
32.3 Import and Export Procedures 32-8
32.3.1 Import 32-8
32.3.2 Export 32-9
32.4 Component Properties Export/Import 32-10
32.4.1 General 32-10
32.4.2 Export/Import Steps 32-10

33 Procedure For Reporting Errors 33-1

33.1 Reporting Errors 33-1
33.2 Procedure For Sending Files 33-1
33.2.1 Structure Data 33-1
33.2.2 Send As An Email Attachment 33-2
33.2.3 Upload to FTP Site 33-2

xiv
 Chapter 1. Overview

1 Overview

PERFORM-3D is a typical Windows program. You can move

around the program by clicking on buttons and folder tabs; select
items by clicking on a drawing of the structure or by choosing from
drop-down lists; and enter data by typing into text boxes. If you
hover the mouse pointer over a button, a tool tip that describes its
purpose will pop up. Some text boxes also have tool tips. In this
version of PERFORM-3D there is no Help button to call up context-
sensitive help.

This chapter provides an overview of the program and covers some

things you should know before you begin.

1.1 Screen Layout

Figure 1.1 shows the screen layout for most tasks. The minimum
required screen resolution is 1024 x768 pixels. Figure 1.1 is the layout
at this resolution. If the resolution is finer the size of the graphics panel
is increased.

Menu bar
Tool bar

Data panel for

data entry.
Usually consists Graphics panel for
of a number of graphic display of
tabbed pages. data or results.

Tools to control
view direction

Figure 1.1 Screen Layout For Most Tasks

PERFORM-3D User Guide 1-1

Chapter 1. Overview

Figure 1.2 shows the screen layout for specifying the properties of
linear and nonlinear components. This screen has diagrams to explain
the required properties.

Menu bar
Tool bar

Graphics panel to show

required data and display
Data panel for properties for checking.
component type
and options.

Consists of a Data panel for

number of component properties.
tabbed pages.
Consists of a number
of tabbed pages.

Figure 1.2 Screen Layout For Component Properties

There are a number of other screens, most of them self-explanatory.

The best way to learn the program is to work through a tutorial, and
also to experiment with the program features.

1.2 Structures

You can set up any number of structures (analysis models). Each

structure must have a unique name. For an explanation of how to
manage structures and data files, see Chapter 2, Structures, Folders
and Files.

When you start PERFORM-3D you have the option of starting a new
structure or opening an existing structure. After you make this choice

1-2 PERFORM-3D User Guide

 Chapter 1. Overview

you can press buttons on the toolbar to move around the program. If
you prefer, you can use the menus rather than the toolbar buttons.

The toolbar buttons are as follows.

Start a new structure from scratch.

Open an existing structure.

Save any changes you have made to the current structure.

Save the current structure as a new structure.

Delete the current structure.

If you are a beginning user, we suggest that you start by opening one of
the example structures and playing with it. If you make changes to an
example, be careful not to save these changes. If you want to make
changes to an example, first save it as a new structure, then make the
changes.

1.3 Phases

There are two phases, namely Modeling and Analysis. The toolbar
buttons for phases are as follows.

Modeling phase (build the structure).

Analysis phase (calculate and process the analysis results).

Each phase has a number of tasks, as described in the following

sections.

1.4 Modeling Phase

The toolbar buttons for tasks in the modeling phase are as follows.

PERFORM-3D User Guide 1-3

Chapter 1. Overview

Umbrella task. Specify overall information for the structure.

After you press the toolbar button the required information is
self-explanatory.

Nodes task. Specify node data, including coordinates, support

conditions, slaving constraints (e.g. rigid floor diaphragms)
and structure masses. This is a relatively simple task, and the
required information is largely self-explanatory. For some key
points see Chapter 3, Nodes.

Component Properties task. Specify properties for nonlinear

and linear components. This is the most critical and complex
task. For an explanation see Chapter 5, Component
Properties.

Elements task. Specify element data, including element types,

locations and properties. For an explanation see Chapter 6,
Elements.

Frames task. The plotted geometry for a 3D structure can be

complex. You can (and usually must) simplify the plot by
showing only selected parts of the structure. A "frame" is any
part of the complete structure. It can be a plane frame, a floor,
or any other part of the structure that you would like to plot
separately from the structure as a whole. You can quickly set
up, modify and delete frames, and then toggle between a full
structure view and a view showing only a single frame. For an
explanation see Chapter 4, Frames.

Load Patterns task. Specify load patterns for node, element

and/or self weight loads. For this version of the program there
is only a limited choice for element loads. For an explanation
see Chapter 7, Load Patterns. In the analysis phase you can
combine load patterns to form load cases. See Chapter 12,
Gravity Load Cases, and Chapter 13, Static Push-Over Load
Cases.

Import Structure Data task. Import nodes, masses, elements

and/or loads from a text file. If you have built an analysis
model in another computer program, and if you can set up the

1-4 PERFORM-3D User Guide

 Chapter 1. Overview

data for this model in a comma-delimited text file, you can

use this task to import much of the structure into PERFORM-
3D. For an explanation see Chapter32, Importing/Exporting
Structure Data.

Drifts and Deflections task. Specify drifts and deflections.

Horizontal drifts are valuable measures of structure
deformation, and you will typically specify a number of them.
For long span structures you may also use vertical deflections
as deformation measures. For an explanation see Chapter 8,
Drifts and Deflections.

Structure Sections task. Specify structure sections. If the

structure has a number of lateral load resisting systems, you
may want to know how the total load is distributed among
them. You can do this by specifying sections through parts of
the structure and examining the forces on these sections. For
an explanation see Chapter 9, Structure Sections.

Limit States task. Specify limit states. This is an extremely

important task. If you make effective use of limit states you
can greatly simplify the decision making process. For an
explanation see Chapter 10, Limit States and Usage Ratios.

Inactive Elements task. Specify elements that are inactive for

gravity loads. For an explanation see Chapter 11, Inactive
Elements.

You do not have to perform these tasks in any particular sequence. For
example, you can specify some nodes, then some elements, then more
nodes, then some component properties, etc.

1.5 Analysis Phase

The tasks in the analysis phase can be divided into Structural Analysis
tasks, Behavior Assessment tasks and Demand-Capacity tasks. The
structural analysis tasks are for defining load cases and running
structural analyses. The behavior assessment tasks allow you to
examine and check the behavior of the analysis model. The demand-

PERFORM-3D User Guide 1-5

Chapter 1. Overview

capacity tasks allow you to calculate demand-capacity ratios, and hence

make decisions about the performance of the structure.

The toolbar buttons for the structural analysis tasks are as follows.

Load Cases task. Specify gravity, push-over, earthquake and

other load cases. For explanations see Chapter 12, Gravity
Load Cases, Chapter 13, Push-Over Load Cases, Chapter 14,
Earthquake Load Cases and Chapter 16, Dynamic Force
Load Cases and Analyses.

Run Analyses task. Run static and dynamic analyses. For an

explanation see Chapter 17, Analysis Series, Chapter 19,
General Load Sequence and Chapter 20, Running Structural
Analyses.

The toolbar buttons for the behavior assessment tasks are as follows.

Mode Properties task. Examine the calculated mode periods

and mode shapes, and examine the results of response
spectrum analyses.. For an explanation see Chapter21, Mode
Shapes. For response spectrum analysis see Chapter 22,
Response Spectrum Analysis.

Energy Balance task. Plot graphs that show how much energy
of different types is dissipated by the structure. For an
explanation see Chapter23, Energy Balance.

Limit State Groups task. Specify limit state groups. In the

modeling phase you may have specified a large number of
limit states. This task allows you to organize the limit states
into related groups, and hence simplify the decision-making
process. For an explanation see Chapter10, Limit States and
Usage Ratios.

Deflected Shape task. Plot deflected shapes, either static or

animated. As an option, color the elements based on their
usage ratios. This can be extremely useful for assessing the
behavior of a structure. For an explanation see Chapter 24,
Deflected Shape Plots.

1-6 PERFORM-3D User Guide

 Chapter 1. Overview

Time History task. For dynamic analyses, plot time history

graphs of a variety of response quantities, including nodal
displacements, velocities and accelerations, element responses
of a variety of types, drifts, and cross section forces. For static
push-over analyses, plot graphs of similar response quantities
against structure drift. For an explanation see Chapter 25,
Time History Plots.

Hysteresis Loop task. For dynamic analyses, plot hysteresis

loops for inelastic components (e.g. plot loops of bending
moment at a plastic hinge versus hinge rotation). For an
explanation see Chapter 26, Hysteresis Loop Plots.

Moment and Shear Diagrams task. For beams, columns and

walls, plot bending moment and shear force diagrams. For
beams and columns draw detailed deflected shapes. For an
explanation see Chapter 27, Moment and Shear Diagrams.

The toolbar buttons for the demand-capacity tasks are as follows.

General Push-Over task. Plot capacity and demand curves,

and hence assess the structural performance. This task covers
a number of different push-over methods. For an explanation
see Chapter 28, General Push-Over Plot.

Target Displacement task. Plot the capacity curve and

calculate the target displacement for push-over analysis using
the FEMA 356 Coefficient (Target Displacement) method.
This task has been superseded by the preceding General Push-
Over task, but has been kept for historical reasons. For an
explanation see Chapter 29, Target Displacement Plot.

Usage Ratio task. For any analysis and limit states, plot graphs
of usage ratio versus time, drift or load factor. For an
explanation see Chapter 30, Usage Ratio Plots.

Combinations and Envelopes task. Combine the results from a

number of analyses, and calculate usage ratios based on
maximum or mean values, using a variety of combination
methods. If desired, color the elements based on their usage
ratios. See Chapter 31, Load Case Combinations and
Envelopes.

PERFORM-3D User Guide 1-7

Chapter 1. Overview

1.6 View Direction and Perspective

The tools for controlling the view direction are at the bottom left of the
screen. These tools are shown in Figure 1.3.

Figure 1.3 Tools for View Direction and Perspective

You can view the structure from any direction in plan, and at any
vertical angle from directly downwards to directly upwards. To change
the view direction, either click in the diagrams to set the directions of
the axes, or enter angles in the text boxes. To change the viewing
distance for the perspective projection, choose one of the standard
distances or enter a distance in the text box. Then press OK to change
the view, or Cancel to retain the current view.

The "Basic" view is the default. If you have changed the view
direction, press the "Basic" button to return to this view. Press the
"Plan", "H1" or "H2" button to get a plan or elevation view, with no
perspective (the view distance is infinity). You can add perspective to
plan and elevation views by choosing a view distance.

1-8 PERFORM-3D User Guide

 Chapter 1. Overview

1.7 Reports

1.7.1 Printed Reports

You can obtain printed reports in both the Modeling and Analysis
phases. For each task there is a button with a printer symbol. For most
tasks in the Modeling phase this button is near the right end of the
toolbar. For a few tasks it is in body of the form. To get a hard copy of
the data in the current form, press the printer button. If the button is
disabled you can not print the data. Either the task is not yet complete,
or there is no print option for the current data.

You can change the paper size and orientation by choosing Printer set-
up from the File menu. Some reports are printed only in portrait
orientation.
1.7.2 Saving Results to a File
In the Analysis phase you can save results to a text file for several of
the tasks. For example, you can save time history results in the Time
History task, and you can save usage ratios in the Usage Ratio task.
You can then process the results using a spreadsheet program.

There is a file button near the right end of the toolbar. If this button is
enabled (it will be green) you can save the current results to a file. You
will be asked to give a name and other information for the file. The file
will have header lines that explain its contents.
1.7.3 ECHO File
If you want a printed report that covers the entire Modeling phase (i.e.,
a report that describes the entire analysis model), you must go through
each of the modeling tasks, print one or more reports for each task, and
organize the separate parts into a single report. This gives you a lot of
flexibility in setting up a report, but it can take a fair amount of work.
As an alternative you can use the PERFORM-3D ECHO file.

As PERFORM-3D processes the structure data, it writes a record of the

data to the ECHO file. This is a text file. It is not as easy to interpret as
the printed reports, but it provides a full description of the structure and
loading. If you wish you can examine this file (e.g., using the Windows
WordPad utility program) and also print it out.

PERFORM-3D User Guide 1-9

Chapter 1. Overview

For the location of the ECHO file, see Chapter 2, Structures, Folders
and Files.

1.7.4 Direct Access to Results Files

PERFORM-3D provides several tools for processing analysis results,
and also allows you to save most results to text files for processing by a
spreadsheet program. However, you may have special processing needs
that are not included in these processing options.

If this is the case, it is possible for you to access the PERFORM-3D

results files directly, assuming that you are able to write your own
computer code to read the files and process the results. The results files
are mostly binary direct access files (simple byte streams). In order to
use these files you must know their record lengths and the data
contained in each record. If you need this information, please contact
us.

1.8 Reporting Errors

We do our best to keep PERFORM-3D reliable and free of errors, but

we are not always successful. If you find an error in the program,
please report it to us. For the procedure see Chapter 33, Reporting
Errors.

1.9 Some Things to Keep in Mind

PERFORM-3D provides you with tools to build nonlinear analysis

models, analyze them, and process the results to help you make design
decisions. You will find that you do not need to know much about the
details of nonlinear analysis, because PERFORM-3D largely automates
the analysis process. However, nonlinear behavior is more complex
than linear behavior, and deformation-based design is different from
strength-based design. To use the program effectively you must
understand nonlinear structural behavior and performance-based
design.

1-10 PERFORM-3D User Guide

 Chapter 1. Overview

Keep in mind that the goal of structural analysis is not to get an

accurate simulation of a structure's behavior, but to get information to
help in making design decisions. Once an analysis model has been set
up, it is easy to be seduced into believing that the model is an exact
representation of the real structure. It probably is not, and it does not
need to be. The model must be sufficiently accurate to provide useful
design information, but the analysis results are almost certainly
approximate, no matter how sophisticated the analysis model.

Also keep in mind that analysis is not an end in itself, but merely a tool
to help in decision making. It can be entertaining to plot deflected
shapes, time histories of node and element response, and hysteresis
loops for nonlinear components. This is not a trivial activity, since such
plots are certainly useful for checking the behavior of the analysis
model. Ultimately, however, decision making must be based on limit
states and demand-capacity ratios.

PERFORM-3D User Guide 1-11

 Chapter 2. Structures, Folders and Files

2 Structures, Folders and Files

A PERFORM-3D analysis model is a "structure". For each

structure PERFORM-3D sets up a number of files. Since you may
create and analyze a large number of structures, you should know a
few things about how the files are organized. This chapter gives an
overview. It is assumed that you are familiar with Windows folders
(directories) and files.

2.1 List of Folders

PERFORM-3D makes use of a number of “standard” folders, for

storing the executable program, the user manuals, ground acceleration
records, and other things. For Version 4, all default folders were
located in the Microsoft Windows ProgramFiles folder. Starting with
Version 5, several of the standard folders have been relocated because
Microsoft has added security features to Windows, which generally
prevent the original folder locations from being used. See the Important
Note in Section 2.2.1 Saving a Structure with respect to the relocation
of existing files when changing from Version 4 to Version 5 or later.

The standard folders include the following.

(1) Program folder. This folder contains the PERFORM-3D executable

program, plus several files that are used by the program.
PERFORM-3D installs, by default, in the following Windows
folder (directory) :

C:\ProgramFiles\Computers and
Structures\PERFORM\PERFORM-3D\Program

You may have chosen some other folder when you installed the
program. Call this the Program Folder.

(2) Manuals folder. This folder contains the PERFORM-3D user

documentation. This folder is stored as a subfolder in the
PERFORM-3D Folder.

PERFORM-3D User Guide 2-1

Chapter 2. Structures, Folders and Files

(3) Key Driver folder. This folder contains drivers for the PERFORM-
3D security hardware. This folder is stored as a subfolder in the
PERFORM-3D Folder.

(4) Structures folder. This folder is the default location for

PERFORM-3D structures. See Section 2.2.1 for a description of its
use and location.

(5) Records folder. This folder contains the ground acceleration

records that are available to PERFORM. See Section 2.5.1 for an
explanation of its use and default location.

(6) RecordsF folder. This folder contains the dynamic force records
that are available to PERFORM. See Section 2.5.2 for an
explanation of its use and default location.

(7) Spectra folder. This folder contains the response spectra that are
available to PERFORM. See Section 2.5.3 for an explanation of its
use and default location.

(8) User folder. You can use this as the default folder for exporting
and importing data. Alternatively you can specify a folder in a
different location. Usually you will specify a different location.
See Section 2.5.4 for more information about its use and an
explanation of its default location.

2.2 Saving and Opening a Structure

2.2.1 Saving a Structure

When you first save a new structure, you have the option of putting it in
the Default Structures Folder (choose the "Default" option) or in a
User Structures Folder (choose the "User defined" option). For the user
defined option you can type in the path to the folder or (more likely)
browse to find the folder. If the folder does not yet exist, it will be
created when you save the structure (see Section 2.5.4 for more
information about its use and an explanation of its default location).

For Version 4, the Default Structures Folder set up during installation

is located here:

2-2 PERFORM-3D User Guide

 Chapter 2. Structures, Folders and Files

C:\ProgramFiles\Computers and Structures\

PERFORM\PERFORM-3D\Structures

Starting with Version 5, the location depends on Windows, which can

be a bit quirky, so the Default Structures Folder is set up here:

For computers using Windows 7 or Vista:

C:\Users\(UserNameAtInstallation)\My Documents\
Computers and Structures\PERFORM

For computers using Windows XP:

C:\Documents and Settings\(UserNameAtInstallation)\

My Documents\Computers and Structures\PERFORM

where (UserNameAtInstallation) is the logged-on user at the time that

PERFORM-3D was installed. For a computer that has several users,
you may wish to set up a special user for the installation, and to log in
as that user when running PERFORM-3D.

Unless you specify otherwise, when you save a structure it will be

placed in the Default Structures Folder. When you re-save a structure
that has already been saved, it is saved automatically in its folder. You
can use Save As to save a structure as a new structure with a different
name. Do not try to achieve the same result by copying and renaming
the folder for a structure. This will not work.

For each structure in the Default Structures Folder there will be a

subfolder that has the same name as the structure. For example, if you
have a structure named "Structure-1", it will be in the folder :

Version 4
C:\ProgramFiles\Computers and Structures\PERFORM\
PERFORM-3D\Structures\Structure-1
or
Version 5 or later and Windows or Vista
C:\Users\(UserNameAtInstallation)\My Documents\
Computers and Structures\PERFORM\Structures\Structure-1
or

PERFORM-3D User Guide 2-3

Chapter 2. Structures, Folders and Files

Version 5 or later and Windows XP

C:\Documents and Settings\(UserNameAtInstallation)\
My Documents\Computers and Structures\
PERFORM\Structures\Structure-1

If you set up a lot of structures, the Default Structures Folder can get
crowded. To avoid this you have the option of setting up your own
User Structures Folders. For example, if you have a number of
structures in a project name Project A, you might set up a User
Structures Folder Project-A. The path to this folder, using Version 5 or
later and Windows 7 or Vista might be :

C:\Users\(UserNameAtInstallation)\My Documents\
Computers and Structures\PERFORM\Project-A

However, you can choose any path you wish.

IMPORTANT NOTE: If you are replacing Version 4 with Version 5

or later, you will have existing Structure/Default Structures Folder
containing subfolders and files, as well as User, Records, RecordF, and
Spectra folders (see Section 2.5), that you probably want to keep.
These folders are not moved to the new locations during the install, and
they can not be used by Version 5 or later. You must move the folders
to the new locations.

To do this you will presumably use Windows Explorer. You must

know the locations of the existing files and the new locations as
explained in Sections 2.2.1 and 2.5. You may need Administrator status
to move the folders.

It goes without saying that you should copy, rather than move, the
folders, and that you should delete the existing folders only after you
are sure that they have been copied correctly.

2.2.2 Opening a Structure

When you open an existing structure you again have the "Default" and
"User" options. If you choose "Default", the structures in the Default
Structures Folder are listed in a table. You can sort this table by date or
by name. To open a structure, click on it in the table and press the Open
button (or alternatively double click on it in the table). If you choose
"User" you can type in the folder name or find it by browsing. When

2-4 PERFORM-3D User Guide

 Chapter 2. Structures, Folders and Files

you select a User Structures Folder, the structures in it will be listed in

the same way.

There is also a third option when you open an existing structure,

namely the "Recent" option. If you choose this option the table will
contain up to the 10 most recently opened structures. This is the option
that you are likely to use most often.

2.3 Location of ECHO file

The file ECHO.txt contains a record of the structure data, loading and
analyses. The location of this file is as follows.

Suppose that you have set up a structure "Structure-1" The Structures

Folder will contain a folder with this name. If you use the Default
Structures Folder, the path is :

Version 4
C:\ProgramFiles\Computers and Structures\PERFORM\
PERFORM-3D\Structures\Structure-1
or
Version 5 or later and Windows or Vista
C:\Users\(UserNameAtInstallation)\My Documents\
Computers and Structures\PERFORM\Structures\Structure-1
or
Version 5 or later and Windows XP
C:\Documents and Settings\(UserNameAtInstallation)\
My Documents\Computers and Structures\
PERFORM\Structures\Structure-1

The folder Structure-1 has a number of files that define the structure,
plus a subfolder for each Analysis Series (see Chapter 17, Analysis
Series). For example, if you have an Analysis Series "Series-A" and
you use the Default Structures Folder, the folder for the Analysis Series
is :

Version 4
C:\ProgramFiles\Computers and Structures\PERFORM\
PERFORM-3D\Structures\Structure-1\Series-A
or

PERFORM-3D User Guide 2-5

Chapter 2. Structures, Folders and Files

Version 5 or later and Windows or Vista

C:\Users\(UserNameAtInstallation)\My Documents\
Computers and Structures\PERFORM\Structures\
Structure-1\Series-A
or
Version 5 or later and Windows XP
C:\Documents and Settings\(UserNameAtInstallation)\
My Documents\Computers and Structures\
PERFORM\Structures\Structure-1\Series-A

This folder contains a number (usually a lot) of files. Never delete,

rename or otherwise change any of these files. One of the files is the
ECHO.txt file for the Analysis Series. Also see the Important Note in
Section 2.2.1 Saving a Structure with respect to the relocation of
existing files when changing from Version 4 to Version 5.

2.4 Moving or Copying a Project or Structure

To move an entire project or a single structure, move the corresponding

folder using Windows Explorer.

To copy a project or a structure from one computer to another, copy the

corresponding folder. If you do this, and if you are running dynamic
analyses, you must make sure that the earthquake acceleration records
are available on the second computer. For the procedure see Chapter
15, Earthquake Records.

You can rename the folder for a project. However, do not rename the
folder for a structure. The folder name for a structure must be the same
as the structure name.

Also see the Important Note in Section 2.2.1 Saving a Structure with
respect to the relocation of existing files when changing from Version 4
to Version 5 or later.

2-6 PERFORM-3D User Guide

 Chapter 2. Structures, Folders and Files

2.5 Other Folders

2.5.1 Records Folder

The first time that you execute PERFORM-3D it creates a Records
folder to hold ground acceleration records for dynamic earthquake
analysis. All acceleration records are stored in this folder. For the
procedure see Chapter 15, Earthquake Records.

For Version 4, if you installed the program in the default location, the
Records folder is:

C:\ProgramFiles\Computers and Structures\PERFORM\Records

Starting with Version 5, the Records folder location depends on

Windows, which can be a bit quirky, and is probably here:

For computers using Windows 7 or Vista:

C:\Users\(UserNameAtInstallation)\My Documents\Computers and

Structures\PERFORM

For computers using Windows XP:

C:\Documents and Settings\(UserNameAtInstallation)\My

Documents\Computers and Structures\PERFORM

where (UserNameAtInstallation) is the logged-on user at the time that

PERFORM-3D was installed. For a computer that has several users,
you may wish to set up a special user for the installation, and to log in
as that user when running PERFORM-3D. See Section 2.2.1 for more
information, including the Important Note concerning relocation of
files when upgrading from Version 4 to 5 or later.

2.5.2 RecordsF Folder

The first time that you execute PERFORM-3D it creates a RecordsF
folder to hold force time-history records for dynamic force analysis. All
force time history records are stored in this folder. For the procedure
see Chapter 16, Dynamic Force Load Cases and Analyses, Section
16.3. You cannot change the location of this folder.

PERFORM-3D User Guide 2-7

Chapter 2. Structures, Folders and Files

For Versions 4, if you installed the program in the default location, the
RecordsF folder is:

C:\ProgramFiles\Computers and Structures\PERFORM\RecordsF

Starting with Version 5, the RecordsF folder location depends on

Windows, which can be a bit quirky, and is probably here:

For computers using Windows 7 or Vista:

C:\Users\(UserNameAtInstallation)\My Documents\Computers and

Structures\PERFORM

For computers using Windows XP:

C:\Documents and Settings\(UserNameAtInstallation)\My

Documents\Computers and Structures\PERFORM

where (UserNameAtInstallation) is the logged-on user at the time that

PERFORM-3D was installed. For a computer that has several users,
you may wish to set up a special user for the installation, and to log in
as that user when running PERFORM-3D. See Section 2.2.1 for more
information, including the Important Note concerning relocation of
files when upgrading from Version 4 to 5 or later.
2.5.3 Spectra Folder
The first time that you execute PERFORM-3D it creates a Spectra
folder to hold response spectra for response spectrum analyses.
Response spectra are stored in this folder. For the procedure see
Chapter 22, Response Spectrum Analysis, Section 22.1.

For Version 4, if you installed the program in the default location, the
Spectra folder is:

C:\ProgramFiles\Computers and Structures\PERFORM\Spectra

Starting with Version 5, the Spectra folder location depends on

Windows, which can be a bit quirky, and is probably here:

For computers using Windows 7 or Vista:

2-8 PERFORM-3D User Guide

 Chapter 2. Structures, Folders and Files

C:\Users\(UserNameAtInstallation)\My Documents\Computers and

Structures\PERFORM

For computers using Windows XP:

C:\Documents and Settings\(UserNameAtInstallation)\My

Documents\Computers and Structures\PERFORM

where (UserNameAtInstallation) is the logged-on user at the time that

PERFORM-3D was installed. For a computer that has several users,
you may wish to set up a special user for the installation, and to log in
as that user when running PERFORM-3D. See Section 2.2.1 for more
information, including the Important Note concerning relocation of
files when upgrading from Version 4 to 5 or later.

2.5.4 User Folder

The first time that you execute PERFORM-3D it creates a User folder.
This folder is used for certain operations as follows.

(1) For some analysis tasks you can save the results in one or more
text files that you can then process using a spreadsheet program.
The default location for these text files is the User folder.
(2) To set up a new earthquake ground acceleration record you must
have the record in a text file, and process this file using
PERFORM-3D. For the procedure see Chapter 15, Earthquake
Records. The process is a little simpler if you first copy the text file
to the User folder.

For Versions 4, if you installed the program in the default location, the
User folder is :

C:\ProgramFiles\Computers and Structures\PERFORM\

PERFORM-3D\User

Starting with Version 5, the User folder location depends on Windows,

which can be a bit quirky, and is probably here:

For computers using Windows 7 or Vista:

C:\Users\(UserNameAtInstallation)\My Documents\
Computers and Structures\PERFORM

PERFORM-3D User Guide 2-9

Chapter 2. Structures, Folders and Files

For computers using Windows XP:

C:\Documents and Settings\(UserNameAtInstallation)\My

Documents\Computers and Structures\PERFORM

where (UserNameAtInstallation) is the logged-on user at the time that

PERFORM-3D was installed. For a computer that has several users,
you may wish to set up a special user for the installation, and to log in
as that user when running PERFORM-3D. See Section 2.2.1 for more
information, including the Important Note concerning relocation of
files when upgrading from Version 4 to 5 or later.

2.6 Do Folder Locations Matter?

For most users of PERFORM-3D, the locations of the standard folders

are not important. The folders that have been relocated are used by
PERFORM-3D, but most users will rarely need to read, write or copy
any files to those folders.

You may, however, need to use the folders for the following reasons.

(1) You may need to copy subfolders and files to the Records folder.
Earthquake records can be added to this folder in two ways. The
first is to use PERFORM-3D to import a record by processing a
text file of the record. This allows you to check that the record is
imported correctly. To do this you do not need to know the
location of the Records folder. The second is to copy a set of
records from one computer to another. This saves time by avoiding
the import operation. To do this you will use Windows Explorer,
and you must know the location of the Records folder.

(2) You may need to do the same for the RecordsF and/or Spectra
folders.

(3) If you save a structure in the Structures folder (which is usually

not recommended), you may want to open the ECHO.txt or
M000.txt files (see Section 2.3), using a utility such as Notebook.
To do this you must know the location of the Structures folder.

2-10 PERFORM-3D User Guide

 Chapter 2. Structures, Folders and Files

(4) If you use PERFORM-3D to write analysis results to the User

folder (as text files), you must know the location of this folder to
use the files. The User folder is the default for writing such files. It
is usually better to specify a different folder.

(5) The user documentation files are installed in the Manuals folder.

Also, see Section 2.2.1 for an Important Note concerning relocation of

files when upgrading from Version 4 to 5 or later.

PERFORM-3D User Guide 2-11

 Chapter 3. Nodes

3 Nodes

A PERFORM-3D analysis model consists of nodes connected by

elements. You can specify nodes and elements in any order (e.g.,
some nodes, then some elements, then more nodes, etc.) This
chapter describes how to specify nodes and node-related data.

3.1 Methods for Specifying Nodes

To specify nodes press the toolbar buttons for the Modeling phase and
the Nodes task.

The axes of the 3D coordinate system are H1, H2 (both horizontal) and
V (vertical upwards). These axes form a right-handed coordinate
system. Nodes are identified by their coordinates. Nodes are not
numbered.

To show or hide the node coordinates in a structure plot, press the

"H,V" button on the toolbar. When the coordinates are shown, the order
at each node, from top number to bottom number, is V, H2, H1.

If you show the coordinates, the structure plot can be crowded. As an

alternative, to show the coordinates of a single node, right click on the
node. To hide the coordinates. Left click anywhere, or right click
another node.

To add, move and/or delete nodes, choose the Nodes tab in the data
panel. You can then choose from several methods for specifying nodes.

For some methods you must select one or more existing nodes. There
are two ways to do this, as follows.

(1) Click on a single node in the graphics panel. The selected node
changes color (to green or red, depending on the method).
(2) Click, hold and drag a rectangle to enclose one or more nodes (a
"box" or "fence" select). This may not work for some methods.

In some cases, clicking on a selected node deselects it. To clear all

selected nodes, press the Clear button.

PERFORM-3D User Guide 3-1

Chapter 3. Nodes

After you have entered the required data in the data panel, press the
Test button to show the effect. Any additions or changes are shown in
yellow. You then have the choice of pressing OK (which accepts the
new or changed nodes) or Undo, which undoes the operation.

The available methods for specifying nodes are as follows.

(1) Create a regular grid of nodes. Choose the Grid tab. It is usually
convenient to begin with this method, then use other methods to
manipulate the grid. You can create a 3D grid, but if you do you
may initially get a jumble of nodes in the structure plot. If you
create a 3D grid it is may be useful to view the structure in plan, so
that the node layout makes sense. You should then set up a number
of frames (see Chapter 4, Frames), so that when you define
elements you can work in a frame-only view and show only the
nodes in one frame. An alternative is to create a 2D grid, add
elements to it, then use the Duplicate tab (see below) to duplicate
that grid. You can also use the Frames task to duplicate nodes and
elements.

(2) Move one or more nodes. Choose the Move tab. There are three
options for moving nodes, namely Translate, Tilt and Rotate. If
you have already specified elements, the elements are also moved.

(3) Delete one or more nodes. Choose the Delete tab. If you have
already specified elements, any elements connected to the deleted
nodes are deleted.

(4) Add new nodes one at a time. Choose the Single tab. There are
three options for adding single nodes, namely Total H,V,
coordinates, H,V Offsets from an existing node, and Polar Coords
using an existing node as the origin.

(5) Interpolate one or more nodes along a straight line between two
existing nodes. Choose the Interpolate tab. You can also
extrapolate.

(6) Duplicate a line or block of nodes. Choose the Duplicate tab. This
operation duplicates nodes only. To duplicate both nodes and
elements see the Duplicate operation in the Frames task.

3-2 PERFORM-3D User Guide

 Chapter 3. Nodes

3.2 Zooming

You can zoom in by pressing the zoom-in button on the toolbar. After
pressing this button, click in the graphics panel, hold, and drag a
rectangle covering the region you want to view. You can zoom in
multiple times, to get progressively more detailed views.

To zoom out, press the zoom-out button on the toolbar. Pressing this
button returns to the un-zoomed view.

3.3 Closely Spaced Nodes

You can not have nodes with identical coordinates. In the Umbrella
task you can specify a minimum node spacing, with a default of 6
inches or 15 centimeters. If you attempt to specify a new node that is
closer than this spacing to an existing node, in the H1, H2 or V
direction, this is an error.

You can change the minimum spacing but it can not be zero. This
means that you can not have zero length elements. If you make the
minimum spacing too small, you may not be able to distinguish closely
spaced nodes.

If you specify some nodes then change the minimum spacing, the new
minimum applies only to any new or changed nodes.

3.4 Supports

To specify rigidly supported (restrained) nodes choose the Supports

tab, set the support type, select nodes singly or using a box select, and
use the Test, Clear, OK and Undo buttons in the usual way. You can
add, change and delete supports at any time.

When you are in the Supports page, support symbols are always
shown. When you are in other pages you can show or hide the support
symbols by pressing the support symbol button in the toolbar.

PERFORM-3D User Guide 3-3

Chapter 3. Nodes

You do not need to specify supports for nodal displacements that are
completely unrestrained. For example, if your structure is a truss you
do not have to restrain the rotational displacements. PERFORM-3D
does this automatically. There will be a warning message in the ECHO
file.

3.5 Masses

To specify masses choose the Masses tab. In the current version of

PERFORM-3D you must specify all masses as nodal properties. There
is no provision for element masses.

You are not restricted to a single mass distribution. Instead, you can
specify up to 6 mass patterns, each identified by a name. For example,
you might specify a mass pattern for dead load masses and a second
pattern for live load masses. Then, in the Analysis phase, you can
specify the masses for the analysis as a linear combination of the mass
patterns (e.g., 1.0 times dead load mass plus 0.25 times live load mass).
For a more detailed discussion see Chapter 17, Analysis Series.

To start a new mass pattern press the New button, specify the pattern
name, then specify the masses, select nodes singly or using a box
select, and use the Test, Clear, OK and Undo buttons in the usual
way.

To change the masses for an existing pattern, choose the pattern name
from the list and specify changes as needed. You can also delete any
mass pattern.

If you have a rigid floor diaphragm (with a rigid floor constraint – see
the next section) you can specify the floor mass in either of two ways,
as follows.

(1) Calculate the center of mass for the floor, place a node at this point
(be sure to include it in the rigid floor constraint), and put the
translational mass and rotational inertia for the entire floor at the
point (be sure to include the rotational inertia). This can be the best
method if you want to consider different mass locations, to account
for "accidental" mass eccentricity. If you define a series of nodes
that have masses but are not connected to any elements in the

3-4 PERFORM-3D User Guide

 Chapter 3. Nodes

structure, you can change the mass locations by moving these

nodes.

(2) Do not calculate the center of mass. Instead, specify a mass at each
beam-column intersection in the floor, using tributary areas. When
you specify a rigid floor constraint, PERFORM-3D will calculate
the floor rotational inertia and the center of mass, and place the
floor mass at this center. You probably do not need to calculate the
tributary rotational inertias at the individual mass points, since in
most cases they will be small. However, you can do so if you wish.

3.6 Slaving Constraints

3.6.1 General

Each constraint constitutes a Constraint Set, and must affect at least

two nodes. As an example, if you have a three story frame and use rigid
floor slaving, you will have three constraint sets, one for each floor or
roof level. If you have four column lines along each of the H1 and H2
directions, each constraint set will affect 16 nodes.

To start a new constraint set press the New button, specify the
constraint set name, and choose the constraint type. You can either add
or delete nodes. Select the affected nodes, singly or using a box select,
and use the Test, Clear, OK and Undo buttons in the usual way. The
order in which you select the nodes does not matter. It is not necessary
to identify a "master" node and a number of "slave" nodes.

3.6.2 Rigid Floor Constraint

In a rigid floor constraint set, all nodes must have the same V
coordinate (i.e. the floor must be level). The constraint makes the H1
and H2 translations and the V rotation (rotation about the V axis) the
same at all nodes in the set. The V translation and the H1 and H2
rotations are not affected.

Note that all of the beam elements in a rigid floor will have zero axial
deformation. Hence if you use a typical beam model the beams will all
have zero axial force. However, if you use a beam model that has fiber

PERFORM-3D User Guide 3-5

Chapter 3. Nodes

sections or concrete type P-M-M hinges, the beam element may want to
extend. Since you are restraining this extension, there may be
compression forces in the beams.

You can not include a node in a rigid floor constraint if the node has an
H1 or H2 translational support or a V rotational support. Conversely,
you can not add an H1 or H2 translational support, or a V rotational
support, to a node that is part of a rigid floor constraint.

3.6.3 Full Rigid Link Constraint

In a rigid link constraint set, a rigid link is added that connects all of the
nodes in the set. The H1, H2 and V translations and rotations are all
affected. If you were to specify a rigid link constraint for the nodes at a
floor level, the floor would be rigid both axially and in bending.

You can not include a node in a rigid link constraint if the node has a
support of any kind. Correspondingly, you can not add a support to a
node that is part of a rigid link constraint.

You will rarely need to use this constraint type. If you do, be careful
with links that are not horizontal. In effect, the constraint links are rigid
elements, and they will generally have axial forces, bending moments
and shear forces. These forces are not calculated. If a rigid link is
vertical or has a vertical component, and if it has an axial force, it may
exert a significant P- effect. This effect is NOT accounted for in the
analysis, and it could introduce significant errors.

3.6.4 Eccentric Connection Constraint

In some frame structures, the beams may be connected eccentrically to
the columns, for example as shown in Figure 3.1.

3-6 PERFORM-3D User Guide

 Chapter 3. Nodes

Eccentricity

(a) Beam and Column (b) Analysis Model

Figure 3.1 Eccentric Connection of Beam to Column

Figure 3.2 shows a more complex example, where both the beams and
columns have eccentricities.

(a) Beams and Columns (b) Analysis Model

Figure 3.2 More Complex Eccentric Connection

If you want to take this type of eccentricity into account, you can model
the connection in three different ways as follows.

(1) Connect the nodes by short, stiff elements. This requires that you
specify the extra elements and element properties, and is usually
not a very convenient method.

(2) Connect the nodes using a rigid link constraint. You can use this
constraint only if the nodes are not part of a rigid floor constraint,
since any displacement can be a part of only one constraint. Rigid
link and rigid floor constraints both constrain the H1 and H2
translations and the V rotation.

PERFORM-3D User Guide 3-7

Chapter 3. Nodes

(3) Connect the nodes using an eccentric connection constraint. This

constraint is specifically intended for the case where the nodes are
already part of a rigid floor constraint. It constrains the V
translation and the H1 and H2 rotations. You should use this
constraint only if the nodes are already part of a rigid floor
constraint.

Note that you must have a separate constraint set for each eccentric
connection.

3.6.5 Equal Displacement Constraint

With an equal displacement constraint you can specify that a number of
nodes (usually only two nodes) have equal displacements for any of the
H1, H2 and V translations and rotations.

As with all constraints, you should use this type carefully, otherwise
you can introduce spurious external forces and violate the equilibrium
requirements for the structure. If you were to impose equal V
displacements on two nodes that have identical coordinates, this would
introduce no equilibrium errors, for two reasons. First, since the H
coordinates are equal, any vertical forces applied on the nodes by the
constraint are equal and opposite in all respects. Second, since the V
coordinates are equal, there are no P- effects in the constraint.
However, if the H coordinates of the two points are not the same, the
forces applied on the nodes by the constraint are equal and opposite in
magnitude and direction but not in location, so the constraint imposes
an external moment on the structure. Also, if the V coordinates of the
two points are not the same, and if you are considering P- effects in
the analysis, there can be implied P-forces but these are not
considered by PERFORM-3D.

In PERFORM-3D you can not have nodes with identical coordinates. If

you have two points with identical coordinates, you must use nodes that
are separated by a small distance. If you create two such nodes and
separate them by the minimum allowed distance, it will usually be OK
to impose equal displacements, since any error is small. In all other
cases, be sure that you understand the physical significance of the
constraint.

3-8 PERFORM-3D User Guide

 Chapter 4. Frames

4 Frames

The plotted geometry for a 3D structure can be complex. You can

(and usually must) simplify the plot by showing only selected parts
of the structure. A "frame" is any part of the complete structure. It
can be a plane frame, a floor, or any other part of the structure that
you would like to operate on separately from the complete
structure.

You can quickly set up, modify and delete frames, and then toggle
between a full structure view and a view showing only a single
frame. This chapter explains the procedure.

4.1 Plotting of Nodes and Elements

In PERFORM-3D, the analysis mode for a structure is a node-element

model. A plot of the structure usually shows both nodes and elements.
Nodes are shown as small squares and elements as lines (for 2-node
elements) and quadrilaterals (for 4-node and 1-node elements).
Elements are usually drawn shrunk (i.e., with gaps between the element
ends or corners and the nodes). PERFORM-3D draws only a "wire
frame" plot of a structure. You can not show the elements as solid
bodies.

Plots of large structures can obviously have many lines for elements.
To help clarify the plots, the elements are plotted in different colors.
For example, elements must be divided into groups (see later). In a plot
of the structure, the elements in the current group are shown light blue
to make them stand out.

For relatively simple structures, the use of color combined with a

perspective projection can be sufficient to make the plot clear. For
complex structures, however, this is usually not enough. To simplify
the plots when you work with complex structures you should make
liberal use of "frames". A frame is any part of the complete structure. It
can be a plane frame, a floor, an elevator core, or any other part of the
structure that you would like to plot separately from the complete
structure. You can quickly define, modify and delete frames, and then

PERFORM-3D User Guide 4-1

Chapter 4. Frames

toggle between a "structure" view that shows the complete structure

and a "frame" view that shows only a single frame.

In a structure view, the elements in the current frame are highlighted to

help them stand out. Any elements that are in both the current frame
and the current element group are plotted light blue. In a frame-only
view, only the nodes and elements in the current frame are plotted. You
can add elements, specify loads, define structure sections, etc. in either
a structure view or a frame view.

4.2 Defining a Frame

To define a new frame or modify an existing frame, press the toolbar

buttons for the Modeling phase and the Frames task. The steps for
adding a new frame are as follows.

(1) Press the New button.

(2) Input a name for the frame, then press OK.
(3) Choose the Add Nodes tab.
(4) In the graphics panel, select the nodes to be added to the frame,
using either single node or box selection. If you are defining a
vertical frame it is usually best to change to a plan view, so that
you can select the nodes with a box select. If you are defining a
horizontal floor it is usually best to change to an elevation view.
(5) If the selected nodes are OK, press OK to add them to the frame. If
not, press Clear. Any elements that are connected to nodes in the
frame are also a part of the frame, including any elements that you
specify later.
(6) Repeat Steps (4) and (5) until you have selected all the desired
nodes.

To delete nodes from a frame, choose the Delete Nodes tab. You can
return to a previously defined frame at any time to add or delete nodes.
To delete a complete frame press the delete (X) button next to the New
button.

To toggle between structure and frame-only views, press the S-F

(Structure-Frame) button. When you are in the Frames task, this button
and the list of frames is in the data panel. After you exit this task, the
list of frames and the S-F button is at the right end of the toolbar. You

4-2 PERFORM-3D User Guide

 Chapter 4. Frames

can choose a frame from the list and toggle between structure and
frame views at any time

4.3 Duplicating a Frame

If you have a number of similar frames in a structure, you can save time
by defining one frame then duplicating it. When you duplicate a frame
you duplicate both the nodes and the elements in the frame (if you use
the Duplicate option for nodes, you duplicate only the nodes).

If you have similar frames, you may duplicate a frame after you have
specified the elements but before you have assigned properties to the
elements. If you have frames that are more nearly identical, you can
duplicate a frame after you have assigned properties to the elements.

When you duplicate a frame you can specify H1, H2 and/or V distances
between the original frame and the duplicate, which means that the new
frame is parallel to the original. If you also want the new frame to be
rotated, you must first create it parallel to the original frame, then use
the Move operation in the Nodes task to rotate the new frame.

To duplicate a frame, press the toolbar buttons for the Modeling phase
and the Frames task, and choose a frame from the frame list. The steps
for duplicating the frame are as follows.

(1) Choose the Duplicate tab.

(2) Specify the name for the duplicate frame.
(3) Specify the H1, H2 and V translations from the current frame.
(4) Press Test to show the new frame.
(5) Press either OK or Undo.

In the current frame the elements will belong to element groups.

Corresponding elements in the new frame belong to the same element
groups.

PERFORM-3D User Guide 4-3

 Chapter 5. Component Properties

5 Component Properties

Elements are made up of components. A simple element, such as a

truss bar, may consist of a single component. A more complex
element, such as a column, can consist of a number of components
of different types. Before you can assign properties to elements, you
must first specify properties for components.

This is the most complex task for nonlinear modeling, and usually
the task that will require most of your time. This chapter covers the
essential concepts, and the procedures for specifying component
properties. For more details, see the document "PERFORM
Components and Elements".

5.1 Elements and Components

PERFORM-3D has several element types (bar, beam, column, wall,

seismic isolator, etc.). To specify the properties for an element you
must first specify the properties of one or more components that make
up the element.

There are several component types, as follows.

(1) Materials. These include steel, concrete and shear materials.

(2) Cross sections. These include beam, column and wall sections.
The properties of some cross sections can depend on the properties
of materials.

(3) Basic structural components. These include bars, plastic hinges,

connection panel zones, seismic isolators and many other
components. The properties of some basic structural components
can depend on the properties of cross sections and/or materials.

Basic components are divided into inelastic and elastic

components. Inelastic components can yield and dissipate energy.
Elastic components do not dissipate energy. Elastic components
are mostly linear, but can be nonlinear (e.g., a gap-hook bar).
Inelastic components are always nonlinear, and are much more

PERFORM-3D User Guide 5-1

Chapter 5. Component Properties

complex than elastic components. A fluid damper dissipates

energy and is classified as inelastic, although its behavior is
different from that of other inelastic components.

(4) Strength sections. These are not structural components. Their

purpose is to allow you to calculate strength demand-capacity
ratios at interior points in beam and column elements.

(5) Compound components. A compound component is made up of a

number of cross section, basic and/or strength section components.
For example, a compound component for a beam might consist of
an elastic end zone component, a plastic hinge component, an
elastic segment with a uniform cross section, a second plastic
hinge, and a final end zone. It might also have shear strength
sections at the plastic hinge locations, to check that the shear
strength is not exceeded.

Some elements consist of one compound component. These include

frame type elements (beam, column, diagonal brace), and also shear
wall, general wall, buckling restrained brace and viscous bar elements.
Other elements consist of one basic component.

5.2 Procedure for Specifying Components

5.2.1 Scope
This section describes the basic steps for specifying components.

Nonlinear components can have complex properties, and the forms for
input of component properties may appear to be complex. You do not,
however, have to specify values for all possible properties. For
example, you may choose to specify elastic-perfectly-plastic behavior
rather than the more complex trilinear behavior; you may not need to
consider brittle strength loss; and unless you are an advanced user you
will not need to specify upper/lower bounds. It is usually a good idea to
start with relatively simple component properties, and to add more
complex properties later, as the analysis proceeds.

If you are a new user, you might like to experiment with the component
properties task before starting a complete structure. The forms for input

5-2 PERFORM-3D User Guide

 Chapter 5. Component Properties

of component properties contain a lot of information, so be sure to look

at them carefully.

For more information on component properties see the following

sections.

(1) If you are not familiar with the PERFORM action -deformation (F-
D) relationship, see the brief description in Section 5.5. For more
detailed information see the document "PERFORM Components
and Elements".
(2) For brittle strength loss, see Sections 5.6, 5.7 and 5.8.
(3) For deformation and strength capacities, and an overview how they
are used for limit states and usage ratios, see Section 5.9.
(4) For cyclic degradation and hysteresis loops see Sections 5.10, 5.11
and 5.12.
(5) For the role of cross sections and how to use them see Sections
5.13 and 5.14.
(6) For "auto" end zone and connection panel zone components, see
Sections 5.15.
(7) For upper and lower bounds on strength and stiffness see Section
5.16.
5.2.2 Component Types and Names
To specify or change component properties, press the toolbar buttons
for the Modeling phase and the Components task.

You can choose from the following component types : Inelastic

(inelastic basic components), Elastic (elastic basic components), Cross
Sects (cross sections), Material (elastic and inelastic materials)
Strength Sects (strength sections) and Compound (compound
components).

For each component that you define, you must assign an identifying
name (max. 40 characters). For a large structure you can have many
components, including compound components that consist of several
lower level components. Be sure to plan component names carefully, so
that you can easily identify them.

See Section 5.3 for some methods of managing component properties.

PERFORM-3D User Guide 5-3

Chapter 5. Component Properties

5.2.3 Materials
To specify a material the steps are the same as for a basic structural
component, as described later.

5.2.4 Beam and Column Cross Sections

To specify a beam or column cross section the steps are as follows.

(1) Choose the Cross Sects tab.

(2) Choose the cross section type from the Type list. Take care
whether you choose a beam type or column type section. You can
use either type of cross section in a frame compound component,
but you must use beam sections for beam type plastic hinge
components, and column sections for column type hinge
components.
(3) Press the New button to start a new cross section.
(4) Enter the component name and choose the units, then press OK.
Plan the names carefully (although you can change the name later
if you wish). Be sure to choose the correct units.
(5) Choose the appropriate Symmetry option (symmetric or non-
symmetric) if it is available.
(6) Enter the section properties. The Section Stiffness properties are
required (they are obtained automatically for a standard section).
The properties in the Inelastic Strength tab are optional. If you
specify some or all of these properties, they are used when you
associate the cross section with an inelastic basic component such
as a plastic hinge. The properties in the Elastic Strength tab are
also optional. If you specify some or all of these properties, they
are used if you associate the cross section with a strength section.
(7) Press the Check button to check the data. You will be told if there
are any errors. Keep an eye on the Status box.
(8) When the data is correct press the Save button to save the data.
Again, keep an eye on the Status box.
(9) To cancel any changes press the UnChange button. This restores
the last saved data.

To review or change an existing cross section, choose the section type

from the Type list and the name from the Name list. If you make
changes, keep an eye on the Status box.

5-4 PERFORM-3D User Guide

 Chapter 5. Component Properties

You can press the Save As button to save the current cross section as a
new section with a different name. It can be a time saver to do this if
you have cross sections that are similar. Set up and save one section,
then use Save As to initialize each similar section, and make just the
needed changes. Be sure to press Save after you have made the
changes.

You can press the Delete button to delete the current cross section
provided that you have not yet saved it. Do this if you have started a
new section and find that you do not want to complete it.

You can delete a saved cross section by pressing the delete (X) button
(next to the New button). If a cross section is used by a basic
component or a compound component, you will not be able to delete it.

You can change the name of a saved cross section by pressing the
Rename button.

5.2.5 Fiber Cross Sections for Walls

Cross sections for shear walls and general walls are always fiber
sections. There are four cross sections of this type, as follows, all of
them requiring similar data.

(1) Shear wall, inelastic fiber section.

(2) Shear wall, elastic fiber section.
(3) General wall, inelastic fiber section.
(4) General wall, elastic fiber section.

For each cross section there are two options for specifying the fibers, as
follows.

(1) Fixed Size. You must specify the area and coordinate location for
each fiber in the cross section. The width of the cross section is
fixed, based on the maximum and minimum fiber coordinates.
When you assign this cross section to an element, you must make
sure that the element width is consistent with the cross section
width. With this type of section you can account for thickness
variations in the wall. You can also have different reinforcement
areas in different part of the section.

PERFORM-3D User Guide 5-5

Chapter 5. Component Properties

(2) Auto Size. You must specify the wall thickness and the number of
fibers. The width of the cross section is not fixed. When you assign
the cross section to an element, PERFORM-3D makes the cross
section width equal to the element width, and hence calculates the
fiber areas and coordinates. With this type of section you can not
vary the wall thickness over the element width, and the
reinforcement percentage must be constant. You can, however,
account for changes in thickness and additional reinforcement,
using additional Concrete Strut and Steel Tie elements.

The steps in defining a fiber wall section are as follows.

(1) Choose the Cross Sects tab.

(2) Choose the cross section type from the Type list.
(3) Press the New button to start a new cross section.
(4) Enter the section name and choose the units. Plan the names
carefully (although you can change the name later if you wish). Be
sure to choose the correct units.
(5) Choose the "Fixed Size" or "Auto Size" option.
(6) Press OK.
(7) Enter the data for structural fibers. Also enter the data for out-of-
plane bending. Use the Add, Insert, Replace and Delete buttons
to add or modify fibers.
(8) If you wish, enter data for monitored fibers. This is usually a good
idea, since it allows you to define limit states based on strain (for
an inelastic section) or stress (for an elastic section). You can also
monitor deformations of wall elements using separate Deformation
Gage components (these are elastic basic components).
(9) Press the Check button to check the data. You will be told of any
errors. Keep an eye on the Status box.
(10) When the data is correct press the Save button to save the data.
Again, keep an eye on the Status box.
(11) To cancel any changes press the UnChange button. This restores
the last saved data.

To review or change an existing cross section, choose the section type

from the Type list and the name from the Name list. If you make
changes, keep an eye on the Status box.

You can press the Save As button at any time to save the current cross
section as a new section with a different name. It can be a time saver to
do this if you have cross sections that are similar. Set up and save one

5-6 PERFORM-3D User Guide

 Chapter 5. Component Properties

section, then use Save As to initialize each similar section, and make
just the needed changes. Be sure to press Save after you have made the
changes.

You can press the Delete button to delete the current cross section
provided that you have not yet saved it. Do this if you have started a
new section and find that you do not want to complete it.

You can delete a saved cross section by pressing the delete (X) button
(next to the New button). If a cross section is used by a shear wall or
general wall compound component, you will not be able to delete it.

You can change the name of a saved cross section by pressing the
Rename button.

5.2.6 Fiber Cross Sections for Frame Elements

There are two cross sections of this type, namely Inelastic Fiber Section
for a Beam and Inelastic Fiber Section for a Column. The data is
similar to that for a fiber wall section, with the main difference that for
a column section you must specify fiber locations along both Axis 2
and Axis 3 of the cross section.

Please be careful using fiber column cross sections. You may be

tempted to specify cross sections with large numbers of fibers. If you
do, and if you use these cross sections in a large number of elements, it
can take a lot of computer time to run the analysis. This cross section is
intended mainly for applications such a bridge piers, where there are
relatively few pier elements. It is not intended for typical columns in
buildings, where there may be many column elements.

If you use this type of section, we recommend that you analyze single
element examples to determine the smallest number of fibers that you
can use and still get reasonable results. You are likely to find that you
do not need a large number of fibers.

5.2.7 Basic Structural Components

To specify an inelastic or elastic basic structural component the steps
are as follows. The steps for a material are the same, and the steps for a
strength section are very similar.

PERFORM-3D User Guide 5-7

Chapter 5. Component Properties

(1) Choose the Inelastic or Elastic tab.

(2) Choose the component type from the Type list.
(3) Press the New button to start a new component.
(4) Enter the component name and choose the units, then press OK.
Plan the names carefully (although you can change the name later
if you wish). Be sure to choose the correct units.
(5) Choose the options (symmetry, whether the component uses a
cross section, whether there is strength loss, etc.). You can change
these later if you wish.
(6) Enter the component properties. You will usually have to fill in
data on a number of tabbed pages in the form. Diagrams in the
form explain the required data. If you choose "Yes" for the "Use
Cross Section" option, you must select a cross section type and
name from the lists. If you do this, the properties that are obtained
from the cross section are shown green. You can not change these
properties. If you have specified cross section strengths that can be
used by the component, these strengths will be shown green. If you
have not specified such cross section strengths, the strength text
boxes will be white and you must enter the properties.
(6) Press the Check button to check and (in most cases) plot the data
so that you can review it graphically. You will be told of any
errors. Keep an eye on the Status box.
(7) When the data is correct press the Save button to save the data.
Again, keep an eye on the Status box. After you have saved the
data, you can plot hysteresis loops for most inelastic components.
This can be valuable for checking the component behavior. See the
next section for the procedure.
(8) To cancel any changes press the UnChange button. This restores
the last saved data.

To review or change an existing component, choose the component

type from the Type list and the component name from the Name list. If
you make changes, keep an eye on the Status box.

You can press the Save As button at any time to save the current
component as a new component with a different name. It can be a time
saver to do this if you have components that are similar. Set up and
save one component, then use Save As to initialize each similar
component, and make just the needed changes. Be sure to press Save
after you have made the changes.

5-8 PERFORM-3D User Guide

 Chapter 5. Component Properties

You can press the Delete button to delete the current component
provided that you have not yet saved it. Do this if you have started a
new component and find that you do not want to complete it.

You can delete a saved component by pressing the delete (X) button
(next to the New button). If a basic component is used in a compound
component, you will not be able to delete it. If a basic component is
assigned directly to an element, you will be able to delete the
component, but you will need to assign a new component to the
element.

You can change the name of a saved component by pressing the

Rename button.

5.2.8 Plot Hysteresis Loops

When you check the properties for an inelastic component or an
inelastic material, PERFORM plots the force-deformation relationship.
This provides a useful graphic view of the properties. However, the
graph does not show what happens to a component when it is deformed
cyclically. If you wish, you can plot hysteresis loops for cyclic
deformations. This can gives you a much better idea of how the
component behaves.

The procedure is as follows.

(1) If you are defining a new component or changing an existing one,

specify the component properties in the usual way, and press the
Check and Save buttons. If you are reviewing an existing
component, just go to that component to display its properties.
(2) Press the Graph button to draw the force-deformation relationship.
When the relationship is plotted you will see a Plot Loops button
(in the same location as the Save button). Press the Plot Loops
button. Note that this button is present only when the force-
deformation relationship is plotted. Some components do not have
this button.
(3) A form with spaces for deformation ranges will be shown. Enter as
many deformation ranges as you wish, and check the ones that you
want to plot.
(4) Press Plot to plot the loops. Uncheck or check deformation ranges
to add or delete them in the plot. Press Print to print the current
loops, or Save to save the loops to a text file. Press Close to exit.

PERFORM-3D User Guide 5-9

Chapter 5. Component Properties

5.2.9 Strength Sections

To specify a strength section the steps are very similar to those for a
basic structural component.

5.2.10 Frame Compound Components

Frame compound components are used for beams, columns, braces, and
other elements of frame type. To specify a frame compound component
the steps are as follows.

(1) Choose the Compound tab and choose the Frame Compound
Component type from the Type list. Figure 5.1 shows the form.
(2) Press the New button to start a new component, and enter the
component name in the usual way.
(3) The compound component is made up of a number of basic
components. Use the COMPONENT TO BE ADDED section to
add components. For each basic component choose the type and
name from the lists, choose the length type and specify the length
value. Then position the basic component along the compound
component using the Add, Insert, Replace and Delete buttons.
(4) Do essentially the same for the strength sections, if any.
(5) Press the Check button to check and plot the data so that you can
review it graphically. In the COMPONENT LIST, if you click on
the row for any basic component, the corresponding component in
the graphical view will be shown yellow. This applies also for
strength sections.
(6) If you want to change a component, the best way is to double click
on it in the COMPONENT LIST. Its properties will be set up in the
COMPONENT TO BE ADDED section. Edit these properties,
then press Replace. This sequence (double click, change the data,
press Replace) can save time when you need to change a
component in the list.
(6) When the data is correct press the Save button. Keep an eye on the
Status box.
(7) To cancel any changes press the UnChange button. This restores
the last saved data.

You can return to examine or change a compound component at any

time. If you make changes, keep an eye on the Status box, and be sure
to press Save to save the component.

5-10 PERFORM-3D User Guide

 Chapter 5. Component Properties

Use the Save As, Delete, UnChange, Rename and X buttons as for
basic components.

Figure 5.1 Form for a Frame Compound Component

5.2.11 Shear Wall Compound Components

Shear wall compound components, and shear wall elements, are
intended for modeling relatively slender walls or shear cores. The
important actions are shear force and axial-bending action along the
vertical direction. However, shear wall elements can also be used for
structures such as box section girders ). The transverse (usually
horizontal) direction is assumed to be a secondary direction, similar to
the transverse direction in a beam or column.

A shear wall element can also bend out of plane. This is assumed to be
a secondary mode of behavior. Out-of-plane bending is assumed to be
elastic.

PERFORM-3D User Guide 5-11

Chapter 5. Component Properties

For a shear wall compound component the overall steps are similar to
those for a frame compound component, but the form is very different.
For the primary bending behavior you must choose a fiber wall cross
section. For shear you must choose a shear material and specify an
effective wall thickness. For transverse behavior you must specify an
effective thickness and an elastic modulus. For out-of-plane bending
you must specify an effective plate bending thickness and a modulus.

5.2.12 General Wall Compound Components

Whereas a shear wall element has a primary axial-bending direction
(usually the vertical direction), a general wall may not have a primary
direction, and may be just as likely to have inelastic axial-bending
behavior horizontally as vertically. Also, a general wall element can
resist shear in two ways, namely by conventional shear ("concrete
shear" in a general wall element) and by diagonal compression action
("diagonal shear").

A general wall element can bend out of plane. As with shear wall
elements, this is a secondary mode of behavior, and is assumed to be
elastic.

The steps and the form for a general wall component are similar to
those for a shear wall component. However, you must choose two fiber
cross sections, one each for the vertical direction and horizontal
directions. You must also specify material and thicknesses for both
concrete shear and diagonal shear. If you wish you can ignore one of
the shear types by specifying zero thickness. For out-of-plane bending
you must specify an effective plate bending thickness and a modulus.

The general wall compound element is a complex element. Before you

sue it you should study its description in the document "PERFORM
Components and Elements".

5.2.13 Viscous Bar and BRB Compound Components

The overall steps for viscous bar and BRB (buckling restrained brace)
compound components are the similar to those for other compound
components, and the form is self-explanatory. A viscous bar compound
component uses fluid damper and elastic bar basic components. A BRB
compound component uses BRB and elastic bar basic components.

5-12 PERFORM-3D User Guide

 Chapter 5. Component Properties

5.3 Managing Component Properties

5.3.1 General
In a large structure there can be a lot of components. This section
describes some tools that can be useful for managing the components
and their properties.
5.3.2 Filter
In PERFORM, components are selected from drop-down lists. If there
are many components of a given type, the drop-down list can be long,
and finding a particular component can involve searching. Also, the
components are listed in the order that they are defined, so that related
components may not be adjacent to each other in the list. You can use
the “filter” feature to shorten the list.

In the Component Properties task, below the list of component names

there is a “Filter” button as shown in Figure 5.2.

Figure 5.2 Purge, Rename and Filter Buttons

To filter the list of component names, type a text string into the “Text
for filter” box, then press the Filter button. Only component names that
include the text string will be listed. The text string is case sensitive.

To use this feature effectively you must use distinctive text strings for
component names. If you wish you can press the Rename button to
change component names.
5.3.3 Purge
A list of component names may contain components that are not used
in any higher level components (such as a compound component) or in
any elements. To purge such components from the list, press the Purge
button (Figure 5.2).

PERFORM-3D User Guide 5-13

Chapter 5. Component Properties

5.3.4 Use Cross Section Strengths

If an inelastic basic component depends on a cross section, it is usually
a good idea to specify the strength data with the cross section. The
inelastic component then gets its basic strength from the cross section,
and if you change the cross section the inelastic component strength
also changes.
5.3.5 Save As
Rather than start each new component by pressing the New button, use
the Save As button and edit the properties.
5.3.6 Copy-Paste
In the Component Properties task, most component property pages
have Copy and Paste buttons as shown in Figure 5.3.

Figure 5.3 Copy-Paste Buttons

You can use these buttons to copy properties from one component to
another, as follows.

(1) Set up the properties for a component. Go to one of the pages that
has Copy-Paste buttons.
(2) Press the Copy button. The name of the component will be shown
in the box between the buttons, as a reminder, and also to show
that properties for that page can be pasted.
(3) Choose another component of the same type, and go to the same
page. Press the Paste button. All of the properties on that page
from the copied component will be pasted into the current
component.

You can press Copy in multiple pages if you wish, then Paste to
multiple pages (one page at a time). The main restriction is that all of
the properties on a page are copied. You can not copy selected
properties.

5-14 PERFORM-3D User Guide

 Chapter 5. Component Properties

5.4 Printed Reports

To get a printed report for any component, press the button with the
printer symbol. In the current version of PERFORM-3D you can print
the properties for only the current component. If you want to print the
properties for all components, you must print each one in turn.

The component properties are also included in the ECHO file.

5.5 F-D Relationship

5.5.1 Actions and Deformations

Each material and each basic structural component has one or more
actions (or forces) and corresponding deformations. For example, for a
simple material the action is stress and the deformation is strain, and for
a simple plastic hinge the action is bending moment and the
deformation is hinge rotation. The relationship between the two is the
F-D relationship.

This section is concerned with the nonlinear F-D relationship for

inelastic components, and the PERFORM-3D procedure for modeling
this relationship. Elastic components are simpler, and do not need any
general explanation.
5.5.2 Actual Components
For the components in a real structure there will usually be a great deal
of uncertainty about the actual behavior. In most cases the best we can
do is capture those aspects of nonlinear behavior that have a substantial
influence on the result of the analysis. Common sense suggests that if
minor differences in the nonlinear behavior of the components have
major effects on the analysis results, then the design is too sensitive,
and we will have difficulty getting analysis results that are reliable
enough for design purposes. Simplicity is also important. Structural
design is a complex task, and it can be made impossibly complex if we
have to use elaborate analysis models.

PERFORM-3D User Guide 5-15

Chapter 5. Component Properties

5.5.3 PERFORM F-D Relationship

Most of the inelastic components in PERFORM-3D have the same
form for the F-D relationship. This is a trilinear relationship with
optional strength loss, as shown in Figure 5.4.

ACTION Zero No strength

slope loss

U L X
Strength
Y loss
Hardening
stiffness Maximum
deformation
Initial R X
stiffness Optional full
strength loss

DEFORMATION

Figure 5.4 PERFORM Action-Deformation Relationship

The key points in the relationship are as follows.

(1) Y Point. This is the first yield point, where significant nonlinear
behavior begins.
(2) U Point. This is the ultimate strength point, where the maximum
strength is reached.
(3) L Point. This is the ductile limit point, where significant strength
loss begins.
(4) R Point. This is the residual strength point, where the minimum
residual strength is reached.
(5) X Point. This is usually at a deformation that is so large that there
is no point in continuing the analysis. You can continue an analysis
beyond this point if you wish, but usually you will stop the
analysis if any component is deformed beyond its X point. For
some components you can specify that the strength reduces to zero
at the X point. In this case the analysis does not stop.

For most components the relationships can be different for positive and
negative deformations (i.e., the component can be unsymmetrical).

5-16 PERFORM-3D User Guide

 Chapter 5. Component Properties

5.5.4 E-P-P and Bilinear Relationships

For many components an elastic-perfectly plastic (e-p-p) relationship,
rather than a trilinear relationship may be adequate. In this case the Y
and U points are the same.

To get a bilinear relationship with no strength loss, specify a large

value for the deformation at the U point. To get a bilinear relationship
with strength loss, specify a U point deformation that is slightly smaller
than the L point deformation.
5.5.5 Additional Parallel Stiffness
Some components may continue to strain harden without reaching an
ultimate load. PERFORM-3D allows this for some components, by
allowing you to specify an additional parallel stiffness as shown in
Figure 5.5.

ACTION Relationship with

added stiffness
Without
Basic relationship strength loss

With
strength loss
Added parallel
stiffness

DEFORMATION

Figure 5.5 Additional Parallel Stiffness

Note that the parallel stiffness is added to all segments of the F-D
relationship, including the initial elastic stiffness. You may have to
adjust the initial elastic stiffness and the component actions to allow for
this.

PERFORM-3D User Guide 5-17

Chapter 5. Component Properties

5.6 Strength Loss

5.6.1 General
In a structural component, "brittle" strength loss can be caused by a
number of effects, including tensile fracture, concrete crushing,
concrete shear failure, and buckling. When a component loses strength,
the lost strength is redistributed to adjacent components (if this is
possible), and the resulting behavior can be complex. The behavior can
also be sensitive to small changes in the loads and/or the component
properties.

In PERFORM, strength loss is optional, As a general rule, you should

specify it only if it is essential.

Usually it is not permissible to deform an inelastic component beyond

the L point (i.e., the deformation capacity is usually smaller than the L
point deformation). For example, the FEMA 356 criteria generally
allow deformation beyond the L point only for certain secondary
members at the collapse prevention performance level. Hence, what
happens after the L point is reached may be academic, and there may
be no point in doing the calculations. For an exception to this rule, see
Section 5.8.

Figure 5.6 shows the action-deformation relationships for FEMA 356

(Q- relationship) and PERFORM (F-D relationship).

Figure 5.6 Strength Loss

5-18 PERFORM-3D User Guide

 Chapter 5. Component Properties

In the FEMA 356 relationship there is sudden strength loss at Point C,

and total strength loss at Point E. In the PERFORM relationship,
strength loss begins at Point L and can be sudden or gradual. It is likely
that strength loss in an actual structure will be gradual, and hence
sudden strength loss is not realistic. As a general rule you should avoid
specifying sudden strength loss. See the Section 5.7 for some problems
that can occur when there is sudden or rapid strength loss.

In the FEMA 356 relationship there is total strength loss at Point E. In

the PERFORM relationship total strength loss at Point X is optional, as
explained in the next section.
5.6.2 Total Strength Loss at X Point
In the PERFORM relationship there is no automatic strength loss
corresponding to FEMA 356 Point E. The PERFORM relationship has
Point X, but this is usually intended to be a maximum deformation,
beyond which the component is so badly deformed that there is no
point in continuing the analysis. For some components, however, there
is an option to reduce the strength to zero at this point. At the time of
writing this is limited to the following components.

(1) FEMA beam, steel type or concrete type.

(2) Moment hinge.
(3) Inelastic steel material (non-buckling).
(4) Inelastic shear material.

For the FEMA beam and moment hinge components, the moment
strength of the component reduces to zero at Point X. However, the
shear strength is not affected. For these components the effect of total
strength loss is the same as inserting a moment release.

If this option is used for an inelastic shear material (in a wall element),
the shear strength and stiffness reduce to zero at Point X.

The procedure for specifying total strength loss is as follows.

Choose a component type, and go to the Strength Loss page. If the

option for total strength loss is available for the component type, there
will be a Total Strength Loss at Point X frame. To get total strength
loss, choose the “Yes” option.

PERFORM-3D User Guide 5-19

Chapter 5. Component Properties

If you specify total strength loss at the X point, and if the component
has an added parallel stiffness (see Figure 5.5), this parallel stiffness is
still present. Hence, the strength does not actually reduce to zero.
5.6.3 Strength Loss Interaction
If a component loses strength in one direction, the strength in the other
direction may or may not be affected. For example, if a concrete beam
loses strength for negative bending because the concrete crushes, the
strength for positive bending may not be affected. For other
components this may not be the case. For example, if a component
factures in tension it may be best to assume that the strength in
compression is also reduced.

You can account for this type of interaction by specifying a value for
the strength loss interaction factor. A value of 0 means that strength
loss in one direction causes no strength loss in the opposite direction. A
value of 1 means that strength loss in one direction causes the same
amount of strength loss in the opposite direction (if the positive and
negative strength losses are different, strength loss in one direction
causes the same proportional strength loss in the opposite direction). A
value of 0.5 means that strength loss in one direction causes one half
that loss (or one half of the corresponding proportion) in the opposite
direction.

The procedure for specifying the interaction factor is as follows.

Choose a component type, and go to the Strength Loss page. If

strength loss interaction is available for the component type, there will
be a Strength Loss Interaction frame. Specify the interaction factor in
this frame. The default factor is zero.

To see the effect of this factor, it can be useful to plot hysteresis loops,
as described earlier in this chapter.

5.7 A Warning on Strength Loss

5.7.1 General
If you specify strength loss, you should be aware of the consequences
and pitfalls. Some of these are considered in this section.

5-20 PERFORM-3D User Guide

 Chapter 5. Component Properties

5.7.2 Components in Parallel

Figure 5.7(a) shows a simple structure consisting of an elastic
component in parallel with a yielding component that loses strength.
Figure 5.7(b) shows the load-displacement relationship for the
structure.

Figure 5.7 Effect of Strength Loss. Parallel Case.

Figure 5.7(b) shows the structure behavior for two cases, as follows.

(1) If the strength loss is gradual, so that the stiffness Ku is

numerically smaller than the elastic stiffness Ke, the structure load-
displacement relationship increases monotonically.
(2) However, if Ku is numerically larger than Ke, the structure strength
decreases. If Ku, is infinite, corresponding to sudden strength loss,
the structure loses strength suddenly.
5.7.3 Components in Series
In Figure 5.7 the components are in parallel, so that when the yielding
component loses strength it transfers load to the elastic component.
Figure 5.8 shows a structure with the components in series rather than
in parallel.

PERFORM-3D User Guide 5-21

Chapter 5. Component Properties

Figure 5.8 Effect of Strength Loss. Series Case.

In this case the structure load-displacement relationship always

decreases as the yielding component loses strength. If Ku is numerically
larger than Ke, the load-displacement relationship reverses, as shown in
Figure 5.8(b). If the strength loss is sudden, Ku is infinite and the
relationship reverses with a stiffness equal to Ke.
5.7.4 Effect on Analysis Method
Nonlinear analysis models for a structure often have behavior that is a
combination of the parallel and series cases. Experience indicates that
PERFORM is able to obtain solutions reliably when there is strength
loss, for both static and dynamic behavior. Dynamic analysis tends to
be less sensitive than static, because sudden strength loss can be
balanced by inertia and damping forces.

However, if reversal occurs, as in Figure 5.8(b) with Ku > Ke, the

behavior can be very sensitive numerically. PERFORM is usually able
to obtain a solution even for this type of behavior. However, this may
not always be the case, and the analysis may fail to converge.

If reversal does occur, it will usually be local rather than global. This is
a reason why it is always a good idea to use multiple “controlled drifts”
in static push-over analysis. If only a single drift is controlled, it may be
badly behaved, with reversal of the relationship between load and drift.
If multiple drifts are controlled, it is unlikely that all of them will be
badly behaved, even if there is local reversal.

Sudden strength loss is probably not a good idea. It rarely occurs in real
structural components, and as a general rule should be avoided.

5-22 PERFORM-3D User Guide

 Chapter 5. Component Properties

Progressive strength loss as shown in Figure 5.9 is suggested as an

alternative, with a strength at R equal to, say, 0.001 times the strength
at U.

Figure 5.9 Suggested Alternative to Sudden Strength Loss

5.7.5 Hinge Strength Loss in a Compound Component

Although complete structures usually have behavior that is a
combination of the parallel and series cases, there can be purely series
behavior at the element level, specifically in PERFORM compound
components.

If a frame compound component consists of elastic beam segments in

series with plastic hinge components that can lose strength, the hinges
and beam segments act in series, and the moment-rotation relationship
for the compound component could reverse as in Figure 5.8(b). This
does not happen when FEMA Beam or FEMA Column components are
used, because each FEMA component is automatically divided into a
rigid-plastic hinge component and an elastic beam component, and the
value of Ku for the hinge is chosen to give the specified Ku for the beam
as a whole. This means that Ku for the hinge is numerically smaller
than Ku for the beam as a whole. However, if you use plastic hinge
components directly, rather than FEMA components, you must specify
the Ku value for the hinge. If you make this value too large, the
moment-rotation relationship for the beam as a whole can reverse. You
should choose the value of Ku for the hinge so that this does not happen
(Ku for the hinge should be numerically smaller than 6EI/L for the
beam as a whole).

If Ku is too large and the moment-rotation relationship for the beam as

a whole reverses, PERFORM is usually able to obtain a solution.
However, this may not always be the case, and the analysis may fail to

PERFORM-3D User Guide 5-23

Chapter 5. Component Properties

converge. Behavior of this type also tends to increase the energy

balance error.

5.8 Column Ductile Limit Depends on Axial Force

5.8.1 Effect to be Modeled

If a column has a large axial force and is cycled inelastically, its ductile
limit can be expected to be smaller than for the same column with a
smaller axial force. This is illustrated in Figure 5.10.

Figure 5.10 Effect of Axial Force on Ductile Limit

Figure 5.10(a) shows a steel column with equal end moments, M, and a
chord rotation . Figure 5.10(b) shows the P-M interaction surface, and
two compression forces, PL and PU. Figure 5.10(c) shows an
elastically-perfectly-plastic moment-rotation relationship with strength
loss at the ductile limit. The yield moment is smaller at PU than at PL,
because of the shape of the interaction surface. The ductile limit is
smaller at PU than at PL, because the column is less ductile at the
larger compression force.

As noted earlier, if the deformation capacity is smaller than the ductile

limit, then there may be no need to model strength loss, since any
column with strength loss will have a deformation D/C ratio larger than
1.0, and hence will not satisfy the performance requirement. However,
FEMA 356 Section 3.4.3.2.1 states the following.

5-24 PERFORM-3D User Guide

 Chapter 5. Component Properties

“Primary and secondary component demands shall be within the

acceptance criteria for secondary components at the selected
Structural Performance Level.”

This section applies to the Nonlinear Static Procedure (NSP), where

demand/capacity ratios for component deformations are evaluated at
the estimated displacement demand (target displacement).

The FEMA 356 commentary on this section states the following.

“When all components are explicitly modeled with full backbone

curves, the NSP (Nonlinear Static Procedure) can be used to
evaluate the full contribution of all components to the lateral force
resistance of the structure as they degrade to residual strength
values. When degradation is explicitly evaluated in the NSP,
components can be relied upon for lateral-force resistance out to
the secondary component limits of response.”

Some engineers interpret this to mean that the deformation capacities

(acceptance criteria) for secondary members can be used for steel
columns. These capacities are larger than for primary members, and
also larger than the ductile limit deformation. If these capacities are
used, a steel column can be deformed beyond its ductile limit, and
hence strength loss must be modeled in the analysis, including the
effect of axial force on the ductile limit.
5.8.2 FEMA 356 Requirements
FEMA 356 Table 5-6 gives values for the plastic rotation, a, in Figure
5.10(c). For a compact section the FEMA 356 equations are as follows.

For P/PCL  0.20 , a  9θ y (5.1a)

 P 
For 0.20  P/PCL  0.50 , a  111  1.7 θ y (5.1b)
 PCL 

where P = axial compression force, PCL = compression strength of the

column, and θ y = yield rotation. The yield rotation is given by

L  P  
  θ y0 1  P


θ y  ZFy 1 (5.2)

6EI  AFy   AFy 
  

PERFORM-3D User Guide 5-25

Chapter 5. Component Properties

where Fy = material yield stress and Z, A, E, I and L have their usual

meanings, and θ y0 = yield rotation when P = 0.

Note the following points.

(1) In Equation 5.1b the axial strength is PCL, which includes the effect
of slenderness ratio and hence is a property of the member as a
whole, whereas in Equation 5.2 the axial strength is AFy, which is
a property of the cross section only.

(2) From Equations 5.1 and 5.2, the rotation at the ductile limit varies
with axial compression force essentially as shown in Figure 5.11a.
The ductile limit rotations in this figure are as follows, where the
approximate values are calculated assuming that PCL  AFy .

 0  9 y 0 (5.3a)
 0.2PCL 
1  9 y0 1    7.2 y0 (5.3b)
 AFy 
 
 0.2PCL 
 2  11 y0 1  1  1.7 0.2PCL   5.8 y0 (5.3c)

 AFy  PCL 
 0.5PCL 
 3  11 y0 1  1  1.7 0.5PCL   0.8 y0 (5.3d)

 AFy  PCL 

5-26 PERFORM-3D User Guide

 Chapter 5. Component Properties

Figure 5.11 Rotation at Ductile Limit

This can be simplified with no practical loss of accuracy, as shown in

Figure 5.11(b). In this figure the reduction in the ductile limit rotation
is based on AFy rather than PCL, with effective length effects accounted
for by considering P- effects in the analysis. Also, all rotations are
multiples of the yield rotation at zero axial force, which is constant, not
a yield rotation that varies with the axial force.

PERFORM allows the ductile limit to vary only as shown in Figure

5.11(b). If FEMA 356 is used, the curve in Figure 5.11(a) must be
approximated.

It may be noted that the FEMA 356 equations for rotation capacities in
columns are similar to the above equations, and they can also be
simplified. PERFORM allows the rotation capacities to vary with axial
force, but using only the simplified form.

5.8.3 PERFORM Implementation of Variable Ductile Limit

In PERFORM, the effect of axial force on the ductile limit can
currently be specified only for the following components.

(1) FEMA Column, Steel Type.

(2) PMM Hinge, Steel Rotation Type.
(3) PMM Hinge, Steel Curvature Type.

PERFORM-3D User Guide 5-27

Chapter 5. Component Properties

In PERFORM, the equivalent of the FEMA 356 rotation “a” is the

rotation, DL, at the ductile limit, Point L. This is different for a FEMA
column component and a PMM hinge component as follows.

(1) For a FEMA Column, Steel Type component the deformation D is

the total end rotation expressed as a multiple of the yield rotation.
Hence, DL  1  a/θ y .

(2) For a PMM Hinge, Steel Rotation Type component the

deformation D is the hinge plastic rotation. Hence DL = a.

(3) For a PMM Hinge, Steel Curvature Type component the

deformation D is the total curvature. Hence DL = ψ y + a, where
ZFy  
ψy  1  P  (5.4)
EI  AFy 


During each analysis (static or dynamic), the rotation and axial force
for each component are monitored at each step. Based on the axial
force, the ductile limit rotation, DL, is calculated. If the rotation
demand exceeds DL, the ductile limit has been reached, and the
component begins to lose strength. After this point, the rotation DL is
“frozen” (kept constant), and monitoring of the axial force ceases.
5.8.4 Procedure
The procedure for specifying the component properties is the same for
FEMA Column, Steel Type components and P-M2-M3 Hinge, Steel
Type components.

In the Strength Loss page you have the option of specifying that
strength loss depends or does not depend on the axial compression
force. If you choose the “No” option the strength loss properties are
fixed, and you specify only one set of properties. If you choose “Yes”
you must specify two sets of properties, one set at an upper
compression force, PU, and a second set at a lower compression force,
PL. Typically, PU will be 0.5PCL (or 0.5AFy in the simplified form) and
PL will be 0.2PCL (or 0.2AFy). For axial forces between PU and PL,
PERFORM interpolates linearly.

5-28 PERFORM-3D User Guide

 Chapter 5. Component Properties

As noted in the preceding section, at each step of an analysis (static or

dynamic), PERFORM calculates the DL value based on the current
axial force, until DL is reached. When DL is reached, the strength loss
properties are “frozen”, and do not change in later analysis steps.

5.9 Deformation and Strength Capacities

5.9.1 General
PERFORM-3D is a tool for both nonlinear analysis and performance
based structural design.

To use the performance based design capabilities, you must specify

deformation capacities for inelastic components and/or strength
capacities for elastic components. You can then define limit states to
make use of these capacities, and process the analysis results to give
demand-capacity ratios (or limit state usage ratios). This section gives
an overview of the procedure.

5.9.2 Demand and Capacity

Structural design decisions are made by comparing demand and
capacity. To do this we need (a) demand-capacity measures (or
engineering demand parameters), (b) demand values and (c) capacity
values.

Demand-capacity measures can be classified as deformation measures

(e.g., plastic hinge rotation) and strength measures (e.g., bending
moment).

Strength measures have been used for many years, and methods for
obtaining demands (usually linear structural analysis) and capacities
(usually design code equations) are well established.

Some deformation measures, such as drift and deflection, have also

been used for many years. However, engineers have less experience
with deformation measures such as plastic hinge rotation. On the
demand side, to obtain rotation demands for a plastic hinge we must
use nonlinear structural analysis. On the capacity side, we do not yet
have definitive design code equations for rotation capacities (although

PERFORM-3D User Guide 5-29

Chapter 5. Component Properties

FEMA 356 will become a design code). Hence, the calculation of

deformation demand-capacity ratios for deformation-based design is
more complex, and requires greater judgment than the calculation of
strength demand-capacity ratios for conventional strength-based design.

PERFORM-3D has extensive capabilities for calculating demand-

capacity ratios, considering both deformation and strength. You can
save a lot of time if you make effective use of these capabilities.

5.9.3 Deformation Capacities

If a component is allowed to be inelastic, its structural adequacy is
determined by its ductility rather than its strength. For all inelastic
components in PERFORM-3D you can specify deformation capacities
for up to 5 performance levels, identified as Level 1 through Level 5.
Often there will be 3 levels, with Level 1 corresponding to Immediate
Occupancy, Level 2 to Life Safety, and Level 3 to Collapse Prevention.
However, this is not required.

Most components have constant deformation capacities, but for some

components the capacity can be affected by interaction. For example, in
a reinforced concrete beam a large shear force can reduce the bending
ductility. Hence it may be necessary to specify the rotation capacity of
a plastic hinge component as a function of the shear force at the hinge.
In a number of components, PERFORM-3D allows you to account for
interaction of this type when you specify capacities.

5.9.4 Drift Capacities

If you specify deformation capacities at the component level you can
get detailed information about the deformation distribution in the
structure, and hence identify "hot spots" that have the largest demand-
capacity ratios. You may decide, however, that this is not needed, and
that you can assure satisfactory performance by considering only
structure drifts, not individual element deformations. You can do this
by specifying one or more drifts, and assigning drift capacities (use the
Drifts and Deflections task to define drifts, and the Limit States task
to specify drift capacities). Most often you will use both drifts and
component deformations as demand-capacity measures.

5-30 PERFORM-3D User Guide

 Chapter 5. Component Properties

5.9.5 Strength Capacities

If a component is required to remain essentially elastic, its structural
adequacy is determined by its strength. If the strength demand exceeds
the strength capacity, then the component does not remain essentially
elastic.

For most elastic components in PERFORM-3D you can specify

strength capacities. You can then calculate strength demand-capacity
ratios for any elements that use these components.

The procedure is different for Frame Compound Components. For

these components, strength is checked using Strength Section
components. In a compound component for, say, a beam element, you
can place strength sections for bending and/or shear at a number of
points along the element. For example, you might assume that plastic
hinges form in the beam only at the column faces, and that the beam
remains essentially elastic in shear. Hence, the only inelastic structural
components would be plastic hinge components, placed at the column
faces. You could then place shear strength sections at the column faces,
and one or more moment strength sections near midspan. If the
demand-capacity ratios for the strength sections are less than one, the
behavior assumptions for the beam are correct. If not, there may be
shear failure near the beam ends or moment hinging near midspan.

As with deformation capacities, you can specify strength capacity

values for up to 5 performance levels. However, these values are now
capacity factors. You can treat them as capacity reduction factors
applied to a nominal strength ( factors in conventional design) or as
factors to account for ductility (m factors in FEMA 356). You first
specify a nominal strength, then specify up to 5 capacity factors,
corresponding to different performance levels.

5.9.6 Overview of Limit States and Usage Ratios

If you have a large structure you can have a large number of component
capacities and hence a large number of demand-capacity ratios. Limit
states allow you to organize the demand-capacity ratios into
manageable packages.

For example, you might create a structure with a number of drifts, a

number of plastic hinges with rotation capacities for three performance

PERFORM-3D User Guide 5-31

Chapter 5. Component Properties

levels, and a number of shear strength sections with strength capacities

for one performance level. For this structure you might set up the
following limit states.

(1) Drift limit state for Immediate Occupancy (IO), considering all
drifts.
(2) Drift limit state for Life Safety (LS), considering all drifts.
(3) Drift limit state for Collapse Prevention (CP), considering all
drifts.
(4) Hinge rotation limits state for IO, considering all beams and using
the Level 1 deformation capacities.
(5) Hinge rotation limits state for LS, considering all beams and using
the Level 2 deformation capacities.
(6) Hinge rotation limits state for CP, considering all beams and using
the Level 3 deformation capacities.
(7) Hinge rotation limits state for IO, considering all columns and
using the Level 1 deformation capacities.
(8) Hinge rotation limits state for LS, considering all columns and
using the Level 2 deformation capacities.
(9) Hinge rotation limits state for CP, considering all columns and
using the Level 3 deformation capacities.
(10) Shear strength limit state, considering all beams.

During each structural analysis, PERFORM-3D calculates all needed

demand-capacity ratios, and expresses the results as usage ratios for the
limit states. For example, for Limit State (4) in the above list,
PERFORM-3D calculates the demand-capacity ratios for all plastic
hinge components in all beams, and saves the maximum value at each
step of the analysis. This maximum value is the usage ratio for the limit
state.

For details on setting up Limit States see Chapter 10, Limit States and
Usage Ratios.
5.9.7 Note on Maximum Demand-Capacity Ratio
In the Analysis phase, the maximum value for any demand-capacity
ratio is 30. In the unlikely event that a calculated ratio exceeds 30, a
value of 30 is used. If you are not sure what to specify for a
deformation or strength capacity, you may be tempted to specify an
arbitrary value and to see what happens. If you specify too small a
value, you may hit the 30 limit, and you will not get useful information.
To avoid unrealistic demand-capacity ratios, you should take care to

5-32 PERFORM-3D User Guide

 Chapter 5. Component Properties

specify reasonably realistic values for deformation and strength

capacities.

5.10 Beam or Column Shear Strength Depends on

Hinge Rotation

5.10.1 General
In a reinforced concrete beam or column, the shear strength is
usually smaller in plastic hinge regions than outside those regions.
If the behavior in shear is required to remain essentially elastic, the
shear strength can be checked using a shear force strength section.

Usually the shear strength will be a fixed value. However, a shear

force strength section can be linked to an adjacent flexural plastic
hinge component, and the shear strength can then vary, depending
on the rotation of the hinge.

This feature can be used for Shear Force Strength Section and V2-V3
Shear Strength Section components. It does not apply to inelastic shear
hinges.

5.10.2 PERFORM Implementation

To use this feature, properties must be specified in two places, as
follows. Details are considered in the next section.

(1) In the strength section properties. See the Rotation Effect page.
The properties on this page specify how the shear strength of the
strength section component is affected by hinge rotation. The
specific hinge is not considered at this stage, only the effect of
rotation.
(2) In the properties of any Frame Member Compound Component
that uses the strength section. See the Strength Sections page.
When you use a shear strength section, you can specify which
rotational hinge component affects its shear strength. The strength
section and the hinge will usually have the same location, but
PERFORM does not check this.

PERFORM-3D User Guide 5-33

Chapter 5. Component Properties

At each step in an analysis, PERFORM checks whether a shear strength

section is affected by the rotation of a hinge component. If so,
PERFORM gets the maximum rotation of the hinge up to the current
step (positive or negative rotation, whichever is larger), uses this
rotation to calculate the strength capacity of the shear strength section,
and hence calculates the shear strength Demand/Capacity ratio for the
strength section.

5.10.3 Procedure for Strength Section Properties

Consider a V2-V3 Shear Strength Section, for a column. A Shear Force
Strength Section, for a beam, is similar.

The steps for specifying the strength section properties are as follows.

(1) In the Component Properties task go to the Strength Sections

page and the V2-V3 Shear Strength Section or Shear Force
Strength Section type. Either start a new strength section or
choose an existing one.
(2) Go to the Rotation Effect page.
(3) Choose the Option for Rotation Effect. The shear strength of a
concrete section will usually have contributions from the concrete
and the shear reinforcement. Choose whether hinge rotation affects
only the concrete contribution (which is usually the case) or the
total shear strength.
(4) Figure 5.12 shows how the shear strength varies with hinge
rotation. Specify the rotations r0, r1 and r2 (in radians) and the
strength reduction factors s1 and s2.
(5) Specify the Axis 2 Rotation Factor. The default is 1.0, which
means that the effective rotation of a P-M-M hinge in a column is
the resultant rotation. If a given rotation about Axis 2 of a column
causes a smaller shear strength reduction than the same rotation
about Axis 3, specify an Axis 2 Rotation Factor larger than 1.0.
The effective rotation is the resultant of (a) the Axis 3 rotation and
(b) the Axis 2 rotation divided by the Rotation Factor. If an Axis 2
rotation causes a larger shear strength reduction, specify a Rotation
Factor smaller than 1.0. This factor is not needed for a shear
strength section in a beam.
(6) Save the properties in the usual way.

5-34 PERFORM-3D User Guide

 Chapter 5. Component Properties

The above steps and Figure 5.12 refer to hinge rotation. As noted in the
next section, the rotation is not limited to the rotation of a plastic hinge
component.

Figure 5.12 Effect of Rotation on Shear Strength

5.10.4 Procedure for Frame Compound Component

The steps for associating a shear strength section with a hinge or other
rotational component are as follows.

(1) In the Component Properties task go to the Compound page and

the Frame Member Compound Component type. Either start a
new component or choose an existing one.
(2) Specify the basic components in the usual way. Some of these
components will be of rotational type. These can include FEMA
beams, FEMA columns, steel or concrete moment hinges, steel or
concrete PMM hinges, and also fiber beam or column segments.
Any of these can be associated with shear strength sections, not
just hinge components.
(3) Go to the Strength Sections page. If you add a shear strength
section, in addition to the usual name and location properties, you
can specify a range of basic components for calculating rotation.
For example, if the second basic component is a PMM hinge, and
if you specify a range from 2 to 2, the rotation of the hinge will be
used. The shear strength section in this case should be at or close to
the PMM hinge component, but PERFORM does not check this.
As a second example, if the second and third components are fiber
segments, and if you specify a range from 2 to 3, the sum of the
rotations across the segments will be used. You can specify any

PERFORM-3D User Guide 5-35

Chapter 5. Component Properties

range, covering components of any type or types. The rotation is

the sum of the rotations over all of the components in the range.

Be sure that you define one or more Strength Limit States that use the
shear strength sections.

If you specify a compound component and later edit it by changing its

basic components, be sure to check that the rotational component
ranges for any strength sections are still correct, since the rotational
components are identified by sequence number, and these numbers may
have changed.

5.11 Cyclic Degradation

5.11.1 Backbone Relationship

Under cyclic loading, most structural components will undergo
stiffness and strength degradation. Note that strength degradation
caused by cyclic loading is different from "brittle" strength loss at the
ductile limit.

As a general rule the degradation will be progressive. In effect, a

component will have an initial force-deformation relationship that
applies for the first loading cycle, or for "monotonic" loading where
there is no cycling. As the number of cycle increases, the effective
force-deformation relationship will degrade, with smaller stiffness,
strength, energy dissipation and/or ductility.

Ideally the force-deformation relationship would account for this

progressive degradation, and it is possible that this will be done in
future nonlinear analyses. At present, however, it is accepted practice to
use a single force-deformation relationship that accounts implicitly for
cyclic loading, assuming a certain amount of cycling. This is usually
referred to as the "backbone" relationship.

The F-D relationship for analyses in PERFORM is the backbone

relationship. FEMA 356 suggests backbone relationships for
components of a variety of types, and also provides guidelines for
constructing this relationship from experimental results (see FEMA 356
Section 2.8).

5-36 PERFORM-3D User Guide

 Chapter 5. Component Properties

The backbone relationship accounts implicitly for cyclic degradation on

the effective strength and ductility. It does not, however, account for
energy degradation in the hysteresis loop. This is considered in the next
section.

5.11.2 Hysteresis Loops

Figure 5.13 shows two hysteresis loops . In one loop the stiffness
degrades and in the other loop it does not. The amount of energy that is
dissipated (the area of the loop) is smaller for the degraded loop. The
amount of energy degradation is related to the amount of stiffness
degradation.

ACTION

No stiffness degradation

With stiffness degradation

DEFORMATION

Figure 5.13 Hysteresis Loop With Stiffness Degradation

Since the seismic response of a structure can be sensitive to the amount

of energy dissipation, it may be important to account for stiffness
degradation in the hysteresis loop. Different methods are needed for
dynamic analysis and static push-over analysis, as considered in the
following sections.

For more details see the document "PERFORM Components and

Elements".

PERFORM-3D User Guide 5-37

Chapter 5. Component Properties

5.11.3 Dynamic Analysis

For step-by-step dynamic analysis, if stiffness and energy degradation
is important it must be accounted for directly, by changing the shape of
the hysteresis loop as indicated in Figure 5.13.

In PERFORM-3D you can do this by specifying Energy Degradation

Factors for inelastic components. In Figure 5.13 the Energy
Degradation Factor is the ratio between the area of the degraded
hysteresis loop and the area and the non-degraded loop. For a typical
component, this area ratio is 1.0 for small deformation cycles (no
degradation) and gets progressively smaller as the maximum
deformation increases (increasing degradation).

In PERFORM, energy degradation is an option. If you choose "yes" for

this option you must specify the relationship between the maximum
deformation of the component and the energy degradation factor.
During an analysis, for each inelastic component and at each time there
is an unloading event for that component (i.e., each time a new
hysteresis loop begins) PERFORM determines the maximum
deformation, considering all cycles up to that point, not just the current
cycle. PERFORM then calculates the energy degradation factor for that
deformation, and adjusts the component stiffnesses to give this area
ratio for the new loop.

A good way to see the effect of degradation is to create a component

with energy degradation factors, and use the Plot Loops feature to draw
hysteresis loops. You can also do this if you are using component
properties that are based on experimental results, and you want to
calibrate the degrading behavior.

5.11.4 Procedure for Degradation Factors

Use the Component Properties task and the Inelastic component
page. Choose a component type, and go to the Cyclic Degradation
page.

You have two options for specifying the energy degradation factors, as
follows.

(1) If the component F-D relationship is trilinear with strength loss

you can specify the degradation factor at each of the Y, U, L, R

5-38 PERFORM-3D User Guide

 Chapter 5. Component Properties

and X points. This is the "YULRX" option. You can also use this
option for e-p-p behavior with strength loss (specify the same
factor for the Y and U points).
(2) You can specify degradation factors at the Y point, the X point,
and at 3 intermediate points. This is the "YX+3" option. If there is
no strength loss, this is the only available option.

5.11.5 Push-Over Analysis

In a static push-over analysis there are no cyclic loads or hysteresis
loops, so energy degradation must be accounted for indirectly. In most
cases you will use the degradation method that is built into the push-
over method. For an explanation of push-over methods, including how
they account for energy degradation, see Chapter 28, General Push-
Over Plots.

An alternative, which can be used only with the Capacity Spectrum

push-over method, is to use the PERFORM energy degradation factors.
If you choose this alternative, PERFORM uses the following
procedure.

(1) At each step of the push-over analysis, and for each inelastic
component, get the maximum deformation experienced by the
component, and hence calculate the energy degradation factor.
(2) For each component estimate the dissipated energy for the non-
degraded loop. Note that the area of this loop is estimated, not
explicitly calculated, since there is no actual loop. Given the
current deformation for the component, PERFORM assumes a
loop that has the same deformation in the opposite direction.
(3) For each component, multiply the non-degraded energy by the
degradation factor.
(4) Sum these dissipated energies over all components.
(5) Calculate the effective push-over strain energy in the structure. If
the push-over forces are H  and the corresponding nodal
displacements are  , this energy is 0.5H T .
(6) Given the dissipated and strain energies, calculate the equivalent
damping ratio = (dissipated energy)/(4 (strain energy)). This
equation assumes that dissipated energy from an inelastic
hysteresis loop has the same damping effect as the same amount of
energy dissipated by viscous damping.

PERFORM-3D User Guide 5-39

Chapter 5. Component Properties

(7) Use this damping ratio to calculate the spectral acceleration

demand in the push-over analysis.

Since this method takes into account the properties of each specific
structure, it may have an advantage over the "one size fits all" formulas
in the capacity spectrum method.

5.12 Control of Hysteresis Loop Shape

5.12.1 Alternative Loops

Given an energy degradation factor, the unloading and reloading
stiffnesses in a hysteresis loop can be chosen to give the required area
ratio. There are, however, many ways in which the stiffnesses can be
adjusted. PERFORM gives you some control over the way that this is
done.

Figure 5.14 shows possible hysteresis loops for a component with a

bilinear force-deformation relationship, all with the same amount of
energy degradation.

In this figure the loop with no energy degradation is shown dashed, and
three alternatives are shown for degraded loops. The degraded loops all
have the same area, and hence the same ratio of degraded to non-
degraded energy (about 0.55 in the figure, where 1.0 = no degradation).
However, the loops have very different shapes. The differences are as
follows,

(1) In Loop A the unloading stiffness is the same as in the non-

degrading case. The unloading elastic range is reduced to provide
the required energy degradation.
(2) In Loop B the elastic range on unloading is a maximum. The
unloading stiffness is reduced to provide the required energy
degradation.
(3) Loop C is midway between Loops A and B.

5-40 PERFORM-3D User Guide

 Chapter 5. Component Properties

Figure 5.14 Loops with Same Energy Degradation

In PERFORM, the default option corresponds to Loop C. If you wish,

you can specify an "unloading stiffness factor" that controls the loop
shape. A value of zero for this factor (the default) corresponds to Loop
C, a value of +1 corresponds to Loop A (maximum stiffness, smallest
elastic range), and a value of -1 corresponds to Loop B (minimum
stiffness, largest elastic range).

5.12.2 E-P-P Relationship

Figure 5.14 shows hysteresis loops for the bilinear case. These loops do
not apply for an elastic-perfectly-plastic F-D relationship. For an e-p-p
relationship the hysteresis loop degrades as shown in Figure 5.15.

This is the same as Loop B in Figure 5.14. The Loop A or Loop C type
of relationship can not be obtained for a truly e-p-p component. If you
specify a truly e-p-p F-D relationship, the unloading stiffness factor is
not used.

You can, however, get loops of Types A and C if you specify a trilinear
F-D relationship that is nearly e-p-p, with a small hardening stiffness.

PERFORM-3D User Guide 5-41

Chapter 5. Component Properties

Figure 5.15 E-P-P Relationship

5.12.3 Trilinear Relationship After U Point

The loops in Figure 5.14 apply for a trilinear F-D relationship before
the U point is reached. If a trilinear component is deformed past the U
point, the hysteresis loop degrades as shown in Figure 5.16.

Figure 5.16 Trilinear Relationship Beyond U Point

Figure 5.16(a) shows the non-degraded loop. This is divided into a

parallelogram and a pair of triangles, as shown. The parallelogram is
degraded as for the E-P-P component in Figure 5.15. The two triangles
make up a bilinear loop as in Figure 5.14, which is degraded using the
unloading stiffness factor. The result for the case of Loop A (unloading
stiffness factor = +1) is shown in Figure 5.16(b).

5-42 PERFORM-3D User Guide

 Chapter 5. Component Properties

5.12.4 Procedure for Unloading Stiffness Factor

Use the Component Properties task and the Inelastic component
page. Choose a component type, and go to the Cyclic Degradation
page. If the option to control the loop shape is available for the
component type, there will be an Unloading Behavior frame.

In this frame, specify the value for the Unloading Stiffness Factor. You
must also specify energy degradation factors, using either the YULRX
or YU+3 option.

After you have specified the component properties and pressed Save , it
is a good idea to press the Plot Loops button, and to plot hysteresis
loops to check the cyclic behavior. If the Plot Loops button is not
shown, first press Graph to plot the component properties.

This feature is currently limited to the following components.

(1) FEMA beam, steel type or concrete type.

(2) Moment hinge.
(3) Inelastic steel material.
(4) Inelastic shear material.

5.13 Cross Sections

5.13.1 The Role of Cross Sections

In a computer program for linear analysis you usually specify the
properties of an element by assigning it a cross section and an elastic
material. In most cases the element properties are then completely
defined. It is not so simple for nonlinear analysis, because more
properties are needed. Linear analysis needs only stiffness properties,
whereas nonlinear analysis needs both stiffness properties and strength
properties. Nevertheless, PERFORM-3D uses cross sections in
essentially the same way as a linear analysis program.

If an element is based on a cross section and consists of one basic

component (e.g., a panel zone element), it typically requires three steps
to assign the element properties, as follows.

(1) Define cross sections.

PERFORM-3D User Guide 5-43

Chapter 5. Component Properties

(2) Define basic components, using cross section properties.

(3) Assign basic components to element.

If a beam or column element consists of a compound component it

could require up to six steps to assign the element properties. For
example, if a beam element has plastic hinges, the steps might be as
follows.

(1) Define beam cross sections, with stiffness and strength properties.
(2) Define plastic hinge components, using cross section properties.
(3) Define shear strength sections, using cross section properties.
(4) Define compound components, using plastic hinge components
(for hinges), beam cross sections (for beam segments between
hinges) and strength sections (to check shear strength).
(5) Assign compound components to elements.

As another example, if a beam element uses fiber cross sections the

steps might be as follows.

(1) Define materials.

(2) Define fiber cross sections, using material properties.
(3) Define non-fiber beam cross sections.
(4) Define shear strength sections, using non-fiber cross section
properties.
(5) Define compound components, using fiber cross sections (for
hinge regions), non-fiber cross sections (for beam segments
between hinge regions) and strength sections (to check shear
strength).
(6) Assign compound components to elements.

The process is thus more complex than for a linear analysis, and
requires more careful planning.

5.13.2 Beam and Column Sections

In a linear analysis you may not need to distinguish between cross
sections that are used for beam elements and those that are used for
column elements. In PERFORM-3D it is necessary to make this
distinction, because the strength properties required for a column are
more complex than those required for a beam. For example, a beam
element usually has a small axial force, and is bent about only one axis.
Hence, in defining the bending strength for a beam section it is

5-44 PERFORM-3D User Guide

 Chapter 5. Component Properties

sufficient to consider uniaxial bending with no P-M interaction.

However, a column element can have large axial forces and is usually
bent about both axes. Hence, in defining the bending strength for a
column section it is necessary to consider biaxial bending and P-M-M
interaction.

For any beam or column section, you must define specify the stiffness
properties, but it is optional to specify the strength properties. If you
specify only the stiffnesses, there is no difference between a beam and
column section. However, it is usually a good idea to specify cross
section strengths, because these strengths can be used to define the
strengths of other components. For example, if you specify a P-M-M
yield surface for a column cross section, and if you create a P-M-M
plastic hinge component that uses the cross section, the hinge
component gets its yield surface from the cross section.

For steel beam and column sections you can specify the cross section
properties directly, or you can make use of standard steel sections.
Details of the procedure are presented later.
5.13.3 Fiber Cross Sections for Beams and Columns
In a fiber section for a beam or column, the cross section shape is
divided into a number of fibers, and material properties are assigned to
each fiber. Several different materials can be used, including concrete,
non-buckling steel, and buckling steel. For beams, short segments with
fiber sections can be used to model hinge regions and also connections,
including connections with fracture. For columns, fiber sections can be
used to model hinges with P-M-M interaction that do not rely on a
yield surface and plasticity theory. It is, however, substantially more
work to use a fiber section than a standard beam or column section, and
the computer time required for analysis may be larger.
5.13.4 Fiber Cross Sections for Walls
Shear wall and general wall compound components have stiffness and
strength in shear and in combined axial-bending. For the shear
properties you must first specify a shear material, then specify an
effective wall thickness for shear. For axial-bending you must define a
cross section. This is a "fiber" cross section, usually made up of steel
and concrete fibers. Essentially, you must define steel and concrete
material properties, then specify the area, location and material for each
fiber in the cross section.

PERFORM-3D User Guide 5-45

Chapter 5. Component Properties

5.13.5 Other Sections

Other types of cross section may be used, including sections for elastic
membrane and elastic shell elements.

5.14 Using Cross Section Dimensions

5.14.1 General
When you create a beam or column cross section, you can specify its
breadth (B) and depth (D) (for standard sections the dimensions are
obtained automatically). When you create certain other components
(e.g., plastic hinges) you can associate the component with a cross
section, and the component then inherits the dimensions from the cross
section. You can use these dimensions in the following ways.

(1) To assign lengths to basic components in a compound component.

(2) To determine the properties of connection panel zones.
(3) To assign lengths to stiff end zones for beam and column elements.

The first of these uses is considered in this section. Panel zones and
stiff end zones are considered in later sections.

5.14.2 Length of Uniform X-Section Segment

Frame compound components will usually consist of segments with
uniform cross sections, separated by components such as hinges. You
specify the beam or column cross section for each segment, and this
defines the segment stiffness. If you wish, you can also use the cross
section dimensions to specify the segment length.

For example, consider a beam member with a Reduced Beam Section

(RBS) located at some distance from the column face. This distance is
often specified as a multiple of the beam flange width. Part of a
possible frame compound component for this member is shown in
Figure 5.17.

5-46 PERFORM-3D User Guide

 Chapter 5. Component Properties

Specified multiple of
beam flange width

Reduced beam section

Uniform X-section segment.

Length = multiple of
section B dimension.

Stiff end zone. Length can Zero length

be based on column size. plastic hinge.

Figure 5.17 Model Using Component Dimensions

The beam member will be modeled using a frame compound

component. At each end of the compound component there is (a) an
end zone that extends from the end node to the to the column face, (b) a
uniform X-section segment that extends to the center of the RBS, and
(c) a zero-length plastic hinge to represent the RBS. The flange width
in the uniform X-section segment is the B dimension for the cross
section. One of the options for the length of a uniform X-section
segment is to specify a fixed length, which you could calculate as a
multiple of the cross section breadth. If you choose this option, and if
some time later you change the size of the cross section, you must also
change the length for the segment. This requires changes in two places,
and can make it difficult to maintain consistency. Another option for
the length of a uniform X-section segment is to specify the length as a
multiple of the cross section breadth. If you choose this option, and if
some time later you change the size of the cross section, the segment
length is automatically updated. Hence, by using the cross section
dimensions you can save substantial effort.

5.14.3 Tributary Length for a Curvature Hinge

In PERFORM-3D a plastic hinge can be of "rotation" type, in which
case it automatically has zero length, or of "curvature" type, in which

PERFORM-3D User Guide 5-47

Chapter 5. Component Properties

case you must specify a tributary length. In some cases the tributary
length may be a multiple of the beam or column depth, D. As in the
preceding section, if you associate the plastic hinge component with a
cross section, the depth D becomes a property of the hinge component.
You can then specify the tributary length of the hinge as a multiple of
the D dimension. If you change the cross section, the tributary length is
automatically updated.

Note that in this case you probably will get not only the hinge
component dimensions from the cross section but also some of the
hinge component strength data. If you change the cross section, the
strength data is also updated.

5.15 "Auto" Components

5.15.1 Auto End Zones for Beams and Columns

In PERFORM-3D, end zones are elastic components that are stiff but
not rigid. They can be used at the ends of frame compound
components.

In most cases you will not need to define any end zone components
(although you can if you wish). The reason is that PERFORM-3D
automatically sets up a default end zone component that can be used for
beam and column elements. This component has a stiffness that is 10
times larger than the body of the component, and an "auto" length that
is obtained from the dimensions of the adjacent beams and columns.
Default end zones in frame compound components that are used for
beam elements get their lengths from the depths of the adjacent
columns, and default end zones in components that are used for column
elements get their lengths from the depths of the adjacent beams. For
brace/other frame elements (e.g., actual bracing members), you can not
use default end zones.
5.15.2 Auto Panel Zones in Steel Frames
The properties of a connection panel zone in a steel frame depend on
the properties of the connected beams and columns. When you specify
a "regular" panel zone component in a steel frame, you have the option
of specifying cross sections for the beam and column (these must be I-
shaped sections). If you use this option, PERFORM-3D uses the cross

5-48 PERFORM-3D User Guide

 Chapter 5. Component Properties

section properties to calculate the panel zone stiffness and strength. If

you change a cross section, the panel zone properties are automatically
updated.

You must, however, be careful. PERFORM-3D uses the cross sections

that you specify with the properties of the panel zone component, not
the cross sections of the connected beams and columns. At each beam-
to-column connection you must be careful to specify the same cross
sections for the panel zone as for the for the beam and column. If there
are several beam and column sizes in the structure, you will also need
several panel zone components, one for each different combination of
beam and column size.

An alternative is to use the "auto" panel zone component. With this

component, PERFORM gets the beam and column sizes at each
connection, and creates a different panel zone component for each
different combination. This can greatly reduce the effort required to
specify panel zones, at the expense of some loss of control over the
process.

The steps for specifying an Auto Panel Zone component are as follows.

(1) In the Component Properties task go to the Inelastic page and

the Auto Connection Panel Zone component type. Either start a
new component or choose an existing one.
(2) Specify the component properties in much the same way as for
other components. In the Formulas page you can specify
parameters that control the stiffness and strength. Only the basic F-
D relationship is calculated automatically. You must still specify
strength loss, deformation capacities and cyclic degradation as for
a standard panel zone component.
(3) Use Save and SaveAs in the usual way.

If all panel zones in a structure have the same column yield stress, the
same doubler plate thickness and yield stress, and the same data for
strength loss, deformation capacities and cyclic degradation, you will
need only one Auto Panel Zone component.

The F-D relationship is not drawn for this component type (if you press
Graph, nothing happens), because the F-D relationship is not yet
known. This means that you can not immediately check that the

PERFORM-3D User Guide 5-49

Chapter 5. Component Properties

component properties are correct, and that you have to exercise more
care. You can (and should) check the properties as follows.

(1) Run one or more analyses, in order to create an ECHO.txt file.

(2) Open the ECHO.txt file. Go to the section for inelastic panel zone
component properties (search for "IZ1A").
(3) If you have any "regular" panel zone components, their properties
are listed first. A set of properties is then listed for each different
beam-and-column combination where auto panel zone components
are used. The name for each set of properties consists of the
number of the column section (search for "COLM" in ECHO.txt to
see the section details), followed by the depth of the beam section,
and followed by the name of the Auto Panel Zone component. If
this name is long it may be truncated, since a component name is
limited to 40 characters.
(4) Check that the properties are correct.
(5) After you have run some analyses. use the Time History and/or
Hysteresis Loops tasks to plot results for typical panel zone
elements, to check that the behavior is correct.
5.15.3 Panel Zone Limit States
For defining limit states, non-auto and auto panel zone components are
different components. Be sure to specify the correct component type
when you define limit states. If you use both auto and non-auto
components in a structure, you can have separate limit states for each
type if you wish.

5.16 Upper and Lower Bounds

5.16.1 Overview
There can be substantial uncertainty about the strengths and stiffnesses
of components and elements, and you may want to run analyses with
different strength and/or stiffness values to assess sensitivity. One way
to do this is to set up different analysis models with different strengths
and stiffnesses. An easier way, in most cases, is to specify upper and
lower bounds. You can then vary strength and stiffness without setting
up multiple analysis models.

5-50 PERFORM-3D User Guide

 Chapter 5. Component Properties

Upper and lower bounds can be specified for certain component

properties. Inelastic components and materials can have bounds on both
strength and stiffness. Elastic components and materials can have
bounds on stiffness. Strength sections can have bounds on strength.

Upper and lower bounds can be specified for most, but not all,
components.

There are two steps in the process, as follows.

(1) When you specify component properties, you have the option of
specifying upper and lower bounds. For a typical inelastic
component, upper and lower bounds can be specified as multiples
of the nominal strength and stiffness. The multiple for upper bound
values must be between 1.0 and 2.0 (with a few exceptions). The
multiple for lower bound values must be between 0.5 and 1.0.

(2) When you define a new Analysis Series (see Chapter 17, Analysis
Series) , you can specify that strengths and stiffnesses other than
the nominal values are to be used. For any inelastic component you
can specify U/L Bound Ratios for strength and stiffness. A zero
ratio means that the nominal values are to be used. A ratio of 1.0
means the upper bound values, and a ratio of -1.0 means the lower
bound values. A ratio between -1.0 and 1.0 means interpolation
between the upper and lower bounds.

The default is to use the nominal values.

5.16.2 Procedure for Specifying Upper/Lower Bounds
Use the Component Properties task. If a component allows upper and
lower bounds, there will be an Upper/Lower Bounds option (“Yes” or
“No”) and an Upper/Lower Bounds properties page.

A relatively simple component such an inelastic bar allows upper and

lower bound values for axial strength and axial stiffness. A more
complex component such as a P-M-M hinge has separate strength
bounds for bending and axial force.

To specify bounds, choose the “Yes” option for Upper/Lower Bounds,

then go to the Upper/Lower Bounds page. In addition to the upper and
lower bound values, there are options that define the effects of the
bounds on the following parameters.

PERFORM-3D User Guide 5-51

Chapter 5. Component Properties

(1) Deformation at U Point, DU. This applies for trilinear F-D

relationships only. The options are (a) keep DU constant as the
strength and/or stiffness are varied, (b) keep the ratio DU/DY
constant, and (c) keep the deformation DU–DY constant.
(2) Deformation at L Point, DL. This applies only for cases with
strength loss. The options are (a) constant DL, (b) constant
DL/DY, (c) constant DL-DY, (d) constant DL/DU (for trilinear
relationships only) and (e) constant DL-DU.
(3) Deformation at X Point, DX. The options are the same as for DL.
(4) Deformation capacities. The options are the same as for DL.

To show the effects of the bounds you can specify U/L Bound Ratios in
the Upper/Lower Bounds page. These are the same as the ratios that
can be specified for an Analysis Series (i.e., values between -1 and 1).
To see the effect, specify U/L Bound Ratios for strength and/or
stiffness, and check the Show Effect box. Then, when you press Check
or Graph, the plotted properties show the U/L bound effects. If you
plan to use U/L bounds, be sure to use this feature to check that the
scaled F-D relationships and other properties are as expected.

Upper and lower bounds are not considered if you use the Plot Loops
feature. For this feature, only the nominal values are used.
5.16.3 Procedure for Using Upper/Lower Bounds
This section has described how to specify Upper/Lower bounds as
component properties. The procedure for using these bounds is
described in Chapter 17, Analysis Series, Section 17.4.

5-52 PERFORM-3D User Guide

 Chapter 6. Elements

6 Elements

This chapter describes how to specify elements. There are three

main parts to the data for an element, namely overall location, local
axis orientation and component properties. If you wish you can
specify element locations before you specify component properties.
However, you will not be able to assign properties to elements until
you have specified the required components.

6.1 Element Types

6.1.1 General
This section presents brief descriptions of the available element types.
For more detailed descriptions see the document PERFORM
Components and Elements.

6.1.2 Simple Bar Elements

Use Simple Bar elements to model members that resist axial force only.
Each element connects two nodes and consists of one basic component.
Bar elements do not need local axes.

You can use a number of different basic components for bar elements,
including simple elastic bars, inelastic bars, elastic gap-hook bars,
nonlinear elastic bars, concrete strut components that use concrete
materials, and steel bar/tie/strut components that use non-buckling or
buckling steel materials.

You can specify initial strains in bar elements (essentially the same as
thermal expansion effects). This can be useful for modeling bars that
have initial stresses from prestressing or other causes.

6.1.3 Beam Elements

Use Beam elements for members that would typically be identified as
beams or girders.

PERFORM-3D User Guide 6-1

Chapter 6. Elements

Beams will usually have relatively small axial forces, so that it is

usually not necessary to account for P-M interaction. Also beams will
usually be bent in only one plane, about the cross section strong axis.
Hence it is usually not necessary to consider biaxial bending or biaxial
shear. However, you can consider P-M interaction, biaxial bending
and/or biaxial shear if you wish (i.e., you can model a beam element
using column type components).

Each beam element connects two nodes and consists of one frame
compound component. In most cases you can use a single beam
element to model a complete beam member, from column centerline to
column centerline. The exception is if a member such as a diagonal
brace connects to the beam within its span. In this case you must locate
a node at the connection point and divide the beam into two or more
elements.

You must assign local axis orientations to beam elements. Local Axis 1
is along the element, from element End I to element End J. Local Axes
2 and 3 are normal to Axis 1, with Axes 1, 2 and 3 forming a right
handed coordinate system. Bending in the 1-2 plane (i.e., bending about
Axis 3) is usually strong axis bending. Bending in the 1-3 plane (about
Axis 2) is usually weak axis bending. Axis 2 is usually directed
vertically upwards, and Axis 3 is usually horizontal. However, the axes
may be inclined.

When you add beam elements to a structure, be sure to use a consistent

method for choosing Ends I and J. If the beams are essentially along the
H1 and H2 directions, we suggest that you always direct I-J along the
corresponding axis.

You can specify distributed and/or concentrated loads along the

element length (these are element loads, as distinct from nodal loads).
In the current version of PERFORM-3D, only gravity loads are allowed
(usually loads in the negative V direction).

6.1.4 Column Elements

Use Column elements for members that would typically be identified as
columns.

Columns will usually have relatively large axial forces, so that it is

necessary to account for P-M interaction. Also columns will usually be

6-2 PERFORM-3D User Guide

 Chapter 6. Elements

bent in about both cross section axes, so that it is necessary to consider

biaxial bending and biaxial shear. However, if a column has small axial
forces and is bent mainly about one axis, you can model it using beam
type components.

Each column element connects two nodes and consists of one Frame
Type compound component. In most cases you can use a single column
element to model a complete column member, from beam axis to beam
axis. The exception is if a member such as a diagonal brace connects to
the column within its height. In this case you must locate a node at the
connection point and divide the column into two or more elements.

You must assign local axis orientations to column elements. Local Axis
1 is along the element, from element End I to element End J, and is
usually vertical. We suggest that you consistently make End I the lower
end and End J the upper end. Local Axes 2 and 3 are normal to Axis 1
and are usually horizontal. You must specify the orientations in plan of
Axes 2 and 3.

You can specify distributed and/or concentrated loads along the

element length. However, since only gravity loads are allowed (usually
loads in the negative V direction) you probably will not do this for
vertical column elements.

6.1.5 Brace/Other Frame Elements

Use Brace/Other Frame elements for members such as diagonal braces
that are not clearly beams or columns, and can not be modeled as bar
elements that resist only axial force.

You could use brace/other frame elements for beam and column
members. However, if you want to use default end zones or panel zone
elements, you must model beam and column members using beam and
column elements. The reason is that PERFORM-3D uses beam
elements to get the "auto" end zone lengths for columns and the depths
for panel zones, and similarly uses column elements to get end zone
lengths for beams and widths for panel zones. If you do not use beam
and column elements, you must specify end zone lengths and panel
zone dimensions explicitly.

You can model diagonal bracing members using simple bar elements or
brace/other frame elements. If you can ignore the bending stiffness of a

PERFORM-3D User Guide 6-3

Chapter 6. Elements

bracing member, the simplest approach is to use a simple bar element.

If you must consider the bending stiffness you must use a brace/other
frame element. Each element consists of one frame compound
component. You can use beam type basic components if you do not
need to account for P-M interaction. You must use column type basic
components if P-M interaction is important.

Do not model bracing members using beam or column elements. As

noted earlier, PERFORM-3D bases the lengths of default end zones for
columns on the dimensions of the adjacent beams, and the lengths of
default end zone for beams on the dimensions of the adjacent columns.
If you use beam or column elements for bracing members, PERFORM-
3D may get confused and use the wrong dimensions.

As with beams and columns, you must assign local axis orientations to
brace/other frame elements. Local Axis 1 is along the element, from
element End I to element End J. Axis 2 will usually be in the vertical
plane, and Axis 3 will usually be horizontal.

You can specify distributed and/or concentrated gravity loads along the
element length.

6.1.6 Shear Wall Elements

Use Shear Wall elements to model most shear walls and shear cores,
including coupled walls.

Shear wall elements act essentially as beams, with bending, axial and
shear deformations. They can be elastic or inelastic for shear, and
elastic or inelastic for bending, using fiber cross sections. Each element
connects four nodes and consists of one Shear Wall compound
component.

You must assign local axis orientations to shear wall elements. Local
Axis 2 is usually vertical, along one element edge. Local Axis 3 is
usually horizontal, also along an element edge (elements will usually be
rectangular). Axis 1 is normal to the plane of the element (all four
nodes for an element must be in one plane).

6-4 PERFORM-3D User Guide

 Chapter 6. Elements

6.1.7 General Wall Elements

Use General Wall elements to model squat walls with irregular
openings.

General wall elements are nonlinear finite elements specifically

intended for modeling reinforced concrete walls that have irregular
openings, and hence have complex bending and shear behavior. Each
element connects four nodes and consists of one General Wall
compound component.

You must assign local axis orientations to general wall elements. Local
Axis 2 is usually vertical, along one element edge. Local Axis 3 is
usually horizontal, also along an element edge (elements will usually be
rectangular). Axis 1 is normal to the plane of the element.

6.1.8 Infill Panel Elements

Use Infill Panel elements to model masonry type infill in frames. Infill
panel elements can be elastic or inelastic. Each element connects four
nodes and consists of one basic component. You can use basic
components with shear type or diagonal strut type behavior for infill
panel elements.

Infill panel elements do not need local axes.

6.1.9 Connection Panel Zone Elements

Use Panel Zone elements to model shear deformations in beam-to-
column connections. Panel zone elements can be elastic or inelastic.
Each element connects to only one node (at the beam-column
connection) and consists of one basic component.

You must assign local axis orientations to panel zone elements. Local
Axis 1 is normal to the plane of the panel zone, Axis 2 is along the
column, and Axis 3 is along the beam (assuming the usual case of
beams and columns at right angles).

6.1.10 BRB Elements

Use BRB element to model buckling restrained brace members. BRB
elements resist axial force only. Each element connects two nodes and

PERFORM-3D User Guide 6-5

Chapter 6. Elements

consists of one compound component. This component consists of an

inelastic BRB component in series with an elastic bar component and
an optional rigid end zone.

BRB elements do not need local axes.

6.1.11 Viscous Bar Elements

Use Viscous Bar elements to model fluid dampers. Viscous bar
elements resist axial force only. Each element connects two nodes and
consists of one compound component. This component consists of a
fluid damper component in series with an elastic bar component.

Viscous bar elements do not need local axes.

6.1.12 Seismic Isolator Elements

Use Seismic Isolator elements to model isolators of rubber and friction
pendulum type. You can also use rubber-type isolator elements to
model other energy absorbing devices, such as the ADAS device
(alternatively you could use inelastic bar elements).

Seismic isolator elements resist bearing and shear forces. In bearing the
element is elastic, with an option for different tension and compression
stiffnesses. In shear the element is inelastic. Each element connects two
nodes and consists of one basic component.

You must assign local axis orientations to seismic isolator elements.

Local Axis 3 is in the bearing direction, usually vertical. Axes 1 and 2
are in the shear directions, usually horizontal

6.1.13 Slab/Shell Elements

The slab/shell element is an elastic 4-node element with membrane (in-
plane) and plate bending (out-of-plane) stiffnesses. It can be used to
model slabs and curved shells.

For PERFORM-3D the most likely application is to model deformable

floor diaphragms. In this application the membrane behavior of the element
accounts for in-plane effects in the diaphragm, and the plate bending
behavior can be used for applying and distributing gravity loads. A floor
diaphragm can be modeled using wall elements, but this requires four steps

6-6 PERFORM-3D User Guide

 Chapter 6. Elements

to define the element properties, namely material, cross section, wall

compound component, and element. For a slab/shell element this is
reduced to three steps, namely material, cross section, and element.

This element is a special case of the inelastic slab/shell element that is

available in PERFORM-COLLAPSE. That element is intended primarily
for modeling two-way slab floors. It has two cross section options, namely
elastic (as in the PERFORM-3D element) and reinforced concrete (using a
layered model similar to a fiber model for a beam section). The reinforced
concrete version of this element is not currently available in PERFORM-
3D.

You must assign local axis orientations to slab/shell elements. Axis 1 is

normal to the plane of the element, and Axes 2 and 3 are in the element
plane, similar to a wall element.

6.1.14 Support Spring Elements

Each support spring element connects to one node and consists of one
support spring component. In effect, each element connects the supported
node to a point that is fixed in space. Support spring elements have no P-
effects. They are plotted as small circles at the supported nodes.

Support spring elements can have translational and/or rotational

stiffnesses, with up to a 6x6 stiffness matrix. Strength capacities and
strength based limit states can be specified.

You must assign local axes to support spring elements. These axes can
be inclined to the H1-H2-V axes if desired.

Support spring elements can be used to impose displacements at

support points. This is done by making the elements very stiff and
specifying that they have initial deformations (similar to thermal
expansions).

6.1.15 Deformation Gage Elements

PERFORM allows you to specify deformation capacities for single
components. In some cases, however, you may want to measure
deformations over a number of elements, and specify corresponding
limit states. For example, in a wall structure you may want to calculate
the average strain over a story height, where there are multiple

PERFORM-3D User Guide 6-7

Chapter 6. Elements

elements over that height. You can do this with a Strain Gage element
that extends over the story height.

Deformation gage elements do not add any stiffness to the structure.

They act literally like strain gages. You can specify gage elements of
the following types.

(1) 2-node extensional strain gage.

(2) 2-node rotation gage of beam type.
(3) 4-node rotation gage for walls.
(4) 4-node shear strain gage for walls.

Only the beam type rotation gage requires local axes. These are the
same as for a beam element.

6.2 Specifying Elements

6.2.1 Element Groups

Elements must be organized into Element Groups. The elements in any
group must all be of the same type. However, the elements of any type
need not all be in one group.

In a small frame you may need only two element groups, one
containing all beam elements and one containing all column elements.
In a larger frame you are likely to use more groups. For example, you
might put the exterior columns in one group and the interior columns in
another. One reason is that deformation and strength limit states are
based on element groups, and you may want to have the exterior and
interior columns in different limit states. If you have a structure with a
variety of element types, you will need at least one element group for
each element type.

Before you begin specifying a structure, spend some time planning the
element groups.

If you need to, you can combine or split element groups after they have
been created.

6-8 PERFORM-3D User Guide

 Chapter 6. Elements

6.2.2 Steps
To specify elements press the toolbar button for the Modeling phase
and the Elements task. The overall steps are as follows.

(1) Create an element group. Each group is identified by a name.

(2) Add elements to the group.
(3) Assign orientations and properties to the elements.

You can edit an existing element group at any time, to add elements,
delete elements or change element properties. You can also delete
entire groups.

6.2.3 Start a New Element Group

To start a new group press the New button, choose the element type,
input the group name, and specify whether you want to consider or
ignore geometric nonlinearity.

Choose "None" for geometric nonlinearity effects if you are sure that
the elements in the group contribute no significant geometric
nonlinearity effects. This will often be the case for beam elements,
since they are horizontal and typically have small axial forces. If in
doubt choose "P-delta". For some elements the option is automatically
"None". The “Large displacements” option is currently disabled for all
elements (this option can be chosen in PERFORM-COLLAPSE).

You also have the option of specifying a scale factor for "K" viscous
damping. For an explanation of the available damping models see
Chapter 18, Elastic Viscous Damping. If you wish, you can increase or
decrease the K damping for an element group by specifying a scale
factor that is larger or smaller than 1.0. If you believe that the
calculated  value will overestimate the energy dissipation for the
elements in a group, specify a factor smaller than 1.0 for that group. If
you believe that the calculated  value will underestimate the energy
dissipation, specify a factor larger than 1.0. If you want to ignore K
damping for an element group, specify a factor of 0.

PERFORM-3D User Guide 6-9

Chapter 6. Elements

6.2.4 Add Elements

To add elements choose the Add Elements tab. To add one or more
elements, select a set of nodes and press the Test button (use the Test,
Clear, OK and Undo buttons in the usual way).

For elements that connect two nodes you have the choice of Sequence,
Series, Straight Line or Grid generation. There are diagrams that
illustrate these options. You can save time by using the appropriate
generation option.

For elements that connect four nodes you have the choice of Sequence,
Column, Row or Grid generation. There are diagrams that illustrate
these options.

As noted earlier, take care with the element I and J ends. Each 2-node
element has an I end and a J end. When the elements are plotted the I
ends are marked with a small dot. Each 4-node element has I, J, K and
L corners. The I end of edge IJ is marked by a dot.

6.2.5 Parallel Elements

There may be occasions where you want to have two or more elements
in parallel (i.e., connecting the same nodes). You can not do this within
an element group. You can have parallel elements, but they must be in
different element groups.

6.2.6 Delete Elements

To delete elements choose the Delete Elems tab. You can select
elements one at a time by clicking on an element or, for a 4-node
element, an element edge. Alternatively you can select multiple
elements with a box select, by dragging a rectangle that contains the
middle parts of the elements to be selected.

Selected elements will change color. If you select an element a second

time, it is deselected. Use the Test, Clear, OK and Undo buttons in the
usual way.

6-10 PERFORM-3D User Guide

 Chapter 6. Elements

6.3 Element Orientations

To assign local axis orientations to elements choose the Orientations

tab. There are diagrams to illustrate how Axes 1, 2 and 3 are defined.
The steps are as follows.

(1) Choose the element group from the Current Group list. Elements in
the current group are light blue. Elements that already have
orientations assigned have axis symbols (see below for explanation
of these symbols).

(2) Select one or more elements. You can select elements in the
following ways. Selected elements are red.

(a) Select singly or using a box select.

(b) Double click on an element. This selects all elements that have
the same orientation data.
(c) Double click in blank space. This selects all elements that do
not yet have assigned orientations.

If all of the selected elements have the same orientation data, the
data in the Orientations tab will be set. You can use this feature to
review the orientation of an element. If the selected elements have
no orientations, or if they have different orientations, the data in
the Orientations tab is not set.

(3) Choose the orientation type, and input any needed data.

(4) Press Test to show the new orientations, then either OK or Undo.

Diagrams in the form illustrate the elements axes (Axes 1, 2 and 3) and
explain how the orientations are assigned.

For panel zone elements you have two options for assigning the
orientations, namely "Standard" and "Nonstandard". If you choose the
Standard option, the orientation of a panel zone element depends on the
orientation of the adjacent column elements. This applies only to cases
where the columns are I sections, and only if there is no ambiguity. If
you choose the Nonstandard option you must input the orientation data
directly. If you assign "auto" panel zone components to any elements,
you must use the Standard option for those elements.

PERFORM-3D User Guide 6-11

Chapter 6. Elements

The plotted orientation symbols have the following meanings.

(1) For a frame type element the symbol is plotted near the element
End I. Axis 1 is along the element from End I to End J. The
orientation symbol is a pair of lines along Axes 2 and 3, with a dot
on the positive Axis 2 side.
(2) For a panel zone element Axis 1 is normal to the element, and has
a dot at its end. Axes 2 and 3 are in the plane of the element. Axis
2 has a dot at its end.
(3) For a shear wall or general wall element the symbol is plotted at
Node I. Axis 1 is normal to the element, and has a dot at its end.
Axes 2 and 3 are in the plane of the element. Axis 2 has a dot at its
end.
(4) For a seismic isolator element, Axis 3 is the bearing direction and
Axes 1 and 2 are the shear directions. The 1-2 plane is shown as a
rectangle. The axes have dots at their ends.
(5) For a support spring element, Axes 1 and 2 are usually the
structure H1 and H2 axes, and Axis 3 is usually the V axis.
However, the axes can be rotated if desired. The 1-2 plane is
shown as a rectangle. The axes have dots at their ends.
(6) For a beam-type rotation gage element, the axes are the same as for
a frame type element.

6.4 Element Properties

To assign properties to elements choose the Properties tab. You must

have specified the properties for all needed components. The steps are
as follows.

(1) Select one or more elements. You can select elements in the
following ways. Selected elements are red.

(a) Select element singly or using a box select. The properties that
are currently assigned to the selected elements, if any, will be
shown in the Component Currently Assigned part of the
form.
(b) Double-click on an element. This selects that element plus all
other elements in the group that have the same properties. The

6-12 PERFORM-3D User Guide

 Chapter 6. Elements

currently assigned properties will be shown in the Component

Currently Assigned part of the form.
(c) Double click in blank space. This selects all elements in the
group that do not yet have assigned properties.

(2) If you want to assign a component to the selected elements, or

change the currently assigned component, press the Assign
Component button. The properties that you can assign will be
shown in the Component to be Assigned part of the form. Select
the component, then press Assign.

The Properties tab has a Show Properties button. If you press this
button, you can review the properties of the current component. You
can not change the properties in this mode. To change the properties
you must use the Component Properties task.

6.5 Moving Elements Between Groups

To move elements from one group to another, choose the Change

Group tab. You can then move elements from the current group to
another group. If you want to split an element group, first create one or
more new groups, with no elements. Then use Change Group to move
elements into the new groups. You can only move elements that have
no element loads.

You can also change the name of any element group, by pressing
Rename in the Group Data tab.

6.6 P-, P- and Large Displacement Effects

6.6.1 General
P-, P- and large displacement effects can cause nonlinear behavior of
elements and hence of complete structures. This is usually referred to
as "geometric" nonlinearity. PERFORM-3D gives you options to
include or ignore geometric nonlinearity.

PERFORM-3D User Guide 6-13

Chapter 6. Elements

This section reviews the different types of geometric nonlinearity, and

summarizes the PERFORM-3D options.

6.6.2 P- vs. True Large Displacements

In small displacements analysis there are two key assumptions, as
follows.

(1) The geometric relationship between node displacements and

element deformations is a linear relationship.
(2) The equilibrium equations can be formed in the undeformed
position of the structure.

In reality neither of these assumptions is correct.

Mathematically, the first assumption is correct only in the limit as the

displacements tend to zero. As the displacements of the nodes (or, more
correctly, the rotations of the elements) increase, the relationship
between node displacements and element deformations becomes
progressively more nonlinear.

The second assumption is not correct for the simple physical reason
that equilibrium must be satisfied in the deformed position. As the
element rotations become progressively larger, this assumption
becomes progressively less correct.

True large displacements analysis takes into account both types of

nonlinearity. P- analysis retains assumption (1), but considers
equilibrium in the deformed position (actually it does not do this
exactly, but this is not a critical point).

Figure 6.1 illustrates the difference for a simple bar.

6-14 PERFORM-3D User Guide

 Chapter 6. Elements

  
P P P
H H H
Top moves Top moves
horizontally horizontally Top moves
in an arc
h h h

(a) Small Displs (b) P- (c) Large Displs

Figure 6.1 Geometric Nonlinearity

Assume for this example that the axial extension of the bar is negligible
(EA is very large). The three parts of the figure are as follows.

(a) Small displacements theory. This theory says (i) that the top of the
bar moves horizontally (this is small displacements geometry,
which also predicts that the bar extension is zero), and (ii) that
equilibrium can be considered in the undeformed position. Hence,
the force H is zero for all values of  (take moments about the base
of the bar).

(b) P- theory. This theory says (i) that the bar moves horizontally
and the bar extension is zero (small displacements geometry), and
(ii) that equilibrium is considered in the deformed position. Hence
H = P/h.

(c) Large displacements theory. This theory says (i) that the top of the
bar moves in an arc, so that it moves vertically as well as
horizontally, and the bar extension is indeed zero, and (ii) that
equilibrium is considered in the deformed position. Hence
H=P/hcos.

The difference between the value of H from P- theory and from large
displacements theory is small up to quite large rotations. For example
for /h = 0.05 (a large drift for most structures), P- theory gives H =

PERFORM-3D User Guide 6-15

Chapter 6. Elements

0.05V and large displacements theory gives H = 0.05006V, which is a

negligible difference. Also, for /h = 0.05 the vertical displacement in
case (c) is 0.00125h. P- theory predicts zero, which is not a significant
error in most cases. Hence, for most structures it is accurate enough to
use P- theory.

Consider, however, the simple bar structure in Figure 6.2.

L L
V

Figure 6.2 Case Where P- Theory Is Not Accurate

For this structure, small displacements theory says that the structure has
zero stiffness, since the theory predicts no extension of the bars as the
deflection increases, and hence no axial force. Hence, force V is zero
for all deflections.

If the initial force in the bars is zero, P- theory also says that the force
V must be zero, since the theory again predicts no extension of the bars.

Large displacements theory, however, predicts that the bars extend, and
that there is a progressively increasing V force as the deflection
increases.

If the initial force in the bars is P in tension, P- theory says that this
force stays constant, and that there is a linear relationship V = 2P/L
between vertical force and vertical displacement (apply equilibrium as
in the preceding example). The stiffness 2P/L is the "geometric" or
"initial stress" stiffness of the two bars. Large displacements theory
correctly predicts a progressively increasing stiffness, with an initial
stiffness equal to 2P/L.

In most building structures subjected to earthquake-type loads, the

behavior is more closely analogous to Figure 6.1 than to Figure 6.2. P-
 theory works well in this case, and has the advantage that it is both
simpler to apply than large displacements theory, and requires less
computation.

6-16 PERFORM-3D User Guide

 Chapter 6. Elements

6.6.3 P- Effect

Figure 6.3(a) shows a cantilever column with vertical and horizontal
loads.


P
H

P
part
h
Small
displs
P
part
part

Hh P

(a) Column (b) Bending Moments

Figure 6.3 P- and P- Effects

If the column remains elastic, it deforms as shown. Hence, considering

equilibrium in the deformed position, the bending moment diagram is
as shown in Figure 6.3(b) (not exactly, if we consider true large
displacements, but to a high degree of accuracy).

The bending moment diagram has three parts, as follows.

(1) A small displacements part, with a moment Hh at the base. This is

the moment from small displacements theory.
(2) A P- part, with a moment P at the base. This depends on the
lateral displacement at the top of the column,
(3) A P- part. This depends on the bending of the column within its
length.

Computationally, it is easy to account for the P- part of the moment,

since it depends on only the overall rotation of the column. It is more
difficult to account for the P- part, since it depends on the bending

PERFORM-3D User Guide 6-17

Chapter 6. Elements

deformation of the column (which in turn depends on the moments and

on whether the column yields or remains elastic).

It is possible to account for the P- effect in structural analyses.

However, it is important to be careful when considering this effect. In
Figure 6.3, the column is elastic. Figure 6.4 shows the same column,
but now it yields and forms a plastic hinge at the base.


P
H

P
part
h
Small
displs
P
part
part

Hh P

(a) Column (b) Bending Moments

Figure 6.4 P- Effect When Column Yields

As the figure shows, for a given  the P- moments are the same as
before, but the P- moments are now much smaller. The P- theory
must account for this. If the P- moments are calculated based on the
elastic deformation of a column, these moments can be substantially in
error.

Figure 6.4 also shows why P- effects reduce the strength of a column.
Let the moment capacity at the plastic hinge be M. This is a constant
quantity – it is not affected by P- effects. If we use small
displacements theory, the plastic hinge forms when M= Hh, and the
predicted horizontal strength of the column is H = M/h. If we consider
P- effects, the hinge forms when M= Hh + P, and the predicted
strength is H = (M – P)/h.

6-18 PERFORM-3D User Guide

 Chapter 6. Elements

6.6.4 Do You Need To Consider P- Effects?

If a column or brace forms plastic hinges only at its ends, it is unlikely
in any practical case that the P- moments will be significant. If a
column is stiff enough to attract substantial moments, its elastic
bending deformations will usually be so small that P- effects are
insignificant. If a column is flexible enough that if could have
significant P- effects, it will usually not attract much moment and its
elastic bending deformations will again be small. In most cases P-
effects can be ignored.

This applies, however, only for columns or braces that yield only at
their ends. P- effects can be substantial if a column or brace forms a
plastic hinge within its length, since the deformations that contribute to
the P- effect now include the inelastic as well as the elastic
deformations.

Finally, note that if you divide a column member into, say, two
elements, with a node in the middle of the member, the P- effect
applies only within each element, and will almost certainly be very
small. Any effects associated with displacement of the middle node are
now P- effects. This is one way to account for P- effects (i.e., add
extra nodes and elements, and convert them to P- effects).
6.6.5 Axial Shortening Due to Bending
As a column bends, the distance between its ends gets slightly shorter,
because the distance along the curved column is slightly longer than the
straight line between its ends. This is a geometric nonlinearity effect. It
is the large displacements counterpart of the P- effect (i.e., P- theory
considers P- moments, but ignores the shortening due to bending,
whereas large displacements theory considers both). This shortening
effect can be considered, but it adds a lot of complexity to the analysis.

If large displacements theory is used for column elements, strictly

speaking it should include the overall effect on axial deformations, as
considered for the simple bar examples, and also the axial shortening
due to bending. It is not necessary, however, to consider both effects.
We can choose to consider the "bar" effect and ignore the "bending"
effect. This is equivalent to ignoring P- effects.

PERFORM-3D User Guide 6-19

Chapter 6. Elements

You can convert P- effects to P- effects by adding nodes within a
member length. Adding nodes also converts "bending" effects to "bar"
effects.
6.6.6 PERFORM Options
PERFORM-3D gives you the option of including or ignoring geometric
nonlinearity.

As noted, for structures subjected to earthquake type loads, it is

sufficient to consider only P- effects. In the present version of
PERFORM-3D you can consider P- effects but not true large
displacement effects. This option is included in PERFORM-
COLLAPSE, where true large displacement effects may be needed for
accurate analysis of "catenary" effects in floor systems.

PERFORM-3D does not currently consider P- effects (i.e. it does not
consider geometric nonlinearity within the length of a column or brace
element). Hence, if you use a single element to model a brace member,
PERFORM-3D will not model buckling of the brace within its length.
You can, however, model this type of buckling by dividing a brace
member into a number of short elements, and specifying that P-
effects are to be considered. Buckling of this type can be sensitive to
initial out-of straightness in the member, and you may have to make the
member deliberately crooked in order to initiate buckling. If you want
to consider this type of behavior, we recommend that you first test the
member model in a small sub-assemblage, to make sure that you get the
expected behavior.

A simpler alternative for a buckling bar is to use the buckling type steel
material.

6-20 PERFORM-3D User Guide

 Chapter 7. Load Patterns

7 Load Patterns

Load definition involves two steps. In the Modeling phase you set
up load patterns. In the Analysis phase you set up load cases, by
combining load patterns and adding other data. This chapter
describes how to set up load patterns.

7.1 Pattern Types and Restrictions

Load patterns are used to construct load cases for static analyses. There
are three load pattern types, namely nodal load patterns, element load
patterns and self weight patterns.

Nodal load patterns are straightforward, with H, V and/or R loads

applied directly on nodes.

Element load patterns are more complex. In principle an element load

pattern can include a variety of loads, including distributed loads, point
loads and thermal expansion effects, applied on elements of a variety of
types. PERFORM-3D currently includes only a few element load types,
including gravity (vertical) loads on beam and column elements, and
thermal expansion effects in bar elements.

Self weight patterns automatically calculate gravity nodal loads using

component weights and element lengths or areas.

Nodal load patterns are usually used for either gravity load cases or
push-over load cases. Gravity loads usually have V loads only, and
push-over loads usually have H loads only. However, PERFORM-3D
does not prevent you from specifying nodal load patterns that include
both H and V loads.

Nodal load patterns are also used for Dynamic Force load cases.

PERFORM-3D User Guide 7-1

Chapter 7. Load Patterns

7.2 Nodal Load Patterns

To specify a nodal load pattern press the toolbar buttons for the
Modeling phase and the Load Patterns task, then choose the Nodal
Loads tab.

Nodal load patterns are identified by name. To start a new pattern press
the New button and follow the instructions. Select nodes and use the
Test, Cancel, OK and Undo buttons in the usual way.

You can review, add, delete or change nodal loads at any time. You can
also delete complete load patterns.

7.3 Element Load Patterns

To specify an element load pattern press the toolbar buttons for the
Modeling phase and the Load Patterns task, then choose the Element
Loads tab.

Element load patterns are identified by name. To start a new pattern

press the New button and input the pattern name. The steps for
specifying element loads are then as follows. These steps are rather
complicated, so proceed carefully. It can be a good idea to use Frames
to show only parts of the structure.

(1) Choose an element group from the list.

(2) Within the element group set up element subgroups, where all
elements in a subgroup have identical element loads. At one
extreme, if all of the elements in the element group have identical
loads there is only one subgroup. At the other extreme, if every
element has a different load each subgroup consists of only one
element. You will usually have a number of subgroups, each with a
number of elements. To define a subgroup choose the Loaded
Elems tab and the Define new subgroup option. Then define the
subgroup as follows.

(a) The elements that are not yet assigned to subgroups are shown
light blue. Select the elements for the subgroup from these

7-2 PERFORM-3D User Guide

 Chapter 7. Load Patterns

elements. As you select elements for the current subgroup they

will change to green. To deselect an element, select it a second
time. You can use single element or "box" selects.

(b) When the subgroup is complete, press Done.

This completes the subgroup. You can now either define another
subgroup, by repeating Steps (a) and (b), or you can specify
element loads and return to add more subgroups later. If you are
ready to specify element loads, go to Step (3). If you wish you can
change any subgroup, by adding or deleting elements, as noted
later.

(3) Add element loads. To do this choose the Add Loads tab. Define
the loads as follows.

(a) Choose the subgroup number from the list. The elements in the
subgroup will be shown green. The "No. of elems" box shows
the number of elements in the subgroup, and the "No. of loads"
box shows the number of load components currently defined
for this subgroup. If there are any loads they will be plotted in a
floating window. If this window obscures the model, drag it out
of the way (click and hold inside the window, drag, then
release).
(b) Choose the load type from the list. A diagram in the form will
show the required data.
(c) Enter the required data, and use the Test, Cancel, OK and
Undo buttons in the usual way.

Repeat Steps (a), (b) and (c) for as many load components as you
want to apply. You can change subgroups by choosing the
subgroup number from the list. You can define new subgroups by
returning to the Loaded Elems tab.

Repeat Steps (1), (2) and (3) for as many element groups as you want to
include in the load pattern.

To review, add, change or delete a subgroup, choose the Loaded Elem

tab and the View/Edit existing subgroups option, and follow the
instructions.

PERFORM-3D User Guide 7-3

Chapter 7. Load Patterns

To delete load components, choose the Delete Loads tab, and follow
the instructions.

7.4 Self Weight Load Patterns

To specify a nodal load pattern press the toolbar buttons for the
Modeling phase and the Load Patterns task, then choose the Self
Weight tab.

Self weight load patterns are identified by name. To start a new pattern
press the New button and follow the instructions.

To specify a self weight load pattern you select the elements that are to
be included in the pattern. You can add as many elements as you wish,
from any number of element groups, provided that the elements are of a
type that allow self weight loads.

To add elements choose the Add Elements tab and follow the
instructions. To delete elements choose the Delete Elements tab.

The self weight for a 2-node element is applied as nodal loads at its I
and J end, 50% at each end. The self weight for a 4-node element is
applied as nodal loads at its I, J, K and L corners. For a rectangular
element 25% of the weight is applied at each corner. For an element
that is not rectangular the weight is distributed among the nodes in a
reasonable way.

You can show the calculated loads at any time by pressing Plot Loads.
These loads are based on the current self weights of the components. If
you change these weights, the calculated loads will change. For
analysis of the structure, the loads are calculated using the component
weights at the time the analysis is run.

7-4 PERFORM-3D User Guide

 Chapter 8. Drifts and Deflections

8 Drifts and Deflections

Horizontal drifts are useful measures of deformation for lateral

loads. Vertical deflections can be useful measures of deformation
for long span structures. You must specify at least one drift, as a
Reference Drift for the Analysis phase. This chapter explains how to
specify drifts and deflections.

8.1 Drifts

8.1.1 Simple Drift

A simple drift is the horizontal displacement of one node relative to a
second node that is lower down in the structure, divided by the height
between the nodes (i.e., in PERFORM, a "drift" is a dimensionless drift
ratio). If the upper node is the roof of a building and the lower node is
at ground level, the drift value is the roof displacement divided by the
building height. This is the average drift over all stories. If the upper
node is at one floor and the lower node is at the floor below, the drift is
for one story of the building.

Note that a drift value of, say, 0.02 is a 2% drift.

8.1.2 Distortion Drift

In some frame structures the effective drift can be different in different
bays of the frame. This is shown in Figure 8.1.

In this case the maximum shear deformation in a story can be

substantially larger than the simple drift. You can account for this using
distortion drifts, each of which is defined by four nodes. The drift
magnitude is the shear deformation, in radians. Again, this is
dimensionless, and a value of, say, 0.02 corresponds to a drift of 2%.

PERFORM-3D User Guide 8-1

Chapter 8. Drifts and Deflections

Outer columns have relatively

small axial deformations.

Shear deformations in outer bays

are larger than story average.

Figure 8.1 Effective Drift Exceeds Story Drift

8.1.3 Procedure for Specifying a Drift

To define one or more drifts press the toolbar buttons for the Modeling
phase and the Drifts and Deflections task, then choose the Drifts tab.

Drifts are identified by name, with a maximum of 20 characters. To

define a new drift press the New button and enter the drift name. Then
choose either the Simple Drift or Distortion Drift option. For a simple
drift choose the drift direction, select the upper and lower nodes as
indicated in the form, and press Test to show the drift. For a distortion
drift, select the four nodes in the sequence shown in the form, and press
Test to show the drift. Then use the OK, Undo and Clear buttons in
the usual way.

Once you have defined a drift you can not change it. You must delete it
and replace it with a new one.

8.1.4 Reference Drift

You must define at least one drift, to use as the Reference Drift in the
Analysis phase. This must be a simple drift, and will usually be a drift
with the upper node at the roof level and the lower node at ground
level. The drift value is then the average drift for all stories in the
structure, and the roof displacement is the drift value multiplied by the
structure height. You will usually (and you usually should) define other

8-2 PERFORM-3D User Guide

 Chapter 8. Drifts and Deflections

drifts, possibly one for each story, to use in drift limit states and to
serve as "controlled" drifts for static push-over analysis.

8.2 Deflections

To define one or more deflections press the toolbar buttons for the
Modeling phase and the Drifts and Deflections task, then choose the
Deflections tab.

A deflection is the downwards V displacement of one node (the

deflected node) relative to a second node (the reference node). The
reference node is not necessarily at a support. Deflections have the
units of length.

Deflections are identified by name, with a maximum of 20 characters.

To define a new deflection press the New button and enter the
deflection name. Then select the nodes, and press OK to save the
deflection.

Once you have defined a deflection you can not change it. You must
delete it and replace it with a new one.

PERFORM-3D User Guide 8-3

 Chapter 9. Structure Sections

9 Structure Sections

A Structure Section is a cut through all or part of a structure.

Structure Sections have a number of uses.

The forces on structure sections can be useful for checking the

behavior of a structure. For example, if a structure has more than
one lateral load resisting system, you may want to know how the
total load is distributed among them. You can do this by specifying
sections through parts of the structure and comparing the forces on
these sections.

In tall shear wall structures it is common to allow inelastic

behavior in bending, but to require that the wall remain essentially
elastic in shear. The shear strength in a wall can be checked either
(a) in individual wall elements, or (b) for a wall cross section that
extends over several elements. Because there can be stress
concentrations, it may be better to check the strength for the wall
section as a whole, rather than for individual elements. You can do
this by defining structure sections through walls and assigning
shear strengths to these sections.

You can plot diagrams of bending moment and shear force over the
height of a tall shear wall, by cutting sections at different levels
over the height and using the forces on these sections.

This Chapter explains the procedures.

9.1 Concept

A structure section is defined by cutting through a number of elements.

You can cut elements only at their ends. To define a structure section
you must specify the cut elements and the cut location for each element
(whether at the I or J end for a 2-node element, and whether at the IJ,
KL, IK or JL edge for a 4-node element).

Figure 9.1(a) shows two alternative sections through a story of a frame.

For Section A the element cut ends (shown by rectangles) are at the top

PERFORM-3D User Guide 9-1

Chapter 9. Structure Sections

of the story, whereas for Section B the element cut ends are at the
bottom of the story.

50

50
Section A

50 Section B

(a) Alternative Structure Sections in a Story

50

50
Section A

(b) Forces on Section A

50 Section B

(c) Forces on Section B

Figure 9.1 Forces on a Structure Section

To get the section forces the cut elements are removed, exposing the
cut ends. The cut elements exert forces and moments on the section as
shown in Figures 9.1(b) and (c). The H1, H2 and V forces and
moments on the section are the resultants of the element H1, H2 and V
forces and moments.

9-2 PERFORM-3D User Guide

 Chapter 9. Structure Sections

Structure sections are usually plane, but this is not essential, and
PERFORM-3D does not check. Sections will most often be horizontal,
and you will usually be interested in the horizontal (shear) forces. For
the loads shown in Figure 9.1, the H force on section A is minus 100
units, and the H force on Section B is plus 100 units. The V force is
zero for both sections.

9.2 Specifying Element Cuts

To specify a structure section press the toolbar buttons for the

Modeling phase and the Structure Sections task. Choose the Define
Sections page.

Structure sections are identified by name. To start a new section press

the New button and input the section name. The steps for specifying the
cut elements are then as follows.

(1) Choose an element group from the list. Uncut elements in the
group are shown white. Cut elements are shown red, with the cut
ends marked by rectangles.
(2) Choose the Add Cuts tab.
(3) Choose the End I or End J option, depending on the desired cut
location. For a 4-node element End I means at the element IJ edge,
and End J means at the element KL edge.
(4) Select the new elements to be cut, using single element or box
select. The selected elements are yellow and the cut ends are
marked by green squares.
(5) Press OK to add the new cuts to the section. Press Cancel to clear
the selected elements.
(6) Repeat from Step (1) for as many element groups as needed.

To delete element cuts, choose the Delete Cuts tab, select the cut
elements to be deleted from the section, and press OK.

You can change an existing structure section at any time, adding or

deleting element cuts. You can also delete complete structure sections.

PERFORM-3D User Guide 9-3

Chapter 9. Structure Sections

9.3 Shear Strength for a Wall Section

Figure 9.2 shows a plan view of a shear wall core.

Figure 9.2 Plan of a Shear Core

Suppose it is required that the wall remain essentially elastic in shear. The
wall has four sides, usually with a number of wall elements along each
side. You may want to check the shear strength for each side separately. To
do this the steps are as follows.

(1) Calculate the shear strength of the wall, considering contributions

from the concrete and the steel reinforcement in the usual way. The
strength can be the strength of the entire wall, the strength per unit
length of the wall, or the strength per unit area. There will probably be
several different strengths, as the reinforcement percentage varies over
the height of the wall.

(2) Using the Component Properties task and the Materials tab, define
one “Elastic Shear Material for Walls” material for each different
strength. You will use these materials to define the properties of the
Shear Wall Compound Components as well as the structure section
strengths. If you wish you can specify that the shear strength depends
on the axial stress.

If you have calculated the shear strength as the strength per unit
area, the material strength is in terms of stress. The wall strength in
this case is the material strength multiplied by the effective wall
area, where the area is specified in Step (4). If you have calculated
the shear strength as the strength per unit length of wall, the wall
“area” in Step (4) is the effective wall length. If you have
calculated the shear strength as the strength of the entire wall, the
wall “area” in Step (4) is 1.0.

9-4 PERFORM-3D User Guide

 Chapter 9. Structure Sections

(3) Using the Structure Sections task and the Define Sections tab, define
one Structure Section for each side of the wall, at each story where
you want to check the shear strength (usually all stories). You may
also define a structure section for the entire wall section at each story.
You probably will not specify strengths for these entire wall sections,
but they can be useful for plotting moment and shear diagrams for the
wall.

(4) Using the Structure Sections task and the Strengths tab, specify a
shear strength for each section at which you want to check the
strength. Do this by choosing the section and associating it with an
Elastic Shear Material.

You must also specify the section “area”. As noted in Step (2) this
may be the actual area, the length of shear wall in the section, or a
unit value. The section shear strength is the material strength
multiplied by the section area. Be careful to use consistent units.

This completes the structure section definition and the strength

specification. The next section explains how to set up shear strength limit
states.

9.4 Section Groups and Strength Limit States

Strength limit states for structure sections are specified not in terms of
individual sections but in terms of Structure Section Groups (in the same
way that limit states for elements are in terms of Element Groups). The
steps are as follows.

(1) Using the Structure Sections task and the Groups tab, arrange the
structure sections in groups. You can put all of the sections in one
group if you wish, but it may be more useful to use a number of
groups. For example, for the shear core in Figure 9.2 you may have
specified structure sections at several stories (possibly all stories), with
sections for each side of the wall and for the entire wall at each story.
In this case you would probably set up five Structure Section Groups,
one for each side of the wall and one for the entire wall. When you set
up structure section groups, also consider whether you want to plot
bending moment and shear force diagrams, as described in a later
section.

PERFORM-3D User Guide 9-5

Chapter 9. Structure Sections

(2) Using the Limit States task (see Chapter 10), define one or more limit
states of Struct Sectn type. The concept is similar to that for element
strength limit states that are based on element groups.

Usage ratios are calculated for all Struct Sectn limit states. In Usage Ratio
plots, the usage ratios are shown in the same way as for limit states of other
types.

9.5 Moment, Shear and D/C Ratio Diagrams

9.5.1 Procedure
It can be useful to plot bending moment and shear force diagrams for shear
walls. Diagrams showing the variation of shear strength D/C ratio over the
height of a shear wall can also be useful, because the usage ratio for a shear
strength limit state gives the D/C ratio for only the worst structure section
covered by the limit state. If you plot D/C variations over the height of a
wall you can see which sections are most highly stressed.

The procedure is as follows.

(1) Use the Moment and Shear Diagrams task (see Chapter 27) and the
Section Group tab.
(2) Choose the result type and whether to plot envelopes only or time
histories. Press Plot to begin the plot. Time histories can be animated.

To get a sensible plot, the structure section group must be a sensible group,
such as a series of horizontal sections over the height of a wall. The group
should not contain multiple sections at any level, otherwise the plot will be
strange.

This feature is most likely to be used for horizontal sections in shear walls,
but it can be used for other purposes. For example, you can cut horizontal
sections through a frame structure and plot story shear and overturning
moment envelopes and time histories. You can also cut vertical sections
through long span structures such as bridges.

Currently the moment and shear diagrams are plotted on a horizontal axis,
even for a group of sections along a vertical line.
9.5.2 Save D/C Values to a File
After you have plotted envelope values for D/C ratios, you can save
these values to a text file. The “save these results to a file” button on

9-6 PERFORM-3D User Guide

 Chapter 9. Structure Sections

the toolbar will be activated. To save the D/C ratios, press this button
and follow the instructions.

Currently you can save the results only for envelope values and only
for D/C ratios.

The most likely application is when you have run analyses for a
number of earthquakes and you want to get the maximum or average
D/C ratio for each section. You can save a text file for each analysis
(and possibly a number of structure section groups), then process the
results using a spreadsheet program.
9.5.3 Signs and Sign Reversal
For a structure section, the section forces are on the section after the cut
elements are removed. If you create a horizontal section by cutting the
elements at the bottom of a story, the elements above the cut are removed.
Hence, a compression force on the section corresponds to a downwards
force. However, if you cut the elements at the top of a story, the elements
below the cut are removed, and a compression force on the section is now
an upwards force.

Usually, a group of structure sections will all be of the same type, with all
cuts at the story bottom or all at the story top. Sometimes, however, you
may want to put sections of different types in a group. For example, you
may cut at the bottom of each story until you reach the roof, then make the
final cut at the top of the story. To get correct moment and shear diagrams
in this case you must change the signs of the forces on this last section. To
do this, when you add this section to the structure section group, check the
“Change signs for M,V diagrams” box.

If you forget to check this box, delete the section from the group, and re-
add it with the box checked.

PERFORM-3D User Guide 9-7

 Chapter 10. Limit States and Usage Ratios

10 Limit States and Usage Ratios

A nonlinear analysis can produce a huge volume of analysis results.

If you make effective use of limit states you can distill the results
down to a few usage ratios that can tell you whether the structure
satisfies the performance requirements.

Capacities, limit states and usage ratios for components are

described in Chapter 5, Component Properties. Strength capacities
for structure sections are described in Chapter 9, Structure
Sections. This chapter explains how to define limit states and how
usage ratios are calculated.

10.1 Limit State Types

You can specify limit states of the following types.

(1) Deformation limit states, based on deformation capacities for

inelastic components.
(2) Strength limit states, based on strength capacities for elastic
components and strength sections.
(3) Drift limit states, based on drifts.
(4) Deflection limit states, based on deflections.
(5) Shear strength limit states for structure sections.

Limit states are identified by name. To start a new limit state, press the
toolbar buttons for the Modeling phase and the Limit States task, then
choose the limit state type from the list, press the New button, input the
limit state name, and define the limit state conditions as described in
the following sections.

Keep an eye on the Status box. The Save, Save As, Delete and
UnChange buttons operate in the same way as for component
properties.

You can edit an existing limit state at any time. You can also delete any
limit state.

PERFORM-3D User Guide 10-1

Chapter 10. Limit States and Usage Ratios

10.2 Deformation Limit States

A deformation limit state consists of a number of deformation

conditions. A deformation condition is defined by the following.

Element group. Choose this from the Element Group list.

The condition applies to all elements in the group. This is one

reason why you might divide the elements of a given type into
several groups.

Component type. Choose this from the Component Type list.

By default the list contains only those component types for which at
least one component has been specified. If there are no components
of a particular type, that type is not listed. You may, however, want
to list all possible component types that could be used by the
element group. To do this, check the List all possible types box.
You might do this because you plan to add components later.

Some elements consist of one basic component (e.g., simple bar

elements). If so, different elements in the group may have different
component types (e.g., some bar elements may be assigned elastic
bar components, and some may be assigned inelastic bar
components). The limit state condition applies only to those
elements that have the specified component type. If you choose a
component type and none of the elements in the group has a
component of this type, the condition does not apply to any
elements, and the demand-capacity ratio for the condition will be
zero. PERFORM-3D does not check this, so be careful when you
specify limit state conditions.

Other elements consist of compound components (e.g., beam

elements). In this case an element can have more than one basic
component of a given type (e.g., a beam element can have multiple
plastic hinge components). The limit state condition applies to all
components of the chosen type. For example, if a beam element has
plastic hinge components at its I and J ends, the demand-capacity
ratio for the condition is based on the larger of the ratios for these
two hinges.

10-2 PERFORM-3D User Guide

 Chapter 10. Limit States and Usage Ratios

As before, if you choose a component type and none of the

elements in the group has a component of this type, the condition
does not apply to any elements. Again, PERFORM-3D does not
check this.

Deformation type. Select this from the Deformation Type list.

A component typically has only one deformation, but this

deformation can be positive or negative. Usually you will choose to
consider the deformation regardless of its sign. However, you can
choose to consider only positive deformations or only negative
deformations. For some components you will have more choices.

Performance level. Select this from the Performance Level list.

The list is always 1 through 5, even if you have defined component

capacities at fewer than 5 levels. If you choose, say, Level 4 and
none of the components has capacities for that level, the condition
will not apply to any elements. PERFORM-3D does not check this.
(The reason is that you can edit the component properties at any
time, and PERFORM-3D would have to check for consistency each
time you make a change. Also, you might want to define a limit
state before you specify the component properties.)

Limit on the demand-capacity ratio. Input this value in the box.

Usually this value will be 1.0 (i.e., the usage ratio is 1.0 when the
demand equals the capacity for the chosen performance level). You
may, however, specify some other value. For example, suppose that
when you defined the component properties you specified a
deformation capacity of 0.01. You then decide that for this limit
state it is justified to increase this capacity by 20%, to 0.012. You
can do this by specifying a limit of 1.2 on the demand-capacity ratio
for this condition.

At any step in a structural analysis the demand-capacity ratio for a

component is the calculated demand divided by the capacity at the
specified performance level. The usage ratio for a limit state condition
is the demand-capacity ratio for the "worst" component, divided by the
limit on the demand-capacity ratio (e.g., divided by 1.2 in the above
example). The usage ratio for the limit state as a whole is the largest
usage ratio for the "worst" condition.

PERFORM-3D User Guide 10-3

Chapter 10. Limit States and Usage Ratios

To add a condition to a limit state, specify the above properties for the
condition, then press the Add button. You can also Insert, Replace and
Delete conditions from a limit state. When you have specified all
conditions, press the Save button to save the limit state.

You can return to this task to edit an existing limit state at any time. Be
sure to keep an eye on the Status box, and to press Save to save any
changes.

10.3 Strength Limit States

A strength limit state consists of a number of strength conditions. A

strength condition is defined by essentially the same properties as a
deformation condition, with the following differences.

(1) The component may be a basic component or a strength section.

(2) Instead of the component deformation type you choose the strength
type.
(3) The strength capacity is the nominal strength multiplied by the
capacity factor for the chosen performance level.

To add a condition to a limit state, specify the properties for the

condition, then press the Add button. You can also Insert, Replace and
Delete conditions from the limit state. When you have specified all
conditions, press the Save button.

You can return to edit an existing limit state at any time.

10.4 Drift Limit States

A drift limit state is defined by the following.

(1) A positive drift limit.

(2) A negative drift limit. Usually the positive and negative limits are
the same.
(3) A set of drifts. This set can include all of the drifts or only selected
drifts. Each drift in the set constitutes a drift condition.

10-4 PERFORM-3D User Guide

 Chapter 10. Limit States and Usage Ratios

Specify the positive and negative limits, and select the All drifts or
Highlighted drifts only option. For the highlighted option, click the
drifts that you want to include. Then press the Save button.

10.5 Deflection Limit States

A deflection limit state is like a drift limit state, except with deflection
limits and a set of deflections. Also, the deflection limits have units of
length, whereas drifts are dimensionless.

10.6 Structure Section Limit States

A structure section limit state is similar to a strength limit state, with

the following differences.

(1) Instead of an element group, choose a structure section group. You

must have defined some structure section groups in the Structure
Sections task.
(2) Instead of a component type and a strength type, choose only a
strength type. The only strength type currently available is shear
strength. You must have defined some structure section shear
strengths in the Structure Sections task.

To add a condition to a limit state, specify the properties for the

condition, then press the Add button. You can also Insert, Replace and
Delete conditions from the limit state. When you have specified all
conditions, press the Save button.

You can return to edit an existing limit state at any time.

10.7 Limit State Groups

You must define limit states in the Modeling phase, since the usage
ratios are calculated during the structural analysis, not as a post-
processing operation. For a large structure you can have many limit

PERFORM-3D User Guide 10-5

Chapter 10. Limit States and Usage Ratios

states with different types and performance levels, and hence a large
number of usage ratios. In the Analysis phase you can organize the
limit states into related groups, to decrease the number of usage ratios
and simplify the decision making process.

Often you will organize by performance level. For example, if you have
limit states for the Immediate Occupancy, Life Safety and Collapse
Prevention levels you might place all IO states in one group, all LS
states in a second group, and all CP states in a third group. For any load
case you can then quickly determine when the IO performance level is
exceeded, by considering only the IO group. The usage ratio for the
group is the largest usage ratio for any limit state in the group. The
results have thus been reduced to a single usage ratio that indicates
whether the structure satisfies all of the IO design criteria. You might
also set up a limit state group containing all IO limit states that affect
beam elements, and another group containing all IO limit states that
affect column elements. You can then quickly see whether the beams or
columns are more critical for the IO limit state.

Limit state groups are identified by name. To start a new group, press
the toolbar buttons for the Analysis phase and the Limit State Groups
task, press the New button, and input the group name. Then choose the
limit states for the group using the type and name lists and the Add,
Insert, Delete and Clear buttons. When you have chosen the limit
states, press the Save button to save the group.

You can edit an existing group at any time. When you do this, keep an
eye on the Status box. You can also delete an existing group.

If you change a limit state, this is a change in the analysis model, and
you must re-run the analyses. If you add, delete or change a limit state
group, you do not have to re-run the analyses.

10-6 PERFORM-3D User Guide

 Chapter 11. Inactive Elements

11 Inactive Elements

PERFORM-3D allows you to specify that some elements are

inactive for gravity load. This chapter explains the procedure.

11.1 Inactive Elements

For gravity load analysis it is assumed, in effect, that the structure is

built in a weightless environment and the loads are then applied.
Consequently, some members that have small or zero gravity forces in
the actual structure (e.g., diagonal braces) may have substantial forces
in the analysis model. This may not be a serious error for linear
analysis, but it can lead to incorrect behavior for nonlinear analysis. For
example, a gravity load analysis may calculate significant compression
forces in the elements used to model diagonal braces. These braces may
then yield prematurely in subsequent push-over or dynamic earthquake
analyses.

It may be more reasonable to assume that such elements have zero

forces under gravity load. You can make this assumption by making the
elements inactive for gravity load. PERFORM-3D deletes inactive
elements at the beginning of the gravity load analysis, and restores
them immediately afterwards.

11.2 Procedure

The only element types that you can make inactive are Simple Bar,
Brace, Infill Panel and BRB elements.

To specify inactive elements press the toolbar buttons for the Modeling
phase and the Inactive Elements task, then follow the instructions. Be
careful to ensure that the structure without the inactive elements is
stable, otherwise the gravity load may cause collapse.

PERFORM-3D User Guide 11-1

 Chapter 12. Gravity Load Cases

12 Gravity Load Cases

You can define and save any number of gravity load cases. You can
then use these load cases to run structural analyses.

This chapter describes how to set up gravity load cases. For an

explanation of how to use these load cases in structural analyses
see Chapter 20, Running Analyses.

12.1 Gravity Loads

A gravity load case is a combination of nodal, element and/or self

weight load patterns. You should usually apply only vertical loads,
usually in the downwards (minus V) direction. PERFORM will warn
you if a gravity load case has horizontal loads.

To specify a load case you must choose the load patterns and specify a
scale factor for each pattern. You must also specify whether the
analysis is to be linear or nonlinear.

In most cases you will expect the structure to remain linear and elastic
under gravity loads, and hence you will usually specify linear analysis.
Some structures can have nonlinear behavior under gravity load even
though they remain essentially elastic. For example, a gap may close or
a concrete fiber in a shear wall may crack. In this case you must specify
nonlinear analysis. If you specify linear analysis and there is nonlinear
behavior, you will get an error message during the analysis. If this
happens, change the analysis option to nonlinear and re-run the
analysis.

Usually you will apply loads that represent actual gravity loads on the
structure. However, if you wish you can apply gravity loads that
deform the structure well beyond yield, for example to assess the
gravity load capacity of a structure. In this case you must choose
nonlinear analysis, and specify a load that exceeds the strength of the
structure.

PERFORM-3D User Guide 12-1

Chapter 12. Gravity Load Cases

12.2 Procedure

12.2.1 Steps
To start a new gravity load case press the toolbar buttons for the
Analysis phase and the Load Cases task. In the Load Cases form
choose Gravity as the Load Case Type, press the New button, and enter
the load case name. Keep an eye on the Status box.

The steps are as follows.

Choose the option (linear or nonlinear) for the analysis method.

(1) If you have chosen nonlinear analysis complete the Control

Information for Nonlinear Analysis. You must specify the
following. For more details on each of these items see later in this
section.

(a) Number of equal load steps. See Section 12.2.2.

(b) Maximum number of nonlinear events allowed in any step. See
Section 12.2.3.
(c) Whether to use equal steps throughout, or to have an initial step
to the first event then equal steps after that. See Section 12.2.4.

(2) Specify the load patterns and scale factors. Use the Add, Insert,
Replace and Delete buttons in the usual way to build up a load
pattern list.

(3) When the load case is complete press the Save button to save it.
Use the Delete button if you want to delete a load case before
saving it.

You can edit a saved load case at any time. If you do, keep an eye on
the Status box. Use the Save, Save As and UnChange buttons in the
usual way. You can also delete saved load cases.

12.2.2 Number of Load Steps

For linear analysis there is always only one load step. For nonlinear
analyses you will usually apply actual gravity loads, and you can
specify just one load step. If you want to increase the gravity load until

12-2 PERFORM-3D User Guide

 Chapter 12. Gravity Load Cases

the structure collapses, a reasonable number of load steps is about 50.

The analysis results are saved at the end of each step. For plotting
analysis results, 50 steps gives a reasonable number of points along a
graph for a collapse load analysis.

12.2.3 Maximum Number of Events

PERFORM-3D uses an event-to-event solution strategy, where the
structure properties are re-formed each time there is a nonlinear event
(a change in stiffness). The program automatically divides each load
step into a number of substeps, with a new substep for each event. The
analysis stops if the number of events (substeps) in any load step
exceeds the specified maximum.

If there is a very large number of events in any load step, this could
indicate that the analysis has become numerically unstable and that
PERFORM-3D can not find a solution (experience suggests that this is
unlikely, but it could happen). In this case you would like the analysis
to stop, otherwise it might waste a lot of computer time. On the other
hand, if you have a large structure with a lot of nonlinear components
you could have a large number of real events in a load step. In this case
you do not want the analysis to stop prematurely.

For a structure of moderate size a reasonable maximum for the number

of events in any load step is about 200. It is unlikely that there will be
that many real events in any load step, so it is unlikely that the analysis
will stop prematurely, and if the analysis becomes unstable you will not
waste a lot of computer time before the analysis stops.

If a gravity load analysis stops because it reaches the maximum number

of events in any load step, this is an abnormal exit. Since the specified
load has not been applied, the load case can not be used as a preceding
load case for subsequent push-over or dynamic earthquake load cases.
You can still process the analysis results, provided at least one load step
has been completed.

12.2.4 Initial Step to First Event

In a typical structure, first yield does not occur until a substantial part
of the collapse load has been applied. In this case it is more efficient to
go to first yield in one initial step, then apply equal load increments

PERFORM-3D User Guide 12-3

Chapter 12. Gravity Load Cases

until collapse, rather than have several linear steps before first yield.
Usually, therefore, you can choose the Yes option.

There may, however, be cases where it is better to use equal steps

throughout. For example if there are large P- effects there can be
substantial nonlinear behavior before first yield is reached. Since the
analysis assumes linear behavior in the initial step, if you choose the
Yes option the calculated behavior at first yield can be incorrect.

In most practical cases the P- effect at first yield is small.

Nevertheless, for analyses that include P- effects you may prefer to
choose the No option to be safe.

If you switch from one option to the other, note that the load step size
will change. For example, if the load at first yield is, say, 40% of the
total gravity load (i.e., if the load factor at first yield is 0.4), and if you
specify 50 steps with the Yes option, the first 40% of the load is
applied in one step, and the last 60% is applied in 50 equal steps. If you
choose the No option with the same 50 steps, it takes 20 steps to reach
first yield, with only 30 steps for the last 60% of the load.

12.2.5 Analysis Limit State

In a nonlinear analysis a point may be reached where the structure is so
severely deformed that there is no more useful information to be
obtained, and hence there is no reason to continue the analysis. You can
stop the analysis by defining a limit state that corresponds to such a
severe deformation. This is the Analysis Limit State.

To specify the analysis limit state, choose the limit state type and name
from the lists.

If you do not specify an analysis limit state, a default state is used. For
gravity loads the default is to stop the analysis if the deformation for
any inelastic component exceeds its U point deformation. This is based
on the assumption that actual gravity loads should never be large
enough to load any member beyond its ultimate strength.

For gravity load analysis you will almost always apply the actual
gravity load, and expect the structure to remain essentially linear. If
there is substantial yielding you have probably made an error and
applied too much load. Hence, if a gravity load analysis stops because

12-4 PERFORM-3D User Guide

 Chapter 12. Gravity Load Cases

the analysis limit state is reached, this is regarded as an abnormal exit.

The load case can not be used as the preceding load case for subsequent
push-over or dynamic earthquake cases. You can still process the
analysis results, provided at least one load step has been completed.

PERFORM-3D User Guide 12-5

 Chapter 13. Static Push-Over Load Cases

13 Static Push-Over Load Cases

You can define and save any number of static push-over load cases.
You can then use these load cases to run push-over analyses.

This chapter describes how to set up push-over load cases. For an

explanation of how to use these load cases in structural analyses
see Chapter 20, Running Analyses.

13.1 Push-Over Loads

13.1.1 Load Distribution and Magnitude

To perform a static push-over analysis you must specify the distribution
of horizontal loads over the structure height. The current version of
PERFORM-3D allows fixed distributions only (i.e., you can not
consider "adaptive" load distributions that depend on the deflected
shape). As explained later you can specify load distributions using
nodal load patterns, or you can specify loads based on the structure
mass and a deflected shape.

When you specify a push-over load case you need to be concerned only
with the load distribution and direction. You do not have to worry
about the load magnitude. For the load steps in a push-over analysis the
structure is deformed in equal drift increments, and in each step
PERFORM-3D calculates the required load increment. The load
increment per step typically gets smaller as the structure yields, and can
be negative if the structure loses strength or becomes unstable because
of P- effects.

13.1.2 Choice of Load Distribution

One of the most difficult issues for push-over analysis is choosing the
push-over load distribution. During an actual earthquake, the effective
loads on a structure change continuously in magnitude, distribution and
direction. The distribution of story shears over the height of a building
can thus change substantially with time, especially for taller buildings
where higher modes of vibration can have significant effects. In a static
push-over analysis the distribution and direction of the loads are fixed,

PERFORM-3D User Guide 13-1

Chapter 13. Static Push-Over Load Cases

and only the magnitude varies. Hence, the distribution of story shears
stays constant. To account for different story shear distributions it is
necessary to consider a number of different push-over load
distributions.

One option in FEMA 356 is to use uniform and triangular distributions

over the building height. Note that a uniform distribution usually
corresponds to a uniform acceleration over the building height, so that
the load at any floor level is proportional to the mass at the floor.
Similarly, a triangular distribution usually corresponds to a linearly
increasing acceleration over the building height.

PERFORM has the following options for push-over loads.

(1) Load distributions based on nodal load patterns.

(2) Load distributions based on masses and a specified displacement
(acceleration) variation over the structure height. Examples are
linear and triangular variations.
(3) Load distribution based on masses and the mode shapes of the
structure. For more details on this option see the next section.

13.1.3 Load Distribution Based on Mode Shapes

Load distributions can be based on the structure mode shapes. For the
purposes of discussion consider a 2D structure. The procedure can be
extended to 3D structures as shown later in this chapter.

For a low-rise structure that is dominated by its first mode response, a

load distribution based on the first mode may be reasonable. For a
structure with significant second mode response a possible approach is
to specify three load distributions as follows (requiring three push-over
analyses).

(1) Mode 1. Use a force distribution based on mode 1.

(2) Mode 1 + Mode 2. Use force distributions based on modes 1 and 2,

and set the sign of the mode 2 distribution so that the mode 1 and 2
base shears are in the same direction. This tends to give larger
story shears in the lower stories than for mode 1 alone, and smaller
shears in the upper stories.

13-2 PERFORM-3D User Guide

 Chapter 13. Static Push-Over Load Cases

(3) Mode 1 - Mode 2. Use force distributions based on modes 1 and 2,

with the mode 1 and 2 base shears in opposite directions. This
tends to give smaller story shears in the lower stories than for
mode 1 alone, and larger shears in the upper stories.

PERFORM-3D allows you to specify load distributions of this type.

The procedure is described later in this chapter.

13.1.4 Linear Analysis Option

In the current version of PERFORM-3D you can not specify a linear
push-over analysis (the option has been disabled).

If you want to apply a specified lateral load, you can do so using a

Gravity load. In this case the analysis can be linear.

13.2 Procedure

13.2.1 Steps
To start a new push-over load case press the toolbar buttons for the
Analysis phase and the Load Cases task. In the Load Cases form
choose Static Push-Over as the Load Case Type, press the New button,
and enter the load case name. Keep an eye on the Status box.

The steps are as follows.

(1) Choose the option for the analysis method. In the current version
this must be Nonlinear Analysis.

(2) Complete the Control Information for Nonlinear Analysis. You

must specify the following. For more details on each of these items
see later in this section.

(a) Number of load steps. See Section 13.2.3.

(b) Maximum number of nonlinear events allowed in any step. See
Section 13.2.4.
(c) Whether to use equal steps throughout, or to have an initial step
to the first event then equal steps after that. See Section 13.2.5.
(d) The maximum allowable drift. See Section 13.2.3.

PERFORM-3D User Guide 13-3

Chapter 13. Static Push-Over Load Cases

(e) A limit state to stop the analysis. See Section 13.2.6.

(f) The reference drift. See Section 13.2.2.

(3) Choose the Load Type option (based on nodal load patterns, mode
shapes or a displacement pattern).

(4a) If the load type is based on nodal load patterns, specify the load
patterns and scale factors. To push the structure in, say, the +H1
direction specify scale factors that give +H1 loads. To push in the
–H1 direction specify scale factors to give –H1 loads. For push-
over along an inclined direction, specify load patterns and scale
factors to give forces in the required direction. Use the Add,
Insert, Replace and Delete buttons in the usual way to build up a
load pattern list. Only horizontal loads are allowed for push-over
loads, and hence you can use only load patterns that have no
vertical loads.

(4b) If the load case type is based on a displacement pattern, specify the
displacement pattern over the structure height. For a displacement
pattern choose the base and roof levels of the structure, and from
zero to 4 intermediate levels, for a total of up to 6 levels. Specify
the V coordinate and the H1 and H2 displacements at each level.
For a uniform pattern specify the same H displacements at the base
and roof levels. For a linear pattern, specify zero H displacement at
the base and unit H displacement (or some other non-zero value) at
the roof. The V coordinates must be in the structure coordinate
system and the structure length unit. The H displacements are
relative values only, and do not have units. For each mass in the
structure, the H1 and H2 loads are proportional to the mass
multiplied by its H1 and H2 displacements.

(4c) If the load case type is based on mode shapes, specify the push-
over direction, the modes to be used, a scale factor for each mode,
and a base shear direction for each mode. For an explanation see
Section 13.3. Use the Add, Insert, Replace and Delete buttons in
the usual way to build up a modal load list. Only horizontal
translational masses are considered for push-over loads. Vertical
translational masses are ignored, and rotational masses do not
affect push-over loads.

(5) Choose the controlled drifts. For an explanation see Section 13.2.3.
The default is to use all drifts as controlled drifts. To over-ride this

13-4 PERFORM-3D User Guide

 Chapter 13. Static Push-Over Load Cases

default choose the Controlled Drifts tab and choose the controlled
drifts. As a general rule you should use the default. It is not a good
idea to run an analysis with only a single controlled drift,
especially if there is strength loss.

(6) When the load case is complete press the Save button to save it.
Use the Delete button if you want to delete a load case before
saving it.

You can edit a saved load case at any time. If you do, keep an eye on
the Status box. Use the Save, Save As and UnChange buttons in the
usual way. You can also delete saved load cases.

13.2.2 Reference Drift

In push-over plots and other graphs for the results of push-over
analyses, the horizontal axis in the plot is reference drift.

FEMA 356 uses the horizontal displacement at the roof as a key

measure of building deformation. In PERFORM-3D the equivalent
measure is the drift of the roof relative to a supported point at the base.
As a general rule you should use this drift as the reference drift. For a
Target Displacement Plot the reference drift must be the roof drift.

13.2.3 Number of Steps

A reasonable number of steps for a push-over analysis is about 50. The
analysis results are saved at the end of each step, and 50 steps gives a
sufficient number of points for plotting push-over curves. It is not a
good idea to use a large number of steps, such as 500. Not only does
this waste computer time, but the analysis can be more sensitive if the
load increments are very small.

In Step 2(d) you specify a maximum allowable drift, and in Step 5 you
choose the controlled drifts. The allowable drift per load step is the
maximum allowable drift divided by the number of steps. In any step
the load increment is chosen so that the largest increment for any of the
controlled drifts is equal to the allowable drift per step (actually, there
is a tolerance and the largest drift increment is not exactly equal to the
drift per step). The analysis stops when the largest of the controlled
drifts exceeds the maximum allowable drift.

PERFORM-3D User Guide 13-5

Chapter 13. Static Push-Over Load Cases

This procedure prevents large increases in any drift in a time step,

which could cause inaccuracies or other problems. This is why it is
advisable to include all drifts in the set of controlled drifts. If you
control the analysis using only the roof drift (i.e., if you choose only
one controlled drift), and if the structure has a weak story, it is possible
for the (uncontrolled) drift in the weak story to increase rapidly even
though the(controlled) roof drift increases slowly (or decreases). This
can cause inaccuracies in the solution process.

If you choose only one controlled drift, and if this is the same as the
reference drift, then the reference drift will have equal increments in
each analysis step. However, if you choose a number of controlled
drifts the reference drift generally will not increase in equal increments.
This is not a problem – in fact it is often desirable. For example, if the
reference drift is the roof drift, and if the structure has a weak story,
you do not want the reference drift to have uniform increments. The
reference drift should increase more slowly as the weak story yields,
and may even decrease in some steps if the weak story loses strength
and load is redistributed through the structure.

The reference drift at the end of the analysis may be smaller than the
maximum allowable drift, and the actual number of analysis steps may
be larger than the specified number. This is because the reference drift
is not necessarily the largest drift, and because the controlling drift is
not necessarily the same drift in all analysis steps.

As a general rule, we suggest that you use the roof drift as the reference
drift, that you use a number of controlled drifts, and that you specify
about 50 load steps with a maximum drift that is rather larger than the
maximum that you actually want to consider.

13.2.4 Maximum Number of Events

PERFORM-3D uses an event-to-event solution strategy, where the
structure properties are re-formed each time there is a nonlinear event
(a change in stiffness). The program automatically divides each
analysis step into a number of substeps, with a new substep for each
event. The analysis stops if the number of events (substeps) in any step
exceeds the specified maximum.

If there is a very large number of events in any analysis step, this could
indicate that the analysis has become numerically unstable and that

13-6 PERFORM-3D User Guide

 Chapter 13. Static Push-Over Load Cases

PERFORM-3D can not find a solution (experience suggests that this is

unlikely, but it could happen). In this case you would like the analysis
to stop, otherwise it might waste a lot of computer time. On the other
hand, if you have a large structure with a lot of nonlinear components
you could have a large number of real events in a step. In this case you
do not want the analysis to stop prematurely.

For a moderately large structure a reasonable maximum for the number

of events in any analysis step is about 1000. It is unlikely that there will
be that many real events in any analysis step, so it is unlikely that the
analysis will stop prematurely, and if the analysis becomes unstable
you will not waste a lot of computer time before the analysis stops.

If a push-over analysis stops because it reaches the maximum number

of events in any load step, this is an abnormal exit. However, the results
up to this point will usually be OK.

13.2.5 Initial Step to First Event

Often, first yield will not occur until there is a substantial amount of
structure drift. In this case it is more efficient to go to first yield in one
initial step, then apply equal drift increments until the maximum
allowable drift is reached, rather than have several linear steps before
first yield. Usually, therefore, you can choose the Yes option.

13.2.6 Analysis Limit State

In a push-over analysis, a point will usually be reached where the
structure is so severely deformed that there is no more useful
information to be obtained, and hence no reason to continue the
analysis. You can stop the analysis by defining a limit state that
corresponds to such a severe deformation. This is the Analysis Limit
State.

To specify the analysis limit state, choose the limit state type and name
from the lists.

If you do not specify an analysis limit state, a default state is used. For
push-over analysis the default is to stop the analysis if the deformation
for any inelastic component exceeds its X point deformation. This is
based on the assumption that the analysis is no longer meaningful if

PERFORM-3D User Guide 13-7

Chapter 13. Static Push-Over Load Cases

any element is deformed past its maximum allowable deformation. As a

general rule, this is the best choice for the analysis limit state.

13.3 Loads Based on Mode Shapes

13.3.1 Symmetrical Structure

If you choose loads based on mode shapes, and if you use more than
one mode shape, you must specify a scale factor for each mode. Only
the relative values of the scale factors are important, since the concern
is the load distribution over the building height, not the load magnitude.

Consider first the case of a symmetrical 3D structure with mode shapes

that are clearly oriented along the H1 and H2 directions. The following
procedure is suggested.

(1) Get the mode shapes and periods. This usually means that you
must run an analysis to calculate the periods before you can define
the push-over load case.

(2) Examine the mode shapes and determine, for each of the H1 and
H2 directions, which is the "first" mode and which is the "second"
mode. For example, modes 1 and 4 might be the first and second
H1 modes, and modes 2 and 5 might be the first and second H2
modes. You may also consider the "third" modes. Ignore torsional
modes.

(3) Choose a design response spectrum. This is a separate task, and

you can not do it in PERFORM-3D. Only the shape of the
spectrum is important, not the intensity.

(4) For each mode, get the mode period and hence get the spectral
acceleration from the response spectrum. This is the scale factor
for the mode.

(5) Decide on a push-over direction, say H2. Specify the angle to this
direction (= 90 degrees for H2).

(6) Specify the mode number for the "first" mode in the push-over
direction (= 2) and the scale factor for this mode from Step (4).
Choose "+D" for the base shear direction. Since the structure in

13-8 PERFORM-3D User Guide

 Chapter 13. Static Push-Over Load Cases

this example is symmetrical, the results will be the same for +D

and –D push-over.

(7) Specify the mode number for the "second" mode in the push-over
direction (= 5) and the scale factor for this mode from Step (4). If
you want the base shears for modes 1 and 2 to be in the same
direction choose "+D" for the direction, else choose "-D".

(8) If you wish, repeat Step(7) for the "third" mode.

This defines the load distribution and direction for this load case.

You may ask why it is necessary to specify the angle to the push-over
direction. For the symmetrical structure in this example, PERFORM-
3D could figure out the direction, since both mode 1 and mode 2 have
base shears only along H2. Also, since the structure is symmetrical it is
necessary to consider push-over only in the +H2 direction. In general,
however, it is not that simple. This is considered in the next section.

13.3.2 Unsymmetrical Structure or Diagonal Push-Over

There are three reasons why the push-over direction must be specified,
as follows.

(1) For a general structure the base shears for the "first" and "second"
modes may not be in exactly the same direction. Hence, the push-
over direction can not necessarily be inferred from the modal base
shear direction.
(2) You may want to push the structure in a diagonal direction that is
not related to a base shear direction for any particular mode.
Again, the push-over direction can not be inferred from the modal
base shear directions.
(3) An unsymmetrical structure may have different behavior for
positive and negative push-over. You must define the positive
push-over direction.

In the general case it is also more complex to calculate the push-over

forces. In the simple example, where the push-over direction is H2 and
the modes have base shears only in this direction, the load for each
mass is the product of the following values.

(1) The H2 mass.

PERFORM-3D User Guide 13-9

Chapter 13. Static Push-Over Load Cases

(2) The H2 displacement of the mass for unit modal amplitude.

(3) The modal (mass) participation factor of the mode for ground
acceleration in the H2 direction.
(4) The spectral acceleration value from the response spectrum.

In the general case there are loads in both the H1 and H2 directions. For
each mass the H1 force is the product of the following values.

(1) The H1 mass.

(2) The H1displacement of the mass for unit modal amplitude.
(3) The modal (or mass) participation factor of the mode for ground
acceleration in the specified push-over direction. In the simple
example this is zero.
(4) The spectral acceleration value from the response spectrum.

Similarly, for each mass the H2 force is the product of the following
values.

(1) The H2 mass.

(2) The H2 displacement of the mass for unit modal amplitude.
(3) The participation factor of the mode for ground acceleration in the
push-over direction.
(4) The spectral acceleration value from the response spectrum.

In general there will be both H1 and H2 loads at each mass point. For
equal H1 and H2 masses (which is required in PERFORM-3D), the
resultant force at any mass point is along the push-over direction.
Hence the resultant base shear is also along the push-over direction.
13.3.3 Other Key Points
The following are two other key points.

(1) Be sure to choose a reference drift that is along the push-over

direction. In the above simple example with H2 forces, if you
choose a reference drift that is in the H1 direction the reference
drift will be zero.

(2) Be sure that at least one of the controlled drifts is along the push-
over direction. If not, the specified load will not cause any drifts
and the analysis will stop. It is usually wise to have several
controlled drifts, along both directions.

13-10 PERFORM-3D User Guide

 Chapter 13. Static Push-Over Load Cases

PERFORM-3D User Guide 13-11

 Chapter 14. Dynamic Earthquake Load Cases

14 Dynamic Earthquake Load Cases

You can define and save any number of dynamic earthquake load
cases. You can then use these load cases to run structural analyses.

This chapter describes how to set up earthquake load cases. For an

explanation of how to use these load cases in structural analyses
see Chapter 20, Running Analyses.

14.1 Earthquake Loads

For dynamic step-by-step analysis the earthquake load is defined by H1

and/or H2 and/or V ground acceleration records (ground
accelerograms). Experience has shown that nonlinear analysis results
can be sensitive to relatively small changes in the ground motion.
Hence, it is usually necessary to run analyses for a number of records.
It is not usually necessary to consider vertical ground accelerations.

14.2 Procedure

14.2.1 Steps
To start a new earthquake load case press the toolbar buttons for the
Analysis phase and the Load Cases task. In the Load Cases form
choose Dynamic Earthquake as the Load Case Type, press the New
button, and enter the load case name. Keep an eye on the Status box.

The steps are as follows.

(1) If you have not already set up the needed earthquake records
(ground accelerograms) press the Add/Review/Delete
Earthquakes button. See Chapter 15, Earthquake Records for
instructions on how to set up earthquake records.

(2) Complete the Control Information for Dynamic Analysis. You

must specify the following. For more details on these items see
later in this section.

PERFORM-3D User Guide 14-1

Chapter 14. Dynamic Earthquake Load Cases

(a) Total time. See Section 14.2.2.

(b) Time step. See Section 14.2.3.
(c) Maximum number of nonlinear events allowed in any step. See
Section 14.2.4.
(c) Interval for saving results. See Section 14.2.5.
(d) A limit state to stop the analysis. See Section 14.2.6.
(e) The reference drift. For processing the analysis results,
PERFORM-3D uses the reference drift as an overall measure
of the structure response. As a general rule you should use the
drift at the roof as the reference drift.

(3) Specify the angle from the structure H1 axis to the earthquake Q1
direction. This allows you to apply an earthquake in any horizontal
direction.

(4) Choose the Q1 Earthquake, if any, from the list of earthquake

records. The peak acceleration and duration for the record will be
shown in the boxes. Also specify acceleration and time scale
factors. The accelerations and times in the earthquake record are
multiplied by these factors. The time factor should usually be 1.0.
You can plot the earthquake record before you specify the scale
factors (See Chapter 15, Earthquake Records), but to plot it after
scaling you must run an analysis and use the Time History task.
Hence, take care with the scale factors.

(5) Similarly choose the Q2 and V Earthquakes, if any. The default is

none.

(6) Press the Save button to save the load case. Use the Delete button
if you want to delete a load case before saving it.

You can edit a saved load case at any time. If you do, keep an eye on
the Status box. Use the Save, Save As and UnChange buttons in the
usual way. You can also delete saved load cases.

14.2.2 Total Time

The total time is the duration of the earthquake for the analysis. This
may be longer or shorter than the length of the earthquake record. If it
is longer the record is padded with zeros. If it is shorter only a part of
the record is used.

14-2 PERFORM-3D User Guide

 Chapter 14. Dynamic Earthquake Load Cases

14.2.3 Time Step

The analysis is performed using step-by-step integration through time,
using the constant average acceleration method (also known as the
trapezoidal rule or the Newmark  = 1/4 method). You must specify the
integration time step. The number of steps is the total time divided by
the time step, unless the analysis terminates before the end of the
earthquake.

Since dynamic earthquake analyses can be time consuming, it is natural

to want to use as large a time step as possible. There are two things to
keep in mind when choosing a time step, as follows.

(1) You must accurately capture the structure response. For a linear
single degree-of-freedom system, the constant average acceleration
method gives sufficiently accurate results for most practical
purposes if the time step is smaller than about 1/12 th of the
structure period. For a linear multi-degree-of-freedom structure,
one method for choosing the time step is to identify the highest
mode that contributes significantly to the response, and use a time
step equal to 1/12th of the period for that mode. Since most
structures increase in period as they yield, a time step based on
linear behavior should usually be short enough for calculating
inelastic response.

(2) You must accurately capture the ground motion. For example, if
the earthquake record is at 0.02 second intervals, you should not
use a time step larger than 0.02 seconds, even for a very long
period structure, otherwise you risk distorting the ground motion.

You may need to experiment by progressively reducing the time step

until there are no significant changes in the analysis results.

Note that you may not save much time by increasing the time step.
PERFORM-3D uses an event-to-event solution strategy, where the
structure properties are re-formed each time there is a nonlinear event
(a change in stiffness). The program automatically divides each time
step into a number of substeps, with a new substep for each event.
Increasing the time step has no effect on the total number of substeps,
and hence may not have much effect on the total computation time.

PERFORM-3D User Guide 14-3

Chapter 14. Dynamic Earthquake Load Cases

14.2.4 Maximum Number of Events

As noted, PERFORM-3D uses an event-to-event solution strategy. The
analysis stops if the number of events (substeps) in any step exceeds
the specified maximum.

If there is a very large number of events in any time step, this could
indicate that the analysis has become numerically unstable and that
PERFORM-3D can not find a solution (experience suggests that this is
unlikely, but it could happen). In this case you would like the analysis
to stop, otherwise it might waste a lot of computer time. On the other
hand, if you have a large structure with a lot of nonlinear components
you could have a large number of real events in a step. In this case you
do not want the analysis to stop prematurely.

For a structure of moderate size, a reasonable maximum for the number

of events in any time step is about 200. It is unlikely that there will be
that many real events in any time step, so it is unlikely that the analysis
will stop prematurely, and if the analysis becomes unstable you will not
waste a lot of computer time before the analysis stops.

If an analysis stops for this reason it is an abnormal exit, but the results
up to that point will usually be OK.

14.2.5 Save Interval

By default the analysis results are saved every time step. If you wish,
you can reduce the lengths of the results files by saving the results
every N steps, where N > 1.

If you do not save the results every step, you may lose detail in time
history results plots. This does not affect the values of demand-capacity
ratios and usage ratios. These are calculated at every step (actually,
every substep).

14.2.6 Analysis Limit State

During the analysis a point may be reached where the structure is so
severely deformed that there is no more useful information to be
obtained, and hence no reason to continue the analysis. You can stop
the analysis by defining a limit state that corresponds to such a severe
deformation. This is the Analysis Limit State.

14-4 PERFORM-3D User Guide

 Chapter 14. Dynamic Earthquake Load Cases

To specify the analysis limit state, choose the limit state type and name
from the lists.

If you do not specify an analysis limit state, a default state is used. For
dynamic earthquake analysis the default is to stop the analysis if the
deformation for any inelastic component exceeds its X point
deformation. This is based on the assumption that the analysis is no
longer meaningful if any element is deformed past its maximum
allowable deformation. As a general rule, this is the best choice for the
analysis limit state.

PERFORM-3D User Guide 14-5

 Chapter 15. Earthquake Records

15 Earthquake Records

You can set up and store any number of earthquake records. You
can then use any record to define an earthquake load case. This
chapter describes how to set up earthquake records.

15.1 Organization

15.1.1 Overall Steps

The overall steps for setting up a new earthquake record are as follows.

(1) Obtain a text (formatted) file containing the earthquake ground

accelerations. This file can have a variety of formats, as explained
later in this chapter.
(2) Decide on a group and name for the earthquake record. This is
explained later.
(3) Convert the ground accelerations file into an earthquake record.
This is explained later.

You can then use the record in any earthquake load case, for the current
structure or any other structure.

15.1.2 Organization of Earthquake Records

PERFORM-3D is installed, by default, in the following folder.

C:\ProgramFiles\Computers and Structures\PERFORM\PERFORM-

3D\Program

During installation you may have specified a different folder.

The first time that you execute PERFORM-3D it creates a Records

folder (if the program is distributed with example earthquake records,
the Records folder is created when the program is installed). If you
installed the program in the default location, the Records folder is :

C:\ProgramFiles\Computers and Structures\PERFORM\Records

PERFORM-3D User Guide 15-1

Chapter 17. Analysis Series

Note that the parent folder is PERFORM, not PERFORM-3D. This is

done so that the earthquake records are available to other programs in
the PERFORM series.

All earthquake records are (and must be) placed in the Records folder.
If they are not, PERFORM-3D will not be able to find them.

Within the Records folder you can organize the earthquake records into
related groups, where the records in any group are in a subfolder of the
Records folder. Since you can have a large number of records, it is a
good idea to organize them in groups.

Each time you add a new earthquake record you must specify the group
name (max. 12 characters). If you specify a new group, PERFORM-3D
creates a record group folder of the same name, as a subfolder of the
Records folder. For example, if you name the group "Group-A", the
default location of the folder for this group is :

C:\ProgramFiles\Computers and
Structures\PERFORM\Records\Group-A

Each time you add an earthquake record to a group you must specify a
file name (max. 12 characters) and a record name (max. 40 characters).
The earthquake record is stored as a file in the folder for the group,
with the specified file name. For example, if you name the earthquake
record "Eqke-A" and put it in Group-A, the default location of the file
is :

C:\ProgramFiles\Computers and
Structures\PERFORM\Records\Group-A\Eqke-A

When you use an earthquake record in an earthquake load case you

must identify the record by its group name and record name (see the
preceding chapter). You do not use the file name for this purpose. The
file name is important only if you need to copy earthquake records
from one computer to another.

Once you have set up a record group folder and the files for a number
of earthquake records in the folder, do not change the folder or file
names. If you do, PERFORM-3D may not be able to find the
earthquake record.

15-2 PERFORM-3D User Guide

 Chapter 15. Earthquake Records

15.1.3 Copying Earthquake Records

If you have PERFORM-3D installed on a number of computers, you
may wish to set up a number of earthquake records on one computer,
using the procedure described in the next section, then copy these
records to another computer, rather than go through the set-up
procedure again.

You can copy any of the following. If the Records folder does not exist
on the second computer, you may need to create it.

(1) The complete Records folder.

(2) The folder for any record group.
(3) The file for any record.

15.2 Setting Up a New Earthquake Record

15.2.1 Text File Format

The ground accelerations file must be a text file. It can have a number
of header lines before the accelerations begin. The rest of the file must
then contain one of the following.

(1) Accelerations only. There must be a fixed number of values per

line at a constant time interval.
(2) Time-Acceleration pairs. There must be a fixed number of pairs
per line.
(3) Acceleration-Time pairs.
(4) Acceleration-Velocity-Displacement sets. There must be a fixed
number of sets per line at a constant time interval.
(5) Time-Acceleration-Velocity-Displacement sets. There must be a
fixed number of sets per line.
(6) Acceleration-Velocity-Displacement-Time sets.

The values in any line can be separated by spaces or commas.

The file can be on a diskette or CD, or it can be on your hard drive.

PERFORM-3D User Guide 15-3

Chapter 17. Analysis Series

15.2.2 Adding A New Record

To read a ground accelerations text file and set up a corresponding
earthquake record, press the toolbar buttons for the Analysis phase and
the Load Cases task, choose Dynamic Earthquake for the Load Case
Type, and press the Add/Review/Delete Earthquakes button. You can
then add one or more new earthquake records, review (plot) any
existing record, and delete any existing record.

The steps to add a new record are as follows.

(1) Enter the name of the ground accelerations text file. If you have
copied this file to the User folder you need to enter only the file
name. Otherwise you must give the complete path to the file. The
file can be on a diskette or CD if you wish.

(2) Choose the file contents type from the list.

(3) If the contents type is Accelerations only or Accn-Veloc-Displ

sets, specify the time interval and whether the first value or set of
values is at time step 0 or time step 1. If in doubt specify time step
1, and the initial value will be set to zero.

(4) Choose the acceleration unit from the list.

(5) Specify the duration to be used for the resulting earthquake record.
This may be longer or shorter than the maximum duration in the
file. If it is longer the resulting earthquake record is padded with
zeros. If it is shorter only a part of the file is read.

(6) Specify the number of acceleration values per line (or the number
of acceleration-velocity-displacement sets).

(7) Specify the number of header lines to be skipped before the

beginning of the acceleration data.

(8) Specify whether the values are separated by spaces or commas.

(9) If the record is to be placed in an existing record group, choose the

name of the group from the list. To create a new group, press the
New button and specify the group name.

15-4 PERFORM-3D User Guide

 Chapter 15. Earthquake Records

(10) Specify a file name and a record name for the earthquake record.
Choose the record name carefully because you must use it to
identify the record when you review it or use it in a load case.

(11) To read the file and set up the earthquake record press the Read
File button. If the file is read successfully the earthquake record
will be plotted. If the record is OK press Save to save the record,
else press Cancel, revise the above data, and try again.

When you have set up all of the needed records, press Return to
Earthquake Load Case.

15.3 Reviewing Existing Records

To review any record go to the Load Cases task, choose Dynamic

Earthquake for the Load Case Type, press the Add/Review/Delete
Earthquakes button, and choose the group and record names from the
lists in the Review/Delete Earthquakes section. To review the record
press Review. To delete it press Delete.

PERFORM-3D User Guide 15-5

 Chapter 16. Dynamic Force Load cases and Analyses

16 Dynamic Force Load Cases and Analysis

This chapter describes the procedures for setting up Dynamic Force

Load Cases. Dynamic force analysis calculates the response of the
structure to a dynamically varying force, such as wind force or blast
load.

This chapter also shows how dynamic force analysis can be used for
earthquake analyses of long-span structures with multi-support
excitation, where different support points have different ground
motions. This is, however, an indirect method and it is an interim
solution. A better and more direct method is to impose dynamic
displacements using support spring elements with initial deformations.
This option is currently available only for static imposed
displacements. It will be extended to dynamic imposed displacements in
a future version.

16.1 Purpose and Overall Procedure

16.1.1 Purpose
A dynamic force analysis calculates the response of the structure to a
dynamically varying force, such as wind force or blast load. This type of
analysis can also be used for dynamic earthquake analysis where different
supports for the structure have different ground motions (i.e., for
earthquake analysis with multi-support excitation).
16.1.2 Overall Procedure
A dynamic force load has two parts, as follows.

(1) A nodal load pattern. This is defined in exactly the same way as a
nodal load pattern for gravity or static push-over analysis. It defines
the spatial distribution of the load, as nodal loads of specified
magnitudes.
(2) A dynamic force record. This defines the time variation of the loads. A
dynamic force record is actually a time history of load factors that are
to be applied to the nodal load pattern. It is set up in much the same
way as an earthquake ground acceleration record. However, it defines
a load factor history rather than a ground acceleration history.

PERFORM-3D User Guide 16-1

Chapter 17. Analysis Series

The overall steps for setting up a dynamic force load case are as follows.

(1) Define the required nodal load patterns.

(2) Set up the required dynamic force records.
(3) Define one or more dynamic force load cases. In any load case you
can apply up to 40 load patterns, each with a corresponding force
record. Hence, if you use a dynamic force load case for an earthquake
analysis with multi-support excitation, you can apply up to 40
different motions.

16.2 Organization of Dynamic Force Records

16.2.1 RecordsF Folder

PERFORM-3D is installed, by default, in the following folder.

C:\ProgramFiles\Computers and Structures\PERFORM\PERFORM-

3D\Program

During installation you may have specified a different folder.

The first time that you execute PERFORM-3D it creates a RecordsF folder
(if the program is distributed with example dynamic force records, the
RecordsF folder is created when the program is installed). If you installed
the program in the default location, the RecordsF folder is:

C:\ProgramFiles\Computers and Structures\PERFORM\RecordsF

Note that the parent folder is PERFORM, not PERFORM-3D. This is done
so that the dynamic force records are available to other programs in the
PERFORM series.

All dynamic force records must be placed in the RecordsF folder. If they
are not, PERFORM-3D will not be able to find them.

Within the RecordsF folder you can organize the records into related
groups, where the records in any group are in a subfolder of the RecordsF
folder. If you have a large number of records, it is a good idea to organize
them in groups.

Each time you add a new dynamic force record you must specify the group
name (max. 12 characters). If you specify a new group, PERFORM-3D
creates a folder for the group, as a subfolder of the RecordsF folder. For

16-2 PERFORM-3D User Guide

 Chapter 16. Dynamic Force Load cases and Analyses

example, if you name the group "Group-A", the default location of the
folder for this group is:

C:\ProgramFiles\Computers and
Structures\PERFORM\RecordsF\Group-A

Each time you add a dynamic force record to a group you must specify a
file name (max. 12 characters) and a record name (max. 40 characters). The
record is stored as a file in the folder for the group, with the specified file
name. For example, if you name the record "Wind-A" and put it in Group-
A, the default location of the file is:

C:\ProgramFiles\Computers and
Structures\PERFORM\RecordsF\Group-A\Wind-A

When you use a dynamic force record in an dynamic force load case you
must identify the record by its group name and record name. You do not
use the file name for this purpose. The file name is important only if you
need to copy records from one computer to another.

Once you have set up a record group folder and the files for a number of
records in the folder, do not change the folder or file names. If you do,
PERFORM-3D may not be able to find the records.
16.2.2 Steps for Setting Up Dynamic Force Records
The steps for setting up a new dynamic force record are as follows.

(1) Set up a text (formatted) file containing the record. This file can have
a variety of formats, as explained later in this chapter.
(2) Decide on a group and name for the record, as explained in the
preceding section.
(3) Convert the text file into a PERFORM force record file. This is
explained later.

You can then use the record in any dynamic force load case, for the current
structure or any other structure, and for PERFORM-COLLAPSE as well as
PERFORM-3D.
16.2.3 Copying Dynamic Force Records
If you have PERFORM-3D installed on a number of computers, you may
wish to set up a number of dynamic force records on one computer, then
copy these records to another computer, rather than go through the set-up
procedure again.

PERFORM-3D User Guide 16-3

Chapter 17. Analysis Series

You can copy any of the following. .

(1) The complete RecordsF folder.

(2) The folder for any record group.
(3) The file for any record.

16.3 Setting Up a New Dynamic Force Record

16.3.1 Text File Format

The dynamic force file must be a text file. It can have a number of header
lines before the force values (load factors) begin. The rest of the file must
then contain one of the following.

(1) Forces only. There must be a fixed number of values per line at a
constant time interval.
(2) Time-Force pairs. There must be a fixed number of pairs per line.
(3) Force-Time pairs. There must be a fixed number of pairs per line.

The values in any line can be separated by spaces or commas, but not by
tabs.

The file can be on a diskette or CD, or it can be on your hard drive.

16.3.2 Adding a New Record
To read a dynamic force text file and set up a corresponding dynamic force
record, press the toolbar buttons for the Analysis phase and the Load
Cases task, choose Dynamic Force for the Load Case Type, and press the
Add/Review/Delete Force Records button. You can add one or more new
records, review (plot) an existing record, and/or delete an existing record.

The steps to add a new record are as follows.

(1) Enter the name of the dynamic force text file. If you have copied this
file to the User folder you need to enter only the file name. Otherwise
you must give the complete path to the file. You can browse to find
the file (press the Browse button). The file can be on a diskette or CD
if you wish.

(2) Choose the file contents type from the list (Forces Only, Time-Force
pairs, or Force-Time pairs).

16-4 PERFORM-3D User Guide

 Chapter 16. Dynamic Force Load cases and Analyses

(3) If the contents type is Forces Only, specify the time interval.

(4) Specify the duration to be used for the imported record. This may be
longer or shorter than the maximum duration in the file. If it is longer
the imported record is padded with zeros. If it is shorter only a part of
the file is read.

(5) Specify the number of force values per line.

(6) Specify the number of header lines to be skipped before the beginning
of the force data.

(7) Specify whether the values are separated by spaces or commas.

(8) If the record is to be placed in an existing record group, choose the

name of the group from the list. To create a new group, press the New
button and specify the group name.

(9) Specify a file name and a record name. Choose the record name
carefully because you must use it to identify the record when you
review it or use it in a load case.

(10) To read the text file and set up the record press the Check button. If
the file is read successfully the record will be plotted. If the record is
OK press Save to save the record, else press Cancel, revise the above
data, and try again.

When you have set up all of the needed records, press Return to Dynamic
Force Load Case.

Note that you do not specify the force unit. The reason is that a dynamic
force record is a series of load factors, not actual forces. When you specify
a dynamic force load case you must associate each record with a nodal load
pattern. The actual forces at any time in the record are the load pattern
forces multiplied by the current dynamic force (load factor). The time unit
is always seconds.

16.4 Reviewing Existing Records

To review a record go to the Load Cases task, choose Dynamic Force for
the Load Case Type, press the Add/Review/Delete Force Records button,

PERFORM-3D User Guide 16-5

Chapter 17. Analysis Series

and choose the group and file names from the lists in the Review or Delete
section. To review the record press Review. To delete it press Delete.

16.5 Dynamic Force Load Cases

16.5.1 Load Cases

To start a new dynamic force load case press the toolbar buttons for the
Analysis phase and the Load Cases task. In the Load Cases form choose
Dynamic Force for the Load Case Type, press the New button, and enter
the load case name. Keep an eye on the Status box.

The steps are as follows.

(1) If you have not already set up the needed dynamic force records, press
the Add/Review/Delete Force Records button and add one or more
records as described earlier.

(2) Specify the Control Information for Dynamic Analysis. This is

essentially the same as for a Dynamic Earthquake load case (see
Chapter 14, Dynamic Earthquake Load Cases, Section 14.2). For the
Reference Drift or Reference Deflection, choose a drift or deflection
that is excited by the loading.

(3) Specify the load list, up to a maximum of 40 loads. For each load
specify a combination of a Nodal Load Pattern and a Dynamic Force
Record. You can also specify a scale factor and a time delay. If the
time delay is not zero, the dynamic force record is delayed by that
amount until it begins to be applied (in effect, zero load factors are
added to the beginning of the record). Create the list using Add, Insert,
Replace and Delete in the usual way.

(4) Press the Save button to save the load case. Use the Delete button if
you want to delete a load case before saving it.

You can edit a saved load case at any time. If you do, keep an eye on the
Status box. Use the Save, Save As and UnChange buttons in the usual
way. You can also delete saved load cases.

16-6 PERFORM-3D User Guide

 Chapter 16. Dynamic Force Load cases and Analyses

16.6 Multi-Support Earthquake Analysis

Note: We believe that the method described in this section is reasonable

for this type of analysis. However, we are continuing to study the details
and to make improvements. If you plan to use this method, we
recommend that you contact us to get the latest developments.
16.6.1 Purpose
During an earthquake the ground displaces dynamically. The ground
displacements, velocities and accelerations can all be important in
determining the response of a structure.

When a structure and the adjacent ground are instrumented, the

accelerations are measured directly, and the velocities and displacements
are usually obtained by integration. For most building structures, it is
reasonable to assume that the ground motions are the same at all support
points. For long structures such as bridges, however, different support
points may have substantially different ground motions. This will be
referred to here as “multi-support excitation”.

Using PERFORM-3D dynamic force analysis, it is possible to run analyses

for multi-support excitation. This section describes the procedure. The
method is indirect, and it has some weaknesses as noted in this section. It is
provided as a temporary method. A more direct method, based on dynamic
imposed support displacements, may be added in a future version.
16.6.2 Analysis Method
For earthquake analysis where all supports have the same ground motions,
it is usual to use the earthquake ground acceleration record to calculate
time histories of the effective inertia forces on the structure. The analysis is
then carried out for these forces. In this type of analysis the displacements
at the support points are zero, and hence the calculated displacements,
velocities and accelerations at the nodes are relative to the ground. To get
the absolute accelerations at any node (which are needed to calculate the
“floor” response spectrum), the ground accelerations must be added to the
calculated relative accelerations.

For linear structural analysis with multi-support excitation, it is usual to use

a similar analysis method. In this method the ground accelerations at the
support points are used to calculate effective inertia forces, and an analysis
is carried out for these forces. This analysis assumes fixed supports, so the
displacements at the support points are zero. This is not correct, since
different supports can have different displacements, and hence the support

PERFORM-3D User Guide 16-7

Chapter 17. Analysis Series

points can displace relative to each other. This effect is accounted for by a
second analysis in which the relative support displacements are considered.
The results of the two analyses are added together. This approach can be
used for linear time history analysis and also for response spectrum
analysis.

The time history version of this method can be extended to nonlinear time
history analysis. However, it is complex. A simpler method is to impose
dynamic support displacements on the structure, where the displacement
time history at any support is the earthquake ground displacement record.
This method is used in PERFORM-3D.

In this method, ground displacements are applied in the analysis, not

ground accelerations. One consequence of this is that the calculated node
displacements during the analysis are absolute displacements, not
displacements relative to the ground. Also, the ground accelerations follow
from the ground displacements by double differentiation, which can lead to
a loss of accuracy. These aspects are considered later in this section.
16.6.3 Imposed Dynamic Support Displacements
PERFORM-3D currently does not provide the ability to apply dynamic
support displacements directly (this will be added in a future version).
Support displacements can, however, be imposed as follows.

(1) At each node where a ground displacement is to be imposed, add a

Support Spring element. This element must be very stiff relative to the
structure stiffness, but avoid astronomically large stiffnesses. Put these
elements in one element group (this is not essential but it is usually a
good idea), and specify zero for the Scale Factor for Beta-K Damping.

It is unlikely that you will want to impose rotational displacements.

Hence, you probably do not need rotational support springs. It is
usually easier to specify translational spring stiffnesses only, and to
provide rotational support using rigid restraints.

(2) For each different ground displacement record, set up a nodal load
pattern. In each pattern, specify a large load at each affected node, in
the direction of the required displacement. If the stiffness of the
support spring element along a displacement direction is K, specify a
load with a magnitude equal to K. This will cause a unit displacement
at the node (actually slightly less, but if K is large enough the
difference is negligible).

16-8 PERFORM-3D User Guide

 Chapter 16. Dynamic Force Load cases and Analyses

(3) For each different ground displacement record set up a Dynamic Force
Record that defines the ground displacement time history. See the next
section for a method to get ground displacement records. Read the
records into PERFORM as described earlier in this Chapter.

Each dynamic force record is actually a time history of load factors

that are to be applied to a load pattern. If the forces in the load pattern
cause unit displacements, a unit load factor in the dynamic force
record corresponds to a unit displacement. Hence, the values in any
force record are numerically equal to the ground displacements for
that record.

(4) Define one or more dynamic force load cases as described earlier in
this Chapter. For each imposed displacement there will be a load
pattern and an associated dynamic force record. If you wish you can
specify a scale factor to adjust the displacement magnitude. If the load
pattern causes a unit displacement, this factor will be 1.0.

(5) Run analyses with the load cases.

16.6.4 Ground Displacements
Displacement records are usually not measured directly, but are obtained
from the ground acceleration records by double integration through time. A
displacement record obtained by double integration can be sensitive to
small errors in the acceleration record, and can show drift over the duration
of the record. That is, even if the actual ground displacement at the end of
the earthquake is zero, the displacement obtained from the acceleration
record can show a non-zero final value. It is usual (and usually necessary)
to correct for this. It is typically assumed that the displacement error varies
quadratically during the earthquake. Hence, the displacement record is
typically corrected to a base line that varies quadratically over the duration
of the record. After this correction the acceleration and displacement
records are no longer completely consistent (unless they are also
corrected), but any inconsistency is small.

Commonly, earthquake acceleration records are discretized at time

intervals of 0.01 or 0.02 seconds, which means that information on high
frequency ground accelerations is not captured accurately. Time history
analyses in PERFORM-3D use the Constant Average Acceleration (CAA)
Method of step-by step integration (also known as the Trapezoidal Rule or
the Newmark  = ¼ Method). This method assumes that the acceleration is
constant in each time step, equal to the average of the values at the
beginning and end of the step. Hence the velocity is assumed to vary
linearly in a time step. Theoretically the displacement varies quadratically,

PERFORM-3D User Guide 16-9

Chapter 17. Analysis Series

but if there are events in a step they are calculated assuming a linear
displacement variation.

The CAA method is the simplest integration method, and arguably the best
for nonlinear analysis. Like all integration methods it is approximate. In
particular, it can calculate inaccurate response for vibration modes with
periods that are shorter than about 12 times the time step. For a time step of
0.02 seconds this means that the calculated response for any mode with a
period shorter than about 0.25 seconds can be inaccurate. For stiff
structures a time step shorter than 0.02 seconds should be used. Note,
however, that earthquake acceleration records are usually discretized at
0.02 second intervals, and that when a time step shorter than 0.02 seconds
is used, the accelerations are interpolated linearly. This is not strictly
correct, so the discretization of the acceleration record can cause
inaccuracy for short period structures.
16.6.5 Ground Displacement Records
If an earthquake ground displacement record is available it will typically be
discretized at 0.01 or 0.02 second time intervals. For time history analysis
it is usually assumed that the displacements vary linearly in each interval.
This means that the implied ground accelerations (which, in effect, are now
obtained by double differentiation of the displacement record) can be
substantially inaccurate for high frequency vibrations. Fortunately,
structures that have multi-support excitations are always large structures
such as bridges, where the important vibrations tend to have longer periods.

If you have only a ground acceleration record, you must double integrate it
to get the ground displacement record. You can do this directly, for
example using a spreadsheet program. Alternatively you can use a simple
PERFORM structure that models a displacement gage. This structure is
shown in Figure 16.1.

16-10 PERFORM-3D User Guide

 Chapter 16. Dynamic Force Load cases and Analyses

Figure 16.1 Displacement Gage for Converting Ground

Accelerations to Ground Displacements

To get the ground displacements you can analyze this structure using a
dynamic earthquake load case, and plot the relative displacement of the
mass. Since the mass is very large it does not move in space, so its
displacement relative to the ground is minus the ground displacement. Save
the displacement time history to a text file, and if necessary apply a
baseline correction using a spreadsheet program. Then read the time history
into PERFORM-3D as explained earlier in this Chapter.

If you double integrate an acceleration record, you must choose an

integration method and an integration time step. If you analyze the
structure in Figure 16.1, the integration method is the CAA method in
PERFORM, but you must still choose a time step. The following method is
suggested.

(1) Decide on the time step that you will use for the dynamic force
analyses of the structure (say 0.02 seconds).

The ground acceleration record will usually be at constant time

intervals, often 0.02 seconds. Since structures that have multi-support
excitation will tend to be larger structures with longer periods, the
time step will usually be this value. As a general rule, do not use a
time step that is larger than this value, since it may not capture the
ground motion accurately. If you use a time step that is smaller, make
it a whole number subdivision (e.g., for an acceleration record at 0.02
second intervals, use time steps such as 0.02, 0.01. 0.005, 0002
seconds.

(2) Use this time step to analyze the displacement gage in Figure 16.1,
and output the displacements at every time step. This is the
displacement record.

This method is suggested because it should be the most consistent. For the
case where all supports move in phase, this method should give the same
ground motions on the structure whether you run (a) a dynamic force
analysis with the calculated ground displacement record or (b) a dynamic
earthquake analysis using the ground acceleration record.

PERFORM-3D User Guide 16-11

Chapter 17. Analysis Series

16.6.6 Alpha-M Damping

“Alpha-M” damping implies that there is a linear viscous damper that
connects each mass to a fixed point. For dynamic earthquake analysis,
these fixed points are fixed to the ground, and the deformation rates of the
viscous dampers depend on the velocities of the nodes relative to the
ground. For a dynamic force analysis, however, the points are fixed in
space, while the ground moves, and the deformation rates of the viscous
dampers depend on the absolute velocities of the nodes. Hence, Alpha-M
damping has a different physical meaning for the two types of analysis.
“Beta-K” viscous damping depends on element deformations (i.e., relative
displacements), so this damping is not affected.

It is not clear whether this has much effect. In PERFORM-3D, one way to
assess the effect of viscous damping is to plot the energy balance and
examine the amount of energy dissipated by viscous damping.
Unfortunately, a weakness of using force time-history analysis for
earthquake excitation is that large deformations are imposed on the stiff
support spring elements, and hence the strain energy in these elements is
very large. When you plot an energy balance graph, you will find that
virtually all of the energy is strain energy in the support springs, and you
can not tell how much energy is dissipated by viscous damping. You can
still see the relative amounts of energy dissipated by inelastic hysteresis in
the element groups. Also, you can get the amounts of M and K damping
from the ECHO.txt file (from the printed energy balance at the end of each
analysis). However, this is not as convenient as using an energy balance
graph.
16.6.7 Results Interpretation
The results interpretation is similar to that for a dynamic earthquake
analysis. However, please note the following two aspects.

(1) As noted in the preceding section, you can not get an overall energy
balance, because the strain energy in the support springs is much
larger than the other energies.

(2) If you plot a time history of displacement at a support point, you will
find that it closely matches the input displacement (assuming the
support spring is sufficiently stiff). However, if you plot a velocity
time history you may find that it is more jagged than expected, and if
you plot an acceleration time history you may find that it is very
jagged. The acceleration time history should, of course, match the
original ground acceleration record. You may find that it does this
fairly well for the first second or two of the response, but then

16-12 PERFORM-3D User Guide

 Chapter 16. Dynamic Force Load cases and Analyses

becomes very jagged with large oscillating values. We think that this
is numerical precision problem, and we are looking into
improvements. We do not think that it affects the overall results.

Note that node displacements, velocities and accelerations are absolute

values. However, drifts are still relative.

PERFORM-3D User Guide 16-13

 Chapter 17. Analysis Series

17 Analysis Series

When you have set up an analysis model and some load cases you
are ready to run structural analyses. Before you do this you must
set up one or more Analysis Series. This chapter explains the
procedure.

17.1 Analysis Series Concept

17.1.1 Concept
At this stage you have set up the structure and the load cases. An
"Analysis Series" allows you to specify certain other parameters before
you run any analyses. Each analysis series corresponds to a different
analysis model, so that you can analyze several different models
without creating a complete new structure. For example, you may want
to run one set of analyses ignoring P- effects and a second set
considering these effects. To do this, you can set up one analysis series
with P- effects turned on, and a second series with these effects turned
off. You can also specify different masses, damping and certain other
parameters.

You can have as many analysis series as you like. You will usually run
several analyses for each series. For the steps in running analyses, see
Chapter 20, Running Analyses.

17.1.2 Procedure
Figure 17.1 shows the form for specifying Analysis Series parameters.

PERFORM-3D User Guide 17-1

Chapter 17. Analysis Series

Figure 17.1 Form for Analysis Series Parameters

The steps for starting a new analysis series are as follows.

(1) Press the toolbar buttons for the Analysis phase and the Run
Analyses task.

(2) Choose the Start a new analysis series option. The other options
are to continue or delete an existing series, as explained later.

(3) Input the series name (choose this carefully) and description.

You can specify the following parameters.

17-2 PERFORM-3D User Guide

 Chapter 17. Analysis Series

(1) Type of load sequence (Standard or General).

(2) Whether or not P- effects are to be considered.
(3) Masses.
(4) Number of mode shapes to be calculated.
(5) Modal damping.
(6) Rayleigh (M+K) damping.
(7) Upper/lower bound ratios.
(8) Certain options for simplified analysis.

See Chapter 18, Elastic Viscous Damping for an explanation of Modal

and Rayleigh damping. See the later sections in this chapter for
explanations of the other parameters.

After you have specified the properties for the Analysis Series, press
OK to save the properties and move to the Run Analyses form. See
Chapter 20, Running Analyses.

You can add analyses to an existing series, or change the parameters for
an existing series, using the Continue or change an existing series
option. If you change the parameters for an existing series, the analyses
in that series are no longer valid, and they will be deleted (you will be
warned before this happens). If you make changes to the parameters
and want to cancel those changes, press the Unchange button.

You can delete an existing series, using the Delete an existing series
option. You can delete just the analyses in any series, or you can delete
the series completely.

17.2 Basic Parameters and Masses

To specify the masses and some basic parameters, use the Basic +
Masses page and do the following.

(1) Choose whether or not P- effects are to be considered. If you

choose “Yes”, P- effects are considered, for those element groups
that have P- effects. If you choose “No”, P- effects are ignored.
When you are running analyses for a new structure, always choose
“No” until you are sure the analysis model is OK. Then choose
“Yes” to add P- effects.

PERFORM-3D User Guide 17-3

Chapter 17. Analysis Series

(2) Choose an event overshoot factor. This is explained later in this

chapter. For a large structure you can save computer time by
specifying a value larger than the default value of 1.0.

(3) Choose whether you want a Standard or General load sequence. In

almost all cases you can use the Standard sequence. The General
sequence is for advanced users with special needs. For an
explanation of the General sequence, see Chapter 19, General
Load Sequence.

(4) Specify scale factors for the mass patterns. If you have not
specified any mass patterns, or if you specify zero scale factors for
all patterns, the analysis model has no mass. Hence, you can apply
gravity and static push-over load cases but not dynamic earthquake
loads. Also, you can not use the General Push-Over Plot or Target
Displacement Push-Over Plot task to assess performance. In the
early stages of an investigation you may have an analysis series
with no mass, but in order to get useful results you will need to
specify masses. For a new analysis model it is can be a good idea
to specify zero mass and run only static analyses until you are sure
the model is correct.

(4) Specify the number of mode shapes that are to be calculated. If you
use only Rayleigh damping, you will need only a few mode
shapes. If you use modal damping you will usually need more
shapes.

17.2.1 Event Overshoot Factor

PERFORM-3D uses an event-to-event solution strategy. In this strategy
the program determines when a significant change occurs in the
structure stiffness (an "event"). If an event occurs within a load step or
a time step, the step is subdivided, and the structure stiffness is
modified.

The event-to-event strategy is more reliable than other strategies. It has

the disadvantage, however, that for a large structure the number of
possible events increases, and the time taken to modify the stiffness
after an event also increases. Hence, the computation time increases
rapidly as the size of the structure increases. With modern high speed
computers, experience has shown that this is not a serious problem, and

17-4 PERFORM-3D User Guide

 Chapter 17. Analysis Series

that PERFORM-3D is efficient computationally even for very large

structures. Nevertheless, it can be useful to use the Event Overshoot
Factor to reduce the computation time. This factor works as follows.

If PERFORM-3D were to consider each stiffness change exactly as it

occurs, the number of events could be very large. To reduce the number
of events, PERFORM-3D uses an event overshoot tolerance, so that a
yield event does not occur until the calculated force exceeds the yield
force by a small amount (the overshoot tolerance). If there are closely
spaced stiffness changes, a number of these changes can occur within
this tolerance. If this happens, these stiffness changes are treated as a
single event, and the total number of events is reduced.

The disadvantage of using an overshoot tolerance is that the overshoot

causes an equilibrium error (unbalanced forces) that must be corrected.
If there is an equilibrium error at the end of any substep, PERFORM-
3D corrects for it in the following substeps, by applying the unbalanced
forces as extra loads. Any unbalanced forces are thus temporary, and
usually have little effect on the calculated response. Equilibrium errors
also lead to energy errors, where the work done by the external forces
on the structure is different from the internal work (which is partly
dissipated and partly stored as recoverable strain or kinetic energy).
PERFORM-3D calculates an energy balance, which is shown in the
ECHO file. If the difference between external and internal work is
larger than a few percent, this indicates that the equilibrium errors
could be significant. In PERFORM-3D the energy balance is usually
very close, although the energy error may be larger if there is brittle
strength loss.

For elements that yield, PERFORM-3D uses a default overshoot

tolerance of 1%. This means that any unbalanced forces are relatively
small. It also means that there is a relatively small chance that a number
of stiffness changes will occur within a single event. For structures of
moderate size, it is usually a good idea to use the default tolerance. For
very large structures, however, it is usually acceptable to use a larger
tolerance, and hence reduce the number of events. The reason is that if
an element yields and there are unbalanced forces, these forces are
transferred to the other elements in the structure. In a small structure,
there are relatively few elements to absorb the unbalance, so it is best to
keep the unbalance as a small percentage of the strength of a single
element. In a large structure, however, there are more elements to
absorb the unbalance, and this unbalance can safely be a larger

PERFORM-3D User Guide 17-5

Chapter 17. Analysis Series

percentage of the strength of a single element. A larger unbalance

means a larger chance that a number of stiffness changes will occur
within a single event, so the computation time can be significantly
reduced.

When you set up a new Analysis Series, you can choose an event
overshoot factor. If you choose a factor of, say, 5, you are essentially
specifying a 5% overshoot tolerance for a yielding element. For a very
large structure you should be able to use a factor of 10 or even higher.
You should experiment to make sure that a larger tolerance does not
cause significant differences in the results, and for each analysis you
should check the energy balance in the ECHO file.

In a push-over analysis, if there are any unbalanced loads at the end of

a load step (as distinct from a substep), PERFORM-3D iterates to
eliminate this unbalance before continuing with the next load step.
Hence, a load step typically consists of a number of "advancing"
substeps, where the structure drift progressively increases, followed by
a few "correcting" substeps, where the unbalanced forces are corrected
at essentially constant drift. If you specify an event overshoot factor of
10 or larger, the correcting substeps are omitted, and any unbalanced
forces are carried over to the next load step.

Correcting steps are also used in nonlinear gravity load analyses. They
are not used in dynamic earthquake analyses (any unbalanced loads at
the end of one time step are carried over to the next time step).

17.3 Damping

17.3.1 General
See Chapter 18, "Elastic" Viscous Damping for an explanation of
Modal and Rayleigh damping.

For most structures we suggest (a) using modal damping, with a

number of modes roughly the same as you would use for a linear modal
analysis of the same structure, and (b) using a small amount of
Rayleigh damping to avoid potential problems with undamped degrees
of freedom. This is explained in Chapter 18, Section 18.4.

17-6 PERFORM-3D User Guide

 Chapter 17. Analysis Series

17.3.2 Procedure for Modal Damping

Use the Modal Damping page. The options are (a) no modal damping,
(b) the same damping ratio for all modes, and (c) damping ratios that
depend on the mode period.

If you choose the same damping ratio for all modes, specify this ratio.
This ratio is used for only the modes for which you have calculated
mode shapes. Any higher modes are not damped.

If you choose damping ratios that depend on period, specify from 2 to 6

periods and corresponding damping ratios. The damping ratio for each
mode is determined by interpolation. As before, there is damping only
in the modes for which you have calculated mode shapes.

17.3.3 Procedure for Rayleigh Damping

Use the Rayleigh Damping page.

Use the Basic Values page to specify how the damping ratio varies
with period. See Chapter 18, Section 18.2.4 for an explanation.

Use the Beta-K Options page to specify different amounts of K

damping for different element groups. The default is to use the factors
specified with the element groups.

If you have a structure with seismic isolation, you may want to use the
Alpha-M Options page. See Chapter 18, Section 18.2.6 for an
explanation. However, experience indicates that there is little difference
between the Standard and Base Isolation options. Also, for a base
isolated structure it is probably better to use modal damping.

17.4 Procedure for Using Upper/Lower Bounds

For specifying and using upper/lower bounds as component properties,

see Chapter 5, Component Properties, Section 5.15.

To use these bounds. use U/L Bounds page. Figure 17.2 shows this
page.

PERFORM-3D User Guide 17-7

Chapter 17. Analysis Series

Figure 17.2 Form for U/L Bound Ratios

When you set up a new Analysis Series, you have the option of varying
the strengths and/or stiffnesses of any or all components. The default is
to use nominal strength and stiffness values for all components. If you
want to use larger or smaller values for any component, you must
specify U/L Bound Ratios for that component. These ratios have the
following meaning.

(1) If the ratio = 0, use the nominal value for the corresponding
strength or stiffness.
(2) If the ratio = 1.0, use the upper bound value.
(3) If the ratio = -1.0, use the lower bound value.

17-8 PERFORM-3D User Guide

 Chapter 17. Analysis Series

(4) If the ratio is between 0 and 1.0, interpolate between the nominal
and upper bound values.
(5) If the ratio is between 0 and -1.0, interpolate between the nominal
and lower bound values.

In general you can specify U/L Bound Ratios for axial strength (P),
bending strength (M), shear strength (V) and stiffness (K). However,
most components have upper/lower bound values for only some of
these quantities. For example, a FEMA Beam, Steel Type component
has bounds for moment strength, shear strength and (bending) stiffness.
This component has only moment strength, but shear strength is
included because the rotation capacity can depend on shear force.

The steps for specifying U/L Bound Ratios are as follows.

(1) Choose a component group (Inelastic, Elastic, Material or Strength

Section) from the Group list.
(2) Choose a component type (e.g., Inelastic Bar) from the Type list.
All components of that type will be listed in the table. You can
reduce the length of the list by typing a text string in the “Text for
filter” box and pressing the Filter button. Only component names
that include the text string will be listed.
(3) Specify U/L Bound Ratios.
(4) Click individual components in the table, to highlight them. Click a
second time to remove a highlight. Alternatively, check Select All
to select all components in the table.
(5) Press Apply. The specified ratios will be shown in the table, and
the table will also be saved. You must press Apply for changes to
take effect.
(6) Repeat as necessary for all components that have U/L Bound
Ratios.

You can edit the tables at any time. If analyses have been run in an
Analysis Series and you change any U/L Bound Ratios, the analyses
must be deleted, since the structure has changed.

PERFORM-3D User Guide 17-9

Chapter 17. Analysis Series

17.5 Quick'n'Dirty

It can often be useful to run a simplified analysis, to save time or to

check an analysis model. The "Quick'n'Dirty" page allows you to run
simplified analyses without changing the analysis model.

At present there is only one option on this page, namely ignoring

strength loss. If you have components that have strength loss, it can be
useful to run preliminary analyses without strength loss, to check the
analysis model. We strongly recommend that you do this in all cases
with strength loss, until you are sure that the model is correct. Then
start a new Analysis Series where strength loss effects are included.
(We also recommend that you turn off P- effects until you are sure
that the model is correct.)

For this option, cyclic degradation of stiffness and energy is also

ignored.

17-10 PERFORM-3D User Guide

 Chapter 18. "Elastic" Viscous Damping

18 "Elastic" Viscous Damping

A structure that is essentially elastic dissipates energy by various

mechanisms. This is "elastic" energy dissipation, which is usually
modeled using viscous damping. A common assumption for linear
structural analysis is 5% of critical viscous damping in each
natural mode of vibration. Since the actual damping mechanism is
almost certainly not viscous, this is an approximation.

After a structure yields it dissipates additional energy through

inelastic behavior. In a nonlinear dynamic analysis this is modeled
directly. The total energy dissipation is the sum of the elastic and
inelastic dissipation.

In nonlinear dynamic analysis it is usual practice to assume viscous

damping to account for the “elastic” energy dissipation.
PERFORM allows two types of viscous damping, namely Modal
and Rayleigh damping. This Chapter describes the assumptions and
theory.

18.1 Modal Damping

18.1.1 Damping Matrix

When modal damping is specified, PERFORM uses a damping matrix
based on the modes for which mode shapes have been calculated. You
can specify that mode shapes are to be calculated for up to 50 modes,
but it is usually not necessary to consider that many. As a general rule it
should be sufficient to consider as many modes as you would for a
linear analysis of the same structure.

PERFORM uses the (elastic) mode shapes and periods, and calculates
the implied damping matrix using structural dynamics principles. This
matrix is given by

n N
4 
M n M n  T
C  T
n 1 n
n
nT M n
(18.1)

PERFORM-3D User Guide 18-1

Chapter 18. "Elastic" Viscous Damping

where N = number of damped modes, Tn = mode period,  n = mode

damping ratio, M = mass matrix and n = mode shape.

As noted later, when you use modal damping, we recommend that you
add a small amount of Rayleigh damping.

If you do not understand the following paragraph, don’t worry. The

important thing is that modal damping can cause an increase in the
energy balance error. Experience indicates that this effect is minor.

For solution of the equilibrium equations at each step of an analysis, the

structure stiffness matrix is banded. However, the damping matrix is
not banded. All terms in the damping matrix that lie within the stiffness
matrix band are included in the stiffness matrix (they are “left-hand-
sided”). The remaining coefficients are not included in the stiffness
matrix, but are considered when the resisting force, and hence the
equilibrium unbalance, is calculated at the end of the step (they are
“right-hand-sided”). Because of this, the damping forces cause some
equilibrium unbalance. This unbalance is added to any unbalances that
result from nonlinear behavior. This tends to increase the energy
balance error. Experience indicates that the effect is usually minor.

18.1.2 Physical Interpretation

It is useful to consider the physical meaning of modal damping, and to
understand the difference between modal damping in linear and
nonlinear structures.

As a linear structure deforms under dynamic earthquake load, its

deformed shape continually changes. At any instant of time this shape
can be decomposed into mode shapes, each mode shape vibrating
independently in its natural frequency. If modal damping is assumed,
each mode is independently damped. Decades of experience have
shown that it is reasonable to assume model damping for linear
analysis.

In principle this could be extended to a nonlinear structure. If the

behavior of the structure is linear between well defined nonlinear
"events", such as plastic hinge formation, the mode shapes could be
recalculated at each event (i.e., each time the structure stiffness
changes), and hence modal damping could be assumed. It is

18-2 PERFORM-3D User Guide

 Chapter 18. "Elastic" Viscous Damping

prohibitively expensive, however, to recalculate the mode shapes (and

hence, in effect, the damping matrix) at every event. Hence, when
modal damping is used for a nonlinear structure, it is assumed that the
damping matrix remains constant. At any instant of time the deformed
shape of the structure still contains contributions from the elastic mode
shapes. However, unlike the linear case, the effective periods of
vibration for these shapes are not the linear periods, the shapes are
generally not independent (uncoupled), and the deformed shape
contains shapes other than the linear mode shapes. This has the
following effects.

(1) The mode shapes are still damped, but since the effective period
may have changed (probably increased) while the effective
damping coefficient is unchanged, the amount of damping,
expressed as a proportion of critical damping, generally changes.
(2) The only components of the deformed shape that are damped are
those that correspond to the linear mode shapes. All other
deformations are undamped.

This does not invalidate a damping model based on linear mode shapes.
It just indicates that a physical interpretation is difficult. This type of
model is probably the best that is currently available (with the addition
of a small amount of Rayleigh damping, as noted later).

The next section describes the Rayleigh damping model in some depth.
A later section makes suggestions for a reasonable damping model that
combines modal damping with a small amount of Rayleigh damping.

18.2 Rayleigh Damping

18.2.1 Rayleigh Damping Model

Rayleigh damping assumes that the structure has a damping matrix, C,
given by:

C M  K (18.2)

where M is the structure mass matrix, K is the initial elastic stiffness

matrix, and  and  are multiplying factors.

PERFORM-3D User Guide 18-3

Chapter 18. "Elastic" Viscous Damping

The physical meaning of M+K damping is illustrated in Figure 18.1.

M1  M1
M1 M1


C C C C

M2  M2
M2 M2


C C C C

Figure 18.1 Physical Meaning of M+K Damping

The figure shows a frame with beam stiffnesses KB, column stiffnesses
KC, and horizontal masses M1 and M2. The physical meaning of M
damping is that viscous dampers with damping coefficients M1 and
M2 connect the masses to external fixed points. The physical meaning
of K damping is that internal viscous dampers with damping
coefficients KB and KC act in parallel with the elements. For a truss
bar element the K damper is a simple axial damper. For a frame
element the concept is the same, but physically the damper is a
velocity-dependent bending element.

18.2.2 Implied Modal Damping Ratio for Linear Analysis

For linear analysis, modal damping is almost always used. However,
Rayleigh damping can be used as an alternative. It can be shown that
M damping corresponds to more damping in the lower (longer period)
modes and less damping in the higher (shorter period) modes. The
relationship between damping ratio and mode period is

18-4 PERFORM-3D User Guide

 Chapter 18. "Elastic" Viscous Damping

Ti
i   (18.3)
4

where Ti = period of mode i and  i = proportion of critical damping in

this mode.

It can also be shown that K damping corresponds to less damping in

the lower modes and more damping in the higher modes, with the
relationship:


i   (18.4)
Ti

For linear analysis, it can also be shown that Rayleigh damping does
not cause coupling between mode shapes (i.e., if a structure is
deformed in a mode shape, the damping forces affect only that mode
shape, and have no effects on any other mode shapes). Modal damping
does the same, of course. With modal damping, however, the damping
ratio can be specified independently for each mode, and is usually
specified to be the same in all modes. With Rayleigh damping this is
not the case. There are only two independent parameters for Rayleigh
damping, namely the factors  and . Hence, the damping ratio can be
specified for only two modes. For the other modes the damping ratio
varies with mode period as shown in Figure 18.2

Figure 18.2 Variation of Damping Ratio with Period for

Linear Analysis

PERFORM-3D User Guide 18-5

Chapter 18. "Elastic" Viscous Damping

It is not possible to get the same damping ratio in all modes. However,
by choosing appropriate values for  and  it is possible to have
essentially constant damping over a significant range of periods, as
indicated in Figure 18.2.

When this type of damping is used for linear dynamic analysis, the
longer period, TB , is usually the first mode period (or close to it) and
the shorter period, T A , corresponds to some higher mode. For example,
if  and  are chosen so that the damping is 5% at TB  0.9T1 , where
T1 is the first mode period, and so that the damping is also 5% at
TA  0.25T1 , then the damping is close to 5% over a range of periods
from 0.2T1 to T1 . This might cover the most important modes. Higher
modes are more heavily damped.

18.2.3 Implied Damping for Nonlinear Analysis

As with modal damping, it is possible in principle to recalculate the
mode shapes and periods at each nonlinear event, and hence to
recalculate the damping matrix. The implied modal damping ratios
would then be as in the preceding section. In practice, however, this is
too expensive computationally, and the damping matrix is kept
constant, based on Equation 18.2 with K equal to the elastic stiffness
of the structure.

The physical interpretation is similar to that for modal damping, with

the following differences.

(1) All mode shapes are damped, not just the lower modes.

(2) Almost all structures have more degrees of freedom than mode
shapes (roughly 6 times the number of nodes for the total number
of DOFs, versus 2 or 3 times the number of masses for the number
of mode shapes). The deformed shapes that do not correspond to
mode shapes have zero mass and hence zero period. This does not,
however, mean infinite damping, because the dynamic analysis is
carried out step-by-step with a finite time step, and the response in
any mode with a period smaller than about 10 times the time step

18-6 PERFORM-3D User Guide

 Chapter 18. "Elastic" Viscous Damping

is inaccurate. There is, however, high effective damping in the

short (and zero) period modes.

Hence, Rayleigh damping, with T A and TB chosen as in the preceding

section, is similar to modal damping, but with more damping in the
higher modes shapes.

18.2.4 PERFORM Implementation

In PERFORM-3D you do not specify values for  and  directly.
Instead you must specify two period ratios, TA / T1 and TB / T1 , and the
damping percentages at these ratios (usually the same percentage at
both). PERFORM-3D then calculates the required values of  and .

There are two points of view on how the periods TA / T1 and TB / T1

should be chosen, as follows.

(1) Use the same method as might be used for linear analysis. That is,
specify TA / T1 = about 0.25 and TB / T1 = about 0.9.

(2) Observe that as a structure yields it usually softens, and it hence its
effective vibration periods usually increase. Period is inversely
proportional to the square root of stiffness (based on secant
stiffness). Roughly, therefore, if the stiffness decreases by a factor
of 4 the period increases by a factor of 2. In terms of ductility ratio,
a stiffness decrease of 4 corresponds roughly to a ductility ratio of
4. Hence, a ductility ratio of 4 corresponds roughly to a period
increase of 2 (or, in general, a ductility of N corresponds roughly to
a period increase of N ).

Suppose that you have a structure where the response is dominated

by the first mode, and that you estimate a ductility ratio of 6. You
can expect that the first mode period will increase, and will vary in
the range from the elastic value to about 2.5 times that value.
Hence, you might like to keep the elastic damping essentially
constant over that range of periods. You can do this by specifying
T A / T1 = about 1.2 and at TB / T1 = about 2.2.

The first of these two methods gives a smaller value for  and a larger
value for  than the second method. It is not possible to say which

PERFORM-3D User Guide 18-7

Chapter 18. "Elastic" Viscous Damping

method is better. Intuitively, the first method, with a smaller value for 
sounds preferable, and if you use only Rayleigh damping, we suggest
that you use that method.

18.2.5 Procedure for Specifying Rayleigh Damping

Use the Run Analyses task, the Analysis Series form, and the
Rayleigh Damping and Basic Values pages. The steps are as follows.

(1) Input the period ratios T A / T1 and TB / T1 , and the damping

percentage at each ratio. If you do not intend to run any dynamic
analyses, you can specify zero damping. To do this leave the boxes
blank.

(2) Press the Draw Graph button to plot the variation in damping
ratio. If you have left the boxes blank the damping will be zero.

(3) If the result is not as expected, press the Close Graph button and
try again.

18.2.6 Special Assumption for Concrete Fibers

If a structure has components that use concrete fibers (Fiber Section
and/or Concrete Strut components), these fibers are stiff in compression
and after cracking have no stiffness in tension.

For K viscous damping, the stiffness K is the initial stiffness of the

structure. This is calculated assuming that all concrete fibers are in
compression (i.e., all concrete fibers are assumed to be initially
uncracked). The same value of K is used after the concrete cracks
(i.e., K is constant). Experience indicates that this can lead to
unrealistically large amounts of viscous damping in concrete fibers
after cracking occurs.

This can occur, for example, when inelastic fiber cross sections are
used in shear wall elements. When a fiber cross section cracks, the
neutral axis shifts, and as the wall bends it also deforms axially. If the
axial K damping stiffness is large, this can lead to unrealistically large
axial viscous damping forces in the wall. It can also stiffen the wall
excessively, and lead to unrealistically large amounts of viscous energy
dissipation.

18-8 PERFORM-3D User Guide

 Chapter 18. "Elastic" Viscous Damping

To avoid this problem, PERFORM-3D makes a special assumption

when it calculates K stiffnesses for elements that use concrete fibers.
These include shear wall and general wall elements that use fiber wall
cross sections, frame type elements that use fiber cross sections, and
bar elements that use concrete strut components. For these elements,
the stiffness, K, used for the element K stiffness is based on only 15%
of the concrete material area. In other words, in components where
concrete cracking can occur, the concrete is considered to be only 15%
effective when the K damping stiffness is calculated. Steel fiber areas
are not reduced, and hence steel fibers are assumed to be 100%
effective for the K damping stiffness.

18.2.7 Alternative Procedure for Base Isolation

The procedure described in this section is available as a PERFORM-3D
option. Experience suggests that choosing this option has little effect on
the analysis results. Hence, we suggest that you do not bother.

If you specify M +K damping for a structure with seismic base

isolation, the physical meaning of the model is as shown in Figure
18.3(a). In this case the M dampers connect the structure masses to
the ground level, below the isolation level. This may not be what you
intend, since it can overestimate the amount of damping. An alternative
model is shown in Figure 18.3(b). In this case the M dampers connect
the structure masses to the level immediately above the seismic
isolators, so that the M part of the damping applies only to the
isolated part of the structure.

PERFORM-3D User Guide 18-9

Chapter 18. "Elastic" Viscous Damping

M dampers connect to ground M dampers connect above isolation

level and affect entire structure. level and affect only isolated structure.

(a) Standard Option (b) Base Isolation Option

Figure 18.3 Alternative M Damping Models

PERFORM-3D allows you to choose either the "Standard" or the "Base

Isolation" option. As noted earlier, experience suggest that there is little
difference between these options.

The following are some points on the Base Isolation option.

(1) Only the H masses are connected as shown in Figure 18.3(b). If

there are any V or R masses, the M dampers for these masses are
connected to the ground.
(2) The K damping model does not change.
(3) You must have a rigid floor constraint at the level immediately
above the isolators. The M dampers connect to this level.

The steps for specifying this option are as follows.

(1) Before you begin the analysis series you must calculate the first
mode period for the non-isolated structure. In effect this is the first
mode period for the structure with rigid seismic isolators. You can
do this using PERFORM-3D, by changing the isolator properties to
make them very stiff before yield. However, you will probably
have done the calculation earlier, using a linear analysis program
such as SAP or ETABS. Call this period T1R , where “R” indicates
rigid isolators.

18-10 PERFORM-3D User Guide

 Chapter 18. "Elastic" Viscous Damping

(2) You must also calculate the elastic first mode period for the
structure with elastic isolators (i.e., with the isolators assigned their
pre-yield stiffnesses). Again, you can use PERFORM-3D but you
will probably have done the calculation earlier. Call this
period T1E , where “E” indicates elastic isolators.

(3) In the Run Analyses task and the Analysis Series form, choose
the Rayleigh Damping and Alpha-M Options pages.

(4) Choose the Base Isolation damping option, not the Standard
option.

(5) Choose the level immediately above the base isolators from the list
of rigid floor constraints. You must already have defined the
required constraint.

(6) Specify the ratio T1E / T1R . This ratio will be larger than 1.0.

When you specify the period ratios T A / T1 and TB / T1 , the standard

option calculates the  and  values using the periods T1E (with
elastic isolators). If you choose the base isolation option, PERFORM
adjusts these values so that they are based on T1R (with a fixed base).

18.3 A Case Where Modal vs. Rayleigh Damping May Be

Substantially Different : Coupled Shear Walls

There is an important case where modal and Rayleigh damping may

give substantially different results. This is the case of a shear wall
structure with stiff coupling panels.

When K damping is used, the K damping coefficient for a coupling

panel is based on the initial elastic bending and shear stiffnesses of the
panel, which can be large. After the panel yields (often in shear) the
ductility ratio can be large. Since the K damping coefficients remain
constant, the K energy dissipation might be overestimated, and the
deformations of the panel might be underestimated.

PERFORM-3D User Guide 18-11

Chapter 18. "Elastic" Viscous Damping

In the Energy Balance task you can examine the dissipated K energy
in each element group. We suggest that you always do this. In the case
of a shear wall, if the amount of energy seems to be unreasonable for
the coupling panel elements, you may wish to specify a reduced Beta-K
damping scale factor (possibly zero) for the corresponding element
group(s).

This potential problem may not be present with modal damping. When
modal damping is used, in effect a viscous damper "element" is added
for each mode shape. When the structure deforms, each of these
elements dissipates energy based on the amount of the mode shape that
is present in the total deformed shape (and on the current deformation
rate for this mode). In the case of coupled walls, it is likely that none of
the elastic mode shapes involves large deformations of the coupling
panels relative to the rest of the wall. Hence, when the panels yield,
modal damping should not contribute a lot of added dissipation.

We must emphasize that this is conjecture. We have not run

comparative analyses to compare the effects of the different kinds of
damping for this type of structure.

18.4 Combined Modal and Rayleigh Damping

You can combine Modal and Rayleigh damping if you wish. One
reason why you might do this is as follows. We suggest that you
consider using this method.

As noted earlier, if you use only modal damping, the only displaced
shapes that have damping are those that correspond to the (elastic)
mode shapes. The total number of degrees of freedom in the structure
will always be larger than the number of modes, so if only modal
damping is used there will be many displaced shapes that are
undamped. We recommend that you add a small amount of K
damping to make sure that these “higher mode” displacements are
damped.

To do this, get the period of the highest calculated mode, and specify
K damping (with no M damping) so that the damping ratio at this
period is small, say 0.2%. This will ensure that “higher mode”

18-12 PERFORM-3D User Guide

 Chapter 18. "Elastic" Viscous Damping

displaced shapes have some damping, but that Rayleigh damping in the
lower modes is negligible.

PERFORM-3D User Guide 18-13

 Chapter 19. General Load Sequence

19 General Load Sequence

PERFORM-3D is intended mainly as a tool for performance based

design. However, it can also be used for general purpose nonlinear
analysis. For performance based design it is sufficient to use the
"Standard" load sequence option that allows gravity load analysis
followed by either push-over analysis or dynamic analysis. For
more general nonlinear analysis you may want to use the "General"
option that allows arbitrary load sequences.

This chapter describes the difference between the two options, and
gives guidance on how to use the General option.

1.1 Standard and General Load Sequences

When you begin a new Analysis Series, you have the option of
choosing the "Standard" or "General" load sequence for that series.

The default option is the Standard sequence. This is the sequence that
you will almost always use if the purpose of the analysis is to assess
structural performance. With this option you can apply gravity load,
then add static push-over or dynamic earthquake load (earthquake or
dynamic force). You can consider any number of gravity, push-over,
dynamic earthquake and dynamic force load cases, but the only load
sequences that are allowed are (1) gravity followed by static push-over,
(2) gravity followed by dynamic load, (3) static push-over applied to
the unloaded structure (unlikely), and (4) dynamic load applied to the
unloaded structure (also unlikely).

If you wish, you can apply more than one gravity load case in a load
sequence. For example, if some elements are inactive for gravity load,
you may want to remove them for dead load, and restore them for live
load (for the removal procedure see Chapter 11, Inactive Elements).
You can do this by applying gravity dead load as the first analysis,
followed by gravity live load as the second analysis. If you specify any
inactive elements, they are removed only for the first gravity load case
in any load sequence, then restored for the second load case in the
sequence, whether it is a gravity case, a push-over case or a dynamic
earthquake case.

PERFORM-3D User Guide 19-1

Chapter 19. General Load Sequence

If you choose the General sequence, you can apply loads in any
sequence. This provides you with a great deal of flexibility for
performing nonlinear analyses. However, it also requires that you
exercise a much greater degree of care and skill.

1.2 Points to Note for General Sequence

The following are important point on the use of the general load
sequence option.
19.1.1 General Push-Over and Target Displacement Plots
You can not use the General Push-Over or Target Displacement tasks if
you use the general sequence. However, you can still use the Usage
Ratio task. Note that a usage ratio can never decrease. Hence, for some
analyses the usage ratios may stay constant.
19.1.2 Unload Push-Over Load Case Type
There is a load case type, namely Unload Push-Over, that is not
allowed for the Standard sequence. This allows you to remove push-
over loads on the structure. You can choose this type when you define
load cases.

When you perform a static push-over analysis, you specify the

maximum drift, not the maximum load, and the program calculates the
amount of load that is needed. Hence, if you want to remove the push-
over loads, and return to a state with only gravity loads, you generally
will not know how much load to remove. The Unload Push-Over load
case type removes push-over loads automatically.

Usually you will specify that all push-over loads are to be removed
(this is the "Unload all push-over cases" option). However, if you want
to specify that only the loads for a single push-over case are to be
removed, you can do so (this is the "Unload only one case" option). Be
careful using the latter option. Suppose, for example, that you apply the
following load sequence.

(1) Push to plus 1% drift. This requires that you apply, say, a load case
"Push plus 1%".
(2) Push to minus 1% drift. This requires that you apply, say, a load
case "Push minus 2%".

19-2 PERFORM-3D User Guide

 Chapter 19. General Load Sequence

(3) Push to zero drift. This requires that you apply "Push plus 1%" a
second time.
(4) The last push-over load case was "Push plus 1%". Hence, you may
try to remove the push-over loads by using an Unload Push-Over
load case with the unload single case option, and unloading the
"Push plus 1%" case.

If you do this, at Step (4) you will be adding minus the loads applied in
Steps (1) and (3) (i.e., you will be unloading all "Push plus 1%" loads),
but you will leave in place the loads applied in Step (2). This is
probably not what you intend. Also, the structure may not be able to
support these loads, and the analysis may not converge. The analysis
for Unload Push-Over load cases is load controlled, not displacement
controlled, and if the load is larger than the collapse load there is no
solution.

In the above example, in Step (4) you usually should use an Unload
Push-Over load case with the "unload all" option. This will remove all
previous push-over loads.

For Unload Push-Over load cases we suggest that you specify 5 load
steps and about 500 maximum events (there will usually be only a few
unloading events, but specify a large number to be safe).
19.1.3 Analysis Limit State
When you run a push-over analysis, it may stop because the maximum
specified drift is reached, or because the Analysis Limit State is
reached. The default Analysis Limit State for push-over analysis is
reaching the X point in any inelastic component.

Be careful if a push-over analysis stops because it has reached its

Analysis Limit State. If you specify a following analysis (either another
push-over case or an unloading case), and if you specify the same
Analysis Limit State, the new analysis will stop after the first load step,
because it is already past its Analysis Limit State. To avoid this you
must specify a more lenient Analysis Limit State.
19.1.4 Controlled Drifts
When you perform a static push-over analysis, you must specify a set
of controlled drifts and a maximum drift value. This is the change in
drift, not the total drift. The analysis stops when the change in any of
the controlled drifts reaches the maximum value. You will also have a

PERFORM-3D User Guide 19-3

Chapter 19. General Load Sequence

reference drift, which is used for plotting graphs. The reference drift is
not necessarily the largest drift. Hence, if you specify a maximum drift
of, say, 0.02 (2%), when the analysis stops the change in the reference
drift may be smaller than 2%.

If you know that the structure is well behaved, it is OK to specify only

one controlled drift, and to make it the same as the reference drift. In
general, however, it is a good idea to use several controlled drifts. For
example, if you control only the drift at the roof, and if the structure
develops a brittle weak story, when the weak story fails the roof drift
may have to decrease as the forces are redistributed. If you control only
the roof drift, that drift is not allowed to decrease, and there may be
very large deformations in the weak story as the structure tries to
redistribute the forces (also, the analysis may not be able to converge).

There is a class of problem where you may want to use only a single
controlled drift. This is if you want to run a cyclic push-over analysis,
to calculate hysteresis loops for a complete structure or (more likely) a
sub-assemblage. In this case you will usually want to cycle between
specified values of the reference drift. You must specify only one
controlled drift, and make it the same as the reference drift. Suppose
that you specify a maximum drift of 2%. Assuming that the analysis
limit state is not reached, you will find that the analysis stops when the
reference drift is slightly larger than 2%. The difference will usually be
small, but if you must push to a precise drift you may have to try
different maximum drift values to get the exact actual value that you
want.
19.1.5 Removing Gravity Loads
If you want to remove gravity loads you must set up one or more
gravity load cases that are the same as the applied load cases but have
opposite signs, then add these load cases. Since unloading events
(nonlinear behavior) may occur, you should specify nonlinear analysis.
We suggest applying the load in 5 steps.
19.1.6 Return To Static State After Dynamic Analysis
At the end of a dynamic earthquake analysis the structure will still be
vibrating. If you want to add more loads (e.g., static push-over loads, or
another dynamic earthquake), you must first return the structure to a
static state. PERFORM-3D does not have an automatic way to do this,
and you must do the following.

19-4 PERFORM-3D User Guide

 Chapter 19. General Load Sequence

(1) Define a dummy gravity load case with the nonlinear analysis
option and a very small load.
(2) Apply this load case after the dynamic earthquake. This will return
the structure to a static state without significantly changing the
gravity loads. We suggest that you specify 5 load steps and about
500 maximum events.

If you plan to apply loads after a dynamic earthquake analysis, it is a

good idea to run the dynamic earthquake analysis well beyond the time
that the strong shaking ends, so that the structure is nearly at rest at the
end of the analysis. Then, when you apply the dummy gravity load
there are only small changes in the state of the structure.
19.1.7 Graphs That Cover Several Analyses
When you plot graphs in PERFORM-3D (e.g., a graph of cross section
force vs. reference drift, using the Time History task), each graph will
cover only one analysis. If you want to plot a graph that extends over
several analyses (e.g., a graph showing the hysteresis loop for cross
section force vs. reference drift), you must press the Save to File button
for each analysis, then combine the files in a spreadsheet program, and
use the spreadsheet program to draw the combined graph.

This is not necessary for structure section time histories. See Chapter
25, Time History Plots, Section 25.7.
19.1.8 Inactive Elements
If you specify inactive elements, these elements are removed only for
gravity analyses that begin with the unloaded state. The elements are
restored at the beginning of the second analysis in the sequence, even if
this is a gravity analysis.

If the first analysis in any sequence (i.e., the analysis that begins with
the unloaded state) is not a gravity analysis, elements are not removed.

PERFORM-3D User Guide 19-5

 Chapter 20. Running Analyses

20 Running Analyses

20.1 Steps for Running Analyses

The form for running analyses is shown in Figure 20.1.

Figure 20.1 Form for Running Analyses

You can run as many analyses as you wish in a single analysis session,
and you can add analyses to any analysis series at any time. The steps
for adding analyses and running them are as follows.

(1) Choose the load case type from the Load Case Type list.

(2) Choose the load case name from the Load Case Name list.

(3) Choose the number of the preceding analysis. For a Standard load
sequence you will usually apply gravity load first, then add push-
over or dynamic earthquake load. The Preceding Analysis list
contains only the analysis numbers for preceding gravity load
analyses. The analyses are numbered in the order that they are

PERFORM-3D User Guide 20-1

Chapter 20. Running Analyses

performed. For a General load sequence, all previous analyses are

listed. For either sequence, make sure you choose the correct
analysis.

(4) Use the Add, Insert and Delete buttons in the usual way to build
up an analysis list.

(5) When you have specified the required analyses, press GO to run
the analyses. Press Don't Go to return to the Analyses Series form.

If you have specified masses for the current analysis series, a mode
shape analysis will automatically be set up. You can run this analysis
alone if you wish, without adding any other analyses to the list.

During the analyses a window is displayed showing the analysis

progress. For a new analysis series PERFORM-3D first sets up data for
the analysis model, then calculates the mode shapes, then runs the
analyses.

After the analyses have been run, they are added to the list of
completed analyses. This list shows the status of each analysis, in
particular whether the full load was applied or the analysis terminated
early. You can get additional details on the status of any analysis by
highlighting it, then pressing the Details button.

20.2 Adding and Deleting Analyses

You can add new analyses to an analysis series at any time. You can
also delete a complete analysis series. You can not delete individual
analyses from a series.

If you delete an analysis series, you have the option of deleting that
series completely, or deleting any analyses in the series but keeping the
analysis series settings (masses, damping, etc.). This second option is
the default, and this is the option that you will most commonly use.

If you make changes to the analysis model, you must delete all analyses
in all analysis series before making the change, since the analysis
results are no longer consistent with the changed model. Unfortunately
this is necessary even if the change is a minor one, since it is the only

20-2 PERFORM-3D User Guide

 Chapter 20. Running Analyses

way to ensure consistency. When you make a change to the model, you
will be asked if you want to delete the existing analysis series (the
analysis series settings are not deleted), and if you answer "Yes",
PERFORM-3D will delete them. If you add or change a Limit State,
Drift, Deflection or Structure Section, this is a change in the model,
since PERFORM-3D processes limit states, drifts, etc. during the
analysis. If you add or change a Frame or a Limit State Group, this is
not a change in the model.

PERFORM-3D User Guide 20-3

 Chapter 21 Mode Shapes

21 Mode Shapes

Mode shapes and periods can be useful for checking the behavior of
the structure, and for comparing a PERFORM-3D model with a
linear analysis model. You can also use mode shapes for static
push-over analysis. If you do, you must be careful to choose the
"correct" mode shape(s).

With the Mode Analysis Results task you can plot and animate mode
shapes, and get other mode shape properties. This chapter explains
the procedure.

21.1 Plotting Mode Shapes

If you specify masses when you define an Analysis Series, one or more
mode shapes and periods are calculated. These mode shapes are for the
structure in its initial unloaded, elastic state. With the present version of
PERFORM-3D you can calculate mode shapes for only this state.

The steps for plotting a mode shape are as follows.

(1) Choose the Modal Analysis Results task and the Modes page.
(2) Specify the displacement scale factor, as the ratio of the maximum
displacement for the mode to the maximum dimension of the
structure. The default is 0.1, which is usually about right.
(3) Choose the mode number from the list.
(4) Press Plot to plot the mode shape, then Animate to animate it.

When you choose a mode, certain of it properties are displayed. These

are as follows.

(1) The mode period.

(2) The effective mass factors for the H1, H2 and V directions for this
mode.
(3) The H1, H2 and V effective mass factors summed for all modes up
to this mode.
(4) The principal horizontal direction for this mode, and the
corresponding effective mass (or base shear) factor. The principal
direction is the direction along which the horizontal effective mass

PERFORM-3D User Guide 21-1

Chapter 21. Mode Shapes

is a maximum. For a symmetrical structure the principal directions

for the dominant modes will be along the axes of symmetry of the
structure. For a purely torsional mode the base shear factor will be
zero.

21.2 Using Mode Shapes for Push-Over Analysis

You can use mode shapes for push-over analyses in the following two
ways.

(1) To specify lateral load distributions. You can use a single mode or
combine a number of modes. For details see Chapter 13, Static
Push-Over Load Cases.

(2) To use as the deflected shape (effective mode shape) for push-over
plots. For details see Chapter 28,General Push-Over Plots or
Chapter 29, Target Displacement Plot.

Mode shapes are not used for dynamic earthquake analyses.

21.3 Vertical Masses

PERFORM-3D allows you to specify vertical (V) masses. As a general

rule you should not do so for building structures. For the assessment of
the seismic performance of a building, the horizontal resistance is
usually of greatest concern, and vertical inertia effects are less
important.

The behavior of the structure is also less sensitive to the distribution of

mass when there is only horizontal (H) mass. If you specify vertical
mass, you must be careful how you locate the mass points. For
example, if you have rigid floor diaphragms you can lump the H
masses at the mass centroids of the floors, but you probably should not
do this for V mass. More likely you will be tempted to lump the V
masses at the beam-column connections. If you do this, because of the
large axial stiffnesses of the columns there can be high frequency
vertical vibrations that cause rapid, and relatively small, variations in
the column axial forces. These force variation are unlikely to have

21-2 PERFORM-3D User Guide

 Chapter 21 Mode Shapes

much effect on the overall behavior of the structure, but if there is

inelastic P-M-M interaction, or if the structure has friction-pendulum
isolators, they can cause the nonlinear analysis to be substantially less
efficient, and possibly to be inaccurate. To get an accurate analysis you
may have to choose a very small time step.

For long span structures such as bridges, vertical inertia effects can be
significant. In this case, be sure to distribute the V masses in a
reasonable way, and consider the vertical vibration periods when
selecting the time step for a dynamic earthquake analysis.

If you specify V masses, and if you use mode shapes to define load
patterns for static push-over analysis, the V masses are ignored. Static
push-over loads are horizontal only.

PERFORM-3D User Guide 21-3

 Chapter 22 Response Spectrum Analysis

22 Response Spectrum Analysis

This is not a full-featured response spectrum analysis. Rather, it is

intended to help in making decisions based on certain sections of
FEMA 356 that refer to analysis results of response spectrum type.

This chapter describes the procedure and limitations.

22.1 Response Spectrum Load Cases

22.1.1 Spectra
Before you can run response spectrum analyses you must define one or
more acceleration response spectra. You must first set up one or more
text files containing the spectra, then read these spectra into
PERFORM. The procedure is similar to that for defining earthquake
acceleration records. The spectra are stored in a Spectra folder that is
similar to the Records folder for earthquake records.

The text file for a response spectrum must be a comma-delimited or

space-delimited file containing the following.

(1) Any number of header lines. These are skipped during the reading
process.
(2) One line containing the damping ratios (maximum 6 ratios). When
the file is read, the number of damping ratios for the spectrum set
is obtained from the number of items in this line.
(3) One line for each period, in order of increasing period. Each line
must contain the period followed one spectral acceleration value
for each damping ratio. The number of lines determines the
number of periods. There must be a line break (carriage return)
after the last line, but do not add blank lines.

Each file must contain only one spectrum (actually a spectrum set, for
different damping ratios).

To read a spectrum text file and add a new spectrum to the Spectra
folder, press the toolbar buttons for the Analysis phase and the Load
Cases task, choose Response Spectrum for the Load Case Type, and

PERFORM-3D User Guide 22-1

Chapter 22 response Spectrum Analysis

press the Add/Review/Delete Spectra button. You can then add one or
more new spectra, review (plot) any existing spectrum, and delete any
existing spectrum.

The steps to add a new spectrum are as follows.

(1) Enter the name of the spectrum text file. If you have copied this
file to the User folder you need to enter only the file name.
Otherwise you must give the complete path to the file. The file can
be on a diskette or CD if you wish.

(2) Choose the acceleration unit from the list.

(3) Specify the number of header lines to be skipped before the

beginning of the data.

(4) Specify whether the values are separated by spaces or commas.

(5) If the spectrum is to be placed in an existing spectrum group,

choose the name of the group from the list. To create a new group,
press the New button and specify the group name.

(6) Specify a file name and a spectrum name for the spectrum. Choose
the spectrum name carefully because you must use it to identify the
spectrum when you review it or use it in a load case.

(7) Press the Check button to read the file and set up the spectrum. If
the file is read successfully the spectrum will be plotted. If the
spectrum is OK press Save to save the spectrum, else press
Cancel, revise the above data, and try again.

When you have set up all of the needed spectra, press Return to
Spectrum Load Case.
22.1.2 Response Spectrum Load Cases
To start a new response spectrum load case press the toolbar buttons for
the Analysis phase and the Load Cases task. In the Load Cases form
choose Response Spectrum as the Load Case Type, press the New
button, and enter the load case name. Keep an eye on the Status box.

The steps are as follows.

22-2 PERFORM-3D User Guide

 Chapter 22 Response Spectrum Analysis

(1) If you have not already set up the needed spectra, press the
Add/Review/Delete Spectra button and add spectra as described
in the preceding section.

(2) Specify the angle from the structure H1 axis to the earthquake Q1
direction. This allows you to apply an earthquake in any horizontal
direction.

(3) Specify the number of modes to be used. If this is larger than the
number of modes actually calculated (when the analysis is run), the
smaller number will be used.

(4) Choose the Q1 Spectrum, if any, from the list of spectra. Also
choose the damping ratio (i.e., the particular spectrum that is to be
used if there are spectra for more than one damping ratio), and
specify a spectrum scale factor if desired.

(5) Repeat for the Q2 and V Spectra, if any. The default is none.

(6) Press the Save button to save the load case. Use the Delete button
if you want to delete a load case before saving it.

You can edit a saved load case at any time. If you do, keep an eye on
the Status box. Use the Save, Save As and UnChange buttons in the
usual way. You can also delete saved load cases.

22.2 Response Spectrum Analyses

22.2.1 Run Analysis

To run a response spectrum analysis, go to the Set Up And Run
Analyses form, choose Response Spectrum for the Load Case Type and
choose the load case name from the Load Case Name list. You must
have specified some masses, so that mode shapes will be or have been
calculated. If there are no mode shapes, there will be no load cases in
the Load Case Name list. The Preceding Analysis Number is always
zero for this load case type.

Use the Add, Insert and Delete buttons in the usual way to build up an
analysis list. You can mix analyses of different types. Press the Go
button to run the analyses.

PERFORM-3D User Guide 22-3

Chapter 22 response Spectrum Analysis

22.2.2 Analysis Results

To see the analysis results choose the Modal Analysis Results task,
then chose the Nodes, Drifts or Sections page and follow the
instructions. You can show the analysis results and/or and save the
results to a text file. You can not get element results.
22.2.3 Lateral Load Patterns from Analysis Results
For static push-over analysis, you may want to set up a lateral load
pattern where the distribution of story shears is the same as in a
response spectrum analysis. This is not necessarily a good choice for a
load pattern, but you can set it up as follows.

(1) Cut a structure section through all elements in each story of the
structure. To be consistent, cut at the bottom of each story. You
must cut a section for every story.
(2) In the Modal Analysis Results task choose the Sections tab and
then the Save tab.
(3) Select the story sections in the list, press the Save button, and
follow the instructions to set up a text file with the structure section
forces. These are SRSS values, so they do not have signs.
(4) Using spreadsheet program, process the file to get a lateral load
pattern that that gives the same story shears.
(5) Enter this pattern into PERFORM as a nodal load pattern.

22-4 PERFORM-3D User Guide

 Chapter 23. Energy Balance

23 Energy Balance

The response of a structure to an earthquake can depend on the

amount of energy that the structure is able to dissipate. In the
analysis of elastic structures, it is usually assumed that energy is
dissipated by viscous damping (which is a modeling approximation,
except for structures with actual viscous dampers). In the analysis
of inelastic structures, it is still usually assumed that there is
viscous damping, and in addition energy is dissipated by inelastic
effects (yielding, friction, etc.).

With the Energy Balance task you can plot graphs that show the
amount of each type of energy. This can be useful for checking the
analysis. You can also determine which members in the structure
contribute the most to the inelastic dissipated energy. This can be
useful for assessing the performance of the structure. This chapter
explains the procedure.

23.1 Types of Energy

23.1.1 Dynamic Analysis

For a dynamic analysis there can be six different types of energy, as
follows.

(1) Kinetic energy in the masses.

(2) Recoverable strain energy in the elements.
(3) Non-recoverable inelastic energy dissipated in the elements.
(4) Viscous energy dissipated by M dampers.
(5) Viscous energy dissipated by K damping in the elements.
(6) Viscous energy dissipated by modal damping.
(7) Viscous energy dissipated in fluid damper components.

PERFORM-3D calculates each of these energies at each step of the

analysis (with certain approximations as noted below). PERFORM-3D
also calculates the external work done on the structure. For earthquake
analysis this is the work done by the base shear, although it is actually
calculated as the work done by the equivalent inertia forces.

PERFORM-3D User Guide 23-1

Chapter 23. Energy Balance

23.1.2 Static Analysis

For a static analysis there are only two types of energy, as follows.

(1) Strain energy in the elements.

(2) Inelastic energy dissipated in the elements.

PERFORM-3D calculates of these energies at each step of the analysis

(again with certain approximations). PERFORM-3D also calculates the
external work done on the structure (the work done by the external
static loads).

If the structure has fluid damper components, and if "push-over"

properties have been specified for these components, PERFORM-3D
also estimates the amount of viscous energy dissipated by the dampers.
This value is approximate, but it can provide useful information. For an
explanation of the method see Chapter 28, General Push-Over Plots.

23.1.3 Calculation of Inelastic and Strain Energies

For any structural component, the sum of the strain energy and the
inelastic dissipated energy is equal to the static work done on the
element. Figure 23.1 shows a force-deformation relationship for a
simple component (e.g. an inelastic bar).

FORCE
A B Unloading with no
stiffness degradation
Unloading with
stiffness degradation

O E D C DEFORMATION

Figure 23.1 Loading and Unloading of an Inelastic Component

If the component is loaded monotonically (i.e., with progressively

increasing deformation and no unloading) from its unloaded point O
through yield at A to State B, the work done on the component is equal
to the area OABC under the force-deformation curve. At this state we
do not generally know how much of this work is stored as recoverable

23-2 PERFORM-3D User Guide

 Chapter 23. Energy Balance

strain energy and how much is dissipated as inelastic energy. The only
way we can find out how much energy can be recovered is to unload
the component.

Figure 23.1 shows two of many possible unloading paths, as follows.

(1) Inelastic unloading along BD, with no stiffness degradation. In this

case the unloading stiffness is equal to the original elastic stiffness.
The strain energy is the recovered work, area BCD. The rest is
inelastic dissipated work.
(2) Inelastic unloading with stiffness degradation. In this case the
unloading stiffness is smaller than the original elastic stiffness. The
strain energy is again the recovered work, area BCE, which is
larger than in Case (1). The rest is again the inelastic dissipated
work, which is smaller than in Case (1).

At each step of an analysis, PERFORM-3D calculates the total static

work on each component, equivalent to the area OABC for State B. For
an inelastic component, PERFORM-3D generally does not know how
the total work is divided between strain energy and dissipated energy,
since in general this can be determined only by unloading the
component. PERFORM-3D thus estimates the amount of strain energy
(and hence also estimates the amount of dissipated energy). This
estimate assumes that the component unloads with no stiffness
degradation.

This means that for a component with stiffness degradation,

PERFORM-3D can underestimate the strain energy, and hence
overestimate the dissipated energy, at least temporarily. For example, if
the component in Figure 23.1 is loaded along path OAB then unloaded
to either Point D or Point E, PERFORM-3D calculates the strain energy
and the dissipated inelastic energy as shown in Figure 23.2.

Up to Point A the behavior is elastic, and the strain energy increases

quadratically with deformation. Between Point A and Point B the total
energy increases linearly. If there is no stiffness degradation the strain
energy stays constant and the inelastic energy increase linearly.
PERFORM-3D calculates the energies using this assumption. If there is
actually no stiffness degradation, with unloading along BD, the strain
energy decreases quadratically and the inelastic energy stays constant.
This is correct. However, if there is stiffness degradation, with

PERFORM-3D User Guide 23-3

Chapter 23. Energy Balance

unloading along BE, the change in strain energy is larger and the
inelastic energy decreases.

Total energy
B
ENERGY
D

A E Inelastic
energy

O LOAD PATH

Figure 23.2 Energy Variation for Paths in Figure 23.1

Since inelastic dissipated energy can not decrease, the energies

calculated by PERFORM-3D for the degrading case are not correct.
However, note the following points.

(1) The dissipated energy is correct at all points if there is no stiffness

degradation.
(2) For the case with degradation the dissipated energy is correct at
Point E, after unloading.
(3) As the deformation increases, the error in the dissipated energy
gets smaller.
(4) If the component is loaded cyclically, as in a dynamic earthquake
analysis, the error in the dissipated energy gets smaller.
(5) Usually we will be concerned not with the absolute amount of
dissipated energy but with the relative amounts contributed by
different components (e.g., for a frame structure, the amount of
energy dissipated in the beams compared with the amount
dissipated in the columns). The relative energies will tend to have
smaller errors than the absolute energies.

For a component with stiffness degradation it is possible to get better

(but not necessarily exact) energies by calculating the degraded
stiffness at each step of the analysis and using this stiffness to calculate
the strain energy. However, we have chosen not to do this. For
computational reasons PERFORM-3D calculates the degraded stiffness
for a component only when it unloads. It is possible to calculate the

23-4 PERFORM-3D User Guide

 Chapter 23. Energy Balance

degraded stiffness at every step, but this requires substantial additional

computation, and is not justified simply to obtain a more accurate
energy estimate.

In summary, the inelastic "dissipated" energy calculated by

PERFORM-3D is not exact, and does not necessarily increase
monotonically. However, any errors are usually small.

23.2 Energy Plots

23.2.1 Complete Structure

To draw an energy plot for the complete structure, choose the Energy
Balance task and the Structure tab. Then choose the load case and
press Plot. You can draw energy balance plots only for dynamic
analyses.

23.2.2 Inelastic Energies in Element Groups

It can be useful to determine which element groups contribute most of
the inelastic dissipated energy. To show this choose the Elem Groups
tab, and the Inelastic option. Choose an Element group from the list,
and press Plot. The graph shows the total amount of dissipated energy
and the proportion of this energy that is contributed by the chosen
group. Cycle through the Element group list to show the proportion
for each element group.

If a structure has viscous dampers, the dissipated viscous energy in

these dampers is included in the total dissipated energy. For a static
push-over analysis this viscous energy is an approximate estimate.

23.2.3 -K Energies in Element Groups

If you specify only Rayleigh damping, it can be also useful to
determine which element groups contribute most of the K dissipated
viscous energy. To show this choose the Elem Groups tab, and the
Beta-K viscous damping option. Choose an Element group from the
list, and press Plot. The graph shows the total amount of K dissipated
energy and the proportion of this energy that is contributed by the

PERFORM-3D User Guide 23-5

Chapter 23. Energy Balance

chosen group. Cycle through the Element group list to show the
proportion for each element group.

23.2.4 Energy Error

The energies that are plotted are the internal energies from the
elements, masses, K dampers, M dampers and modal dampers.
PERFORM-3D also calculates the external energy, which is the work
done by the external forces. Because the nonlinear analysis does not
exactly follow the nonlinear behavior of the structure there will be
differences between the internal and external forces (equilibrium
errors) and hence between the internal and external energies. This is the
energy error for the analysis.

The internal and external energies for any dynamic analysis are shown
in the energy plots. The internal and external energies are also shown in
the ECHO file. The energy error should be small (no larger than a few
percent). A larger error (larger than about 5%) indicates that the
analysis may be inaccurate.

The error will tend to be larger for structures with brittle strength loss.
The error may also be larger for structures with friction-pendulum
dampers. These components can be sensitive numerically, and to get
accurate results you may have to use a time step that is shorter than you
would need for other structures.

23.3 Approximate Damping Ratio

23.3.1 Overview
The concept of a (modal) damping ratio, as a percentage of critical
damping, is strictly applicable only for linear structural behavior.
However, it is possible to calculate an approximate equivalent damping
ratio for nonlinear behavior. This section describes the procedure.
23.3.2 Theory
If a linear single DOF structure with viscous damping is subjected to
forced sinusoidal vibration at a constant amplitude, energy is dissipated
in each cycle. The relationship between the dissipated energy in a
complete cycle and the maximum strain energy for the cycle is :

23-6 PERFORM-3D User Guide

 Chapter 23. Energy Balance

Dissipated Energy
 4 (22.1)
Max. Strain Energy

where  = damping ratio.

If a linear single DOF structure is subjected to earthquake loading, the

strain energy will vary with time as indicated in Figure 23.3(a). Each
strain energy peak corresponds to a half cycle. If the dissipated energy
is known for a whole number of half cycles, the damping ratio can be
calculated as

1  Dissipated Energy 
    (22.2)
2N  Mean of Strain Energy Peaks 

where N is the number of half cycles (strain energy peaks).

For a multi-DOF nonlinear structure the strain energy variation is likely

to be more complex, as indicated in Figure 23.3(b). However, a rough
estimate of the effective damping ratio can be calculated as

1  Dissipated Energy 
    (22.3)
2N  2Mean Strain Energy  

which assumes that the mean of the strain energy peaks is 2 times the
mean strain energy.

Figure 23.3 Strain Energy Variation for Earthquake

Loading

PERFORM-3D User Guide 23-7

Chapter 23. Energy Balance

23.3.3 Procedure
The above theory is implemented in PERFORM-3D.

Use the Energy Balance task. Choose the load case and press Plot to
plot the energy balance for the structure. You can then enter data
defining a range of times (from T1 to T2 seconds) and the number of
strain energy peaks over that range.

In the energy plot, choose a strain energy trough and note the
corresponding time (move the cursor to the trough to show the time).
Then count forward a number of strain energy peaks, and note the time
at another trough. Enter the times and the number of peaks, and press
Calculate to show approximate damping ratios over the range. Ratios
are calculated for inelastic energy dissipation and for the four possible
types of viscous damping, namely modal, M, K and fluid dampers.

23-8 PERFORM-3D User Guide

 Chapter 24. Deflected Shape Plots

24 Deflected Shape Plots

Deflected shape plots can be extremely useful for studying the

behavior of a structure. A deflected shape plot is even more useful
when the elements are colored to show usage ratios. This chapter
explains the procedure.

24.1 Deflected Shape

Figure 24.1 shows a deflected shape plot at a point in a dynamic

earthquake analysis. You can also draw deflected shapes for gravity
and push-over analyses. You can draw the deflected shape for any step
in an analysis, and you can animate the deflected shape if you wish.

Figure 24.1 Deflected Shape Plot

You have two overall options, as follows.

PERFORM-3D User Guide 24-1

Chapter 24. Deflected Shape Plots

(1) Draw the deflected shape only. This can be useful for studying
such things as concentration of story drift.

(2) Draw the deflected shape, calculate usage ratios for chosen limit
states, and assign colors to the elements based on their usage
ratios. This is valuable for identifying the critical elements in a
structure. Be warned that it requires a lot of computation to process
the results for this type of plot, and for a large structure there may
be a significant delay before plotting begins.

In both cases you can animate the plot if you wish.

24.2 Procedure

24.2.1 Deflected Shape Only

To start a deflected shape plot press the toolbar buttons for the
Analysis phase and the Deflected Shapes task. In the Plot Scale and
Type page choose the Entire elements option (see later for an
explanation of this). Also choose the analysis series from the Series list
and the load case from the Case list.

When you choose the load case a "thumbnail" view of the structure
response appears. For a dynamic earthquake analysis this is a plot of
reference drift vs. time. For a push-over analysis it is a plot of base
shear vs. drift. For a gravity analysis it is a plot of load factor vs. load
step number.

To plot deflected shapes only, without usage ratio coloring, the steps
are as follows.

(1) Choose the method for scaling structure displacements. Specify

either a displacement scale factor or the ratio of maximum
displacement to maximum structure dimension.

(2) Select the point in the analysis for the deflected shape. The
thumbnail plot has a yellow vertical line. To position this line at a
point in the plot, click that point. You can do this before or after
the deflected shape is drawn. The default is the shape at the
beginning of the analysis.

24-2 PERFORM-3D User Guide

 Chapter 24. Deflected Shape Plots

(3) Press the Plot button to draw a static deflected shape at the
selected point.

(4) To animate the plot use the animation buttons. You can move the
plot forwards or backwards a single step at a time, jump to the
beginning or end of the analysis, and animate forwards or
backwards. To control the speed of the animation use the slider.
On the fastest setting the animation for dynamic analyses is in real
time (but only if your computer is fast enough).

24.2.2 Deflected Shape with Usage Ratio Coloring

To add usage ratio coloring you must choose a set of limit states. To do
this choose the Limit States tab and do the following.

(1) Choose the limit states that are to be considered. Usually you will
choose a limit state group.

(2) Choose a color group. There are 5 groups with pre-set usage ratio
ranges. You can change the ranges for any color group if you wish.

(3) Press the Plot button and continue as before. As noted above, it
may take a little time before the first plot appears.

If you choose the Entire elements option in the Plot Scale and Type
page, the color of each element depends on the usage ratio for that
element. In the common case where a beam or column element has
plastic hinges at both ends, the element usage ratio is the ratio for the
“worst” hinge, and the plot does not show which hinge is critical.

If you choose the Components option in the Plot Scale and Type page
you will get a plot that does show which hinge is critical. With this
option all elements are plotted, but only frame-type (beam, column and
brace) elements are colored, and the basic components that make up
each element are colored rather than the entire element. Moment and
PMM hinges are shown as circles. Shear and axial hinges are shown as
rectangles. Fiber section segments are shown as thick lines.

PERFORM-3D User Guide 24-3

Chapter 24. Deflected Shape Plots

24.2.3 Some Points on Components vs. Entire Element

(1) For the Components option, only deformation limit states are
considered. For the Entire Elements option, deformation and/or
strength limit states are considered.

(2) The plots do not use hidden line removal. If you use the
Components option and many frame-type elements are plotted, the
circles for hinge components may overlap in strange ways. It is
usually best to set up Frames that are either plane or do not include
many elements, and to plot the results for those frames.

(3) For FEMA Beam and FEMA Column components, each

component is divided (by PERFORM) into a hinge component and
an elastic segment, in such a way that the FEMA component has
the required relationship between end moment and end rotation. In
the plot, FEMA components have circles at their ends, as for
moment or PMM hinges. The color of each circle is based on the
usage ratio for the corresponding FEMA component.

(4) You can make printer plots for both options, but you can save the
results to a text file only for the Entire Elements option. Use the
buttons on the tool bar.

(5) You can toggle between the two options without closing the plot.

(6) To zoom in or out, use the buttons below the Plot / Close Plot
button.

(7) Frame-type elements are drawn using straight lines, not their actual
deflected shapes. To see actual deflected shapes, use the Moment
and Shear Diagrams task.

24.2.4 Other Aspects

If you have defined any frames, there will be a list of frames on the
toolbar. If you choose a frame, only the elements in that frame are
plotted.

You can zoom in. The zoom-in and zoom-out buttons are in the Plot
Scale and Type page, not on the toolbar.

24-4 PERFORM-3D User Guide

 Chapter 24. Deflected Shape Plots

In deflected shape plots, the elements are drawn using straight lines
only. You can examine the detailed deflected shapes for individual
beam, column and brace elements using the Moment and Shear
Diagrams task.
24.2.5 Save Usage Ratios to a File
If the save to file button on the toolbar is enabled (colored green), you
can save the usage ratios to a text file by pressing this button.

You will be asked to input a file name and description. By default the
file is saved in the User folder (see Chapter 2, Structures, Folders and
Files). The file contains a description of the data, the element number,
the coordinates of node I, and the usage ratio for each element.

The saved usage ratios are for the current deflected shape. For example,
if you are considering a dynamic earthquake analysis the usage ratios
are for the time that you have currently chosen. To get the final values
for the earthquake, be sure to go to the end of the earthquake and draw
the color-coded deflected shape at that time. For a push-over analysis
you may want to output usage ratios at drifts other than the drift at the
end of the analysis (usually you will output them at the estimated drift
demand as calculated in the General Push-Over Plot or Target
Displacement Push-Over Plot task).

PERFORM-3D User Guide 24-5

 Chapter 25. Time History Plots

25 Time History Plots

Time history plots are useful for checking the behavior of a

structure. You can draw time history plots for a variety of node and
element results, and also for drifts and structure sections. You will
probably use these plots mainly for dynamic earthquake analyses.
However, you can also draw plots for gravity analyses (in which
case "time" is load factor) and push-over analyses (in which case
"time" is reference drift).

This chapter explains the procedure.

25.1 Plot History for a Single Node

25.1.1 Options
For node results you can plot histories of the following. Velocities and
accelerations can be plotted only for dynamic analyses.

(1) The H1, H2 or V translation or rotation of any node. This is

displacement relative to the ground.
(2) The H1, H2 or V translational or rotational velocity of any node.
This is velocity relative to the ground.
(3) The H1, H2 or V translational or rotational relative acceleration of
any node. This is acceleration relative to the ground.
(4) The H1, H2 or V translational or rotational absolute acceleration of
any node. This is the acceleration relative to the ground plus the
ground acceleration.

If you choose absolute acceleration and select a supported node you

will plot the ground acceleration. This enables you to check that the
earthquake load has been specified correctly.

If you plot the absolute acceleration for any node, you can then plot
response spectra for that acceleration record. For a node in a floor, this
gives the floor response spectrum. For a node at the ground it gives the
spectrum for the ground motion.

PERFORM-3D User Guide 25-1

Chapter 25. Time History Plots

25.1.2 Procedure
For a single node history plot, press the toolbar buttons for the
Analysis phase and the Time History task. In the Time History form
choose the Node and Single Node tabs.

In the toolbar choose the analysis series from the Series list and the
load case from the Case list.

The steps are as follows.

(1) Choose the result type to be plotted.

(2) Choose the translation or rotation direction.
(3) Specify the displacement unit.
(4) Select the node, by clicking on it in the graphics panel. To make a
new selection press the Clear button and repeat.
(5) Press the Plot button.

After you press the Plot button for absolute accelerations, you can enter
damping ratios and other data for a set of corresponding response
spectra. Press the Plot Spectrum button to plot the spectra.
25.1.3 Save Plotted History to a File
If the save to file button on the toolbar is activated (colored green) , you
can save the current history to a text file by pressing this button. You
can then process this file using a spreadsheet program.

You will be asked to input a file name and description. By default the
file is saved in the User folder (see Chapter 2, Structures, Folders and
Files). The file contains a description of the data followed by two
columns of X-Y values.

25.2 Save Histories for Multiple Nodes

If you want to save histories for several nodes you can plot and save
them individually, but this can be inconvenient. A quicker method is to
choose the Multiple Nodes tab.

The steps are as follows.

(1) Choose the result to be plotted and the displacement unit.

25-2 PERFORM-3D User Guide

 Chapter 25. Time History Plots

(2) Select the nodes, by clicking on individual nodes or using one or

more box selects. To make a new selection press the Clear button
and repeat.
(3) Choose whether to save complete histories or maximum and
minimum values only.
(4) Press the Save button.

You will be asked to input a file name and description. By default the
file is saved in the User folder (see Chapter 2, Structures, Folders and
Files). The file contains a description of the data followed by the
results.

25.3 Plot History for a Single Element

25.3.1 Procedure
For a single element history plot, press the toolbar buttons for the
Analysis phase and the Time History task. In the Time History form
choose the Element and Single Element tabs.

In the toolbar choose the analysis series from the Series list and the
load case from the Case list.

The steps are as follows.

(1) Choose the element group from the list. The elements in the chosen
group are shown light blue.

(2) Select the element, by clicking on it in the graphics panel. Its color
changes to red. To make a new selection press the Clear button
and repeat.

(3) Choose whether you want to plot results for the element as a
whole, a structural component or a strength section. Some choices
may not be available. For example, you can choose the strength
section option only for frame type elements.

(4) For component or strength section results, choose the component

or strength section type. This depends on the element type, as
follows.

PERFORM-3D User Guide 25-3

Chapter 25. Time History Plots

(a) If the element consists of a single basic component the

component type is shown in the component type list, and you
can choose only this component.

(b) If the element is a frame type it can have several basic

components and several strength sections. These are shown in
the list in their order along the element. Choose the component
or strength section. If the element has, for example, two plastic
hinge components, take care to choose the correct component.

(c) If the element is a shear wall element it has a bending

component and a shear component. Choose the component.

(d) If the element is a general wall element it has two bending

components and two shear components. Choose the
component.

(e) If the element is a viscous bar it has a viscous damper and an

elastic bar component. Choose the component.

(5) Choose the result type for the chosen component or strength
section. The choices depend on the component type.

(6) If desired specify the length and force units.

(7) Press the Plot button to draw the graph. The X axis is time,
reference drift or load factor, depending on the load case type. The
Y axis is the result quantity.
25.3.2 Save Plotted History to a File
If the save to file button on the toolbar is activated (colored green) , you
can save the current history to a text file by pressing this button. You
can then process this file using a spreadsheet program.

You will be asked to input a file name and description. By default the
file is saved in the User folder (see Chapter 2, Structures, Folders and
Files). The file contains a description of the data followed by two
columns of X-Y values.

25-4 PERFORM-3D User Guide

 Chapter 25. Time History Plots

25.4 Save Histories for Multiple Elements

If you want to save histories for several elements you can plot and save
them individually, but this can be inconvenient. A quicker method is to
choose the Multiple Elements tab.

The steps are as follows.

(1) If you are in a Complete Structure view, the elements in the current
Element Group will be shown light blue. If you are in a Frame
view, the elements in the current Element Group and current
Frame will be shown light blue. The elements for which results
will be saved are these light blue elements. That is, you select the
elements by choosing the Element Group and Frame, not by
clicking on elements. If the current view does not show the
elements that you want, you must go to the Add or Delete Frames
task and set up a Frame with the required elements. There is a
reason for this. If you are saving results for multiple elements you
will probably do it several times, for different analyses. If you set
up a Frame with the required elements, you need to define the
elements once only, and you do not have to select the elements by
clicking on them each time.
(2) Choose whether you want to save element level or component
level results. Since all elements in an Element Group have the
same element level results, you can always choose this option.
However, Frame type elements can have different sequences of
basic components along their lengths. You can save component
level results only if all of the selected elements have similar basic
component sequences. If you choose component level results and
the elements are not similar, you will get an error message. You
must go to the Add or Delete Frames task and set up a Frame
with compatible elements.
(3) Choose the result that you want to save, and the units.
(4) Press the Save button and follow the instructions for saving the
results.

PERFORM-3D User Guide 25-5

Chapter 25. Time History Plots

25.5 Drift or Deflection Histories

For drift or deflection histories, press the toolbar button for the
Analysis phase and the Time History task. In the time history form
choose the Drift/Deflection tab, and either the Single or Multiple tab.
In the toolbar choose the analysis series from the Series list and the
load case from the Case list.

For a single Drift or Deflection, choose the drift or deflection from the
list, and press Plot to draw the graph. If the save to file button on the
toolbar is activated (colored green), save the current history by pressing
this button.

For multiple Drifts or Deflections, highlight the required drifts or

deflections in the table, and press Save.

Drifts are dimensionless. For Deflections you can choose the

displacement unit.

25.6 Structure Section Force Histories

To start a structure section plot, press the toolbar button for the
Analysis phase and the Time History task. In the time history form
choose the Structure Section tab, and either the Single Sect or
Multiple Sects tab. In the toolbar choose the analysis series from the
Series list and the load case from the Case list.

For a single Structure Section, choose the result type and units, choose
the section from the list, and press Plot to draw the graph. If the save to
file button on the toolbar is activated (colored green), save the current
history by pressing this button.

For multiple Structure Sections, choose the result type and units,
highlight the required sections in the table, and press Save.

25-6 PERFORM-3D User Guide

 Chapter 25. Time History Plots

25.7 Plot Section Results for Multiple Load Cases

25.7.1 Purpose
You can also plot Structure Section results (e.g. base shears) for a
single section and multiple load cases. This applies only for static push-
over load cases.

You can use this to do the following.

(1) Compare push-over curves for different lateral load cases. One
restriction is that the load cases must be in the same Analysis
Series. This means, for example, that you can not directly compare
push-over curves with and without P-delta effects. To compare
such curves you must save the histories to text files, then combine
and plot these files in a spreadsheet program.
(2) Plot hysteresis loops for story shear. To do this you must have used
a General Load Sequence and applied cyclic push-over loads. If
you then plot the Structure Section force corresponding to the story
shear, you will get a hysteresis loop.
25.7.2 Procedure
The steps are as follows.

(1) In the Analysis phase choose the Time Histories task.

(2) Choose the Structure Section and Multiple Loads tabs.
(3) Choose the result type and units.
(4) Choose the Structure Section from the list.
(5) Highlight the required push-over load cases in the table.
(6) Press Plot.

PERFORM-3D User Guide 25-7

 Chapter 26. Hysteresis Loop Plots

26 Hysteresis Loop Plots

Hysteresis loop plots are useful for checking the behavior of

inelastic components. This chapter explains the procedure.

26.1 Purpose

Hysteresis loop plots are useful for checking the behavior of inelastic
components (and also viscous dampers). Elastic components do not
have hysteresis loops.

You will usually plot hysteresis loops for dynamic earthquake load
cases, where components are subjected to cyclic loading. You can also
plot them for gravity and push-over load cases if you wish. Since there
is usually no cyclic loading for these load cases, the hysteresis loops are
usually just action-deformation plots with monotonically increasing
values.

PERFORM-3D uses a specific model for hysteresis loops. For an

explanation of this model see the document PERFORM Components
and Elements.

26.2 Procedure

To start a hysteresis loop plot press the toolbar buttons for the Analysis
phase and the Hysteresis Loops task. In the Hysteresis Loops form
choose the analysis series from the Series list and the load case from
the Case list.

When you choose the load case a "thumbnail" view of the structure
response appears. For a dynamic earthquake analysis this is a plot of
reference drift vs. time. The steps are then as follows.

(1) Choose the element group from the list. The elements in the chosen
group are shown white.

PERFORM-3D User Guide 26-1

Chapter 26. Hysteresis Loop Plots

(2) Select the element, by clicking on it in the graphics panel. Its color
changes to red. To make a new selection click the Clear button
and repeat.

(3) Choose the component type. This depends on the element type, as
follows.

(a) If the element consists of a single basic component the

component type is shown in the component type list, and you
can choose only this component.

(b) If the element is a frame type element it can consist of several

basic components. These are shown in the list in their order
along the element. Choose the component. If the element has,
for example, two plastic hinge components, take care to choose
the correct component.

(c) If the element is a shear wall element it has a bending

component and a shear component. Choose the component.

(d) If the element is a general wall element it has two bending

components and two shear components. Choose the
component.

(e) If the element is a viscous bar it has a viscous damper and an

elastic bar component. Choose the component.

(4) Choose the force-deformation type for the hysteresis loop. The
choices depend on the component type. There are no choices for
elastic components, since they do not have hysteresis loops. For a
viscous damper you can plot force vs. deformation, which is the
hysteresis loop, or force vs. deformation rate. This is not a
hysteresis loop, but it is a useful plot.

(5) If desired specify the length and force units.

(6) Select the starting and ending points for the plotted loop. The
thumbnail plot has a green vertical line to mark the starting point
and a red line to mark the ending point. To position the green line
at a point in the plot, left-click that point. To position the red line,
right click. You can do this before or after the loop is plotted.

26-2 PERFORM-3D User Guide

 Chapter 26. Hysteresis Loop Plots

(7) Press the Plot button to draw the loops. The starting point for the
selected range is marked by a solid circle, and the ending point by
a solid square.

(8) To expand or contract the loops, left and right click in the
thumbnail plot to move the starting and ending points. For fine
tuning, or to watch the loop develop, move the starting and ending
points up or down one step at a time using the green and red arrow
buttons.

26.3 Cut-Off Corners

A plotted hysteresis loop may have corners as shown in Figure 26.1.

This is not an error. PERFORM-3D saves the analysis results only at
the end of each time step. If there is an event in a step the true force-
deformation path is as shown by the dashed line in Figure 26.1.
However, since only the points at the step ends are plotted, the corner is
cut off.

ACTION

Corners may be cut

off because loop is
plotted only at time
step intervals.

DEFORMATION

Figure 26.1 Loop With Cut-Off Corners

26.4 Save Loops to a File

If the save to file button on the toolbar is activated (colored green), you
can save the current set of hysteresis loops to a text file, by pressing

PERFORM-3D User Guide 26-3

Chapter 26. Hysteresis Loop Plots

this button. You can then process this file using a spreadsheet program.
This can be useful if, for example, you want to plot two sets of loops on
the same graph to compare them.

You will be asked to input a file name and description. By default the
file is saved in the User folder (see Chapter 2, Structures, Folders and
Files). The file contains a description of the data followed by two
columns of X-Y values.

26-4 PERFORM-3D User Guide

 Chapter 27. Moment and Shear Diagrams

27 Moment and Shear Diagrams

The Moment and Shear Diagrams task allows you to draw shear
force and bending moment diagrams and envelopes for a single
element or a line of elements (but not an entire frame), to draw
shear force and bending moment diagrams based on cross section
forces (for example, plot a shear force diagram for a shear wall), to
draw detailed deflected shapes for single frame type elements, and
to show the gravity loads on frame type elements. The diagrams can
be animated if desired.

This chapter explains the procedures.

27.1 Detailed Results for a Single Element

27.1.1 Element Selection

Press the toolbar buttons for the Analysis phase, the Moment and
Shear Diagrams task and the Single Element page. Choose the
analysis series from the Series list in the toolbar, and the load case from
the Case list.

Select an element as follows.

(1) Choose an element group from the list. Only frame type element
groups are listed. In the graphics panel the chosen group is shown
light blue.

(2) Select the element, by clicking on it in the graphics panel. Its color
changes to red. To make a new selection press the Clear button
and repeat.

(3) Choose the units.

(4) Choose the result type. The options are (a) element loads, (b)
moment and shear envelopes, (c) moment and shear histories and
(d) deflected shape. See the following sections for the procedure
for each option.

PERFORM-3D User Guide 27-1

Chapter 27. Moment and Shear Diagrams

27.1.2 Element Loads Option

With the element loads option, to show the loads on the selected
element press the Plot button. For a gravity load case these are the
loads at the end of the analysis. You can apply element loads only in
gravity load cases. Hence, for push-over and dynamic earthquake load
cases the element loads are constant.

In these plots all elements are plotted horizontal. Since only gravity
loads are allowed on frame elements, the loads are in the vertical
direction, not necessarily along Axis 2 or Axis 3 of the element.

To change the element press Clear and select a new element.

27.1.3 Moment and Shear Envelopes Option

With the shear and moment envelopes option, to plot envelopes for the
selected element press the Plot button. For a gravity load case the plot
shows the bending moment and shear force diagrams at the end of the
analysis. For push-over and dynamic earthquake cases the plot shows
the envelopes at the end of the analysis. The diagram in the form shows
the sign convention. Use the option buttons at the top of the plot to
toggle between the element 1-2 and 1-3 planes.

Use the option buttons at the top of the plot to toggle between the
element 1-2 and 1-3 planes.

In these plots all elements are plotted horizontal.

To change the element press Clear and select a new element.

27.1.4 Moment and Shear Histories Option

For the moment and shear history option a "thumbnail" view of the
structure response appears. For a dynamic earthquake analysis this is a
plot of reference drift vs. time. For a push-over analysis it is a plot of
base shear vs. reference drift. For a gravity analysis it is a plot of load
factor vs. load step number.

To draw moment and shear diagrams proceed as follows.

(1) Choose the force and length units.

27-2 PERFORM-3D User Guide

 Chapter 27. Moment and Shear Diagrams

(2) Select a point in the analysis for the diagrams. The thumbnail plot
has a yellow vertical line. To position this line at a point in the
plot, click that point.

(3) Press the Plot button to draw the diagrams at the selected point.

(4) To animate the plot use the animation buttons. You can move the
plot forwards or backwards a single step at a time, jump to the
beginning or end of the analysis, and animate forwards or
backwards. To control the speed of the animation use the slider.

Use the option buttons at the top of the plot to toggle between the
element 1-2 and 1-3 planes.

In these plots all elements are plotted horizontal.

To change the element press Clear and select a new element.

27.1.5 Deflected Shapes Option

In the Deflected Shapes task, elements are plotted as straight lines.
With the deflected shapes option you can draw detailed deflected
shapes, showing element curvatures and hinge rotations.

For the deflected shapes option a "thumbnail" view of the structure

response appears. For a dynamic earthquake analysis this is a plot of
reference drift vs. time. For a push-over analysis it is a plot of base
shear vs. reference drift. For a gravity analysis it is a plot of load factor
vs. load step number.

To draw deflected shapes proceed as follows.

(1) Specify a displacement scale factor.

(2) Specify whether you want the relative end displacement to be

included in the deflected shape. If you are not sure what this
means, experiment.

(3) Select the point in the analysis for the deflected shape. The
thumbnail plot has a yellow vertical line. To position this line at a
point in the plot, click that point.

PERFORM-3D User Guide 27-3

Chapter 27. Moment and Shear Diagrams

(4) Press the Plot button to draw a static deflected shape at the
selected point.

(5) To animate the plot use the animation buttons. You can move the
plot forwards or backwards a single step at a time, jump to the
beginning or end of the analysis, and animate forwards or
backwards. To control the speed of the animation use the slider.
On the fastest setting the animation is in real time (if your
computer is fast enough).

Use the option buttons at the top of the plot to toggle between the
element 1-2 and 1-3 planes.

In the plots, elements that are inclined less than 45 degrees to the
horizontal are plotted horizontal, and elements that are more steeply
inclined are plotted vertical.

To change the element press Clear and select a new element.

27.2 Moment and Shear Diagrams for a Line of

Elements

Press the toolbar buttons for the Analysis phase, the Moment and
Shear Diagrams task and the Line of Elements page. Choose the
analysis series from the Series list in the toolbar, and the load case from
the Case list.

The procedure is as follows.

(1) Choose an element group from the list. Only frame type element
groups are listed. In the graphics panel the chosen group is shown
light blue.

(2) In the graphics panel select a line of elements, by clicking the

elements individually and/or using box selects. The selected
elements must be in a continuous line, usually either a horizontal
or vertical line. The selected elements are shown red.

(3) Choose the units.

27-4 PERFORM-3D User Guide

 Chapter 27. Moment and Shear Diagrams

(4) Choose the result type. The options are (a)moment and shear
envelopes and (b) moment and shear histories.

(5) For the histories option, select a point in the analysis for the first
plot. The thumbnail plot has a yellow vertical line. To position this
line at a point in the plot, click that point.

(6) Press the Plot button.

(7) To animate a history plot use the animation buttons. You can move
the plot forwards or backwards a single step at a time, jump to the
beginning or end of the analysis, and animate forwards or
backwards. To control the speed of the animation use the slider.

The moment and shear diagrams are plotted on horizontal axes, even for a
vertical line of columns.

27.3 Moment and Shear Diagrams Based on Structure

Sections

To draw moment and shear diagrams based on structure sections, press

the toolbar buttons for the Analysis phase, the Moment and Shear
Diagrams task and the Section Group page.

For the remainder of the procedure, see Chapter 9, Structure Sections,

Section 9.5. You must have specified one or more structure section
groups.

PERFORM-3D User Guide 27-5

 Chapter 28. General Push-Over Plots

28 General Push-Over Plots

There are a number of different methods for performance

assessment using static push-over results. FEMA 440 has proposed
a new Linearization Method, and also a number of changes to the
FEMA 356 Coefficient (Displacement Modification) Method. The
ATC 40 Capacity Spectrum Method is outdated, but the Capacity
Spectrum Method in general is still useful, and has some
advantages over other methods.

The General Push-Over Plot task that implements all of these

methods. This Chapter explains the procedures, illustrates the
differences among the methods and covers some essential theory.

This Chapter also shows how the Capacity Spectrum Method can be
used to assess the effects of fluid dampers.

28.1 Overview

28.1.1 Available Methods

You can choose from the following methods.

(1) FEMA 440 Linearization Method.

(2) FEMA 440 modification of the Coefficient Method (also known as
the Displacement Modification Method).
(3) FEMA 356 Coefficient Method.
(4) Capacity Spectrum Method, with options for the ATC 40
procedure or a modified procedure that may be more accurate.

The FEMA 356 Coefficient Method is also implemented, in a

somewhat different way, but with the same results, in the Target
Displacement task (see Chapter29, Target Displacement Plot).
28.1.2 Differences Among Methods
All of the push-over methods have the same main steps, as shown in
Figure 28.1

PERFORM-3D User Guide 28-1

Chapter 28. General Push-Over Plots

Figure 28.1 Main Steps for Static Push-Over

However, the methods differ substantially in their details. Figures 28.2

through 28.5 illustrate the differences. Each method starts with a push-
over curve that relates lateral load, H, to horizontal displacement, . In
each figure, follow Steps 1, 2, 3, etc. for the computation sequence.

28-2 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

Figure 28.2 Steps in FEMA 356 and FEMA 440 Coefficient

Methods

PERFORM-3D User Guide 28-3

Chapter 28. General Push-Over Plots

Figure 28.3 Steps in FEMA 440 Linearization Method

28-4 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

Figure 28.4 Steps in ATC 40 Capacity Spectrum Method

PERFORM-3D User Guide 28-5

Chapter 28. General Push-Over Plots

Figure 28.5 Steps in Modified Capacity Spectrum Method

28.1.3 Some Key Points

The following are some key points.

28-6 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

(1) The Capacity Spectrum and Linearization Methods require a

response spectrum family, with spectra for a number of damping
ratios. The Coefficient Method uses only one spectrum, usually for
5% damping.

(2) The Coefficient and Linearization Methods use empirical formulas

that are obtained by calibration against a large number of dynamic
analyses. Empirical formulas are used in the Coefficient Method
for the coefficients C1, C2 and C3, and in the Linearization
Method for the period Teff and damping ratio eff.

(3) The ATC 40 Capacity Spectrum Method can be interpreted as an

extension of response spectrum analysis for a linear structure, but
using the secant period and damping ratio rather than the elastic
period and 5% damping. With this interpretation, this method is
rational rather than empirical (although the implementation has
empirical features). The Capacity Spectrum Method is also more
flexible than the purely empirical methods. For example, it can be
used to assess the effect of additional energy dissipation in fluid
damper components (this is explained in a later section).

(4) It has been shown that the ATC 40 Capacity Spectrum Method is
not as accurate as the Coefficient and Linearization methods (see,
for example, FEMA 440). It can be argued that this is not a fair
comparison, because the Coefficient and Linearization Methods
have both been calibrated against dynamic analysis, and the
Capacity Spectrum Method has not. The Capacity Spectrum
Method has empirical parameters, and if these parameters were
calibrated against dynamic analysis, it might be as accurate as the
other methods, while still retaining its rational flavor.

One reason why the Capacity Spectrum Method is inaccurate may

be that it uses the secant period and damping. This is equivalent to
assuming that in an earthquake the structure will experience
several cycles with displacements equal to the maximum
displacement. It is more likely that there will be only a few cycles
with amplitudes close to the maximum, with more cycles at
smaller amplitudes. Intuitively, it should be more accurate to use
an effective period and damping, based on a displacement that is
smaller than the maximum.

PERFORM-3D User Guide 28-7

Chapter 28. General Push-Over Plots

This is the main basis of the Modified Capacity Spectrum Method.

In this method, the period and damping are calculated at a
displacement that is smaller than the maximum. Based on limited
experience, a displacement about 0.8 times the maximum appears
to be reasonable. The usual effect is to increase the period and
decrease the damping, and hence to increase the displacement
demand compared with the ATC 40 Capacity Spectrum Method.

Another cause of inaccuracy may be that the Capacity Spectrum

Method explicitly relates dissipated energy to damping ratio. The
equation for this relationship (see later) applies for linear analysis
but may not be accurate for nonlinear analysis. The accuracy of the
method might be improved if the relationship included empirically
calibrated parameters.

(5) This is a philosophical point, but worth noting. When dynamic

analysis is used, it is common to consider several ground
acceleration records, and to use the average result for performance
assessment. The goal of push-over analysis is usually to calculate a
displacement demand that matches this average result. This is
reasonable if push-over analysis is viewed as an analysis method.
However, it is not necessarily wise if push-over analysis is viewed
as a design tool. Push-over analysis is approximate. Hence, when it
is used as a design tool it should tend to err on the conservative
side.
28.1.4 Main Steps
In all of the preceding figures, the displacement demand is calculated
by successive trials, until the displacement in the last step matches the
estimated demand in the first step. The PERFORM-3D implementation
is a little different. In this implementation, a Demand curve is plotted,
and the displacement demand is the point where the Demand curve
crosses the Capacity (push-over) curve. The calculation steps are
essentially the same as in the preceding figures. The end result is a plot
as shown in Figure 28.6.

28-8 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

Figure 28.6 Push-Over Plot

In FEMA 356 the displacement demand is the Target Displacement. In

ATC 40 the displacement demand locates the Performance Point.

As shown later, this plot can provide useful information on the

sensitivity of the displacement demand to changes in the strength and
stiffness of the structure.

The main steps are as follows. Detailed steps are given later.

(1) Choose a push-over load case and plot the Capacity Curve. This is
usually a plot of base shear or base shear coefficient vs. reference
drift. However, it can be a plot of spectral acceleration vs. spectral
displacement.
(2) Select a number of trial points on the Capacity Curve, and for each
point construct a bilinear approximation.
(3) Choose a push-over method. Depending on the method, choose
certain options and specify certain details.
(4) If you have not done so previously, define one or more new
response spectra. If desired, edit existing spectra.
(5) Choose a response spectrum.
(6) Plot the Demand Curve. For each trial point a corresponding
demand point is calculated. The line passing through the demand
points is the demand curve. The point where the demand curve
intersects the capacity curve gives the displacement demand (target
displacement or performance point).
(7) Examine the limit state usage ratios at the displacement demand to
assess the performance.

PERFORM-3D User Guide 28-9

Chapter 28. General Push-Over Plots

28.2 Procedure for Defining Response Spectra

28.2.1 Spectrum Types

Each spectrum is actually a spectrum family, with spectra for a number
of damping ratios. There are four options for the spectrum shape, as
follows.

(1) FEMA 356 spectrum. The shape of this spectrum is as defined in

FEMA 356, and the reduction factors specified in FEMA 356 are
used to convert the 5% spectrum to spectra for other damping
ratios.
(2) ATC 40 spectrum. The shape of this spectrum is as defined in ATC
40, and the reduction factors specified in ATC 40 are used to
convert the 5% spectrum to spectra for other damping ratios.
(3) Standard shape, user reduction factors. The shape of the spectrum
is essentially the same as that in FEMA 356 (with one extra
parameter that can be used to vary the shape), but the user must
specify the reduction factors.
(4) User shape and reduction factors. The user must specify the
spectrum shape for each of the damping ratios (up to 7 ratios).

The FEMA 440 Linearization Method and the Capacity Spectrum

Method require a spectrum family. The Coefficient method uses only
the spectrum for 5% damping.
28.2.2 Add or Edit a Spectrum
Press the toolbar buttons for the Analysis phase and the General Push-
Over Plot task. Go to the Spectra page.

To add a new spectrum, press New, choose the spectrum type, and
press OK. Input the required data, press Graph to plot the spectrum,
and Save to save it.

Alternatively, if there is an existing spectrum that is similar, choose it

from the list, press Save As to copy it with a new name, edit the data,
and press Save.

The data for an existing spectrum can be edited at any time. Choose it
from the list, edit the data, and press Save. To delete an existing
spectrum, choose it and press Delete.

28-10 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

28.2.3 Copy Spectra from a Different Structure

Any spectra that you define for a given structure are stored with that
structure, and are not currently available in other structures. If you have
defined a number of spectra for one structure and you want to use them
for a different structure, you must jump through a hoop, as follows.

The spectra are stored in a file PF3DPOA in the folder for the current
structure. This is not a text file, and it can not be edited. To make the
spectra available to another structure, copy this file to the folder for the
other structure. If there is already a PF3DPOA file for the other
structure, it will be replaced. Hence, if you want to keep the existing
file, rename it.

Response spectra are also used in Response Spectrum load cases. The
response spectra defined for those load cases can not be used for push-
over analysis, and vice-versa.

28.3 Procedure for Capacity and Demand Curves

28.3.1 Choose Task

Press the toolbar buttons for the Analysis phase and the General Push-
Over Plot task.
28.3.2 Plot Capacity Curve
In the Capacity page, choose the curve type. The options are as
follows.

(1) Base shear vs. reference drift.

(2) Base shear coefficient vs. reference drift.
(3) Spectral acceleration vs. spectral displacement.

If you choose the spectral acceleration vs. spectral displacement option,

you must also choose a method for calculating the spectral acceleration
values. See the following sections for an explanation.

You usually will also show limit points on the curve. The limit point
for a limit state is the point where the limit state usage ratio equals 1.0.
The are several options for the limit states that are to be considered,
including limit state groups.

PERFORM-3D User Guide 28-11

Chapter 28. General Push-Over Plots

Press Plot to draw the capacity curve.

If limit points are plotted, they are shown as short red lines. Click on a
line to show the corresponding limit state in the box near the bottom of
the Capacity page.

At the bottom of the Capacity page are boxes for optional Scale
Factors and Plot Ranges. See later for an explanation.
28.3.3 Sa vs. Sd Option
If you choose the spectral acceleration vs. spectral displacement option,
the push-over curve is a plot of spectral acceleration capacity vs.
spectral displacement. Spectral acceleration demand is obtained from
the response spectrum, as considered later.

The spectral acceleration capacity at any point on a push-over curve

depends on the deflected shape of the structure. You have three options,
as follows.

(1) Use the “first” elastic mode shape. For a 2D structure this is indeed
the first mode. For a 3D structure it is the mode that has the largest
base shear factor along the push-over direction. If you choose this
option, PERFORM finds this “best” mode. If you are applying
FEMA 356 you will usually choose this option. It is shown in later
sections that this may not be consistent theoretically.
(2) For each point on the push-over curve, use the current deflected
shape, and otherwise follow FEMA 356. Arguably the current
deflected shape is a better choice than the elastic mode shape. With
this option, the spectral acceleration capacity is calculated using, in
effect, FEMA 356 Equation 3-16. This is explained in more detail
in the next section.
(3) For each point on the push-over curve, use the current deflected
shape, but instead of FEMA 356 Equation 3-16, use an equation
that is theoretically more consistent. Again, this is explained in
more detail in the next section.
28.3.4 Spectral Acceleration Capacity
FEMA 356 Equation 3-16 specifies, in effect, that given the base shear
capacity at a point on the push-over curve (where the base shear
capacity is equal to the applied horizontal load), the corresponding
spectral acceleration capacity is given by

28-12 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

1 H capacity
S a , capacity  (3.1)
Cm W / g 

where H capacity = base shear, W = structure weight, g = gravitational

acceleration, and Cm = dimensionless effective mass factor for the
“first” mode shape. This factor is the effective mass along the push-
over direction divided by the total mass, given by

( T M d ) 2 1
Cm  . (3.2)
 M  
T
W / g 

where   = mode shape, M  = mass matrix, and d  = matrix of

direction cosines for the push-over direction. If the push-over direction
is along the H1 axis, matrix d  has a unit value for each H1 mass in
M  , and a zero value for all other masses. Similarly, if the push-over
direction is along the H2 axis, matrix d  has a unit value for each H2
mass in M  , and a zero value for all other masses.

There are two potential inaccuracies in this equation, as follows.

(1) It is accurate only if the deflected shape of the structure is

proportional to the mode shape,   .
(2) It is accurate only if the push-over load distribution is proportional
to M   .

As shown in Section 3.7, if the applied push-over load at a point on the

push-over curve is R and the deflected shape is r , a more consistent
equation is

R rr M d 1

T T H capacity
S a , capacity 
R d r M r C
T T
mr W / g 
(3.3)

where

Cmr 
r M d  .
T 2
1
(3.4)
r M r
T
W / g 

PERFORM-3D User Guide 28-13

Chapter 28. General Push-Over Plots

For the special case where the loads, R are proportional to M r ,
Equation 3.3 reduces to

1 H capacity
S a , capacity  (3.5)
Cmr W / g 

If the current deflected shape is proportional to the mode shape,   ,

Equation 3.5 becomes Equation 3.1. Since push-over methods are
approximate, Equation 3.1 can be an acceptable approximation.

Equation 3.3 may seem complex. However, since RT d   H , this

equation reduces to

S a ,capacity 
rT R (3.6)
rT M d 
If you choose the first option for the deflected shape (“first” mode
shape), PERFORM calculates the spectral acceleration capacity at any
point on the push-over curve using Equation 3.1. If you choose the
second option (current deflected shape with FEMA 356 Equation 3-16),
PERFORM uses Equation 3.5. If you choose the third option (current
deflected shape with more consistent equation), PERFORM uses
Equation 3.6.
28.3.5 Define Trial Points
Go to the Points page. You can define up to 6 trial points, and you
must define at least 2 points. The properties at the currently defined
points are shown in the table.

There are two ways to define a new point, namely "click-and-drag" and
"right-click". The click-and-drag method is more general. The right-
click method is a lot easier.

To use the click-and-drag method, move the cursor in the graph to the
desired drift or displacement value (D), then left click, hold, and drag to
set up a bilinear approximation. For the FEMA 356 approximation, the
areas under the bilinear and actual curves must be equal, and the initial
slope of the bilinear curve must intersect the actual curve at 60% of the
distance to the yield point. As you drag the bilinear curve, the area ratio
is shown and there is a small circle at the 60% point. Drag until the area

28-14 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

ratio is very close to 1.0 (a value between 0.998 and 1.002 is suggested)
and the small circle lies on the actual curve. After you release the
mouse button the bilinear curve properties are shown highlighted in the
table. If this approximation is OK, press OK (actually, this is optional –
if you wish you can just start the next point). If the approximation is
not OK, press Clear and repeat.

To use the right-click method, move the cursor in the graph to the
desired drift or displacement value (D), then simply right click. A
FEMA 356 bilinear approximation will be drawn automatically (in
some cases this may not work – if so there will be a message). As
before, the bilinear curve properties are shown highlighted in the table.
If the approximation is OK, press OK, and if not press Clear.

To change a point, choose its number, press Clear, and define a new
point (or just choose its number and define a new point).

The points do not have to be in order of increasing D. Often you will

start with only 2 or 3 points to locate the displacement demand
approximately, then add more points close to the displacement demand
to get better accuracy.
28.3.6 Period Calculation
28.3.6.1 Purpose
All push-over methods use a vibration period for each trial point, which
depends on the structure stiffness. This period, together with a damping
ratio, is used to enter the response spectrum to obtain the spectral
acceleration demand. This is illustrated in Figure 28.7.

PERFORM-3D User Guide 28-15

Chapter 28. General Push-Over Plots

Figure 28.7 Spectral Acceleration Demand

28.3.6.2 Stiffnesses
The vibration period depends on the structure stiffness. There are
several stiffnesses that may be used. These stiffnesses can be related to
slopes in the push-over plot, as shown in Figure 28.8.

Figure 28.8 Stiffnesses

This figure shows 5 stiffnesses (slopes), as follows.

28-16 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

(1) Ki = tangent to the push-over curve at the origin. If the structure

remains linear under gravity load, this represents the elastic
stiffness of the structure. However, if there is nonlinear behavior
such as yield or cracking under gravity load, this may not be an
accurate representation of the elastic stiffness.
(2) Ks = secant stiffness at the trial point.
(3) Ke = elastic stiffness for the bilinear approximation at the trial
point. In ATC 40 this stiffness is K0.
(4) Kh = strain hardening stiffness for the bilinear approximation.
(5) Keff = effective stiffness, used by some push-over methods.

28.3.6.3 Relationship Between Period and Stiffness

For a single DOF structure, the vibration period is given by the well-
known equation

M
T  2 (3.7)
K

By analogy, if the stiffnesses at two points, A and B, on a push-over

curve are KA and KB, the corresponding periods are related by

TA KB
 (3.8)
TB KA

For example, in Figure 28.8 the period Ti, corresponding to the initial
stiffness Ki, can be assumed to be the “first” mode elastic period of the
structure, which is known. Hence, the period Te, corresponding to the
stiffness Ke, can be calculated as

Ki
Te  Ti (3.9)
Ke

Alternatively, the period Ts, corresponding to the secant stiffness Ks,

can be calculated (see below). Hence, the period Te can be calculated as

Ks
Te  Ts (3.10)
Ke

PERFORM-3D User Guide 28-17

Chapter 28. General Push-Over Plots

The Coefficient Method uses the period Te to enter the response

spectrum, the ATC 40 Capacity Spectrum Method uses the period Ts,
and the FEMA 440 Linearization Method and the Modified Capacity
Spectrum Method use an effective period Teff.

28.3.6.4 Options for Period Calculation.

In the Points page there is a box with two options for the method of
period calculation. The options are as follows.

(1) Use the elastic mode period. In this case the period Te for each trial
point is calculated using Equation 3.9. This is the method that is
most commonly used. It is specified by FEMA 356, although it is
not necessarily the best method.

As noted above, this method can be inaccurate if there is yield or

cracking under gravity load. To allow for this case, you have the
option of using the initial slope of the push-over curve as the
stiffness Ki, or of specifying Ki directly. If you choose this second
option, you must run a separate push-over analysis for the structure
with no gravity load, get Ki from the initial slope of this curve, and
specify this stiffness (slope) directly.

(2) Use the Rayleigh Quotient. In this case, the secant period Ts is
calculated for each trial point, using the Rayleigh Quotient (see the
next section).The period Te is then calculated using Equation 3.10.
This method has some advantages, and is arguably more consistent
theoretically. The Rayleigh Quotient is calculated as follows.

28.3.6.5 Rayleigh Quotient

Let the deflected shape at a trial point be r . Assume that the structure
vibrates with this shape, at a period determined by the secant stiffness
at the trial point. The maximum strain and kinetic energies are given by

SE  0.5rT K s r  0.5rT R (3.11a)

and
4 2
KE  0.5 rT M r (3.11b)
Te2

28-18 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

where K s  = secant stiffness matrix, R = applied push-over loads at

the trial point, and M  = mass matrix. Equating these energies gives
the secant period as

Ts  2
rT M r (3.12)
rT R
This is the Rayleigh Quotient method. For a given deflected shape it
gives the best estimate for the period. It has the advantage that it does
not depend on the initial slope, Ki, of the push-over curve, so it does not
matter if there is yield or cracking under gravity load.

If you choose the Rayleigh Quotient method, by default the secant

period, Ts, is calculated, and Te is obtained using Equation 3.10. You
will usually use this default. However, if you wish you can specify that
the Rayleigh Quotient is to be calculated at any point between the
bilinear yield point and the trial point. The reasoning is that the
deflected shape at a trial point corresponds to the maximum
displacement of the structure, and that an effective displacement that is
smaller than this maximum may be more representative. A similar
argument is used for the Modified Capacity Spectrum Method (see
Appendix A). If you specify an effective displacement that is smaller
than the displacement at the trial point, Equations 3.12 and 3.10 are
modified to use the stiffness and period at the effective displacement. It
may be noted that if the displaced shape does not change as the
displacements increase, it does not matter which point is chosen.

28.3.6.6 Warning on Relationship Between Period and Stiffness

The relationship between period, T, and stiffness, K, is accurate only if
the structure is well behaved. This is the case if the deflected shape
does not change substantially as the deformations increases. It may not
be the case if the deflected shape changes substantially. This can be the
case if there is brittle strength loss, which can cause local deformations
that change the deflected shape. An example is a structure with a brittle
bottom story, where the deflected shape can be roughly linear over the
building height before the story fails, then can suddenly change to
roughly constant over the height.

PERFORM-3D User Guide 28-19

Chapter 28. General Push-Over Plots

There can also be inaccuracy if the push-over curve is plotted in terms

of reference drift (typically based on roof displacement) and if this drift
is not closely related to the masses and the push-over loads. Figure 28.9
shows an example.

Figure 28.9 Choice of Reference Drift

The structure in this figure is unsymmetrical in plan, and push-over

loads applied at the center of mass might cause significant torsional
deformation. If the reference drift is Drift A, on one side of the
building, this drift may not provide a good representation of the
behavior of the building as a whole, and periods based on this drift may
not be accurate. A drift such as Drift B is likely to provide a better
representation.

As noted later, this effect can be a reason for plotting the push-over
curve in terms of spectral acceleration and spectral displacement.
28.3.7 Choose Push-Over Method and Response Spectrum
Go to the Demand page.

Choose the performance evaluation method from the available options.

Then go to the Details page and specify the details for the chosen
method. These details are essentially as follows.

(1) For the Coefficient Method, choose the method for calculating the
coefficients C1, C2 and C3. There are FEMA 356, FEMA 440 and
User-Specified options.

28-20 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

(2) For the Linearization Method choose the type of hysteresis loop
that most closely approximates the H vs.  loop for structure.
FEMA 440 gives equations for effective stiffness and damping
ratio for three different hysteresis loops, and a fourth set of
equations that are not based on a specific loop. There is also an
option for user specified values.

(3) For the Capacity Spectrum Method, choose an option for energy
degradation. To implement the Modified Capacity Spectrum
Method specify an Effective Drift Factor. If the structure has fluid
dampers, specify an Effectiveness Factor for the energy dissipation
in these dampers. These aspects are explained in more detail in
later sections.

After specifying the details, return to the Demand page. The properties
at the trial points are shown in the table in that page. If any user-
specified values are to be entered, the corresponding cells in the table
are colored beige, and you must enter values for these properties.
Otherwise the properties are calculated by PERFORM.

Choose the response spectrum from the list. You must have set up one
or more spectra, as described earlier in this Chapter. If you wish you
can specify scale factors to be applied to the spectral accelerations and
periods.

If the capacity curve is base shear or base shear coefficient vs.

reference drift (i.e., if the curve is not spectral acceleration vs. spectral
displacement), you must also choose a method for converting spectral
acceleration demand (which is obtained from the response spectrum) to
base shear demand. The options are as follows.

(1) Use the “first” elastic mode shape.

(2) Use the current deflected shape, and otherwise follow FEMA 356.
(3) Use the current deflected shape, but instead of FEMA 356
Equation 3-16, use a more consistent equation.

These are the same options as for the case where the capacity curve is
spectral acceleration vs. spectral displacement, and the theory is exactly
the same. The difference is that it is now necessary to convert spectral
acceleration demand to base shear demand, rather than base shear
capacity to spectral acceleration capacity. For each option the base
shear demand at a trial point is calculated as

PERFORM-3D User Guide 28-21

Chapter 28. General Push-Over Plots

S a , demand
H demand  H capacity (3.13)
S a , capacity

where S a , demand is obtained from the response spectrum, H capacity =

base shear capacity at the trial point, and S a , capacity is calculated as in
Section 3.3.4.

Also choose the option for Lines of Constant Period This affects only
the sensitivity information contained in the demand curve. It does not
affect the displacement demand. The options are (a) lines of constant
period are vertical and (b) lines of constant period are radial. The
difference is explained in the next section.
28.3.8 Plot Demand Curve
To plot the demand curve press Plot.

For performance assessment, the key point is where the demand curve
crosses the capacity curve. The points on the demand curve are
connected by straight lines, and if the trial points are widely spaced the
crossing point may be different from the displacement demand. To get
the displacement demand more accurately, go to the Points page and
modify the trial points, then return to the Demand page and press Plot.

The demand curve also gives useful sensitivity information, as shown

in Figure 28.10.

Figure 28.10 Sensitivity Information

28-22 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

In Figure 28.10(a) the capacity curve is changed by changing both the

strength and the stiffness, in the same proportion. In this case lines of
constant period are vertical. This figure shows how much the strength
(and stiffness) of the structure would have to be increased to reduce the
displacement demand from its calculated value to 1. The figure also
shows how much the strength could be decreased if the displacement
demand could be increased to 3.

In Figure 28.10(b) the capacity curve is changed by changing only the

strength, keeping the stiffness constant. In this case lines of constant
period are radial.

You can assess the sensitivity to both types of change by selecting the
“vertical” or “radial” options for Lines of Constant Period. You can
also explore the effects of strength and stiffness changes more directly,
as explained in the next section.
28.3.9 Capacity Curve Scaling for Sensitivity Studies
The demand curve includes sensitivity information for two special
cases, as described in the preceding section. You can do more general
sensitivity studies by scaling the capacity curve as follows.

In the Capacity page there are two boxes for Scale Factors for
Sensitivity Studies, one for a force scale factor and the second for a
drift or displacement scale factor. To study the effects of changing
strength and/or stiffness, specify values for these factors, press Plot to
redraw the capacity curve, and proceed as before to draw the demand
curve.

Since you have changed the capacity curve, you will have to define a
new set of trial points.

28.4 Performance Evaluation

You will usually show limit points on the capacity curve. As a general
rule, if all pertinent limit points are to the right of the displacement
demand, the performance is satisfactory. In this case the usage ratios
for all limit states are smaller than 1.0 at the displacement demand.

To study the limit state usage ratios in more detail, make a note of the
load case and the drift demand, then go to the Usage Ratio Graphs task.

PERFORM-3D User Guide 28-23

Chapter 28. General Push-Over Plots

For the same load case, enter the drift demand as the maximum
reference drift, and draw the usage ratio graph. From this graph you can
see how the usage ratios vary for all limit states, up to the drift demand.

28.5 Consistency

28.5.1 General
As noted earlier in this Chapter, FEMA 356 Equation 3-16, which
converts spectral acceleration to base shear, is approximate. This can
affect the consistency of the results. This section considers this aspect
in more detail.
28.5.2 Calculation of Spectral Displacement
When the capacity curve is plotted as spectral acceleration vs. spectral
displacement, the spectral displacement is needed for each point on the
curve.

If the displaced shape at a point is q  , where q is the (dimensionless)

modal amplitude, the spectral displacement is given by

q
Sd  (3.14)

where  is the modal mass participation factor along the push-over
direction, given by

 
T M d  (3.15)
T M  
FEMA 356 does not explicitly specify this equation. However, this
equation is specified by ATC 40, and is implied by FEMA 440.

If the displaced shape is r , it is arguably more consistent to use

1
Sd  (3.16)
r
where

28-24 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

r 
rT M d  (3.17)
rT M r
It is, however, not necessarily correct to use either of these equations.
The reason is that the following relationship should exist between S a
and S d :

S a 4 2
 2 (3.18)
Sd T

where T is period. For a point on a push-over curve, T is the secant

period.

For a point on a push-over curve, if S a is calculated using Equation

3.1, S d is calculated using Equation 3.14, and T is calculated using
Equation 3.9 (based on stiffness ratios), then Equation 3.18 is generally
not satisfied. The main reason is that S a and S d will be consistent
only if the push-over load at each point on the push-over curve is
proportional to M r , where r is the deflected shape at the point. In
general this condition is not satisfied. A secondary reason is that T may
not be given accurately by Equation 3.9 or 3.10.

As a consequence, if you do a performance evaluation using the push-

over curve in terms of base shear and reference drift, then repeat the
evaluation in terms of S a and S d , you may not get the same result. For
example, if the displacement demand for one form of the capacity
curve is exactly at a limit point, the displacement demand for the other
form of the curve generally will not be at that limit point. The
difference tends to be larger when the push-over load pattern differs
substantially from the deflected shape (e.g. for a “uniform” load
pattern).

This is the main reason for using “adaptable” push-over analysis, where
the push-over load pattern is adjusted at each step to match the
deflected shape. Adaptable push-over analysis is currently not available
in PERFORM.

PERFORM-3D User Guide 28-25

Chapter 28. General Push-Over Plots

28.5.3 Suggested Methods

The theoretically consistent approach is to use Equation 3.6 for S a ,
Equation 3.16 for S d , and the Rayleigh Quotient for T. In this case the
performance assessment does not depend on the form of the capacity
curve (provided that Equation 3.9 or 3.10 is accurate).

However, this is not the approach specified in FEMA 356, FEMA 440
or ATC 40. For practical analysis, the following methods are suggested.

(1) For FEMA 356 or FEMA 440, for both the Coefficient and
Linearization methods, plot the capacity curve in terms of base
shear or base shear coefficient. If there is a mode shape that closely
matches the push-over direction, convert spectral acceleration to
base shear using this mode shape and FEMA 356 Equation 3-16. If
there is no such mode shape use the current deflected shape and
FEMA 356 Equation 3-16. This is not necessarily a consistent
method, but it follows FEMA 356.

(2) For the ATC 40 Capacity Spectrum Method, plot the capacity
curve in terms of S a and S d . If there is a mode shape that closely
matches the push-over direction, calculate S a and S d using this
mode shape. If there is no such mode shape use the current
deflected shape. This is not necessarily a consistent method, but it
follows ATC 40.

(3) To be most consistent, for any push-over method, convert spectral

acceleration to base shear (or vice versa) using the current
deflected shape and the alternative (consistent) method based on
Equation 3.6. Also choose the Rayleigh Quotient method for
calculating periods.
28.5.4 A Reason for Using Sa vs. Sd Option
The spectral acceleration vs. spectral displacement form of the push-
over curve is sometimes used because it is consistent with the
“Acceleration-Displacement” form of the response spectrum (ADRS).
In this form of the response spectrum, both accelerations and
displacements are shown, and lines of constant period are radial. In the
spectral acceleration vs. period form of the spectrum, displacements are
not shown and lines of constant period are vertical.

28-26 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

The spectral displacement is not particularly useful as a displacement

measure, and the response spectrum is no simpler if lines of constant
period are radial, so there is no inherent advantage in using the ADRS.
However, there can be an advantage to using the spectral acceleration
vs. spectral displacement form of the push-over curve. The reason is
that the spectral displacement is an overall effective displacement of
the structure, not just the displacement at one point (e.g., the roof).
Hence, push-over plots in terms of spectral acceleration and spectral
displacement tend to be better behaved.

As noted in an earlier section, one case where this can be important is

where a structure has components that lose strength. In this case, the
push-over curve in terms of base shear and reference drift (usually
based on the roof displacement) can be badly behaved. A spectral
acceleration vs. spectral displacement tends to give a smoother curve.

28.6 Energy Dissipation and Damping

28.6.1 General
When a structure yields, energy is dissipated through inelastic
hysteresis. This energy dissipation can affect the results of dynamic
analysis, and hence should also be considered in static push-over
analysis.

In dynamic analysis the hysteresis loops, and hence the energy

dissipation, are considered directly. In static push-over analysis there is
no cyclic loading, so energy dissipation must be considered indirectly.
This section considers the procedures used by the different push-over
methods.
28.6.2 Stiffness and Degradation and Energy Ratio
The dissipated energy for an inelastic deformation cycle is related to
the area under the hysteresis loop. A measure of the amount of stiffness
degradation is the area of the degraded loop divided by the area of the
non-degraded loop. This can be termed the “energy ratio”. The larger
the amount of stiffness degradation, the smaller the energy ratio.

Figure 28.11 shows three hysteresis loops for a simple structure. If

there is no stiffness degradation, the loop might be as shown in Figure

PERFORM-3D User Guide 28-27

Chapter 28. General Push-Over Plots

28.11(a). In most cases, however, there will be stiffness degradation,

and the loop might be as shown in Figure 28.11(b) or (c).

For the loop in Figure 28.11(a) the energy ratio is 1.0. For the loop in
Figure 28.11(b) the energy ratio is 0.5 for all ductility ratios if Kh = 0.
If Kh > 0 the ratio increases somewhat as the ductility ratio increases.
The loop in Figure 28.11(c) also has an energy ratio of about 0.5. There
are many possible loops that give the same energy ratio. Other ratios
can be obtained by changing the details of the loop.

Figure 28.11 Hysteresis Loops

The energy ratio applies for full amplitude deformation cycles. The
shaded bands in Figure 28.11 show loops for small amplitude cycles. If
the energy ratio for a small amplitude cycle is the ratio of the degraded
area to the non-degraded area for the same deformation range, this ratio
is not necessarily the same as for a full amplitude cycle. Also, it is not
the same in Figures 28.11(b) and (c). This illustrates the inherent
difficulty in accounting for stiffness and energy degradation.
28.6.3 Equivalent Damping Ratio
In a linear structure with viscous damping, energy is dissipated by the
viscous damping. If the structure is vibrating with a constant amplitude
and period, it can be shown that the damping ratio, as a proportion of
critical damping, is given by

1  Dissipated Energy per Cycle 

Damping Ratio    (27.1)
4  Maximum Strain Energy 

28-28 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

This equation is used in the Capacity Spectrum Method to convert

inelastic dissipated energy to an equivalent damping ratio. The equation
is not used in the Coefficient Method or the Linearization Method.

For the ATC 40 Capacity Spectrum Method (Figure 28.4), the damping
ratio is calculated at the trial displacement. The dissipated energy is the
area of the hysteresis loop, and the maximum strain energy is the area
under the secant stiffness line, or 0.5HFor the Modified Capacity
Spectrum Method (Figure 28.5), the damping ratio is calculated at the
effective displacement. In both cases, the area of the hysteresis loop,
and hence the damping ratio, depends on the energy ratio. The
procedures for specifying the energy ratio for the Capacity Spectrum
Method are described in the next section.
28.6.4 Degradation in Push-Over Methods
Energy degradation is considered in all of the push-over methods, but
they use different procedures, as follows.

(1) In the FEMA 356 Coefficient Method, the coefficient C2 depends

on whether the structure degrades. For a degrading structure, C2
also depends on the performance level (Immediate Occupancy,
Life Safety or Collapse Prevention), increasing from the IO to the
CP level. In effect, this means that the energy ratio decreases as the
ductility ratio increases, but the relationship is not explicitly
specified.

(2) The FEMA 440 Coefficient Method is calibrated for certain

hysteresis loops, two of which are those in Figures 28.11(a) and
(b). The equations for C2 are different in the non-degrading and
degrading cases.

(3) The FEMA 440 Linearization Method is calibrated for certain

hysteresis loops, two of which are those in Figures 28.11(a) and
(b), with different calibration equations for the non-degrading and
degrading cases.

(4) In the ATC 40 Capacity Spectrum Method the damping ratio is

calculated at the trial point, using Equation 3.19. The energy ratio
depends on the non-degraded energy and the structure behavior
type (Type A, B or C). The energy ratios are specified by ATC 40.
For each behavior type the ratio decreases as the non-degraded
energy increases. The energy ratios are largest for Type A, and

PERFORM-3D User Guide 28-29

Chapter 28. General Push-Over Plots

decrease for Types B and C. PERFORM-3D adds a User type,

where the energy ratios are specified by the user.

(5) In the Modified Capacity Spectrum Method the damping ratio is

calculated at the effective displacement, using Equation 3.19.
There are three options for the energy ratio. One option is to use
the ATC 40 ratios. The second option is to specify the energy ratio
as a function of the ductility ratio. This is similar to the ATC 40
procedure, since the non-degraded energy increases as the ductility
ratio increases. The third option is to estimate the degraded energy
from the properties of the structural components and their
deformations, and to calculate the energy ratio directly. This is the
“PERFORM method”. It is described in the next section.
28.6.5 PERFORM Method for Energy Ratio
The “PERFORM method” for calculating energy ratios is more direct
than other methods. The steps in the method are as follows.

(1) The non-degraded energy is estimated for each inelastic

component in the analysis model, based on its current deformation.

(2) Energy degradation is accounted for using the PERFORM “Energy

Factors” (for an inelastic component, see the Component
Properties task and the Cyclic Degradation page). The energy
factor is calculated for each inelastic component based on its
current deformation. The non-degraded energy is multiplied by this
factor.

(3) The non-degraded and degraded energies are summed for all
inelastic components in the structure. The energy ratio is the ratio
of these two energies.
28.6.6 Extra Damping from Fluid Dampers
If a structure has elements that use fluid damper components, these
elements have zero stiffness for push-over analysis. Hence, they
develop no resisting forces and have no effect on the capacity curve.
However, since fluid dampers increase the amount of damping, they
can affect the demand curve, and hence the displacement demand. If
you wish, you can account for this additional damping. The method is
very approximate, but it may be useful for preliminary design.

To use the method you must do two main things, as follows.

28-30 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

(1) Specify “static pushover” properties when you define the

properties for fluid damper components (in the Component
Properties task). The default properties are zero.

(2) Use the Capacity Spectrum Method. This is the only method that
explicitly considers the amount of damping. The other methods
consider damping only implicitly, and can not account for
additional damping due to fluid dampers.

In the Component Properties task for Fluid Damper components,

choose the Static Pushover page. On this page, specify a viscous
coefficient, C, and a damping exponent, n. For energy dissipation
calculations in a static push-over analyses the force, F, in the
component is calculated as

F  C n (3.20)

where  is the deformation rate. The deformation rate is calculated as

follows.

During a push-over analysis, PERFORM-3D calculates the axial

deformation, , of each viscous bar element, at each point on the
capacity curve. All of this deformation is assumed to be concentrated in
the fluid damper component. That is, the elastic bar component that
acts in series with the damper is assumed to be very stiff. PERFORM-
3D then assumes that the structure is vibrating sinusoidally with a
constant amplitude, and that the period of vibration is the current secant
period. Hence, the deformation rate also varies sinusoidally, and the
force variation in the damper can be calculated (it is sinusoidal only if n
= 1). Hence the dissipated energy for one cycle of vibration can be
found. This energy is summed for all viscous bar elements, and
converted to a damping ratio, using Equation 3.19.

When the capacity curve is plotted for a structure that has viscous bar
elements, the damping ratio from the fluid dampers is shown on the
horizontal axis.

When the demand curve is calculated, the additional damping is

considered. The effect of this damping on the displacement demand can
be seen by specifying a value for the “effectiveness factor for fluid

PERFORM-3D User Guide 28-31

Chapter 28. General Push-Over Plots

dampers” in the Details page. To ignore the additional damping

specify 0 for this factor. To use the full damping specify 1. Since the
method is very approximate, experience may show that a factor
between 0 and 1 gives a more accurate estimate.
28.6.7 Strength Loss and Degradation
Strength loss can be caused by brittle behavior, by progressive strength
degradation under cyclic load, and/or by P- effects. For push-over
analysis it is usual to account for brittle strength loss and progressive
strength degradation in the “backbone” curves used for inelastic
components. P- effects are considered rationally, using P- theory.
This theory considers equilibrium of the structure in its deformed shape
whereas linear theory considers equilibrium in the undeformed shape.
True large displacements theory goes one step further, and considers
the nonlinear relationships between node displacements and element
deformations. It is rarely, if ever, necessary to consider true large
displacement effects for earthquake analysis of building structures.

Strength loss can cause a negative value for the stiffness Kh in the
capacity curve. It can cause unpredictable behavior in dynamic
analysis, with some analyses showing satisfactory behavior and others,
for similar earthquakes, showing collapse.

The FEMA 356 Coefficient Method considers strength loss in the

coefficient C3. The FEMA 440 Coefficient Method uses C3 = 1, and
imposes a limit on the amount of strength loss. The FEMA 440
Linearization Method considers negative values of Kh/Ke in its
calibration against dynamic analysis. The Capacity Spectrum Method
does not explicitly consider strength loss.

In general, the effects of strength loss, from whatever cause, are

complex, and it is difficult to get reliable results, from static push-over
analysis or dynamic analysis.

28.7 Theory for Relationship Between Sa and H

28.7.1 General
A key step in the computation is as follows: given the spectral
acceleration demand, S a , demand , calculate the corresponding base shear

28-32 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

demand, H demand . This is Step 3 in Figure 28.2, and Step 4 in Figures

28.3, 28.4 and 28.5.

For a single DOF structure, H demand is simply

W
H demand  S a , demand (3.21)
g

Where W is the structure weight. For a multi-DOF structure, FEMA

356 specifies the equation

W
H demand  Cm S a , demand (3.22)
g

where Cm is the effective mass factor for the “first” elastic mode. For
the general 3D case this means the mode that has the largest effective
mass factor along the push-over direction. As noted earlier in this
Chapter, this equation is accurate only for certain special cases. To
show this it is necessary to review some theory.

One of the important aspects of the dynamic response to an earthquake

is that there are no actual external forces on the structure. Instead, the
earthquake excites a dynamic response, and the forces on the structure
are the inertia forces associated with that response.

The same type of behavior is implied when push-over analysis is used.

In this case the dynamic response is approximated by applying static
external forces. There is implied vibration of the structure, and hence
implied inertia forces. These forces are not necessarily the same as the
applied external forces. This is an important aspect of the theory.
28.7.2 Special Case With Modal Load and Displaced Shape
Consider the following special case.

(1) The elastic mode shape of the structure is   .

(2) The external load Ris proportional to M   , where M  is the
mass matrix. In PERFORM this is a diagonal matrix.
(3) The displaced shape, r , is proportional to   . This will be the
case if the structure remains elastic. In general, as the structure

PERFORM-3D User Guide 28-33

Chapter 28. General Push-Over Plots

yields the displaced shape will change. For this section assume that
the displaced shape does not change.

If the spectral acceleration demand is S a , demand , its effect is similar to

that of a uniform, horizontal gravitational field of this magnitude. In
this gravitational field the forces on the structure are

Rg   S a, demand M d  (3.23)

where d  = matrix of direction cosines for the push-over direction. If

the push-over direction is along the H1 axis, matrix d  has a unit
value for each H1 mass in M  , and a zero value for all other masses.
Similarly, if the push-over direction is along the H2 axis, matrix d 
has a unit value for each H2 mass in M  , and a zero value for all other
masses.

The gravity forces are not the dynamic inertia forces on the masses. In
effect, the structure is deformed in its mode shape,   , and is vibrating
with some period, T, with a modal amplitude  (i.e., the displaced
shape at peak amplitude is    ). The peak accelerations are

2
r  42  (3.23)
T

and hence the peak inertia forces are

2
RI   M r  42  M  (3.25)
T

To obtain the modal amplitude, equate the work done by the gravity
and inertia forces acting through a displacement   :

T Rg   T RI  (3.26)

from which the amplitude is given by

28-34 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots


T2
S
 T M d  (3.27)
4 2
a , demand
T M  
Substitute for  in Equation (3.25) to get

RI   S a , demand  T M d  M 

T
(3.28)
 M 
Hence the inertia force base shear in the push-over direction is

( T M d ) 2
H I  RI T d   S a ,demand (3.29)
T M 
or
H I  S a , demand .M * (3.30)

where

( T M d ) 2
M*  (3.31)
T M 
is the effective mass. For push-over analysis it is usual to use the
dimensionless effective mass factor, Cm , equal to the effective mass
divided by the total mass. That is,

M* ( T M d ) 2 1
Cm   . (3.32)
W / g   M 
T
W / g 
where W is the structure weight. Hence the inertia force base shear
demand is

W
H I  S a ,demand .C m (3.33)
g

In this special case, the applied push-over loads are proportional to

M  , and hence have the same distribution as the inertia forces.
PERFORM-3D User Guide 28-35
Chapter 28. General Push-Over Plots

Hence, the inertia force base shear is equal to the actual base shear in
the push-over analysis. That is,

W
H demand  H I  S a , demand .Cm (3.34)
g

In this case, FEMA 356 Equation 3-16 gives the correct base shear
demand.
28.7.3 General Case Where Load Matches Displaced Shape
In the preceding case, the deflected shape is the elastic mode shape, and
the load pattern corresponds to this shape. It is also possible for the
load pattern to correspond to the deflected shape when this shape is not
a mode shape. This is the goal of “adaptive” push-over analysis, where
the load pattern at each step of the analysis is adjusted to match the
current displaced shape. That is, if the current displaced shape is r ,
the load is proportional to M r .

In this case, the equations of the preceding section still apply, but with
the mode shape,   , replaced by the deflected shape, r . The final
result is

W
H demand  S a ,demand .Cmr (3.35)
g
where
(rT M d ) 2 1
C mr  . (3.36)
rT M r W / g 

Again, FEMA 356 Equation 3-16 gives the correct base shear demand.

28.7.4 Case Where Load Does Not Match Displaced Shape

In most push-over analyses, the displaced shape, r , is not the elastic
mode shape, and the load pattern is not proportional to M r . Hence,
Equation 3.35 does not give the correct base shear demand.

28-36 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

Consider the case with a trial point that has a displaced shape r and a
lateral load R . The equations in this case are as follows.

As before, a spectral acceleration demand of S a , demand is equivalent to

a uniform gravitational field of this magnitude. In this gravitational
field the forces on the structure are

Rg   S a,demand M d  (3.37)

Again, these are not the dynamic inertia forces on the structure. In this
case the structure is vibrating with an effective mode shape r , a
period T and a modal amplitude  (i.e., the displaced shape at peak
amplitude is  r ). Hence the maximum inertia forces are

2
RI   42  M r (3.38)
T

To obtain the modal amplitude, equate the work done by these forces
acting through the displacement r . The work equation is

rT Rg   rT RI  (3.39)

from which the amplitude,  , is given by


T2
S
rT M d  (3.40)
4 2
a , demand
rT M r
Substitute for  in Equation 3.38 to get

RI   S a,demand rT M d M r

T
(3.41)
r M r
Hence the inertia force base shear is

(rT M d ) 2
H I , demand  RI T d   S a , demand (3.42)
rT M r
PERFORM-3D User Guide 28-37
Chapter 28. General Push-Over Plots

or
W
H I , demand  S a , demand .Cmr . (3.43)
g

where

(rT M d ) 2 1
C mr  . (3.44)
rT M r W / g 
This is the inertia base shear demand, not the required base shear for
the analysis, which is the base shear for the push-over loads R. To
calculate the required base shear demand, consider the relationship
between the push-over loads, R, and the inertia forces, RI  , at a
trial point. These are the base shear capacities at the trial point.

The base shear capacity from the push-over loads is

H capacity  RT d  (3.45)

The inertia forces are

2
RI   42 M r (3.46)
T

and the base shear capacity from these inertia forces is

H I , capacity  RI T d  (3.47)

The ratio between these two base shear capacities applies to force
demands as well as capacities. Hence

H capacity
H demand  H I ,demand (3.48)
H I ,capacity

or

28-38 PERFORM-3D User Guide

 Chapter 28. General Push-Over Plots

H demand 
RT d  S a ,demand Cmr
W
(3.49)
2
4
rT M d 
g
2
T

The effective period at the trial point is the secant period. From the
Rayleigh quotient, this period is given by

T  2
rT M r (3.50)
rT R
Substituting this into Equation 3.49 gives

 RT d   rT M r  W

H demand    S C (3.51)
 RT r  rT M d   a , demand mr g
  

If R is proportional to M r , the terms in parentheses = 1.0, and

Equation 3.51 reduces to the equation for the special case.

Equation 3.51 is rather complex. A simpler method is as follows.

From Equation 3.44

(rT M d ) 2 1
C mr  . (3.52)
r M r
T
W / g 
Substitution in Equation 3.51 gives

H demand  S a , demand
rT M d  RT d  (3.53)
rT R
Since the base shear capacity at the trial point is

H capacity  RT d  (3.54)

Equation 3.53 can be written as

PERFORM-3D User Guide 28-39

Chapter 28. General Push-Over Plots

H demand  S a , demand
rT M d  H (3.55)
rT R capacity
or

S a , demand
H demand  H capacity (3.56)
S a , capacity

where

S a ,capacity 
rT R (3.57)
rT M d 

28-40 PERFORM-3D User Guide

 Chapter 29. Target Displacement Plot

29 Target Displacement Plot

The preceding chapter describes a number of methods for

performance assessment using static push-over analysis, including
the FEMA 356 Coefficient (Target Displacement) Method. You can
use this method in either the General Push-Over plot task, or in this
Target Displacement Plot task. The procedures are rather different,
but the results should be the same.

This chapter summarizes the method and explains the procedure.

29.1 Target Displacement Method

29.1.1 Concept
The target displacement (or coefficient) method makes use of the
capacity curve from a static push-over analysis, and uses an empirical
formula to calculate a displacement demand (the target displacement).
You can then use limit states to assess the performance at this
displacement.

PERFORM-3D uses drift rather than displacement. The program

calculates a target value for the reference drift, which is typically the
roof displacement divided by the building height.

Figure 29.1 shows the ingredients of a target displacement plot. These

are as follows.

(1) A capacity curve. This is the relationship between lateral load and
lateral drift from a static push-over analysis.
(2) An estimated (guessed) target drift.
(3) A bilinear approximation of the capacity curve at the estimated
drift.
(4) A calculated target drift based on this bilinear approximation.

PERFORM-3D User Guide 29-1

Chapter 29. Target Displacement Plot

Figure 29.1 Ingredients of a Target Displacement Plot

In Figure 29.1 the calculated target drift is larger than the estimated
drift. A new estimate must be made, until the estimated and calculated
drifts are close to each other. This defines the target drift.

The key parts of the method are as follows.

(1) The bilinear approximation. The FEMA 356 approximation is

shown in Figure 29.1. The areas under the capacity curve and the
bilinear approximation are equal at the estimated target drift.

(2) The calculated target drift. This is as follows.

29.1.2 Target Drift

The formula for the calculated target drift is as follows.

  C0 C1C 2 C3 S 

where:

 = calculated target drift;

29-2 PERFORM-3D User Guide

 Chapter 29. Target Displacement Plot

S  = spectral displacement (drift) demand for the elastic

structure (see below);
C 0 = factor to convert spectral drift to reference drift;
C1 = factor to convert elastic drift to inelastic drift;
C 2 = factor to account for framing type and performance level;
and
C3 = factor to account for P- effects.

29.1.3 Spectral Drift

The spectral drift for the elastic structure is calculated by the procedure
specified in FEMA 356. The calculation requires the capacity curve, a
response spectrum, a mode shape and a mode period. You must specify
the response spectrum. The mode shape is usually the first elastic mode
shape (PERFORM-3D gives you other options). The mode period, Te,
is based on the effective stiffness, Ke.

See FEMA 356 for the formulas.

29.1.4 Factor C0
This factor converts spectral drift to reference (roof) drift. The spectral
drift is essentially an average drift value, and the roof drift is
numerically larger. For a simple lollipop the values are the same. For a
tall building the roof drift is about 1.5 times the spectral drift. The
actual value depends on the mode shape.

You have the option of using the mode shape or the deflected shape at
the estimated target displacement. You can also specify a user value.

29.1.5 Factor C1

This factor converts the drift for an elastic structure with stiffness Ke to
the drift for an inelastic structure with bilinear behavior as shown in
Figure 29.1. This factor is largely empirical. For details see FEMA 356.

You have the option of the FEMA 356 formula or a user value.

PERFORM-3D User Guide 29-3

Chapter 29. Target Displacement Plot

29.1.6 Factor C2
This factor accounts for the quality of the structure and the desired
performance level.

You can choose from two structure types, namely Type 1 (worse, with
more stiffness degradation) and Type 2 (better, with less stiffness
degradation). You can choose from three performance levels, namely
Immediate Occupancy, Life Safety and Collapse Prevention.

29.1.7 Factor C3
If there are substantial P- effects the tangent stiffness of the capacity
curve can become negative before the target drift is reached. For static
load this means that the structure is unstable and would collapse. For
dynamic load this is not necessarily the case, but the possibility of
serious damage is increased.

This factor accounts for this effect. It depends on the hardening

stiffness, Ke, in the bilinear approximation. For details see FEMA
356. You can also specify a user value.

There can also be negative stiffness because of strength loss,

independent of any P- effects.

29.2 PERFORM Implementation

To create a target displacement plot you must have results for at least
one push-over load case and you must have calculated one or more
mode shapes for the structure. To start a plot press the toolbar buttons
for the Analysis phase and the Target Displacement task. In the
Target Displacement form choose the analysis series from the Series
list and the load case from the Case list.

There are three main steps, each with substeps. First complete Step 1 as
follows.

(1) Define an elastic design spectrum. A diagram in the form shows

the required data.

29-4 PERFORM-3D User Guide

 Chapter 29. Target Displacement Plot

(2) Choose the framing type.

(3) Choose the performance level.
(4) Choose the curve type (base shear vs. reference drift or base shear
coefficient vs. reference drift).
(5) Press the Check button to check the data.

It everything is OK the capacity curve will be plotted. Next complete

Step 2 as follows.

(1) Choose the limit points that you want to show on the plot. A plot
with limit points is much more useful than one without.
(2) If you wish, click on any limit point in the plot. Its color changes to
blue and the name of the corresponding limit state is shown in the
box. You can return to this step to do this at any time.
(3) Press OK to go to Step 3.

Complete Step 3 as follows.

(1) Estimate a target drift and draw a bilinear approximation. To do

this follow the Step 3A instructions in the form. This may take
some practice. Press Clear to start over.
(2) Choose options for the C factors. If you choose the User option for
any factor, input the value.
(3) Press Calculate to calculate and plot the target drift.
(4) Repeat from Step (1) until the calculated drift is close to your
estimate.

The instructions for Step 3A are for the "click-and-drag" method for
constructing a bilinear approximation (see Chapter 28, General Push-
Over Plot, Section 28.3.5). For a FEMA 356 bilinear approximation
you can also use the quicker "right-click" method.

You can judge the performance of the structure by comparing the target
drift with the limit point drifts. To get usage ratios in more detail, use
the Usage Ratio Plot task, and plot usage ratios at the target
displacement.

To change the limit states, or to show the limit state for any limit point,
return to Step 2.

PERFORM-3D User Guide 29-5

 Chapter 30. Usage Ratio Graphs

30 Usage Ratio Graphs

Usage ratio graphs are valuable tools for assessing whether a

structure satisfies the performance criteria. This chapter explains
the procedure.

30.1 Usage Ratios

As the load factor increases in a gravity analysis, or as the drift

increases in a push-over analysis, or as the time increases in a dynamic
analysis, the usage ratios for the limit states will progressively increase.
A usage ratio graph shows the variation of usage ratio with load factor,
drift or time, depending on the analysis type.

Figure 30.1 shows the ingredients of a usage ratio graph, for a dynamic
earthquake analysis.

USAGE RATIO

A
1.0 Limit state
B usage ratios
at end of
earthquake
C

0 3 6 9 12 15 TIME (sec)
Usage ratio after
gravity analysis End of earthquake

Figure 30.1 Ingredients of a Usage Ratio Graph

In this figure there are three limit states (there will usually be more).
The upper line is the maximum usage ratio, considering all three limit

PERFORM-3D User Guide 30-1

Chapter 27. Usage Ratio Graphs

states. The usage ratio lines for the individual limit states are also
shown (see below for how to do this). Up to a time of about 6 seconds,
Limit State B governs, and after that Limit State A governs. Limit State
C is never critical.

The maximum usage ratio at the end of the analysis is less than 1.0 for
Limit States B and C, indicating that the structure satisfies the
performance criteria for these limit states. However, it is larger than 1.0
for Limit State A, indicating that the structure does not satisfy the
performance criteria for the complete limit state group.

30.2 Procedure

To start a usage ratio plot press the toolbar buttons for the Analysis
phase and the Usage Ratio task. In the Usage Ratio form choose the
analysis series from the Series list and the load case from the Case list.

To draw the graph choose the limit states that are to be considered, then
press the Plot button. The maximum usage ratio line is drawn, and at
the end of the plot the usage ratios for all limit states in the group are
shown as horizontal red lines. Click on any of these lines to draw the
usage ratio line for the corresponding limit state, and to show the name
of the limit state in the box.

You can save the usage ratios at the end of the graph by pressing the
save to file button at the right end of the toolbar.

30-2 PERFORM-3D User Guide

 Chapter 31. Load Case Combinations and Envelopes

31 Load Case Combinations and Envelopes.

The Usage Ratio Graphs and Deflected Shapes tasks consider

usage ratios for only one load case at a time. The Usage Ratio
Envelopes task allows load case combinations to be specified, and
for usage ratios to be calculated and displayed for these
combinations.

This Chapter describes the procedure. There are a number of

options for setting up the load case combinations and for combining
the usage ratios. It is important to understand the combination
methods that are used, and to choose the appropriate options for
your application.

31.1 Overview

31.1.1 Main Steps

The main steps in the process are as follows.

(1) Run a number of dynamic earthquake or static push-over analyses.

For dynamic earthquake analyses it will be usual to consider
several earthquakes and possibly to apply each earthquake in two
horizontal directions. For 7 earthquakes and 2 direction this is 14
analyses. For push-over analysis it will be usual to consider two or
more lateral load distributions and to push the structure in two or
more horizontal directions. For 2 lateral load distributions and 2
directions this is 4 analyses. For each analysis the usage ratios will
be calculated for a number of limit states. For dynamic analyses
the usage ratios are at the end of the analysis. For push-over
analyses they are at the estimated displacement demand (the target
displacement or performance point) for each analysis.

(2) Specify one or more Load Case Combinations. A combination can

consist of just a set of load cases, for example the 7 earthquake
cases in one horizontal direction. In this case there would be 2
combinations, each with 7 load cases. Alternatively, a combination
can consist of a number of groups of load cases, for example 7
groups of 2 load cases each, with each group containing the

PERFORM-3D User Guide 31-1

Chapter 31. Load Case Combinations and Envelopes

analyses for 2 horizontal directions. In this case there would be a

single combinations, with 14 load cases in 7 groups of 2.

(3) If the combination includes load case groups, choose a

combination method to be applied within each group. The options
are Maximum, SRSS (square root of sum of squares) and Sum.

(4) Choose a combination method to be applied across groups (or

across load cases if there are no groups). The options are “Max-
Max”, “Mean-Max”, “Max-Mean” and "Max-(Mean+NSigma)", as
explained later.

The end result is a combined usage ratio for each limit state. These
usage ratios can be displayed in a diagram that is somewhat similar to
that used in the Usage Ratio Graphs task, except that final usage ratios
are shown, not graphs of usage ratio vs. time or load. Usage ratios for
individual elements can also be displayed in a color-coded drawing
similar to that in the Deflected Shapes task.
31.1.2 Load Case Combinations
A load case combination can consist of up to 50 load cases, divided
into groups if desired. If there are groups, they must all have the same
number of load cases. A combination can contain both static and
dynamic load cases, but this is unlikely.

As one example, consider a structure for which dynamic earthquake

analyses have been carried out for 7 earthquakes applied in each of 2
directions. For each analysis, usage ratios are calculated at the end of
the analysis. FEMA 356 allows the usage ratios to be averaged over the
7 earthquakes. The load case combination in this case could consist of 7
groups of 2 load cases. Within each group (2 earthquake directions) the
usage ratios might be combined using the Maximum method. Across
groups (7 earthquakes) the usage ratios might be combined using the
Max-Mean method, as explained later.

As a second example, consider a structure for which static push-over

analyses have been carried out for lateral load distributions
corresponding to 3 mode shapes, in each of 2 directions. For each
analysis, usage ratios are calculated at the estimated drift demand
(target drift). Usage ratios are to be combined using the Chopra-Goel
multi-mode procedure. The load case combination in this case could
consist of 2 groups of 3 load cases. Within each group (3 modal load

31-2 PERFORM-3D User Guide

 Chapter 31. Load Case Combinations and Envelopes

cases) the usage ratios would be combined using the SRSS method.
Across groups (2 directions) the usage ratios would be combined using
the Max-Max method.
31.1.3 Combination Methods
Within any load case group, the combination options for usage ratios
are Maximum, SRSS and Sum, which are self-explanatory. Usage
ratios are always positive or zero, so signs are not a concern.

Across load case groups the combination options are Max-Max, Mean-
Max, Max-Mean, and Max-(Mean+NSigma). To illustrate the
difference, consider a structure with 2 elements (Elements 1 and 2) and
3 load cases (Load Cases A, B and C). Consider one of the limit states,
with usage ratios for the elements as shown in the following table.

Load Case A Load Case B Load Case C

Element 1 Usage ratio =0.91 0.77 0.87
Element 2 0.67 0.83 0.75

The limit state usage ratio for the structure is calculated as follows, for
each of the four combination methods.

(1) Max-Max.
For each element, usage ratio = Max over all load cases.
For structure, usage ratio = Max over all elements.
Element 1 usage ratio = Max (0.91, 0.77, 0.87) = 0.91
Element 2 usage ratio = Max (0.67, 0.83, 0.75) = 0.83
Structure usage ratio = Max (0.91, 0.83) = 0.91

(2) Mean-Max.
For each element, usage ratio = Max over all load cases.
For structure, usage ratio = Mean over all elements.
Element 1 usage ratio = Max (0.91, 0.77, 0.87) = 0.91
Element 2 usage ratio = Max (0.67, 0.83, 0.75) = 0.83
Structure usage ratio = Mean (0.91, 0.83) = 0.87

(3) Max-Mean
For each element, usage ratio = Mean over all load cases.
For structure, usage ratio = Max over all elements.
Element 1 usage ratio = Mean (0.91, 0.77, 0.87) = 0.85
Element 2 usage ratio = Mean (0.67, 0.83, 0.75) = 0.75
Structure usage ratio = Max (0.85, 0.75 ) = 0.85

PERFORM-3D User Guide 31-3

Chapter 31. Load Case Combinations and Envelopes

(4) Max-(Mean+NSigma)
For each element, usage ratio = Mean + N times the standard
deviation (sigma) over all load cases.
For structure, usage ratio = Max over all elements.
Element 1 usage ratio : Mean = 0.85, Sigma = 0.059
Element 2 usage ratio : Mean = 0.75, Sigma = 0.065
For N = 1, structure usage ratio = Max (0.909, 0.835 ) = 0.909

The Max-Max method is the most conservative. This method is

required by FEMA 356 if fewer than 7 earthquakes are considered. The
Max-Mean method is the least conservative. This method is allowed by
FEMA 356 if 7 or more earthquakes are considered. The Mean-Max
method is between the Max-Max and Max-Mean methods. It is not
included in FEMA 356. The Max-(Mean+NSigma) method is also not
included in FEMA 356, but it may be appropriate for strength-based
limit states when capacity design is used. In this case, strengths based
on the Max-Mean method may not be sufficiently large to ensure
essentially elastic behavior.

The procedure is similar for limit states based on drift, deflection or

structure section results, rather than element results. For example, if a
drift limit state has several drifts, in the above table replace Element 1
by Drift 1, etc.

31.2 Procedure

31.2.1 Load Combinations

Press the toolbar buttons for the Analysis phase and the Combinations
and Envelopes task. The required analyses must have been run. They
do not all have to be in the same Analysis Series.

The steps for setting up or editing a Load Case Combination are as

follows.

(1) To start a new combination, press New. Specify the combination

name, specify the number of load cases per group (1 if there are no
groups) and choose the combination method within groups.
(2) Set up a list of load cases in the combination, by choosing the
Analysis Series and Load Case Name from the lists on the toolbar,
and pressing Add. Each load case group is assigned a number as

31-4 PERFORM-3D User Guide

 Chapter 31. Load Case Combinations and Envelopes

the load cases are added. You can delete load cases, a group at a
time, by clicking in the table to highlight the group and pressing
Delete.
(3) When the list is complete, press Save. You can edit an existing
combination at any time, using Delete and Add.
31.2.2 Usage Ratios for Structure
The steps for calculating and displaying usage ratios for the complete
structure are as follows.

(1) Go to the Structure page.

(2) Choose the Combination Method.
(3) Choose the limit states that are to be considered.
(4) Press Plot to plot the results.

The results are plotted as shown in Figure 31.1.

In this case there are 7 load cases, with only one case per group (i.e., no
combination within groups). For each load case there is a blue vertical
line, with a short horizontal line for each limit state, showing the limit
state usage ratios for the load case. On the left side there is a red
vertical line, also with a short horizontal line for each limit state,
showing the limit state usage ratios for the load combination. If you
click on any of the limit states on the red line, a horizontal red line is
drawn showing the usage ratio for the load combination (as shown in
the figure), a blue line is drawn connecting the usage ratios for the
separate load cases (as shown in the figure), and the name of the limit
state is shown in a box at the bottom of the form (not shown in the
figure).

PERFORM-3D User Guide 31-5

Chapter 31. Load Case Combinations and Envelopes

Figure 31.1 Usage Ratios for Structure

31.2.3 Usage Ratios for Elements

The steps for showing element usage ratios in a color-coded drawing
are as follows.

(1) Go to the Element Colors page.

(2) Choose the Combination Method. This is the combination type
across load cases (or load case groups). The options are Max,
Mean and Mean+NSigma. Usage ratios are not combined across
elements for this plot.
(3) Choose the limit states that are to be considered.
(4) Choose a color group for the element colors.
(5) Press Plot to draw the structure.

31-6 PERFORM-3D User Guide

 Chapter 31. Load Case Combinations and Envelopes

The drawing is similar to that for the Deflected Shape task, but
deflections are not shown. To zoom in, use the two zoom buttons
directly below the Plot button.

PERFORM-3D User Guide 31-7

 Chapter 32. Importing/Exporting Structure Data

32 Importing/Exporting Structure Data

If you have analyzed a structure using a linear analysis program,

you may be able to save time by importing the linear analysis model
into PERFORM-3D. You may also wish to export data from a
PERFORM-3D model, for import into another PERFORM-3D
model or a different computer program.

PERFORM-3D can import and export the data for nodes, masses,
elements and nodal loads. This chapter describes the procedures.

PERFORM-3D can also export and import the data for inelastic,
elastic, etc. components. Among other things, you can use this
feature to set up libraries of commonly used components, and
import them into different PERFORM-3D models. This chapter also
describes this procedure.

32.1 Importing Structure Data

32.1.1 Text Files

PERFORM-3D can not import structure data by connecting directly to
another program. You must export the data from the other program, and
organize it into a number of comma-delimited text files. PERFORM
can then read those files. You may be able to export the files directly
from the other program, or you may need to export the data and modify
its format using a spreadsheet program. You can import the following.

(1) Elements, with or without their nodes.

(2) Nodes only. It is usually better to import nodes with the elements.
(3) Nodal masses.
(4) Nodal loads.

A key point is that nodes must be identified by their H1, H2, V

coordinates, not by node numbers. For example, if you are importing
element data for 4-node wall elements, the text file must have 12
coordinate values for each element, 3 each for nodes I, J, K and L of
the element.

PERFORM-3D User Guide 32-1

Chapter 32. Importing/Exporting Structure Data

32.1.2 File for Elements (with Optional Nodes)

In the PERFORM-3D model, the elements are organized in Element
Groups. In a single import operation you can import elements into only
one group. This does not means that you must have a separate text file
for each element group. If you wish you can put all of the elements in
one file, then read them selectively, as described later. However, you
can put the elements for each element group into a separate file if you
wish. You can also read elements into a single element group, then
divide them into separate groups within PERFORM.

For any element group you have the option of importing only the
elements, or elements plus nodes. The difference is as follows.

(1) If you choose the element-only option, an element is imported only

if all of its nodes currently exist in the PERFORM-3D model. The
tolerance for whether a node exists is one third of the minimum
spacing between nodes for the PERFORM-3D model. This spacing
is typically 6 inches or 15 cm, but you can set it to a different value
for the import operation if you wish. If all of the element nodes
exist with H1, H2 and V coordinates that are within this tolerance
(typically 6/3 = 2 inches or 15/3 = 5 cm), the element is imported.
If not, it is ignored. You will be warned if there are some elements
that are not imported.

(2) If you choose the element-plus-nodes option, an element is

imported even if all of its nodes do not currently exist in the
PERFORM-3D model. If a node does not already exist, a new
node is created. The tolerance for whether a node exists is equal to
the minimum spacing between nodes for the PERFORM-3D model
(typically a tolerance of 6 inches or 15 cm).

The file format is as follows.

(1) First N lines (as many as you wish) : descriptions of the file. Each
line can be of any length, and must end with a line feed (carriage
return). These lines are skipped when the data is imported.

(2) Remaining lines : one line per element, as follows.

32-2 PERFORM-3D User Guide

 Chapter 32. Importing/Exporting Structure Data

(a) Single node element (e.g., a panel zone element). Each line
must contain the H1, H2 and V coordinates of the node,
separated by commas.

(b) Two-node element (e.g., a beam element). Each line must

contain the H1, H2 and V coordinates of the node at End I,
followed by the H1, H2 and V coordinates of the node at End J.

For a frame type element (beam, column or brace/other) there

can be two additional items if desired, describing the
orientation of the local 2, 3 axes for the element (the 1 axis is
along the element). The first item is the global reference axis
for the element. If this item is 1, local Axis 2 of the element is
basically parallel to the global H1 axis; if it is 2, local Axis 2 of
the element is basically parallel to the global H2 axis; and if it
is 3, local Axis 2 of the element is basically parallel to the
global V axis. The second item is the twist angle, for cases
where local Axis 2 is not along a global axis. This item is the
twist angle to the inclined axis, in degrees, measured clockwise
about the element Axis 1.

(c) Four-node element (e.g., a shear wall walls). Each line must
contain 12 items, namely the H1, H2 and V coordinates at the
element nodes. You have two options for ordering the nodes.
The first option is I, J, K, L, where I is usually bottom left, J is
bottom right, K is top left, and L is top right. In this case the
nodes are ordered zigzag fashion, not around the element. The
second option is I, J, L, K, which is counterclockwise around
the element.

The length unit for the coordinates must be the same as the global
length unit for the PERFORM-3D structure.

You can import only the basic element geometry. You must specify the
element properties in PERFORM-3D. You must also specify the local
axis orientations of single-node and four-node elements in PERFORM-
3D.
32.1.3 File for Nodes
Usually you will not need to import nodes directly, since nodes can also
be created as you import elements. However, you can import nodes if
you wish.

PERFORM-3D User Guide 32-3

Chapter 32. Importing/Exporting Structure Data

The file format is as follows.

(1) First N lines (as many as you wish) : descriptions of the file. Each
line can be of any length, and must end with a line feed (carriage
return). These lines are skipped when the data is imported.

(2) Remaining lines : one line per node. Each line must contain the
H1, H2 and V coordinates of the node, separated by commas. The
length unit for the coordinates must be the same as the global
length unit for the PERFORM-3D structure.

If a node already exists the line is ignored. The tolerance for

whether a node exists is equal to the minimum spacing between
nodes for the PERFORM-3D model. This spacing is typically 6
inches or 15 cm, but you can set it to a different value for the
import operation if you wish. You will be warned if there are some
nodes that do not match.
32.1.4 File for Nodal Masses
In order to import masses you must first create or import the nodes.

The file format is as follows.

(1) First N lines (as many as you wish) : descriptions of the file. Each
line can be of any length, and must end with a line feed (carriage
return). These lines are skipped when the data is imported.

(2) Remaining lines : one line per mass. Each line must contain the
H1, H2 and V coordinates of the corresponding node, followed by
6 mass values, namely the H1, H2 and V translational masses and
the H1, H2 and V rotational masses at the node, all separated by
commas. The length unit for the coordinates must be the same as
the global length unit for the PERFORM-3D structure. The masses
must be in mass units (FT2/L). The force and length units must be
the same as the global force and length units for the PERFORM-
3D structure. T is seconds.

If the file contains a node that does not exist (i.e., if the coordinates do
not match the coordinates of an existing node), the masses are not
imported. The tolerance for whether a node exists is equal to the
minimum spacing between nodes for the PERFORM-3D model. This

32-4 PERFORM-3D User Guide

 Chapter 32. Importing/Exporting Structure Data

spacing is typically 6 inches or 15 cm, but you can set it to a different

value for the import operation if you wish. You will be warned if there
are some nodes that do not match.

If a node already has mass, the imported masses are added to the
existing masses. If a node appears two or more times in a file, the
masses at that node are added.

You can also use these files for importing nodes only (using the node
coordinates but ignoring the mass data). When you use a nodal mass
file, this option is available. It can be convenient to use this option to
read nodes first, then read masses, using the same file.
32.1.5 File for Nodal Loads
You can import nodal loads only. Element loads must be specified in
PERFORM-3D.

In the PERFORM-3D model, the loads are organized in Load Patterns.

Load Cases are then built up from the load patterns. You can import
only load patterns.

In a single import operation you can import loads into only one load
pattern. This does not means that you must have a separate text file for
each pattern, although that is probably a good idea.

In order to import loads you must first create or import the nodes.

The file format is as follows.

(1) First N lines (as many as you wish) : descriptions of the file. Each
line can be of any length, and must end with a line feed (carriage
return). These lines are skipped when the data is imported.

(2) Remaining lines : one line per set of loads. Each line must contain
the H1, H2 and V coordinates of the corresponding node, followed
by 6 load values, namely the H1, H2 and V forces and the H1, H2
and V moments, all separated by commas. The length unit for the
coordinates must be the same as the global length unit for the
PERFORM-3D structure. The force and length units for the loads
must be the same as the global force and length units for the
PERFORM-3D structure.

PERFORM-3D User Guide 32-5

Chapter 32. Importing/Exporting Structure Data

If the file contains a node that does not exist (i.e., if the coordinates do
not match the coordinates of an existing node), the loads are not
imported. The tolerance for whether a node exists is equal to the
minimum spacing between nodes for the PERFORM-3D model. This
spacing is typically 6 inches or 15 cm, but you can set it to a different
value for the import operation if you wish. You will be warned if there
are some nodes that do not match.

If a node already has loads, the imported loads are added to the existing
loads. If a node appears two or more times in a file, the loads on that
node are added.

You can also use these files for importing nodes only (using the node
coordinates but ignoring the load data). When you use a nodal load file,
this option is available. It can be convenient to use this option to read
nodes first, then read loads, using the same file.

32.2 Note on Element Import

In PERFORM, elements are divided into Element Groups. If you

export elements from PERFORM you must do so one element group at
a time, and there will be a different file for each group. If you then
import the same elements, you must do so one file at a time. You can
read the files into different element groups, corresponding to the groups
that were exported, or you can import multiple files into one element
group. Since the files are text files you can edit them if you wish. You
could concatenate a number of exported files into a single file.

If the file containing element data is created by a program other than

PERFORM, the elements may not be divided into files that correspond
to element groups in the PERFORM model. For example, in
PERFORM, beams and columns are both frame type elements, but they
must be in separate element groups. If the file is created by another
program, beam and column elements may be in the same file. Also,
PERFORM has different 4-node element types, namely shear wall,
general wall and slab/shell elements. Elements of these types must be
in separate groups. Hence, a file may contain elements that you want to
assign to different element types. These situations can be handled as
follows.

32-6 PERFORM-3D User Guide

 Chapter 32. Importing/Exporting Structure Data

32.2.1.1 Beams and Columns in One File

If you have Beam and Column elements together in a file, the steps are
as follows.

(1) In the Elements task create the element groups that you need, with
no elements.
(2) In the Import/Export Structure Data task, import all of the
elements into one of these groups.
(3) In the Elements task, choose the Change Group tab and move
elements into the required groups. PERFORM allows you to move
elements from Beam element groups to Column element groups
and vice versa.

There is an alternative method, which may be easier if the elements in

the text file are well organized. This is to edit the text file, splitting it
into two or more files where each file contains only elements that you
want to be of Beam or Column type. You can then import each file into
a separate element group.
32.2.1.2 Shear Wall and General Wall Elements in One File
The steps are the same as for beams and columns. PERFORM allows
you to move elements from Shear Wall element groups to General Wall
element groups and vice versa.
32.2.1.3 Elements of Incompatible Types in One File
Suppose you have a file that contains a number of diagonal brace
elements, and you want to model some of them as Brace/Other Frame
elements and the rest as BRB elements. You can not use the preceding
method, because PERFORM does not allow you to copy elements from
a Brace/Other Frame element group to a BRB element group. In this
case you may be able to edit the file to divide the elements. If this is not
convenient you can use the following procedure.

This procedure is more complex than that for separating frame or wall
type elements. Since you can not use the Change Group operation, you
must use the Import/Export task.

First consider the case where you want to move all of the elements
from a Brace/Other Frame group to a BRB group. The steps are as
follows.

PERFORM-3D User Guide 32-7

Chapter 32. Importing/Exporting Structure Data

(1) In the Elements task create the BRB element group that you need,
with no elements.
(2) In the Import/Export Structure Data task, export the
Brace/Other Frame element group to a file. Be sure not to include
the element orientation data, otherwise you will get an error
message in Step (4).
(3) Delete the Brace/Other Frame element group.
(4) In the Import/Export Structure Data task, re-import the elements
into the BRB element group.

Second, consider the case where you have a file with elements that you
want to divide into different types. Suppose that you want to divide the
elements into Brace/Other Frame and BRB groups. The steps are as
follows.

(1) In the Elements task create two element groups, both of

Brace/Other Frame type.
(2) In the Import/Export Structure Data task, import all of the
elements into one of these groups.
(3) In the Elements task, choose the Change Group tab and move
those elements that you want to be BRB elements into the second
group.
(4) Using the above procedure, export the elements in the second
group and re-import them as BRB elements.
(5) If you wish, use the Elements task and Change Group tab to
divide into smaller element groups.

32.3 Import and Export Procedures

32.3.1 Import
You must set up the data in comma-delimited text files, as described in
the preceding section. For a structure of modest size the most direct
method is to create one file for the elements, a second file for the
masses, and a third for the loads. Import the elements first, in blocks
that correspond to the element groups. This will also create the nodes.
Then import the masses and loads. In between import operations you
may wish to do other operations such as adding supports or defining
frames.

32-8 PERFORM-3D User Guide

 Chapter 32. Importing/Exporting Structure Data

For a very large structure it may be better to use a larger number of

files. For example, you might have a separate file for each element
group, and a separate file for each load pattern.

To start an import operation, choose the Modeling phase and the

Import/Export Structure Data task. Choose the Import option and
the Elements, Nodes, Masses or Loads page, and follow the
instructions.

For each import operation you must specify the name of the text file. If
the file is in the standard USER folder you need to give only the name
(including any extensions such as ".txt"). If the file is in some other
location you must give the complete path. You can browse to do this.

You must then specify the number of lines to be skipped at the start of
the file, and the number of lines to be read. You must always skip any
description lines at the start of the file, and you may choose to skip
other lines.

For example, suppose that you have a file that has 3 description lines,
followed by all of the elements in a structure. Suppose that you want to
put the first 100 elements into element Group 1, the next 100 into
element Group 2, and the remainder into element Group 3, and that you
have already created these three groups (with no elements). This
requires three import operations. In the first, skip 3 lines and read 100
lines into Group 1. In the second, skip 103 lines and read 100 lines into
Group 2. In the third, skip 203 lines and read the rest of the file into
Group 3.

For each operation, after you have specified the file name, the numbers
of lines, and possibly some import options, use the Test, Cancel, OK
and Undo buttons in the usual way.
32.3.2 Export
You can export data into the same text files as described in the
preceding sections.

To start an export operation, choose the Modeling phase and the

Import/Export Structure Data task. Choose the Export option and
the Elements, Nodes, Masses or Loads page, and follow the
instructions.

PERFORM-3D User Guide 32-9

Chapter 32. Importing/Exporting Structure Data

32.4 Component Properties Export/Import

32.4.1 General
Component properties can be exported from one PERFORM model and
imported into another PERFORM model. You can do the following
operations.

(1) Export a complete set of component properties from one structure

and import these properties into another structure.
(2) Export the properties for selected components of a particular type
(e.g., selected fiber cross sections, or selected seismic isolator
components) and import them into another structure.
(3) Create libraries of commonly used components, and import
components from these libraries into any structure.

The export/import is done by writing and reading files. These are not
text files and they can not be edited. PERFORM currently does not
allow component properties to be imported from text files.
32.4.2 Export/Import Steps
32.4.2.1 1. Create the components.
First create the components. If you want to set up a library of standard
components you can set up a new structure, define components but no
nodes or elements, then export those components.

As shown below, it is easier to export and import a complete set of

components than selected components, but you can export and import
selected components if you wish. This will require careful organization
to keep track of the files.
32.4.2.2 2. To export a complete set of components :
(1) In the Modeling phase choose the Component Properties task.
(2) At the bottom of the form choose the Export Components tab and
the Components of all types option.
(3) Press the Export button and follow the instructions.

You must create a new file for each set of components that you
export. The file will be assigned a ".PF3CXX" extension.

32-10 PERFORM-3D User Guide

 Chapter 32. Importing/Exporting Structure Data

32.4.2.3 3. To import a complete set of components :

(1) In the Modeling phase choose the Component Properties task.
(2) At the bottom of the form choose the Import Components tab and
the Components of all types option.
(3) Press the Import button and follow the instructions.

You must read a complete file, with a ".PF3CXX" extension.

Usually you will use this import option only for a new structure,
where you have not yet defined any components. If you have
already defined any components for the current structure, these
components will be overwritten by the imported components.
Also, if you have assigned component properties to any elements,
the assignments will all be changed to “NONE”.
32.4.2.4 4. To export selected components :
(1) In the Modeling phase choose the Component Properties task.
(2) Choose the component type that you want to export from the
component type list.
(3) At the bottom of the form choose the Export Components tab and
the Components of this type only option.
(4) Press the Export button and follow the instructions. You can
export all components of the current type, or only selected
components.

You can either create a new file or add components to an existing

file. The file will be assigned a ".PF3CMP" extension. You can
export components of several types to a single file if you wish.
32.4.2.5 5. To import selected components :
(1) In the Modeling phase choose the Component Properties task.
(2) Choose the component type that you want to import from the
component type list. You may or may not already have
components of this type.
(3) At the bottom of the form choose the Import Components tab and
the Components of this type only option.
(4) Press the Import button and follow the instructions.

You can import all of the components of the current type from the
file, or only selected components. If you have already defined
some components, be sure to specify whether you want

PERFORM-3D User Guide 32-11

Chapter 32. Importing/Exporting Structure Data

components with duplicate names to be imported (over-write

current component) or ignored (keep current component).

The file can contain components of several types (see preceding

section).

An important consideration is that some components can depend

on others. For example, a plastic hinge component may depend on
a cross section component, and a fiber section component will
depend on one or more material components. If you want to import
a component (say Component A) that depends on another
component (Component B), before you can import Component A
you must first import Component B. If Component B does not
exist, Component A will not be imported (you will get a warning
message).

32-12 PERFORM-3D User Guide

 Chapter 33. Reporting Errors

33 Procedure For Reporting Errors

We do our best to keep PERFORM-3D reliable and free of errors,

but we are not always successful. If you report an error you may be
asked to send us copies of certain structure files. This chapter
describes the procedure.

33.1 Reporting Errors

To report an error, please send an email to the following address :

[email protected]

Please describe the problem as clearly as you can. If PERFORM-3D

has displayed an error message, please include the text of the message.
If you get a "we goofed" message, please go to the end of the ECHO
file, where there should be an error message. Please zip the ECHO file
and attach it to your email. If the zipped file is long (more than about
1MB) please send just the last few lines of the ECHO file, including the
error message.

You will be contacted and advised on how to proceed.

We may ask you to send us the data files for your structure. The
procedure is described in the next section.

33.2 Procedure For Sending Files

33.2.1 Structure Data

For a description of the way that PERFORM-3D data is stored, see
Chapter 2, Structures, Folders and Files. Find the folder for your
structure.

You can send us the data as an email attachment, or you can upload it
to our ftp site.

PERFORM-3D User Guide 33-1

Chapter 33. Reporting Errors

33.2.2 Send As An Email Attachment

Usually you will send only the structure data, with no analysis results.
The reason is that the amount of data to be sent in this case is not large,
and it can usually be sent as an email attachment. If there are analysis
results the amount of data is much larger, and even a zipped file is
likely to be too large (the limit for an email attachment is about 5MB)..

The steps are as follows.

(1) If you do not have any completed analyses for your structure, go to
Step 4.
(2) If you have analyses and do not need to keep them, open the
structure, go to the Run Analyses task, and delete the analyses in
all analysis series (do not delete the analysis series completely). Go
to Step 4.
(3) If you have analyses and want to keep them, press the Save as a
new structure button on the toolbar. Save the structure without
analysis results. Go to Step 4.
(4) Zip the folder for your structure, including subfolders. There will
be one subfolder for each analysis series. If you have deleted all
analysis results these subfolders will be empty, but include them
anyway.
(5) Send us an email explaining the problem, with the zipped file as an
attachment.
33.2.3 Upload to FTP Site
If the amount of data is too large to email, we may ask that you upload
it to an ftp site. If this is necessary we will send you instructions.

33-2 PERFORM-3D User Guide