Lehigh

Preserve Institutional Repository

Shape Function Limitations within Equivalent

Single-Degree-of-Freedom Analysis for Structural
Elements Subject to Blast Loading
Mullen, Evan
2016

Find more at https://preserve.lib.lehigh.edu/

This document is brought to you for free and open access by Lehigh
Preserve. It has been accepted for inclusion by an authorized administrator
of Lehigh Preserve. For more information, please contact
[email protected].
Shape Function Limitations within Equivalent Single-Degree-of-Freedom Analysis for
Structural Elements Subject to Blast Loading

by

Evan M. Mullen

A Thesis
Presented to the Graduate and Research Committee
of Lehigh University
in Candidacy for the Degree of
Master of Science

in
Structural Engineering

Lehigh University
January 2016
Copyright Page

ii
This thesis is accepted and approved in partial fulfillment of the requirements for the
Master of Science.

_______________________
Date

_______________________________________
Dr. Spencer Quiel, Thesis Advisor

________________________________________
Dr. Panayiotis (Panos) Diplas, Chairperson of Department

iii
 Acknowledgements

I would like to first thank Sara Barker and Harry Nimoityn for their continuous support

throughout my studies and research. They offered great encouragement from start to

finish during this challenging, yet rewarding process.

I would also like to thank my research advisor, Dr. Spencer Quiel. His combination of

practical and theoretical knowledge shaped my research and helped me complete my

thesis with significant conclusions.

iv
 Table of Contents

Acknowledgements ......................................................................................................... iv

List of Tables .................................................................................................................... ix

List of Figures .................................................................................................................... x

Abstract............................................................................................................................. 1

Chapter 1 Introduction ..................................................................................................... 3

1.1 General .................................................................................................................... 3

1.2 Background ............................................................................................................. 3

1.3 Research Objectives ................................................................................................ 4

1.4 Scope of Thesis ........................................................................................................ 5

Chapter 2 Background and Modeling ............................................................................... 6

2.1 General .................................................................................................................... 6

2.2 Numerical Solutions ................................................................................................ 7

2.3 Time Discretization ............................................................................................... 10

2.4 Loads ..................................................................................................................... 10

2.5 Material Properties and Plasticity ......................................................................... 13

2.6 Boundary Conditions............................................................................................. 15

2.6.1 Transformation Factors (For SDOF Analysis).................................................. 16

v
 2.7 Space Discretization .............................................................................................. 21

2.8 Element Type ........................................................................................................ 23

2.9 Geometric Effects.................................................................................................. 24

2.10 Mass Matrix......................................................................................................... 26

2.11 Mode Shapes....................................................................................................... 27

2.12 Damping .............................................................................................................. 29

Chapter 3 Single-Degree-of-Freedom Solver.................................................................. 34

3.1 General .................................................................................................................. 34

3.2 SBEDS .................................................................................................................... 34

3.2.1 Intro Tab ......................................................................................................... 34

3.2.2 Input Tab ........................................................................................................ 34

3.2.3 Results Tab ..................................................................................................... 35

3.2.4 P-I_Diagram Tab ............................................................................................. 36

3.2.5 SDOF Output Tab ............................................................................................ 36

3.2.6 Database Tab .................................................................................................. 37

3.2.7 Beam_Types Tab ............................................................................................ 37

3.3 MATLAB Replica .................................................................................................... 37

3.3.1 Main File ......................................................................................................... 38

3.3.2 Load File.......................................................................................................... 38

vi
 3.3.3 Biggs File ......................................................................................................... 39

3.3.4 Const_vel File ................................................................................................. 39

3.4 Solver Verification ................................................................................................. 40

3.4.1 Inconsistencies ............................................................................................... 42

Chapter 4 Multi-Degree-of-Freedom Solver .................................................................. 50

4.1 General .................................................................................................................. 50

4.2 SAP2000 ................................................................................................................ 50

4.2.1 Inputs .............................................................................................................. 50

4.2.2 Outputs ........................................................................................................... 53

4.3 MATLAB Replica .................................................................................................... 53

4.3.1 Input File ......................................................................................................... 54

4.3.2 Main File ......................................................................................................... 54

4.4 Solver Verification ................................................................................................. 55

4.4.1 Inconsistencies ............................................................................................... 59

Chapter 5 Comparisons .................................................................................................. 69

5.1 General .................................................................................................................. 69

5.2 Results and Analysis .............................................................................................. 70

5.2.1 First Data Set .................................................................................................. 70

5.2.2 Second Data Set.............................................................................................. 80

vii
 5.2.3 Third Data Set ................................................................................................. 84

5.2.4 Individual Deflected Shapes ........................................................................... 89

Chapter 6 Conclusions and Future Research .................................................................. 99

6.1 General .................................................................................................................. 99

6.2 Summary ............................................................................................................... 99

6.3 Conclusions ......................................................................................................... 100

6.4 Future Research .................................................................................................. 102

References .................................................................................................................... 106

Appendix A: MATLAB Code for SBEDS Replica (SDOF Solver) ............................... 108

Appendix B: SDOF Solver Verifications (Time-History Plots) ................................. 123

Appendix C: MATLAB Code for More Accurate Model (MDOF Solver) ................... 177

Appendix D: MDOF Solver Verifications (Time-History Plots) ............................... 246

Appendix E: First Data Set Plots ................................................................................ 264

Appendix F: Second Data Set Plots ............................................................................ 300

Appendix G: Third Data Set Plots ............................................................................... 310

Vita ................................................................................................................................ 320

viii
 List of Tables

Table 2.5.1: Various Strength Increase Factors for Steel (US Army, 2008)………………….13

Table 2.6.1.1a: Transformation Factors for Pinned-Pinned Beam (Biggs, 1964)……….…18

Table 2.6.1.1b: Transformation Factors for Fixed-Fixed Beam (Biggs, 1964)………………18

Table 2.6.1.1c: Transformation Factors for Fixed-Pinned Beam (Biggs, 1964)…………….19

Table 5.1.1: Constant Input Parameters…………………………………………………………………….70

Table 5.1.2: Charges and Standoffs for Data Set 1……………………………………………………..71

Table 5.2.3.1: Experimental Blast Data (Nassr, 2012)………………………………………………...85

Table 5.2.3.2: Timoshenko Solution Calibrations to Better Match Experimental

Data....................................................................................................................86

ix
 List of Figures

Figure 2.4.1: Friedlander Equation Plot and Ramp Approximation with Arbitrary

Values…………………………………..………………………………………………………………………..11

Figure 2.4.2: Pressure-Time Plot to Approximate Friedlander Equation (ASCE/SEI 59-11,

2011)…………………………………………..…………………………………………………………………12

Figure 2.5.1: Stress-Strain Relationship for Two Different Materials………………………….14

Figure 2.6.1.1: Transformation from Continuous System to SDOF System (US Army,

2008)…………………………………………………..…………………………………………………………16

Figure 2.6.1.2: Illustration of Stages of Yielding (US Army, 2008)……………………………….20

Figure 2.6.1.3: General Resistance-Deflection Relationship for SBEDS (US Army,

2008)……………………………..………………………………………………………………………………21

Figure 2.7.1: Example of Space Discretization for Two-Member Frame……………………..22

Figure 3.2.5.1: Displacement-Time Plot for Arbitrary SBEDS Analysis (SBEDS, 2008)….37

Figure 3.4.1: Acceleration Comparison of SBEDS with SDOF MATLAB Program………….41

Figure 3.4.1.1a: Resistance-Deflection Plot for W40X655 Fixed-Pinned with Uniform

Load and 0% Damping…………………………………………..……………………………………….43

Figure 3.4.1.1b: Resistance-Time Plot for W40X655 Fixed-Pinned with Uniform Load

and 0% Damping………………………………....………………………………………………………..43

Figure 3.4.1.2a: Resistance-Deflection Plot for W10X12 Fixed-Fixed with Uniform Load

and 0% Damping……………………………….…………………………………………………………..44

Figure 3.4.1.2b: Resistance-Time Plot for W10X12 Fixed-Fixed with Uniform Load and

x
 0% Damping………..…………………………………………………………………………………………45

Figure 3.4.1.2c: Resistance-Deflection Plot for W10X12 Fixed-Fixed with Uniform Load

and 0% Damping (Magnified)……………………………………………..………………………….45

Figure 3.4.1.3a: Deflection-Time Plot for W10X12 Fixed-Pinned with Point Load and 0%

Damping…………………………………….………………………………………………………………….47

Figure 3.4.1.3b: Acceleration-Time Plot for W10X12 Fixed-Pinned with Point Load and

0% Damping……………………………………………..……………………………………………………47

Figure 3.4.1.3c: Acceleration-Time Plot for W10X12 Fixed-Pinned with Point Load and

0% Damping (Magnified)………………………………………………………. ………………………48

Figure 3.4.1.4: Deflection-Time Plot for W10X12 Fixed-Fixed with Uniform Load and 5%

Damping………………………………………….…………………………………………………………….49

Figure 4.2.1.1: Example of Material Definition Window for SAP2000…………………………51

Figure 4.2.1.2: Graphical Representation of a Pinned-Fixed Beam with 0.35 k/in Load

and 4 Elements……………………………………..………………………………………………….……52

Figure 4.2.2.1: Example Plot of Deflection-Time in SAP2000……………………………….……..53

Figure 4.4.1a: Midpoint Deflection Comparison of SAP2000 with MDOF MATLAB

Program………………………..……………………………………………………………………………….57

Figure 4.4.1b: Midpoint Resistance Comparison of SAP2000 with MDOF MATLAB

Program……………………………..………………………………………………………………………….58

Figure 4.4.1.1 Equations for Reduced Plastic Moment Capacity in SBEDS (US Army,

2008)…………………………………………………………..…………………………………………………62

Figure 4.4.1.2a: Resistance-Time Plot for W10X12 Fixed-Pinned with Point Load, 2%

xi
 Damping, and 0.2 Axial Load…………………………………………..……………………………..64

Figure 4.4.1.2b: Deflection-Time Plot for W10X12 Fixed-Pinned with Point Load, 2%

Damping, and 0.2 Axial Load……………………………………………………..…………………..65

Figure 4.4.1.3a: Deflection-Time Comparison for Different Stiffness and Mass

Models………………………………………………………………..…………………………………………67

Figure 4.4.1.3b: Resistance-Time Comparison for Different Stiffness and Mass

Models…………………………………………………..………………………………………………………68

Figure 5.2.1.1: Deviation in K¬L vs. Z for W14X48……………………………………………………...72

Figure 5.2.1.2: Deviation in K¬L vs. Z for W14X257……………………………………….……………73

Figure 5.2.1.3: Deviation in Bernoulli-SDOF Maximum Deflection vs. Z for

W14X257……………………………………………………………………..………………………………..75

Figure 5.2.1.4: Deviation in Timoshenko-SDOF Maximum Deflection vs. Z for

W14X257………………………………………….……………………………………………………………77

Figure 5.2.1.5: Deviation in Timoshenko-Bernoulli Maximum Deflection vs. Z for

W14X257…………………..…………………………………………………………………………………..79

Figure 5.2.2.1: Deviation in Bernoulli-SDOF Maximum Deflection vs. Z for Pinned-

Pinned W14X109 (Fine Z Increment)…………………………………....……………………….81

Figure 5.2.2.2: Deviation in Timoshenko-SDOF Maximum Deflection vs. Z for Pinned-

Pinned W14X109 (Fine Z Increment)…………………………...............…………………….83

Figure 5.2.3.1: Deviation in Bernoulli-SDOF Maximum Deflection vs. Z for Pinned-

Pinned W14X109 (Calibrated Damping)……………………………………………..………….87

Figure 5.2.3.2: Deviation in Timoshenko-SDOF Maximum Deflection vs. Z for Pinned-

xii
 Pinned W14X109 (Calibrated Damping)…………………………………………..…………….89

Figure 5.2.4.1: Deflected Shape at First Yield for Pinned-Pinned W14X109 with Z=12.4

ft⁄lb 1⁄3……………………………….…………………………………………………………………………………….91

Figure 5.2.4.2: Deflected Shape at First Yield for Pinned-Pinned W14X109 with Z=9.1

ft⁄lb 1⁄3 ………………….....…………………………………………………………………………………..91

Figure 5.2.4.3: Deflected Shape at First Yield for Pinned-Pinned W14X109 with Z=7.7

ft⁄lb 1⁄3 …………………………………….....………………………………………………………………..92

Figure 5.2.4.4: Deflected Shape at First Yield for Fixed-Fixed W14X109 with Z=11.0

ft⁄lb 1⁄3 ………………………………….………………………………………………………………………94

Figure 5.2.4.5: Deflected Shape at First Yield for Fixed-Fixed W14X109 with Z=9.5

ft⁄lb 1⁄3 ……………………….....……………………………………………………………………………..95

Figure 5.2.4.6: Deflected Shape at First Yield for Fixed-Fixed W14X109 with Z=5.8

ft⁄lb 1⁄3 ………………..………………………………………………………………………………………..95

Figure 5.2.4.7: Deflected Shape at First Yield for Fixed-Fixed W18X65 with 45 psi and

45 psi-msec……………………………....…………………………………………………………………..97

Figure 5.2.4.8: Deflected Shape at First Yield for Fixed-Fixed W18X65 with 50 psi and

50 psi-msec…………………………………....……………………………………………………………..98

xiii
 Abstract

This thesis presents an investigation into the potential inaccuracies when analyzing

structural elements under blast loads using equivalent single-degree-of-freedom

(SDOF) systems. An equivalent SDOF system converts a continuous element into a

single mass-spring-damper system, where the displacement of the mass equates to

the largest deflection of the continuous element. The potential inaccuracies arise

from the fact that this equivalent system uses the static deflected shape of the

element. The Single-Degree-of-Freedom Blast Effects Design Spreadsheet (SBEDS) is

an example of an industry accepted software package that is used for this type of

simplification.

To identify inaccuracies in results from SDOF methods, the results were compared to a

more accurate model. This model is a multi-degree-of-freedom (MDOF) system, which

is able to better represent a structural element, due to discretizing the element into

smaller segments. These smaller segments are able to respond with a more realistic

deflected shape when subjected to blast loading. This is more accurate than lumping

all of the information about the element to a single degree of freedom.

Using MATLAB, blast analysis of single structural elements were performed to

replicate the corresponding results from two software packages: SBEDS (for SDOF

analysis) and SAP2000 (for MDOF analysis). Once the MATLAB solutions were verified

1
and calibrated, analyses were performed using both MATLAB scripts for combinations

of input parameters, such as boundary conditions, magnitude of load, and section size.

After results were obtained from these analyses, they were compared, which allowed

for identification of static-shape limitations of SDOF analysis.

2
Chapter 1 Introduction

1.1 General

This thesis presents an investigation into inaccuracies within equivalent single-degree-

of-freedom (SDOF) analysis under blast loading. This chapter introduces the topic of

blast design, and outlines the objectives and scope of this thesis.

1.2 Background

Dynamic loading of a structure requires dynamic analysis to obtain accurate values of

deformation and internal stress. Examples of dynamic loads are live loads,

earthquakes, wind, and blasts. The duration of the load distinguishes blast loads from

the other dynamic loads. While these other loads can last seconds or minutes, blast

loads have much shorter durations on the order of milliseconds. Whether this blast

load is a result of an accidental explosion or an intended terrorist act, the load is

always a fast-acting pulse.

SBEDS is used in the structural engineering industry for the “design of structural

components subjected to dynamic loads using single degree of freedom (SDOF)

methodology” (US Army, 2008). This program is often used for the design of these

components under blast loading. By defining inputs such as the time-history of the

blast load, material properties, and geometric features of the structural element in

question, SBEDS performs this SDOF analysis and outputs several results that the

3
Structural Engineer may use to aid in the design of the component. These outputs

include maximum deflection, maximum support rotation, ductility, and maximum

shear force, as well as time-histories of various output variables. Due to the fact that

SBEDS uses equivalent SDOF methods, the analysis requires little computing time,

which is convenient for the industry.

1.3 Research Objectives

Yokoyama (2014) has shown that for large blasts at close distance, discrepancies arise

between SBEDS solutions and solutions from more accurate models. Scaled distance

(Z) is a common measure used to define the strength of a blast. The conventional Z

cutoff for which SDOF analysis is considered to be no longer applicable is below 1.2

𝑚⁄𝑘𝑔1⁄3 (3𝑓𝑡⁄𝑙𝑏 1⁄3 ). However, Yokoyama has shown cases where applying a scaled

distance above the cutoff produces significantly different results between SBEDS and

more accurate models. This can be a major issue for Structural Engineers who utilize

SBEDS for blast design without knowing the boundaries for which the results from

SBEDS become unacceptably inaccurate.

The research presented in this thesis investigates and establishes more specific cutoffs

for when SDOF analysis becomes too inaccurate. The objectives of this research are

listed below:

4
 (1) Write MATLAB code that replicates the SBEDS solution (Single-Degree-

of-Freedom (SDOF) solver)

(2) Write MATLAB code that uses a more accurate model (Multi-Degree-of-

Freedom (MDOF) solver)

(3) Run both MATLAB programs for various input parameters, and compare

solutions obtained from both solution methods.

(4) Identify the limitations of using single-degree-of-freedom analysis when

analyzing structural components under blast loading based for various

input parameters.

Objective (1) will be validated against like analyses in SBEDS. Objective (2) will be

validated against like analyses in SAP2000, until the limitations of SAP2000 are

exceeded. Objective (3) will be completed by obtaining percent variation on the

outputs for each case of inputs. For Objective (4), a maximum acceptable percent

variation must be established to determine these region boundaries.

1.4 Scope of Thesis

Five additional chapters follow this chapter. Chapters 2, 3, and 4 discuss background

information, modeling techniques, and the creation of both MATLAB solvers. Chapter

5 presents the comparisons of the data frames for each solver. Finally, Chapter 6

provides a summary of the research, identifies boundaries of SDOF analysis limits, and

concludes the paper with future research to be explored on this topic.

5
Chapter 2 Background and Modeling

2.1 General

Modeling refers to the creation of a simulation that replicates an event which occurs in

reality. Modeling is a very complex topic in structural engineering, as well as in other

disciplines. The complexity arises from the vast amount of choices, approximations, or

assumptions that the modeler may make when creating this model. Some options are

better suited for certain events than others. Also, certain combinations of these

choices may interact to cause very different outcomes. The user must judge each

modeling choice and decide which options will best represent a realistic result.

Another problem the modeler faces is creating a model that is not only reasonably

accurate, but efficient as well. Certain modeling options will be chosen that are

simpler than others, in order to decrease computational runtimes. Having the most

sophisticated model, which is most accurate, may be unrealistic for some applications

due to the limited processing power of the modern computer.

For this paper, the events of interest are the behaviors of structural elements (beams

and beam-columns) subjected to blast loading. Models of this type implement various

techniques from structural dynamics, which is a combination of structural analysis and

dynamics.

6
This chapter discusses the modeling techniques for structural dynamics, and how they

specifically relate to structural elements subjected to blast loads. Sections 2.2 through

2.6 and section 2.9 pertain to all models. Section 2.6.1 pertains to SDOF models only.

Sections 2.7, 2.8, and 2.10 through 2.12 pertain to MDOF models only.

2.2 Numerical Solutions

The analysis of any structural dynamics problem starts with the equation of motion,

which is shown below in Equation 2.2.1:

𝑚𝑢̈ + 𝑐𝑢̇ + 𝑘𝑢 = 𝑓(𝑡) (2.2.1)

where:

𝑚=mass

𝑐=damping coefficient

𝑘=stiffness coefficient

𝑓(𝑡)=applied force

𝑢̈ =acceleration

𝑢̇ =velocity

𝑢=displacement

Displacement is the unknown variable, and velocity and acceleration are time

derivatives of displacement. A closed form solution is available for this second-order

7
differential equation, depending on the form of the applied force. To be able to solve

this equation for any form of applied force, numerical methods are often used.

Numerical methods take initial conditions, and extrapolate values incrementally, to

obtain approximate solutions. Depending on the increments, the solutions may have

little error compared to their closed form solutions.

Several numerical methods are available to solve a second-order differential equation.

Some examples are the constant-velocity procedure (Biggs, 1964), the central

difference method (Chopra, 1995), and Newmark’s method (Chopra, 1995). Each

method approaches the problem slightly differently, but they all produce consistent

results assuming an appropriate time step was chosen. An important parameter that

relates to the accuracy and stability of each solution method is the natural period, 𝑇,

shown below in Equation 2.2.2:

𝑚
𝑇 = 2𝜋√ 𝑘 (2.2.2)

For the constant-velocity procedure, Biggs states that a time increment should not

exceed one-tenth of the natural period for accurate results. He also states that the

time increment should be small enough to accurately capture the load. This is

especially important for blast loads, where the total duration of the load is small. For

the central difference method, Chopra states that the solution will only be stable

8
(converge) if the ratio of time step to natural period of vibration is below a certain

value. This value is displayed below in Equation 2.2.3a. He also states the same

accuracy limit that Biggs states, which a smaller increment than the stability limit is.

Newmark’s method is more complex, and involves two additional constants (β, γ),

which control the acceleration within each time step. The stability limit for this

solution method is also shown below in Equation 2.2.3b:

∆𝑡 1
<𝜋 (2.2.3a)
𝑇

∆t 1 1
≤ π√2 ∙ (2.2.3b)
T √γ−2β

SBEDS uses the constant-velocity procedure to solve the equation of motion. The

MATLAB SDOF solver will therefore implement this procedure as well. The MATLAB

MDOF solver will implement Newmark’s method for its versatility and unconditional

stability for certain values of β and γ.

These methods are either classified as explicit or implicit methods. Implicit methods

iterate within each time step. They calculate the internal and external forces on the

structure and drive the unbalanced load (the difference of these forces) below a

certain tolerance by adjusting the displacements, velocities, and accelerations, before

proceeding to the next time step (Newton-Raphson iteration). Explicit methods

neglect this unbalanced force and proceed through the time stepping without

9
analyzing unbalanced forces. The constant-velocity procedure is an explicit solver,

while Newmark’s method with Newton-Raphson iteration is an implicit solver.

2.3 Time Discretization

Discretization refers to the subdivision of a continuous domain into distinct quantities.

Discretization is how numerical solutions can be achieved. For the equation of motion,

time is the continuous domain that is discretized as time steps. As explained above, it

is important to pick a time step that satisfies stability and accuracy limits for the

numerical method chosen, as well as captures the load accurately.

2.4 Loads

Blast loads are very complex loads to model. How to characterize blast loads as

applied to structures is not in the scope of this thesis. However, the basic concepts of

determining a load function are described below. Blasts create fast-acting shock

waves. These shock waves create positive pressure, followed by negative pressure on

any mass in its direction of travel. Friedlander (1946) fit a curve to the pressure of

blast wave over time. The Friedlander Equation is shown below:

−𝑡
𝑡
𝑃 = 𝑃𝑠 ∙ 𝑒 𝑡∗ ∙ (1 − 𝑡 ∗) (2.4.1)

where:

10
𝑃=pressure at time t

𝑃𝑠 =peak pressure

𝑡=time

𝑡 ∗ =time of positive duration of pressure

Figure 2.4.1 below shows a graph of this equation with arbitrarily selected values as

well as a ramp approximation to the curve, which is explained after:

Figure 2.4.1: Friedlander Equation Plot and Ramp Approximation with Arbitrary Values

ASCE/SEI 59-11 (2011) describes a procedure for converting a weight of explosive

material into a pressure-time graph that approximates the Friedlander relationship.

11
This graph becomes the applied load to the element by converting the pressure into a

load. Figure 2.4.2 below shows the diagram depicted in the actual manual:

Figure 2.4.2: Pressure-Time Plot to Approximate Friedlander Equation (ASCE/SEI 59-11, 2011)

Often in practice, the negative portion of the graph is neglected, as a conservative

simplification of the analysis since the negative phase will partially counteract the

effects of the initial positive pressure in most cases. The scope of this thesis is not to

investigate the load itself. Therefore, the loads to be applied to the structural

elements will follow what is most commonly used in practice. This approach consists

of obtaining a value for positive peak pressure and positive impulse from a charge (lb

of TNT) and standoff (ft from structure). With the peak pressure and impulse, a simple

triangular ramp-down time-history is created, with positive pressure only. This

12
pressure is applied to the structural element using its tributary area. Figure 2.4.1 from

before shows this simplified approach.

2.5 Material Properties and Plasticity

Materials can exhibit different properties under dynamic loading, compared to static

loading. Therefore, increase factors can be applied to the yield strengths of the

materials in dynamic analyses. The Methodology Manual for the Single-Degree-of-

Freedom Blast Effects Design Spreadsheets (SBEDS) summarizes these different

increase factors, as shown in Table 2.5.1 below:

Table 2.5.1: Various Strength Increase Factors for Steel (US Army, 2008)

Modeling plastic behavior can be complicated and is material dependent. For

example, hot-rolled steel exhibits a linear stress-strain relationship until its yield point,

at which it gains no increase in stress with additional strain. At a certain strain value, it

then exhibits strain hardening, and has an increase in stress with additional strain.

13
This strain-hardening portion of the curve is strain-rate dependent, which causes much

complexity to modeling this behavior. Reinforced concrete has a less distinct linear

stress-strain relationship, and has no distinct point of yielding. The stress-strain curve

continues to decrease in slope until a maximum stress is observed. Because of the

complexity and variation in stress-strain behavior, assumptions are usually made

about the stress-strain curve that do not introduce significant error into the solution,

and at the same time, make computations significantly simpler. A crude

representation of the stress-strain behaviors of the two materials explained is shown

below in Figure 2.5.1:

Figure 2.5.1: Stress-Strain Relationship for Two Different Materials

14
Blast loads often have enough energy in such a short duration to cause yielding in the

structural elements they are applied to. Due to these conditions, plasticity must be

considered in blast analysis.

For this thesis, only hot-rolled steel is considered. Hardening of the material is

neglected to simplify the analysis without risking much inaccuracy. Therefore, the

stress-strain relationship used in both MATLAB solvers follows an elastic-perfectly

plastic relationship. For the analysis, once the internal moment reaches the plastic

moment, that location is considered fully plastic until rebounding (change in sign of

velocity). After rebounding, the section becomes elastic again, but with residual

displacement. Yielding within the cross-section is neglected. Additionally, all types

and modes of buckling are not considered. All elements considered are assumed to be

non-slender and not susceptible to lateral torsional buckling. Only compact sections

are considered to avoid local buckling considerations as well. Buckling limit states are

outside the scope of this thesis.

2.6 Boundary Conditions

In reality, boundary conditions are never fully rigid nor fully rotatable. Connections

will have partial restraint of rotation. However, for simplicity, connections are often

modeled as either fully moment-connected, or fully pinned. For this thesis, the

boundary conditions considered are fully fixed, fully pinned, and free (no constraint).

15
2.6.1 Transformation Factors (For SDOF Analysis)

SDOF analysis is computationally convenient because the quantities in the equation of

motion are scalars. For MDOF analysis, these quantities becomes matrices, which

requires matrix computations. For this reason, SBEDS makes use of certain

transformation factors that convert beams (MDOF systems with potentially infinite

DOFs) to SDOF systems. Figure 2.6.1.1 below shows this transformation pictorially:

Figure 2.6.1.1: Transformation from Continuous System to SDOF System (US Army, 2008)

As explained in Methodology Manual for the Single-Degree-of-Freedom Blast Effects

Design Spreadsheets (SBEDS), “[a]n equivalent SDOF system is defined as a system that

has the same energy in terms of work energy, strain energy, and kinetic energy, as the

blast-loaded component responding in a given, assumed mode shape.” The

mathematical derivations of these definitions are listed in the following equations:

𝐿
𝑀𝑒 ∫0 𝑚(𝑥)[∅(𝑥)]2 𝑑𝑥
𝐾𝑀 = = 𝐿 (2.6.1.1)
𝑀𝑡 ∫0 𝑚(𝑥)𝑑𝑥

16
where:

𝐾𝑀 =mass factor

𝑀𝑒 =equivalent SDOF mass

𝑀𝑡 =total mass of MDOF element

𝑚(𝑥)=distributed mass along MDOF element

∅(𝑥)=static deflected shape (unit normalized)

𝐿=length of element

𝐿
𝐹𝑒 (𝑡) ∫0 𝑝(𝑥,𝑡)∅(𝑥)𝑑𝑥
𝐾𝐿 = = 𝐿 (2.6.1.2)
𝐹𝑡 (𝑡) ∫0 𝑝(𝑥,𝑡)𝑑𝑥

where:

𝐾𝐿 =load factor

𝐹𝑒 (𝑡)=equivalent SDOF load

𝐹𝑡 (𝑡)=total load on MDOF element

𝑝(𝑥, 𝑡)=distributed load along MDOF element

Using the above equations, these factors are calculated. The mass of the system gets

multiplied by the mass factor, and the load on the system and the resistance of the

system get multiplied by the load factor. These factors vary for each static deflected

shape. Therefore, they change based on the boundary condition, the shape of loading,

and the state of plasticity. By dividing the mass factor by the load factor, a single

17
“load-mass” factor can be used as well. Biggs originally derived these factors for

various conditions, which are shown below in Tables 2.6.1.1a, 2.6.1.1b, and 2.6.1.1c:

Table 2.6.1.1a: Transformation Factors for Pinned-Pinned Beam (Biggs, 1964)

Table 2.6.1.1b: Transformation Factors for Fixed-Fixed Beam (Biggs, 1964)

18
 Table 2.6.1.1c: Transformation Factors for Fixed-Pinned Beam (Biggs, 1964)

Displacement is calculated at each time step of the solution. From the original beam,

a value of displacement can be solved for, corresponding to first yielding of the

section. If the displacement of the equivalent system reaches this value of

displacement, the transformation factors change from the elastic factors to the elastic-

plastic factors or the plastic factors, depending on which boundary condition is used.

After rebound, the factors go back to the elastic factors and the tracking restarts in the

opposite direction.

Certain boundary conditions have two sets of transformation factors, while other

boundary conditions have three sets. This is because depending on the boundary

condition, the beam can develop several locations of sequential yielding before

19
becoming a collapse mechanism (not able to resist any further increase in loading).

For example, pinned-pinned beams yield in the center and then become a collapse

mechanism. Fixed-fixed beams first yield at the supports simultaneously, and the

beam essentially becomes a pinned-pinned beam, until it yields again at the center,

forming the collapse mechanism. An illustration of this process is shown below in

Figure 2.6.1.2:

Figure 2.6.1.2: Illustration of Stages of Yielding (US Army, 2008)

SBEDS makes use of these factors in its analysis. By tracking the displacements at

which yielding occurs, SBEDS creates a load-displacement graph that tracks how much

resistance the system can have at any given displacement. Figure 2.6.1.3 on the

following page shows the general resistance-deflection relationship that SBEDS uses.

This resistance-deflection relationship is equivalent to the stress-strain relationship

mentioned earlier, and takes into account the varying transformation factors at

different points of yielding.

20
 Figure 2.6.1.3: General Resistance-Deflection Relationship for SBEDS (US Army, 2008)

2.7 Space Discretization

For SDOF analysis, the mass of the system is located at a point. There is no need to

discretize space, because the mass is all located at the same point in space. However,

for MDOF analysis, the mass of the system for a structural element is distributed along

the element itself. The solution to the displacement of each point along this element

can be solved for using continuum analysis. However, this is complex and not

computationally efficient. It is therefore necessary to discretize space along the

element, so that mass can be lumped at specific points (nodes) along the element.

This allows matrices to be assembled that represent the mass, stiffness, damping, and

applied loads for the element. These matrices can then be used in a matrix

21
formulation of any numerical method described above. An example of space

discretization for a two-member frame is shown below in Figure 2.7.1.

Figure 2.7.1: Example of Space Discretization for Two-Member Frame

The two continuous elements of lengths 3 and 2 meters are discretized into multiple

elements of length 0.1 meters. The mass (and any applied loads) of the continuous

elements becomes lumped at each node using tributary widths. Each discretized

element has its own stiffness and damping coefficients as well, which relate to the

overall stiffness and damping of the structure.

22
2.8 Element Type

In MDOF analysis, an element type must be selected. Different types of elements are

capable of exhibiting different types of response behavior. For example, the Euler-

Bernoulli Beam only considers deformation due to bending, while the Timoshenko

Beam considers deformation due to both bending and shear. It is important to know

which type of element is appropriate to use. If the element in question is slender,

Euler-Bernoulli Beams give good approximations of behavior. However, for short

elements, shear deformation is significant, and Timoshenko Beams must be used to

obtain more accurate results. In “A First Course in the Finite Element Method” (Logan,

2012), stiffness matrices for both types of elements are derived. Below displays the

elemental elastic stiffness matrix of a Timoshenko Beam:

12 6𝐿 −12 6𝐿
2
𝐸𝐼 6𝐿 (4 + 𝑟)𝐿 −6𝐿 (2 − 𝑟)𝐿2
𝒌𝑒 = 𝐿3 (1+𝑟) ( ) (2.8.1)
−12𝐿 −6𝐿 12 −6𝐿
6𝐿 (2 − 𝑟)𝐿2 −6𝐿 (4 + 𝑟)𝐿2

where:

𝒌𝑒 =elemental elastic stiffness matrix

𝐸=elastic modulus of element

𝐼=moment of inertia of element

𝐿=length of element

12𝐸𝐼
𝑟=
𝐴𝑠 𝐺𝐿2

23
𝐴𝑠 =shear area of element

𝐺=shear modulus of element

By substituting zero for r, this stiffness matrix reduces to the Euler-Bernoulli elastic

stiffness matrix.

After a system is discretized in space, elemental stiffness matrices are calculated for

each discretized element. These elemental stiffness matrices are then assembled into

an overall global stiffness matrix, by utilizing the connectivity of the structure.

2.9 Geometric Effects

When axial load is applied to a beam, it becomes classified as a beam-column, because

it resists axial load like a column as well as transverse load like a beam. Axial load has

an effect on the stiffness of the element. In general, tension in a member will increase

its stiffness, while compression will decrease its stiffness. Different methods may be

implemented to account for this decrease in stiffness in SDOF analysis. One method,

which is implemented by SBEDS, applies additional transverse load to the element for

the following time step, based on the static axial load on the member as well as the

deflection at the current time step. Here, the stiffness is not directly manipulated, but

the end result is the same: a larger deflection.

24
For MDOF analysis, a geometric stiffness matrix can be used to account for this change

in stiffness. This geometric stiffness matrix is added to the elastic stiffness matrix,

either before assembly as element matrices, or after assembly as a global matrix. In

“Structural Element Stiffness, Mass, and Damping Matrices” Gavin (2014) derives

geometric stiffness matrices for both Euler-Bernoulli and Timoshenko elements. By

substituting zero into the shear factor, the Timoshenko geometric stiffness matrix

reduces to the Euler-Bernoulli matrix. Below is the elemental geometric stiffness

matrix for Timoshenko Beam:

(2.9.1)

where:

𝑘𝑔=elemental geometric stiffness matrix

𝑃=applied axial load (tension positive)

𝐿=length of element

𝑟 defined earlier in Equation 2.8.1

25
2.10 Mass Matrix

In MDOF analysis, mass matrices can be modeled differently as well. A simple way to

model mass matrices is to lump half of the elements mass at each node (assuming a

two-node element), thus creating a diagonal mass matrix. A more comprehensive

approach involves the use of shape functions, which come from the Finite Element

Method (not covered in this thesis). “A First Course in the Finite Element Method”

(Logan, 2012) derives mass matrices using these shape functions. This “consistent-

mass matrix” is shown below:

156 22𝐿 54 −13𝐿

𝒎𝑒 = 420 ( 22𝐿
𝜌𝐴𝐿 4𝐿2 13𝐿 −3𝐿2 ) (2.10.1)
54 13𝐿 156 −22𝐿
−13𝐿 −3𝐿2 −22𝐿 4𝐿2

where:

𝒎𝑒 =elemental mass matrix

𝜌=density of element

𝐴=area of element

𝐿=length of element

Przemieniecki (1968) derived the consistent-mass matrix using Timoshenko theory.

Substituting zero for the shear factor in the Timoshenko mass matrix reduces the

matrix to the Euler-Bernoulli mass matrix. The Timoshenko mass matrix requires

26
significantly more space to display than the Euler-Bernoulli consistent-mass matrix,

and is therefore not shown.

After a system is discretized in space, elemental mass matrices are calculated for each

discretized element. These elemental mass matrices are then assembled into an

overall global mass matrix, by utilizing the connectivity of the structure. This process is

identical to the process for assembling the global stiffness matrix.

2.11 Mode Shapes

In MDOF analysis, once the global matrices are assembled, there are two options for

solving the displacements using numerical methods. The first option is to directly

input the matrices into Equation 2.2.1 from earlier, and then use the matrix

formulation of whichever numerical method is chosen. This involves matrix

computations.

The second option is to use transformations to create another set of matrices, called

modal matrices. These matrices are diagonal, which allows Equation 2.2.1 to be

decoupled. This means that instead of having one N by x matrix equation (N is number

of DOFs), there are instead N scalar equations. Each of these scalar equations can be

solved independently using any chosen numerical method. The solution to these

equations are modal displacements, which then get transformed back into physical

displacements.

27
To obtain the modal stiffness and mass matrices, a matrix eigenvalue problem must be

solved, which is shown below in Equation 2.11.1:

𝒌∅𝑛 = 𝜔𝑛2 𝒎∅𝑛 (2.11.1)

where:

𝒌=global stiffness matrix

𝒎=global mass matrix

∅𝑛 =mode shape for mode n of N

𝜔𝑛 =natural frequency of vibration for mode n of N

Once the mode shapes are determined, they can be concatenated into a matrix, Φ.

The global stiffness and mass matrices can then be pre-multiplied and post-multiplied

by Φ to create the modal stiffness and modal mass matrices, which are diagonal as

stated previously. These new matrices can be substituted into Equation 2.2.1, with the

unknown quantities now being modal displacements (𝑞𝑛 ), modal velocities (𝑞̇ 𝑛 ), and

modal accelerations (𝑞̈ 𝑛 ). These form decoupled equations, since the modal matrices

are diagonal. Therefore, the modal displacements can be solved for, for each mode

individually. Once these are determined, a transformation brings the modal

displacements into the physical displacements. The same equation applies to

velocities and accelerations. Equation 2.11.2 shows this transformation:

28
 𝒖(𝒕) = ∑𝑁
𝑟=1 ∅𝑟 𝑞𝑟 (𝑡) (2.11.2)

where:

𝒖(𝒕)=vector of physical displacements

This second method has two benefits. The first benefit is the fact that the differential

equations become decoupled and are simpler to solve independently. The second

benefit is that higher modes can be neglected in the transformation back to physical

coordinates. These higher modes have large corresponding natural frequencies of

vibration. This means the periods of vibration for these higher modes are very short,

which can lead to inaccuracy and instability in the numerical solution, as described in

Chapter 2.2. By neglecting these higher modes, the accuracy and stability limits

become more feasible to satisfy when choosing a time step.

However, Chopra (1995) states that, “[c]lassical modal analysis is […] not applicable to

the analysis of nonlinear systems.” Therefore, a different formulation will be

explained in the following section.

2.12 Damping

Damping refers to the loss of energy due to friction, which causes a decrease in the

response over time. Damping can be strain-rate dependent, which can make it very

29
complex to model. In SDOF analysis, damping is modeled using the damping

coefficient, c, in Equation 2.2.1 from earlier. SBEDs allows users to input a damping

ratio, which is used to calculate c for the analysis. In Methodology Manual for the

Single-Degree-of-Freedom Blast Effects Design Spreadsheets (SBEDS), reference is

made to TM 5-1300, which recommends using damping ratios of 5% for steel, and 1%

for reinforced concrete. Below is another excerpt from the methodology manual:

“UFC 3-340-01 recommends that damping values commonly used for

earthquake analysis and design […] can be used for elastic and elastoplastic
component response only. Therefore, the damping term is set to zero for all
plastic response. This is recommended because the damping forces for
structure response to earthquake loading were developed primarily for elastic
component response, which involves deformation over the whole volume of
the component. After ultimate resistance, deformation consists primarily of
localized hinging at maximum moment regions, which occurs over a much
smaller volume, so that no significant additional damping occurs” (Us Army,
2008).

In MDOF analysis, there are many modeling options for damping, which vary in

complexity. Chopra (1995) presents three main options: Caughey Damping,

superposition of modal damping matrices, and nonclassical damping. The first two

options fall under the category of classical damping.

The formulation of the natural modes explained in the previous section are

constructed in a way so that the mass and stiffness matrices are orthogonal to each

mode. This means that pre-multiplying either matrix by the transpose of any mode

30
shape, and post-multiplying the result by the same mode shape will result in 0. For a

damping matrix to be classically damped, it must also satisfy this mode orthogonality.

In the first method, Caughey Damping, damping ratios for the first J modes can be

specified. The rest of the damping ratios that are not specified are automatically

calculated. Equations 2.12.1a and 2.12.1b below are used to calculate this type of

damping matrix:

𝒄 = 𝒎 ∑𝐽−1 −1 𝑙
𝑙=0 𝑎𝑙 [𝒎 𝒌] (2.12.1a)

1
𝜁𝑛 = 2 ∑𝐽−1 2𝑙−𝑙
𝑙=0 𝑎𝑙 𝜔𝑛 (2.12.1b)

where:

𝒄=global damping matrix

𝜁𝑛 =damping ratio for mode n

𝐽=number of modes to specify damping ratios for

First, damping ratios are specified for J modes. Then, 𝑎𝑙 constants are solved for using

Equation 2.12.1b. Once these constants are determined, the global damping matrix

can be determined. A special case of this is called Rayleigh Damping, where only two

modes have determined damping ratios (J=2).

31
The second method, superposition of modal damping matrices, is very similar to

Caughey Damping. Damping ratios are specified for any number of modes. However,

instead of the modes not specified being automatically calculated, there is no damping

in these modes. Equation 2.12.2 below is used to calculate the classical damping

matrix using superposition of modal damping matrices:

2𝜁𝑛 𝜔𝑛
𝒄 = 𝒎 (∑𝐽𝑛=1 𝜙𝑛 𝜙𝑛𝑇 ) 𝒎 (2.12.2)
𝑀𝑛

where:

𝑀𝑛 =modal mass for mode n (nth diagonal term on modal mass matrix)

The third method, non-classical damping, involves modifying the eigenvalue

formulation for determining mode shapes. In the previous section, the eigenvalue

problem was formulated by neglecting damping. By including damping, the matrix

eigenvalue problem becomes the quadratic eigenvalue problem. Equation 2.12.3

below represents this formulation:

(𝜆2𝑛 𝒎 + 𝜆𝑛 𝒄 + 𝒌)𝜓𝑛 = 𝟎 (2.12.3)

where:

𝜆𝑛 =eigenvalue of problem

32
𝜓𝑛 =eigenvector of problem

𝜆𝑛 can be thought of as equivalent to 𝜔𝑛2 in the matrix eigenvalue problem and 𝜓𝑛 can

be thought of as equivalent to 𝜙𝑛 in the matrix eigenvalue problem. The solution to

this equation yields complex numbers.

Under nonlinear dynamic analysis, damping becomes even more complex. Once an

element yields, with the sudden reduced stiffness, the velocity is likely to increase

suddenly. Since viscous damping is proportional to the velocity, there would be

increased damping at this point of yielding, which is not realistic. To mitigate this

issue, many solvers implement a reduction in damping during yielding. An example of

this is to reduce the elemental damping matrices that correspond to any element that

yields by a factor of 10−3. This reduction would make the damping force associated

with the increased velocity insignificant enough not to overdamp the system during

yielding.

33
Chapter 3 Single-Degree-of-Freedom Solver

3.1 General

This chapter discusses all aspects of the SDOF analysis. This includes the modeling

techniques that SBEDS implements, how SBEDS was replicated using MATLAB, and

some minor inconsistencies between theory and output.

3.2 SBEDS

SBEDS is a Microsoft Excel-based spreadsheet used to design structural elements

under dynamic loading. It is commonly used for blast design. There are several sheets

to the file, which are explained in the following sections.

3.2.1 Intro Tab

The Intro tab allows the user to select a type of component used for the analysis, such

as a steal beam, or a reinforced concrete beam. In this tab, the user can also choose

between English and Metric units. For this thesis, English units will be used, as this is

the standard in practice. The Intro tab also lists instructions and comments about the

program.

3.2.2 Input Tab

The Input tab allows users to input several parameters for the analysis. Specifically,

yellow cells specify user input cells. Element size, braced length, boundary conditions,

34
supported weight, axial load, lateral load functions, time step, damping ratio, and

incident angle are some input parameters in this tab. A recommended time step is

displayed here. This recommended time step is the smallest of several values, which

are listed below (US Army, 2008):

• 10% of the natural period

• 10% of the smallest time increment in a manually input, point-wise l
inear blast load
• 3% of the equivalent triangular positive phase duration or 1.5% of the
equivalent triangular negative phase duration of an input charge
weight-standoff blast load
• 3% of the smallest calculated time between local maxima and minima
points of an input blast load file
• eight natural periods divided by 2900 time steps

This tab also displays the resistance-deflection relationship for the chosen parameters,

as well as yield displacements and other important parameters. Once the input

parameters are entered, a green button labelled “RUN SDOF” runs the analysis.

Results of this analysis include maximum values of displacement, rotation, and

ductility.

3.2.3 Results Tab

This tab displays several plots once the analysis is run. These plots include

displacement-time, load-time, resistance-time, resistance-displacement, and dynamic

shear-time.

35
3.2.4 P-I_Diagram Tab

This tab allows the user to create a pressure-impulse diagram for the selected

element.

3.2.5 SDOF Output Tab

This tab shows the tabulated data for the analysis. These values are the plots in the

Results tab. This output tab displays columns of output, such as deflection, velocity,

acceleration, resistance, and the load-mass factor for each time step. SBEDS uses the

constant-velocity procedure to calculate the output, as explained in section 2.2 earlier.

This method solves the displacement, velocity, and acceleration at each time step.

When the value of the displacement reaches a yield value (calculated in the Input tab),

the load-mass factor changes to the appropriate factor, and the time stepping

continues. On rebound, the load-mass factor changes back to the elastic value and the

process starts over on rebound. The program calculates these values for a total of

2900 time steps, no matter what size time step is chosen. Figure 3.2.5.1 on the

following page shows the displacement-time plot for an arbitrary analysis:

36
 Figure 3.2.5.1: Displacement-Time Plot for Arbitrary SBEDS Analysis (SBEDS, 2008)

3.2.6 Database Tab

The Database tab has various constants that the analysis uses, depending on what the

user chooses as inputs. These constants include the Biggs transformation factors for

each condition, yield stresses and elastic moduli for certain materials, and increase

factors for yield values.

3.2.7 Beam_Types Tab

This tab is similar to the Database tab. In this tab, properties of every cross-section

type are listed for the analysis to reference. For every section type (Wide Flange, Tube

Section, etc.), properties are listed such as weight, plastic section modulus, and

moment of inertia. These are all needed for the analysis.

3.3 MATLAB Replica

The first step in quantifying the inaccuracies in SBEDS is to create a program that takes

the same inputs as SBEDS, performs an equivalent SDOF analysis, and returns the

37
same outputs as SBEDS. The reason this has to be done with programming is because

this program will be used several times for ranges of different input parameters, in

order to create data frames. These data frames will be compared to data frames

created from the MDOF solver. MATLAB was used to write this program.

To keep the program organized like SBEDS is, several function files were created that

have different features. These functions are equivalent to the different tabs that

SBEDS has. For the full set of program files, see Appendix A: MATLAB Code for SBEDS

Replica (SDOF Solver).

3.3.1 Main File

The first program file, called “main.m,” is the main part of the program. From here,

variables are passed to and from all other functions for analysis. This main file starts

with declarations of all inputs. This is equivalent to the Intro and Input tabs from

SBEDS. All of the input parameters are then converted to a consistent set of units so

that all further calculations are in consistent units. The rest of the main file consists of

functions that transfer variables and perform calculations.

3.3.2 Load File

The first function, called “Load.m,” creates the loading function. Four inputs are taken

from the main file (initial applied pressure [psi], length of pressure [ms], time step

38
[ms], and number of steps). This load function creates a triangular load function and

passes the value of the load at each time step back to the main file.

3.3.3 Biggs File

The next function, called “Biggs.m,” specifies which transformation factors will be used

for the equivalent SDOF analysis. This function takes two inputs from the main file

(load type [point load or uniform load], and boundary conditions [pinned-pinned,

fixed-fixed, fixed-pinned, or cantilevered]). From these two inputs, all the

transformation factors are defined, and passed back to the main file.

3.3.4 Const_vel File

The next function, called “const_vel,” performs the bulk of the calculations. Here, the

constant-velocity procedure is implemented. Initial conditions on the displacement

and velocity, the Biggs transformation factors, mass, damping ratio, stiffness, the

loading function, time step, and number of steps are all passed to this function. For

each time step, the value of the resistance is compared to the values of resistance that

would cause yielding, and therefore would cause the transformation factors to

change. Once the value of resistance reaches 95% of maximum resistance, the

damping ratio changes to zero until rebound. SBEDS implements this modeling option

for damping, as explained earlier.

39
For each time step, every value of time, loading, displacement, velocity, acceleration,

stiffness, resistance, load-mass factor, and internal shear are stored into a matrix

called “alldata.” This is the same format as the SDOF Output tab in SBEDS, so that it is

easy to compare the output of SBEDS to the output of this MATLAB program. All of

these calculated variables are then sent back to the main file, where final post-

processing is performed, at which point plots are created of the outputs.

3.4 Solver Verification

To ensure that this MATLAB program functions identically to SBEDS, several sets of

inputs were analyzed with both programs and compared. The element chosen for

comparison was a W10X12 Grade 50 steel beam, with length of 20 ft and tributary

width of 0.5 ft. The duration of the pressure load was 17.8 ms, which is a common

duration length in practice. Every boundary condition and load type were tested, with

damping ratios of 0%, 5%, and 20%. The initial pressure load was also varied for each

combination to get the maximum displacement of the system to be in each region of

plasticity.

Once all of these cases were verified, other parameters were changed as well. This

included changing the duration of the pressure load to 8 ms, adding additional weight

(30 psf, which converted to additional mass) to the system, and changing the beam

shape to a W40X655.

40
An example of how the two analyses were compared is shown below in Figure 3.4.1:

Acceleration Comparison for W10X12

Fixed-Pinned Point Load with 20%
damping, elasto-plastic region
0.14

0.12
MATLAB ACC
0.1
SBEDS ACC
0.08
Acceleration (in/ms^2)

0.06

0.04

0.02

0
0 20 40 60 80 100 120
-0.02

-0.04

-0.06

-0.08
TIme (ms)

Figure 3.4.1: Acceleration Comparison of SBEDS with SDOF MATLAB Program

The data from the SDOF Output tab from SBEDS was pasted into a new Excel

document. The data from the “alldata” matrix was also pasted into this Excel

document. Plots were created for the displacement, velocity, acceleration, resistance,

and internal shear force for each set parameter variations. Since the plots of both sets

of outputs overlapped, the MATALB replica was verified as being identical to SBEDS.

41
For the full set of verifications, see Appendix B: SDOF Solver Verifications (Time History

Plots).

3.4.1 Inconsistencies

While comparing each set of parameters to SBEDS, an inconsistency was found

between the theory stated in the SBEDS Manual and the SBEDS output. Additionally,

some minor inconsistencies between SBEDS and the MATLAB replica were displayed in

the plots, which are the results of certain modeling choices.

The inconsistency between SBEDS theory and output was in the resistance-deflection

relationship. Figure 2.6.1.3 from earlier shows this relationship that SBEDS apparently

always follows. The results from the W40X655 element, however, showed a different

resistance-deflection relationship. Figures 3.4.1.1a and 3.4.1.1b below shows the

resistance-deflection output for this W40X655 element, which underwent a load large

enough to create a fully plastic resistance:

42
 Resistance-Deflection
14000
12000
10000
8000
Resistance (psi)

6000
4000
MATLAB Resistance
2000
SBEDS Resistance
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
-2000
-4000
Deflection (in)

Figure 3.4.1.1a: Resistance-Deflection Plot for W40X655 Fixed-Pinned with Uniform Load and 0% Damping

Resistance-Time for W40X655

14000
12000
10000
MATLAB RESISTANCE
8000
Resistance (psi)

SBEDs RESISTANCE
6000
4000
2000
0
0 5 10 15 20 25
-2000
-4000
Time (ms)

Figure 3.4.1.1b: Resistance-Time Plot for W40X655 Fixed-Pinned with Uniform Load and 0% Damping

43
From Figure 3.4.1.1a, it can be seen that the first yield transition occurs at a resistance

of approximately 8000 psi. Once the mass goes through rebound, and then inbound

again, the resistance reaches this value again. This can be seen in Figure 3.4.1.1b,

since there is no overlap. However, the resistance function does not change slopes

here (the stiffness stays the same). In other cases, the resistance-deflection

relationship is consistent with the SBEDS theory. Figures 3.4.1.2a, 3.4.1.2b, and

3.4.1.2c below show this with the W10X12 element, which also underwent a load large

enough to create a fully plastic resistance:

Resistance-Deflection for W10X12

80

60

40
Resistance (psi)

20

0
0 1 2 3 4 5 6 7 8
-20
MATLAB RESISTANCE

-40 SBEDS RESISTANCE

-60
Deflection (in)

Figure 3.4.1.2a: Resistance-Deflection Plot for W10X12 Fixed-Fixed with Uniform Load and 0% Damping

44
 Resistance-Time for W10X12
80
MATLAB RESISTANCE
60 SBEDs RESISTANCE

40
Resistance (psi)

20

0
0 10 20 30 40 50 60 70
-20

-40

-60
Time (ms)

Figure 3.4.1.2b: Resistance-Time Plot for W10X12 Fixed-Fixed with Uniform Load and 0% Damping

Resistance-Deflection for W10X12

80

60

40
Resistance (psi)

20

0
5.3 5.5 5.7 5.9 6.1 6.3 6.5
-20
MATLAB RESISTANCE
-40 SBEDS RESISTANCE

-60
Deflection (in)

Figure 3.4.1.2c: Resistance-Deflection Plot for W10X12 Fixed-Fixed with Uniform Load and 0% Damping (Magnified)

45
From Figure 3.4.1.2a, it can be seen that the first yield transition occurs at a resistance

of approximately 50 psi. Once the mass goes through rebound, and then inbound

again, the resistance reaches this value again. This can be seen in Figure 3.4.1.2b,

since there is no overlap. For this analysis, the resistance function does change in

slope. Figure 3.4.1.2c shows this (the stiffness changes). In other cases, the

resistance-deflection relationship is consistent with the SBEDS theory.

To account for this inconsistency in the MATLAB replica, a datum resistance value is

stored during the time stepping. This datum resistance is subtracted from the

resistance that is checked against the resistance values to cause the transformation

factors to change (yield regions). This allows the replica to match SBEDS for both

situations. Although this inconsistency does exist in SBEDS, it does not cause any

major issues while using the program. This is because design is based on maximum

response, and most maximum responses occur before rebound, which is before this

issue ever arises.

The modeling choices made while creating the MATLAB replica led to some minor

inconsistencies in the plots when validating each case. Figures 3.4.1.3a, 3.4.1.3b, and

3.4.1.3c on the following pages show one of these discrepancies:

46
 Deflection-Time Discrepancy
80
70
60
Deflection (in)

50
40
MATLAB DISPL
30
SBEDS Displ
20
10
0
0 20 40 60 80 100 120
Time (ms)

Figure 3.4.1.3a: Deflection-Time Plot for W10X12 Fixed-Pinned with Point Load and 0% Damping

Acceleration-Time Discrepancy
0.6
0.5
MATLAB ACC
0.4
Acceleration (in/ms^2)

SBEDS ACC
0.3
0.2
0.1
0
0 20 40 60 80 100 120
-0.1
-0.2
-0.3
Time (ms)

Figure 3.4.1.3b: Acceleration-Time Plot for W10X12 Fixed-Pinned with Point Load and 0% Damping

47
 Acceleration-Time Discrepancy
-0.1
-0.11 15 16 17 18 19 20

-0.12
Acceleration (in/ms^2)

-0.13
-0.14
-0.15 MATLAB ACC
-0.16 SBEDS ACC
-0.17
-0.18
-0.19
-0.2
Time (ms)

Figure 3.4.1.3c: Acceleration-Time Plot for W10X12 Fixed-Pinned with Point Load and 0% Damping (Magnified)

As seen in the above plots, the acceleration in SBEDS is slightly more negative than the

MATLAB replica, which causes the SBEDS displacement to be decreased. This is due to

the time step at which the transformation factors change, as a result of the resistance

reaching the yield value. For this specific case, the SBEDS solution spent a few more

time steps in the elasto-plastic region than the MATLAB solution did. This additional

stiffness for the elasto-plastic time steps is what caused the deflection to be smaller

for the SBEDS solution. The reason for these additional elasto-plastic time steps may

be due to the amount of significant digits that Excel stores compared to the amount

that MATLAB stores, and if SBEDS rounds the deflection values at which the stiffness

changes. However, this is a minor discrepancy (1.3% error), and when damping is

introduced, the plots overlap one another.

48
Another slight discrepancy is shown in Figure 3.4.1.4 below:

Deflection-Time
1.4

1.2

1
Deflection (in)

0.8

0.6 MATLAB DISPL

0.4 SBEDS Displ

0.2

0
0 10 20 30 40 50 60 70
Time (ms)

Figure 3.4.1.4: Deflection-Time Plot for W10X12 Fixed-Fixed with Uniform Load and 5% Damping

This discrepancy is due to the fact that the MATLAB program does not reset the

resistance as the exact yield value when transitioning. Instead, it uses the calculated

resistance from the previous time step, and therefore slightly overshoots the value of

resistance. Since the resistance is larger for one time step, the deflection is slightly

lower for the rest of the solution. However, this decrease is small enough (1.3% error)

to still be considered accurate with respect to SBEDS’s output.

49
Chapter 4 Multi-Degree-of-Freedom Solver

4.1 General

This chapter discusses all aspects of the MDOF analysis. This includes the modeling

techniques that SAP2000 implements, how the single-beam analysis in SAP2000 was

replicated using MATLAB, and some minor inconsistencies between the two solvers as

well as SAP2000 methodology and SBEDS methodology.

4.2 SAP2000

SAP2000 is a software package used to perform structural analysis, structural dynamic

analysis, and structural design. It is commonly used in the structural engineering

industry. The structural dynamic analysis capabilities of SAP2000 are the primary

focus of this verification study. This is a more precise version of the analysis that

SBEDS performs. The increase in accuracy comes from the fact that SAP2000 performs

MDOF analysis, whereas SBEDS performs SDOF analysis. Some of the key features of

this tool are explained in the following sections.

4.2.1 Inputs

SAP2000 has a graphical user interface as opposed to being a spreadsheet like SBEDS.

First, a grid is created for the nodes, or degrees of freedom to be placed. Next,

definitions of materials and section properties are specified, including material

50
nonlinear hinges which be applied to the nodes. Once these are specified, the user

can create the structural element or system with nodes and lines. An example of the

material definition window is shown below in Figure 4.2.1.1:

Figure 4.2.1.1: Example of Material Definition Window for SAP2000

Once the geometry is defined with associated materials and section properties,

boundary conditions and loads can be applied to the structure, as well as mass

51
definitions, which will be used in dynamic analysis. These loads can be linked to time-

history plots, and damping and a numerical method can be specified for the time-

history analysis. SAP2000 is limited by having lumped mass matrices only; it does not

use consistent-mass matrices. However, shear area can be included or neglected,

thereby specifying Timoshenko stiffness matrices or Euler-Bernoulli stiffness matrices,

respectively. Geometric effects due to axial load can be included or neglected as well.

Figure 4.2.1.2 below shows the graphical representation of a defined structure in

SAP2000:

Figure 4.2.1.2: Graphical Representation of a Pinned-Fixed Beam with 0.35 k/in Load and 4 Elements

Damping in SAP2000 is specified with Rayleigh Damping factors. These factors, 𝛼 and

𝛽 are multiplied by the elemental mass and elastic stiffness matrices respectively. An

important note is that SAP2000 uses the initial assembled damping matrix throughout

the whole time-history analysis. If a degree of freedom corresponding to an element

yields, its corresponding elemental stiffness-proportional damping matrix is multiplied

by a factor between 10−2 and 10−3, and then the global damping matrix is

reassembled. This is used until the degree of freedom that yielded reverses direction,

at which the original damping matrix is used again.

52
4.2.2 Outputs

Once the input parameters are chosen, a time-history can be performed, and all the

output data is accessible. This includes time-history data in spreadsheet format for

displacements, velocities, accelerations, internal forces, and reactions. SAP2000 has a

feature for plotting the data as well. Shown below in Figure 4.2.2.1 is an example plot

from SAP2000:

Figure 4.2.2.1: Example Plot of Deflection-Time in SAP2000

4.3 MATLAB Replica

The next step in quantifying the inaccuracies in SBEDS is to create a program that takes

the same inputs as SAP2000 (and SBEDS), performs a MDOF analysis like SAP2000, and

53
returns the outputs of interest. MATLAB was used to write this program as well, as

explained earlier. This code is provided in Appendix C: MATLAB Code for More

Accurate Model (MDOF Solver).

4.3.1 Input File

The first function file, called “input.m,” contains all the input data for the analysis,

specified by the user. This file allows the user to specify the geometry of the structure,

type of mass matrices to be used (consistent or lumped), type of elements to be used

(Timoshenko or Euler-Bernoulli), β and γ factors for Newmark’s method, tolerance of

unbalanced load for Newton-Raphson iteration, shape of the loading time-history

curve, boundary conditions, force values and locations scaled by the time-history

curve, axial load, Rayleigh Damping factors, length of the system, number of elements

(mesh), material parameters (elastic modulus, yield stress, density), section properties

(cross-sectional area, moment of inertia, shear area), and non-structural masses.

4.3.2 Main File

The program file, called “mainNR.m,” is the bulk of the code. It first reads the data

from the input function and performs preprocessing on this data to setup all the

matrices and vectors for time-history analysis. The code follows the steps that a

standard direct stiffness method or finite element method approach would follow, by

assigning all the properties to each element, and then assembling global matrices

based on the boundary conditions. Several small functions are used throughout this

54
process, such as a function that compute equivalent nodal loads for distributed loads

(called “g_element.m”) and a function that calculates the stiffness and mass matrices

for each element passed to it (called “k_element.m”).

Once all of this setup is complete, a “while” loop runs through the time stepping of

Newmark’s method with Newton-Raphson iteration. If an element yields, its stiffness

and mass matrices are passed to a function, called “condense.m,” that performs static

condensation to release the degree of freedom that hinged. After releasing this

degree of freedom, the resistance of this degree of freedom is capped at the plastic

moment capacity, since the solution is incremental. When this element yields, the

rows and columns in the stiffness-proportional damping matrix associated with the

degree of freedom that yielded are scaled by 10−3. SAP2000’s solution scales the

entire stiffness-proportional damping matrix; however, by only scaling the rows and

columns associated with the yielded degree of freedom, the MATLAB solution matches

SAP2000 more closely.

Once the time stepping is complete, some post-processing is performed on the data to

allow the data to be stored in a more efficient manner and plotted.

4.4 Solver Verification

To ensure that this MATLAB program functions as close to SAP2000 as possible,

several sets of inputs were analyzed with both programs and compared. The element

55
chosen for comparison was a W10X12 Grade 50 steel beam, with length of 10 ft. Four

elements were used for the mesh. Tributary width was not used because for these

solvers, force is directly specified, as opposed to specifying pressure and tributary

width. Therefore a ramp-down load was used with a specific value of force or force

per length. The duration of the forcing load was 17.8 ms. Every boundary condition

and load type were tested, with 2% damping in the first two modes, as well as no

damping. Peak loads were also varied for each combination to allow the output to

vary significantly. Axial loading was also incorporated into the verifications, as

proportions of the compressive capacity of the section. These values were 0, 0.2 and

0.4. A case was also performed with additional mass, equal to ten times the self-

weight.

An example of how the two analyses were compared is shown on the following pages

in Figures 4.4.1a and 4.4.1b. A 40k point load was used here as the initial value of the

ramp-down load function.

56
Figure 4.4.1a: Midpoint Deflection Comparison of SAP2000 with MDOF MATLAB Program

57
 Figure 4.4.1b: Midpoint Resistance Comparison of SAP2000 with MDOF MATLAB Program

The output from SAP2000 and the MATLAB program was pasted into a new Excel

document for comparison. Plots were created for deflection and resistance (internal

moment) at midpoint for each set of parameters. Even though the maximum

deflection does not occur at the midpoint for fixed-pinned boundary conditions, the

midpoint was still used as the comparison for convenience. For cantilevered boundary

conditions, the deflection was measured at the free end, and the internal resistance

was measured at the support end. Since the plots of both sets of outputs overlapped

58
well, the MATLAB replica was verified as being a good representation of what SAP2000

would output. For the full set of verifications, see Appendix D: MDOF Solver

Verifications (Time-History Plots).

4.4.1 Inconsistencies

While comparing each set of parameters to SAP2000, some discrepancies were

noticed between the two solutions. This could be the result of modeling choices;

however, these discrepancies are minor and do not cause the MATLAB solution to

deviate enough from the SAP2000 solution to cause significant differences in outputs.

With some of the test comparisons, the maximum deflections at midspan varied, and

the resistance curves were not identical. However, the deflections from the MATLAB

solution never deviated from the SAP2000 deflections by more than 3%. Therefore,

the MATLAB solution was accurate enough to be substituted for SAP2000 for the

analyses. This discrepancy in deflection and resistance could be attributed to several

minor differences in modeling choices between SAP2000 and the MATLAB replica:

1. When an element in SAP2000 undergoes yielding in the time-history

analysis, the time step of yielding is broken up into smaller increments.

This allows the solution find more accurately when the yielding occurs,

since the yielding occurs within a time step. The MATLAB replica does not

increase the time step of yielding; instead, if the resistance of an element

59
 surpasses the plastic capacity, its corresponding resistance vector (all

internal forces in the beam), are scaled back until the highest value of

resistance is the plastic capacity. This is a more inaccurate method,

however since all analyses are performed with very small time increments

to begin with (from the nature of a blast load), the difference in the two

modeling choices is negligible.

2. Another difference in the two solutions is how the damping matrix is

reduced during yielding. As explained previously, SAP2000 reduces any

stiffness-proportional damping matrix that corresponds to a degree of

freedom undergoing yielding. Once a rotational degree of freedom yields,

the stiffness components of the damping matrices corresponding to

elements that share this yielded global degree of freedom are reduced by a

factor between 10−2 and 10−3. By implementing this strategy into the

MATLAB replica, deflections were significantly larger than the deflections in

SAP2000. To better represent SAP2000’s output, more damping was

required during yielding to reduce the deflections. By only reducing the

rows and columns of the stiffness-proportional damping matrices

associated with yielded global degrees of freedom (by a factor of 10−3), the

MATLAB replica matched the SAP2000 output within 3% error (compared

to as much as 300% error before). All verification cases with damping use

60
 this method of row/column-only reduction, to get as close to SAP2000 as

possible.

3. The final differences between the models are related to adding axial load to

the beam, creating a beam-column. The first of these differences is related

to the reduction in the plastic moment capacity of a section due to axial

load. There are several resources that can be used for calculating this

reduction in plastic moment capacity. Some of these include Equation 5-4

from Prestandard and Commentary for the Seismic Rehabilitation of

Buildings (FEMA 356) and Equations H-1a and H1-1b from Steel

Construction Manual Fourteenth Edition (AISC). Since the goal is to

compare SBEDS to other models, it makes the most sense for the model

being compared to SBEDS to have the same reduced plastic capacity

calculation to stay consistent. In the SBEDS Methodology Manual, the

reduced plastic capacity calculation is outlined. Figure 4.4.1.1 on the

following page is taken from the SBEDS manual:

61
Figure 4.4.1.1 Equations for Reduced Plastic Moment Capacity in SBEDS (US Army, 2008)

62
By simply entering a yield stress for the material in SAP2000, and then performing the

time-history analysis with applied axial load, the yielding occurred at an internal

moment value different than expected. This value varied significantly from the value

that SBEDS uses for the same set of inputs. Hand calculations were performed to

identify the plastic moment from Equation 5-4 from Prestandard and Commentary for

the Seismic Rehabilitation of Buildings (FEMA 356) and Equations H-1a and H1-1b from

Steel Construction Manual Fourteenth Edition (AISC), but neither value matched the

value from SAP2000. Unable to determine how SAP2000 calculates reduced plastic

capacity, a different approach was taken. A different yield stress was manually

inputted into the SAP2000 material until the plastic moment capacity matched that of

SBEDS. This is how all of the verification tests for the MATLAB replica were performed.

After the plastic moment capacities matched, there was still a discrepancy between

SAP2000 and the MATLAB replica. This difference arose from including geometric

effects, which would implement the geometric stiffness matrices in the solution. In

the SAP2000 solution, the internal axial force in the beam-column fluctuated. Since

the axial force has an effect on the geometric stiffness matrix, this was changing the

geometric stiffness matrix at every time step. In the MATLAB replica, the axial load is

applied statically before the time-history analysis, and then this axial load is held

constant through the time stepping, and therefore so does the geometric stiffness

matrix. In SAP2000, a static analysis with the axial load was also applied to the system

before the time-history analysis. However, during the time-history analysis, the axial

63
degrees of freedom continued to move, causing this change in axial internal force.

SBEDS holds the axial load constant through the time-history analysis, so this was the

modeling choice chosen for the MATLAB solution as well, even though SAP2000 was

unable to keep the axial load constant through the time-history analysis.

This difference in modeling choice is most likely causing the discrepancy between

SAP2000’s and the MATLAB replica’s internal moment resistance plots. Once the

plastic capacity is reached, the MATLAB solution stays at this value until rebound.

However, the SAP2000 solution oscillates around the plastic capacity. This is shown

below in Figure 4.4.1.2a:

Figure 4.4.1.2a: Resistance-Time Plot for W10X12 Fixed-Pinned with Point Load, 2% Damping, and 0.2 Axial Load

64
The fact that the resistance does not flat-line during yielding changes the rest of the

solution. However, this discrepancy does not result in significant deflection deviation,

as shown below in Figure 4.4.1.2b:

Figure 4.4.1.2b: Deflection-Time Plot for W10X12 Fixed-Pinned with Point Load, 2% Damping, and 0.2 Axial Load

The discrepancy in the resistance plot has little effect on the deflection. For this

particular case, there is 2.5% deviation in the deflection from the SAP2000 solution,

which is acceptable. Also, by varying the yield stress in SAP2000, the deviation can be

further reduced.

65
All of the verifications thus far were performed with lumped mass matrices. SAP2000

is unable to implement a consistent-mass matrix. However, after all these

verifications were performed, a final comparison was performed. This analysis held

everything constant (W10X12 pinned-pinned with 40k point load), and only varied the

mass matrix between a lumped mass matrix and a consistent-mass matrix. The

comparisons to SBEDS are performed with consistent-mass matrices, so a comparison

of the lumped mass solution to the consistent-mass matrix solution provides a

verification that the consistent-mass matrix is properly defined, even though the

consistent-mass matrix exceeds the capabilities of SAP2000. Figures 4.4.1.3a and

4.4.1.3b on the following pages show these final comparisons:

66
Figure 4.4.1.3a: Deflection-Time Comparison for Different Stiffness and Mass Models

67
 Figure 4.4.1.3b: Resistance-Time Comparison for Different Stiffness and Mass Models

The figures above show that by using a consistent-mass matrix, the solution changes

slightly; however the implementation of the consistent-mass matrix is shown as not

significantly affecting the solution because the overall resistance-time and deflection-

time plots remain very similar. It is expected that the period should change slightly

when changing from a lumped mass matrix to a consistent-mass matrix, since the

period is dependent upon the stiffness and mass distribution.

68
Chapter 5 Comparisons

5.1 General

This chapter discusses the different input parameters that were varied for all of the

analyses, and the corresponding output of each case. This includes incorporating the

two created MATLAB codes into one overall code and which output parameters were

saved into the data frames for each analysis.

In section 8.1.2 of Yokoyama’s thesis, he introduces a stiffness factor that compares

the overall deflected shape of SDOF solution to the MDOF solution, for a specified

time step. This “KL factor” takes the effective stiffness of the MDOF solution and

divides by the SDOF stiffness. When the deflected shape of the MDOF solution

matches that of the SDOF solution, this factor will be near the load transformation

factor explained in section 2.6.1 earlier. Deviation of this “KL factor” from the load

transformation factor means that the two solutions develop different overall deflected

shapes (Yokoyama, 2011).

The parameters varied for the analyses were charge of TNT (lb), standoff distance from

charge to structure (ft), axial load as proportion of axial capacity of the element,

boundary conditions, damping ratio, section shape, and added weight (psf). Other

input parameters were held constant for all three data sets. Table 5.1.1 on the

following page summarizes the constant parameters:

69
 Parameter Length of Loading Tributary Elastic Yield Mesh of Type of
Element Width Modulus Stress Element Element
and Mass
Value 15 ft Uniform 15 ft 29,000 ksi 50 ksi 16 consistent-
elements mass
Comment Standard More Conservative Steel Steel Less than Most
for Column applicable for spacing 1 ft per accurate
to blast element

Table 5.1.1: Constant Input Parameters

Additionally, a static and dynamic increase factor is applied to the yield stress, as

SBEDS uses. These values vary from 1.00 to 1.05 (static) and 1.19 to 1.24 (dynamic), as

explained in the following sections. Whenever a Bernoulli element is specified, the

corresponding consistent-mass matrix for a Bernoulli element is used. Whenever a

Timoshenko element is specified, the corresponding consistent-mass matrix for a

Timoshenko element is used.

5.2 Results and Analysis

The following sections contain selections from the full sets of analyses performed,

with comments on the selected plots.

5.2.1 First Data Set

A first set of analyses were performed, with large increments in each varied

parameter, to get a sense of the regions of discrepancy between SDOF and MDOF

analysis. Charges from 10 lb to 2000 lb were paired with standoffs from 3 ft to 250 ft.

70
However, only combinations which resulted in a scaled distance between 0.25 and 10

𝑚⁄𝑘𝑔1⁄3 were included. A summary table of the combinations of charges and

standoffs is shown below:

Standoff (ft) 3 5 10 25 50 100 250

Charges (lb) 10 10 10 10 10
25 25 25 25 25
50 50 50 50 52
200 200 200 200 200
500 500 500 500 500
2000 2000 2000 2000 2000

Table 5.1.2: Charges and Standoffs for Data Set 1

Axial loads were 0, 0.2, and 0.4 proportion of compressive axial capacity. Boundary

conditions were pinned-pinned, fixed-pinned, and fixed-fixed. Damping was set to 2%,

and in the first two modes for the MDOF analysis. Section shapes included were

W14X48, W14X53, W14X61, W14X82, W14X109, W14X132, W14X145, and W14X257.

The added weights used were 0 psf, 15 psf, 40 psf, and 81.25 psf. These values

represent no weight, glazing, brick, and concrete wall. The static and dynamic increase

factors were set to 1.05 and 1.19, respectively.

For this first set, 4 sets of plots were created for each beam section. All graphs are

plotted against scaled distance, Z. The first set of plots displays the deviation in the KL

factor using Bernoulli elements from the appropriate load factor for the SDOF solution.

The time step used for comparison was the time step in which first yield occurred. No

71
axial load is incorporated here because the SDOF solution stiffness does not take into

account axial load, so the plots would not be comparable. Additionally, Timoshenko

elements are not included in these plots because the overall shape for a Timoshenko

beam will always be different than the overall shape for a SDOF model. Two examples

of this set of plots are shown below:

Pinpin KL factor
200
15psf
150
KL deviation (%)

40psf
81.25psf
100

50

-50
0 5 10 15 20 25
(1/3)
Z (ft/lb )
Fixpin KL factor
200
15psf
150
KL deviation (%)

40psf
81.25psf
100

50

-50
4 6 8 10 12 14 16 18 20 22 24
(1/3)
Z (ft/lb )
Fixfix KL factor
200
15psf
150
KL deviation (%)

40psf
81.25psf
100

50

-50
4 6 8 10 12 14 16 18 20 22 24
(1/3)
Z (ft/lb )

Figure 5.2.1.1: Deviation in KL vs. Z for W14X48

72
 Pinpin KL factor
200

KL deviation (%)
15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
(1/3)
Z (ft/lb )

Fixpin KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
(1/3)
Z (ft/lb )

Fixfix KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
(1/3)
Z (ft/lb )

Figure 5.2.1.2: Deviation in KL vs. Z for W14X257

Depending on the boundary condition, the Z at which the MDOF solution begins to

deviate from the SDOF solution will change. Additionally, the beam section size

changes this deviation point as well, since adding stiffness will reduce the potential for

yielding under the same load. Since the Z values had such large increments, the

deviation appears to grow instantaneously. Additionally, the solution may break down

73
at very low Z values near 2 𝑓𝑡⁄𝑙𝑏 1⁄3 , but a limit can clearly be seen. The deviation

curves would be more gradual if finer increments were to be used. From this data, it

can be seen that the supposed 3 𝑓𝑡⁄𝑙𝑏 1⁄3 limit does not necessarily apply universally

for all analyses, and this limit is highly dependent upon the element’s stiffness and

boundary condition. For a pinned-pinned W14X257, this limit may be acceptable, but

by decreasing the section size to a W14X48 and changing the boundary conditions to

fixed-fixed, the limit is at roughly 12𝑓𝑡⁄𝑙𝑏 1⁄3 . From these plots, the added weight

does not seem to have significant effect on changing the limit for a specific set of

conditions.

The second set of plots displays the deviation in maximum deflection using Bernoulli

elements from the SDOF maximum deflection. Axial load was implemented for these

plots. An example of this set of plots is shown on the next page:

74
Max deflection deviation (%)
Pinpin SDOF-Bernoulli Deflection
0
0 axial 15 psf
0 axial 40 psf
0 axial 81.25 psf
-50 0.2 axial 15 psf
0.2 axial 40 psf
0.2 axial 81.25 psf
-100 0.4 axial 15 psf
0 5 10 15 20 25
0.4 axial 40 psf
Z (ft/lb(1/3)) 0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
0 axial 81.25 psf
-50 0.2 axial 15 psf
0.2 axial 40 psf
0.2 axial 81.25 psf
-100 0.4 axial 15 psf
0 5 10 15 20 25
0.4 axial 40 psf
Z (ft/lb(1/3)) 0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
0 axial 81.25 psf
-50 0.2 axial 15 psf
0.2 axial 40 psf
0.2 axial 81.25 psf
-100 0.4 axial 15 psf
0 5 10 15 20 25
0.4 axial 40 psf
(1/3)
Z (ft/lb ) 0.4 axial 81.25 psf

Figure 5.2.1.3: Deviation in Bernoulli-SDOF Maximum Deflection vs. Z for W14X257

In these plots, the significant deviations start to occur under 8.5𝑓𝑡⁄𝑙𝑏 1⁄3 . This is

interesting because Figure 5.2.1.2 indicates consistent overall shape until 12𝑓𝑡⁄𝑙𝑏 1⁄3 .

This shows that even though deflected shapes at first yield may be the same, their

magnitudes can be different. This is still a deviation in solution. Adding axial load

creates more deviation in solutions. This may be due to the fact that the SDOF

75
solution takes axial load and applies additional equivalent transverse load, while the

MDOF solution decreases the stiffness of the element (using the geometric stiffness

matrix). Since the deviation values become negative, the Bernoulli maximum

deflection is smaller than the SDOF solution.

The third set of plots displays the deviation in maximum deflection using Timoshenko

elements from the SDOF maximum deflection. An example of this set of plots is

shown on the following page:

76
Max deflection deviation (%)
Pinpin SDOF-Timoshenko Deflection
50
0 axial 15 psf
0 axial 40 psf
0 axial 81.25 psf
0 0.2 axial 15 psf
0.2 axial 40 psf
0.2 axial 81.25 psf
-50 0.4 axial 15 psf
0 5 10 15 20 25
0.4 axial 40 psf
Z (ft/lb(1/3)) 0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 axial 81.25 psf
0 0.2 axial 15 psf
0.2 axial 40 psf
0.2 axial 81.25 psf
-50 0.4 axial 15 psf
0 5 10 15 20 25
0.4 axial 40 psf
(1/3)
Z (ft/lb ) 0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 axial 81.25 psf
0 0.2 axial 15 psf
0.2 axial 40 psf
0.2 axial 81.25 psf
-50 0.4 axial 15 psf
0 5 10 15 20 25
0.4 axial 40 psf
Z (ft/lb(1/3)) 0.4 axial 81.25 psf

Figure 5.2.1.4: Deviation in Timoshenko-SDOF Maximum Deflection vs. Z for W14X257

The above plots show how different maximum deflections may be when shear

deformation is included. For very high Z values, deviation is at 25% to 50%. Since the

Bernoulli solution starts to undershoot the SDOF solution below a certain Z, and since

Timoshenko elements allow more overall deformation of the system, the two facts

almost negate each other. The absolute values of the deviations in Figure 5.2.1.4 are

77
smaller than those in Figure 5.2.1.3. However, the deviations in Figure 5.2.1.4 are still

significant. A definitive limit on Z cannot be seen because the Timoshenko solution is

always above or below the SDOF solution.

The final set of plots displays the deviation in maximum deflection using Timoshenko

elements from maximum deflection using Bernoulli elements. An example of this set

of plots is shown on the following page:

78
 Max deflection deviation (%)
Pinpin Timoshenko-Bernoulli Deflection
100
0 axial 15 psf
0 axial 40 psf
0 axial 81.25 psf
50
0.2 axial 15 psf
0.2 axial 40 psf
0.2 axial 81.25 psf
0 0.4 axial 15 psf
0 5 10 15 20 25
0.4 axial 40 psf
Z (ft/lb(1/3)) 0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
0 axial 81.25 psf
50
0.2 axial 15 psf
0.2 axial 40 psf
0.2 axial 81.25 psf
0 0.4 axial 15 psf
0 5 10 15 20 25
0.4 axial 40 psf
Z (ft/lb(1/3)) 0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
0 axial 81.25 psf
50
0.2 axial 15 psf
0.2 axial 40 psf
0.2 axial 81.25 psf
0 0.4 axial 15 psf
0 5 10 15 20 25
0.4 axial 40 psf
Z (ft/lb(1/3)) 0.4 axial 81.25 psf

Figure 5.2.1.5: Deviation in Timoshenko-Bernoulli Maximum Deflection vs. Z for W14X257

This set of plots shows that the Timoshenko solution always produces a larger

maximum deflection than the Bernoulli solution, since the deviation is always positive.

Also, the deviations grow as Z decreases, especially for pinned-pinned and fixed-fixed

boundary conditions. This makes sense because these boundary conditions are less

stiff, so the inclusion of shear deformation has a larger effect for these conditions.

79
For this first data set, each curve held boundary conditions, added weight, and axial

load constant. However, different combinations of charge and standoff were lumped

together on this curve. The fact that these curves are not smooth at high Z values

shows that using Z to identify a limit for SDOF analysis is not necessarily a viable

option. Different combinations of charge and standoff that create a similar Z value

have different deviations. This can be seen in Figure 5.2.1.5 for the fixed-pinned and

fixed-fixed conditions, where there is an immediate jump in deviation at a Z of

17𝑓𝑡⁄𝑙𝑏 1⁄3. For two different combinations of charge and standoff that create a Z of

17𝑓𝑡⁄𝑙𝑏 1⁄3, the deviation in maximum deflection changes. This is why the plots in the

first data set seem to oscillate (in addition to the large increment in Z for each data

point).

The full set of plots for the first data set is located in Appendix E: First Data Set Plots.

5.2.2 Second Data Set

The second set focused on the fact that two different sets of charges and standoffs

that create the same Z value can have different results. This data set used a W14X109

(mid-stiffness from the first data set) with 2% damping, and used the same modes for

damping as the first data set for MDOF analysis. With a 25 lb charge and a 200 lb

charge, standoffs were back-calculated to establish a range of Z values from 2.5 to 15

80
𝑓𝑡⁄𝑙𝑏 1⁄3 in smaller increments than the first data set, so that a finer solution set could

be obtained. By separating the curves by charge and by creating a finer Z increment,

the oscillations and jagged nature of the curves from the first data set are removed.

The same combinations of boundary conditions, added weight, and axial loads were

used as the first data set. The same deviations from the first data set were plotted

against Z. An example of one of these plots is shown below:

Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection 0 Axial

0
15 psf 25 lb charge
15 psf 200 lb charge
40 psf 25 lb charge
-50 40 psf 200 lb charge
81.25 psf 25 lb charge
81.25 psf 200 lb charge
-100
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection 0.2 Axial

0
15 psf 25 lb charge
15 psf 200 lb charge
40 psf 25 lb charge
-50 40 psf 200 lb charge
81.25 psf 25 lb charge
81.25 psf 200 lb charge
-100
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection 0.4 Axial

0
15 psf 25 lb charge
15 psf 200 lb charge
40 psf 25 lb charge
-50 40 psf 200 lb charge
81.25 psf 25 lb charge
81.25 psf 200 lb charge
-100
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

Figure 5.2.2.1: Deviation in Bernoulli-SDOF Maximum Deflection vs. Z for Pinned-Pinned W14X109 (Fine Z Increment)

81
Since the different charges are not in the same curve anymore, the plots are smoother

and follow a more distinguishable path. The above plots show that for a pinned-

pinned, W14X109 member, increasing the charge while also increasing Z, decreasing

supported weight, and adding axial load all cause the Bernoulli MDOF solution to

deviate from the SDOF solution at a higher Z value. The curve of various standoffs

with no axial load, 81.25 psf added weight, and a 25 lb charge never significantly

deviates from the SDOF solution, even at Z values lower than the accepted limit of

3𝑓𝑡⁄𝑙𝑏 1⁄3 . Meanwhile, the curve with 0.4 axial load, 15 psf added weight, and a 200

lb charge shows 20% deviation from the SDOF solution at a Z of 12𝑓𝑡⁄𝑙𝑏 1⁄3 . These

two curves represent the extremes for the pinned-pinned condition, but they illustrate

the fact that significant deviations start to develop at very different Z values,

depending on the conditions.

The figure on the following page shows the same plots as Figure 5.2.2.1, but using

Timoshenko elements instead of Bernoulli elements, to see the effect of shear

deformation.

82
Max deflection deviation (%)
Pinpin SDOF-Timoshenko Deflection 0 Axial
50
15 psf 25 lb charge
15 psf 200 lb charge
40 psf 25 lb charge
0 40 psf 200 lb charge
81.25 psf 25 lb charge
81.25 psf 200 lb charge
-50
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Pinpin SDOF-Timoshenko Deflection 0.2 Axial

50
15 psf 25 lb charge
15 psf 200 lb charge
40 psf 25 lb charge
0 40 psf 200 lb charge
81.25 psf 25 lb charge
81.25 psf 200 lb charge
-50
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Pinpin SDOF-Timoshenko Deflection 0.4 Axial

50
15 psf 25 lb charge
15 psf 200 lb charge
40 psf 25 lb charge
0 40 psf 200 lb charge
81.25 psf 25 lb charge
81.25 psf 200 lb charge
-50
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

Figure 5.2.2.2: Deviation in Timoshenko-SDOF Maximum Deflection vs. Z for Pinned-Pinned W14X109 (Fine Z
Increment)

As stated previously in the first data set, the fact that the Bernoulli MDOF solution

undershoots the SDOF solution, and the fact that the Timoshenko MDOF solution

allows more deformation than the Bernoulli MDOF solution, tends to negate one

another. For the plot with 0 axial load, the Timoshenko solution overestimates the

maximum deflection of the SDOF solution by roughly 8%, until a Z of approximately

83
11.5𝑓𝑡⁄𝑙𝑏 1⁄3. At this point, only the curve for 15 psf added weight and 200 lb charge

starts to show increased deviation. By a Z of 9𝑓𝑡⁄𝑙𝑏 1⁄3, the curve for 40 psf added

weight and 200 lb charge starts to deviate more as well. However, all these deviations

stay under 30%, and by very low Z values, the curves tend to stop increasing and start

to decrease – this trend was also shown in the Bernoulli plots.

Due to all these factors, especially since different charges exhibit larger deviations

than other charges for the same scaled distance, it is difficult to attach a single value

on the limit of Z for accuracy of SDOF analysis to blast loads.

The full set of plots for the second data set is located in Appendix F: Second Data Set

Plots.

5.2.3 Third Data Set

Experimental blast testing was performed on various beams and beam-columns, which

was published in “Experimental Performance of Steel Beams under Blast Loading” by

Amr A. Nassr and others. The results from 5 blast shots are shown on the following

page in Table 5.2.3.1. In the same paper, models were created to replicate the blast

and results. In the paper, no static increase factor was used, and a dynamic increase

factor of 1.24 seemed to match the experimental data the best (for holding the

increase factor constant throughout the analysis). For this reason, the dynamic

84
increase factor of 1.24 was used for the third data set, even though this is larger than

the conventional value.

Table 5.2.3.1: Experimental Blast Data (Nassr, 2012)

The third data set was the same as the second set, but with two changes. This third

set re-calibrated the SDOF and MDOF models to better match the experimental data.

For this reason, the static increase factor was set to 1 and the dynamic increase factor

was set to 1.24 for data set three. The first mode was damped at 2%, and the second

chosen mode to be damped by 2% for MDOF analysis was adjusted until the

Timoshenko solution came closest to the experimental data. The Timoshenko solution

is assumed to be the most accurate solution, so once the variables were calibrated

using the Timoshenko solution, these calibrated variables were applied to the

Bernoulli and SDOF solutions, where applicable. Only simply supported data was

available. For the 16 element model in use, damping in the first and 48th modes

resulted in a maximum deflection closest to the experimental data. This represented

the last mode of the system, and therefore fixed-pinned and fixed-fixed conditions had

damping applied to their first and last modes as well. A summary of the adjustments

and results of the adjustments are displayed on the next page in Table 5.2.3.2:

85
 Measured Maximum Timoshenko Maximum Second Mode Set to 2% SIF,DIF
Deflection (mm) Deflection (mm) Damping
Shot 1 6.9 7.1 48th (Last) 1.00, 1.24
Shot 3 33.2 32.7 46th 1.00, 1.24
Shot 5 62.8 68.2 30th 1.00, 1.24

Table 5.2.3.2: Timoshenko Solution Calibrations to Better Match Experimental Data

To best match experimental deflection from each shot, the second mode to be

damped at 2% changed for each shot. Therefore, the last mode was the chosen mode

for data set three. It is difficult to say which chosen mode is most accurate because

there may be other factors that are causing the discrepancy in deflection, such as the

fact that the pressure gages along the length of the beam reported different pressure

values for the same time step (the experimental load was non-uniform). Additionally,

one of Nassr’s conclusions is that, “[u]sing constant dynamic increase factor to

estimate the material strength because of strain rate might not provide a realistic

assessment of the actual effect of the strain rate on the dynamic response of beams.”

The models for this thesis implement a constant dynamic increase factor, so this is why

calibrations were made to other variables in order to best match the experimental

data.

Once the model was re-calibrated, this third data set was performed, and the same

types of plots were created with this new data. The figure on the following page

shows one of these plots:

86
Max deflection deviation (%) Pinpin SDOF-Bernoulli Deflection 0 Axial
40
15 psf 25 lb charge
15 psf 200 lb charge
20
40 psf 25 lb charge
40 psf 200 lb charge
0 81.25 psf 25 lb charge
81.25 psf 200 lb charge

-20
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

Pinpin SDOF-Bernoulli Deflection 0.2 Axial

Max deflection deviation (%)

40
15 psf 25 lb charge
15 psf 200 lb charge
20
40 psf 25 lb charge
40 psf 200 lb charge
0 81.25 psf 25 lb charge
81.25 psf 200 lb charge

-20
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

Pinpin SDOF-Bernoulli Deflection 0.4 Axial

Max deflection deviation (%)

40
15 psf 25 lb charge
15 psf 200 lb charge
20
40 psf 25 lb charge
40 psf 200 lb charge
0 81.25 psf 25 lb charge
81.25 psf 200 lb charge

-20
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

Figure 5.2.3.1: Deviation in Bernoulli-SDOF Maximum Deflection vs. Z for Pinned-Pinned W14X109 (Calibrated
Damping)

From the second data set, these conditions led the Bernoulli solution to undershoot

the SDOF solution. Altering the static and dynamic increase factors has little effect on

the results, since these values are used in both the SDOF and MDOF solutions to define

87
when yielding in an element will occur. Additionally, using a static increase factor and

dynamic increase factor of 1.05 and 1.24 respectively creates a net increase factor of

1.25. This third set of data uses a net increase factor of 1.24, which is very close to

1.25. The damping is the main difference between data set two and data set three.

Now that damping for the MDOF model has been decreased, the Bernoulli solution

overshoots the SDOF solution. This third set of data was calibrated using the

Timoshenko solution, so the following plots are even more relevant:

In these plots, the MDOF solution overshoots the SDOF solution even more. The plots

follow the same trends as every other set of plots mentioned, with larger deviations

depending on the set of conditions used. One again, this supports the fact that it is

not accurate to use a single value of Z to distinguish when SDOF analysis is applicable.

The full set of plots for the third data set is located in Appendix G: Third Data Set Plots.

88
 Pinpin SDOF-Timoshenko Deflection 0 Axial

Max deflection deviation (%)

100
80 15 psf 25 lb charge
15 psf 200 lb charge
60
40 psf 25 lb charge
40 40 psf 200 lb charge
20 81.25 psf 25 lb charge
81.25 psf 200 lb charge
0
-20
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

Pinpin SDOF-Timoshenko Deflection 0.2 Axial

Max deflection deviation (%)

100
80 15 psf 25 lb charge
15 psf 200 lb charge
60
40 psf 25 lb charge
40 40 psf 200 lb charge
20 81.25 psf 25 lb charge
81.25 psf 200 lb charge
0
-20
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

Pinpin SDOF-Timoshenko Deflection 0.4 Axial

Max deflection deviation (%)

100
80 15 psf 25 lb charge
15 psf 200 lb charge
60
40 psf 25 lb charge
40 40 psf 200 lb charge
20 81.25 psf 25 lb charge
81.25 psf 200 lb charge
0
-20
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

Figure 5.2.3.2: Deviation in Timoshenko-SDOF Maximum Deflection vs. Z for Pinned-Pinned W14X109 (Calibrated
Damping)

5.2.4 Individual Deflected Shapes

This fourth section consists of individual chosen analyses where the MDOF solution

deviated significantly from the SDOF solution. Included is the same set of inputs which

Yokoyama (2011) identified as causing significant deviation in the MDOF deflected

89
shape from the SDOF deflected shape. Plots here display the SDOF, Bernoulli MDOF,

and Timoshenko MDOF deflected shapes of the element at first yield. Input

parameters stayed consistent with parameters from the third data set, such as

dynamic increase factors and damping options.

The first individual case investigates a pinned-pinned W14X109 beam with a 200 lb

charge, 15 psf added weight, and no axial load. Cases were taken from the first plots

in Figures 5.2.3.1 and 5.2.3.2, using Z values of approximately 12.4, 9.1, and

7.7𝑓𝑡⁄𝑙𝑏 1⁄3 . In Figure 5.2.3.1, these cases encompass little deviation to large

deviation in maximum deflection for the Bernoulli elements.

Deflected Shape,W14X109,pinpin, with0 Axial,15 psf,72.3305 ft Standoff,200 lb Charge, At First Yield

1.4
SDOF
Bernoulli
1.2
Timoshenko

1
Deflection (in)

0.8

0.6

0.4

0.2

0
0 20 40 60 80 100 120 140 160 180
Distance Along Beam (in)

Figure 5.2.4.1: Deflected Shape at First Yield for Pinned-Pinned W14X109 with Z=12.4𝑓𝑡 ⁄𝑙𝑏 1⁄3

90
Deflected Shape,W14X109,pinpin, with0 Axial,15 psf,53.0937 ft Standoff,200 lb Charge, At First Yield
1.4
SDOF
Bernoulli
1.2
Timoshenko

1
Deflection (in)

0.8

0.6

0.4

0.2

0
0 20 40 60 80 100 120 140 160 180
Distance Along Beam (in)

Figure 5.2.4.2: Deflected Shape at First Yield for Pinned-Pinned W14X109 with Z=9.1𝑓𝑡 ⁄𝑙𝑏 1⁄3

Deflected Shape,W14X109,pinpin, with0 Axial,15 psf,45 ft Standoff,200 lb Charge, At First Yield

1.4
SDOF
Bernoulli
1.2
Timoshenko

1
Deflection (in)

0.8

0.6

0.4

0.2

0
0 20 40 60 80 100 120 140 160 180
Distance Along Beam (in)
Figure 5.2.4.3: Deflected Shape at First Yield for Pinned-Pinned W14X109 with Z=7.7𝑓𝑡 ⁄𝑙𝑏 1⁄3

91
Figure 5.2.4.1 shows little deviation in maximum deflection for the Bernoulli elements.

This is consistent with how the deflected shape looks at first yield; the Bernoulli

solution fits almost exactly on the SDOF solution. The standoff and charge

combination here must not be strong enough to cause the MDOF solution to deviate

the maximum deflection significantly. Both the Bernoulli and SDOF solutions yield at a

midspan deflection of 1.02 inches.

By decreasing the standoff, the solutions deviate from one another, as shown in Figure

5.2.4.2. Now, the Bernoulli solution yields at 1.08 inches, while the SDOF solution

yields at 1.15 inches. Because the SDOF solution assumes a certain deflected shape

through the whole analysis, the only reason for an increase in deflection at first yield

would be the duration and magnitude of the load versus the time step. By decreasing

the duration and/or increasing the magnitude of the load, the change in resistance will

be larger for every time step, and the time step under first yield will overshoot the

theoretical deflection at first yield, depending on the size of the time step. However,

for a MDOF analysis, since each node can deflect independently, a change in the

deflection at first yield can be caused by the relative deflection between nodes, which

may vary different combinations of duration and magnitude of load. In addition, the

MDOF deflection at first yield is dependent upon the time step for the same reason as

the SDOF solution. From Figure 5.2.3.1, this case with a Z of 9.1𝑓𝑡⁄𝑙𝑏 1⁄3 shows a

92
deviation of 11.5%. However, the deviation at first yield is -6.1%. This means that

during the time steps past yield, the Bernoulli solution deflects enough to change its

relative deviation from -6.1% to 11.5% compared to the SDOF solution. This shows

that most of the deviation comes from the time steps post-yield.

Decreasing the standoff even further, the deflected shapes change relative to each

other again, as shown in Figure 5.2.4.3. Here, the Bernoulli deflection at first yield is

1.20 inches, while the SDOF deflection at first yield is 1.17 inches. This is a 2.6%

deviation. From Figure 5.2.3.1, this case with a Z of 7.7𝑓𝑡⁄𝑙𝑏 1⁄3 shows a deviation of

14%. Once again, this shows that the deviation in maximum deflection for the

Bernoulli solution is primarily due to the time steps post-yield.

The second individual case investigates a fixed-fixed W14X109 beam with a 200 lb

charge, 15 psf added weight, and no axial load. Cases were taken from equivalent first

plots in Figures 5.2.3.1 and 5.2.3.2 for a fixed-fixed condition, using Z values of

approximately 11.0, 9.5, and 5.8𝑓𝑡⁄𝑙𝑏 1⁄3 (see Appendix G: Third Data Set Plots for

these plots for a fixed-fixed condition). These cases encompass little deviation to large

deviation in maximum deflection for the Bernoulli elements.

93
 Deflected Shape,W14X109,fixfix, with0 Axial,15 psf,64.3284 ft Standoff,200 lb Charge, At First Yield
0.7
SDOF
Bernoulli
0.6
Timoshenko

0.5
Deflection (in)

0.4

0.3

0.2

0.1

0
0 20 40 60 80 100 120 140 160 180
Distance Along Beam (in)

Figure 5.2.4.4: Deflected Shape at First Yield for Fixed-Fixed W14X109 with Z=11.0𝑓𝑡 ⁄𝑙𝑏 1⁄3

Deflected Shape,W14X109,fixfix, with0 Axial,15 psf,55.5563 ft Standoff,200 lb Charge, At First Yield

0.7
SDOF
Bernoulli
0.6
Timoshenko

0.5
Deflection (in)

0.4

0.3

0.2

0.1

0
0 20 40 60 80 100 120 140 160 180
Distance Along Beam (in)

Figure 5.2.4.5: Deflected Shape at First Yield for Fixed-Fixed W14X109 with Z=9.5 𝑓𝑡 ⁄𝑙𝑏 1⁄3

94
 Deflected Shape,W14X109,fixfix, with0 Axial,15 psf,33.6262 ft Standoff,200 lb Charge, At First Yield
0.4
SDOF
0.35 Bernoulli
Timoshenko

0.3

0.25
Deflection (in)

0.2

0.15

0.1

0.05

0
0 20 40 60 80 100 120 140 160 180
Distance Along Beam (in)

Figure 5.2.4.6: Deflected Shape at First Yield for Fixed-Fixed W14X109 with Z=5.8 𝑓𝑡 ⁄𝑙𝑏 1⁄3

The same remarks from Figures 5.2.4.1 through 5.2.4.3 can be applied to Figures

5.2.4.4 and 5.2.4.5. However, as Z is decreased for fixed-fixed boundary conditions,

the deflected shapes at first yield for the MDOF solutions become drastically different

than for the SDOF solution. Since the SDOF solution assumes a deflected shape, only

the magnitude of the overall shape may change at first yield. However, the MDOF

solutions may have any shape, especially as the number of elements used increases.

From Figure 5.2.4.6, the MDOF solutions yield in locations different than at midspan.

The beam will yield where the relative displacement between consecutive nodes is

large enough to cause a resistance equal to the yield moment. This can occur at any

95
node. For this set of conditions, this happens to occur at approximately quarter-span

(and three-quarter-span). The combination of boundary conditions, load, and

duration of load causes the beam to deflect in a way where the quarter-span (and

three-quarter-span) experiences the largest relative deflection. If the shapes at first

yield look this dissimilar and the majority of the deviation comes from the time steps

post-yield, then these solutions will be vastly different, since they start the post-yield

time step at vastly different shapes already.

The last individual case is a replica of Yokoyama’s case (2011). The inputs for this case

consist of a fixed-fixed W18X65 beam with no added weight and no axial load. The

length and tributary width are 30 ft and 10 ft, respectively. Yokoyama directly

inputted a peak pressure and impulse, instead of obtaining these from a certain charge

and standoff. Yokoyama used a peak pressure of 100 psi and an impulse of 100 psi-

msec. The case here has peak pressures of 45 and 50 psi, and impulses of 45 and 50

psi-msec. This is so the duration of the load stays the same for each pressure-impulse

combination, and is consistent with the duration from Yokoyama’s case. The following

figures show the deflected shapes at first yield:

96
 Deflected Shape,W18X65,fixfix,0 Axial,0 psf,45 psi ,45 psi*ms, At First Yield
1.6
SDOF
1.4 Bernoulli
Timoshenko

1.2

1
Deflection (in)

0.8

0.6

0.4

0.2

0
0 50 100 150 200 250 300 350
Distance Along Beam (in)

Figure 5.2.4.7: Deflected Shape at First Yield for Fixed-Fixed W18X65 with 45 psi and 45 psi-msec

97
 Deflected Shape,W18X65,fixfix,0 Axial,0 psf,50 psi ,50 psi*ms, At First Yield
1.4
SDOF
Bernoulli
1.2
Timoshenko

1
Deflection (in)

0.8

0.6

0.4

0.2

0
0 50 100 150 200 250 300 350
Distance Along Beam (in)

Figure 5.2.4.8: Deflected Shape at First Yield for Fixed-Fixed W18X65 with 50 psi and 50 psi-msec

In Yokoyama’s article, “SDOF Limits on Static Deflected Shape Assumptions,” he plots a

deflected shape for his SDOF analysis and MDOF analysis, using Bernoulli elements

with 100 psi and 100 psi-msec. The shapes from Figure 5.2.4.8 above are consistent

with the results from Yokoyama.

The yielding of the element at locations different than midspan supports the fact that

SDOF methodology may not be the best choice for analysis, under certain

circumstances.

98
Chapter 6 Conclusions and Future Research

6.1 General

This final chapter discusses summarizes the results from the research. Also included in

this chapter are areas for future study.

6.2 Summary

Single-Degree-of-Freedom and Multi-Degree-of-Freedom analyses are two different

methods for obtaining the time-history results (displacement, internal force, etc.) for a

structural element under a dynamic load. They attempt to model reality by

discretizing the continuous domains, time and space. MDOF analysis is generally

accepted as a more precise representation of structural behavior than SDOF analysis.

However, SDOF analysis is more computationally efficient, and is therefore more

commonly used in practice.

Under blast loading, the structural element experiences load which is large in

magnitude and short in duration. These fast-acting loads may cause deflected shapes

that cannot be represented by SDOF analysis. SDOF analysis assumes a deflected

shape for the element, which is based on its static deflected shape. These blast loads

may cause the element to respond with a deflected shape very different than the

static shape. This is where the discrepancy between SDOF and MDOF analysis arises.

99
Scaled distance (Z) is a measure of the intensity of a blast load. The conventional Z

cutoff for which SDOF analysis is assumed to be no longer applicable is below 1.2

𝑚⁄𝑘𝑔1⁄3 (3𝑓𝑡⁄𝑙𝑏 1⁄3 ). However, results from this research show from several sets of

initial conditions that this cutoff is too low, and that using scaled distance may not

even be an appropriate factor for determining a limit on SDOF analysis.

6.3 Conclusions

For this thesis, an absolute deviation larger than 15% is considered too large to still be

considered accurate. Several figures presented in Chapter 5 that display deviation in

maximum deflection reach deviation values larger than this 15% limit before Z

decreases to 3𝑓𝑡⁄𝑙𝑏 1⁄3 .

The results from the analyses displayed in this research point towards the fact that

establishing a simple boundary for when SDOF analysis becomes inaccurate is not as

simple as stating a Z value. Results from Figures 5.2.3.1 and 5.2.3.2 for 15 psf added

weight and a 200 lb charge show that the MDOF solution can deviate 15% to 20% (for

Bernoulli and Timoshenko elements, respectively) from the SDOF solution at a Z of

9𝑓𝑡⁄𝑙𝑏 1⁄3 . This is three times the accepted limit. Using other sets of inputs, the

deviation can be much smaller for a Z value down to 5𝑓𝑡⁄𝑙𝑏 1⁄3 for 81.25 psf added

weight and a 25 lb charge. Several factors affect how MDOF analysis may deviate

from SDOF analysis. Different combinations of charges and standoffs may produce the

100
same Z value, but will have different values for deviation in maximum deflection. This

fact alone is enough to prove that using Z as a limit is not appropriate. A better way

may be to create charge-standoff surfaces. This will eliminate the issue of different

combinations of charge and standoff that have the same Z leading to different

amounts of deviation. However even this surface method may be too simple, since

this neglects other factors that are important to consider. This includes strain rate

effects, which will change the dynamic increase factor, and damping, which will

change the results.

The effect of shear deformation and distributed mass are two factors that work

against each other, but can produce very different results from SDOF analysis as well.

Conventional SDOF analysis does not take into account the effects of shear

deformation, but it may in fact be a significant proportion of the deformation of the

element, which would increase the overall deflection. With the implementation of an

MDOF system, the mass becomes distributed along the element, instead of lumped

into one point, as is done in SDOF analysis. This distribution will tend to decrease the

overall deflection of the element, due to the inertial forces being distributed to each

degree of freedom. These forces balance with each stiffness force at each degree of-

freedom, which will decrease each degree of freedom’s movement compared to an

equivalent SDOF system. Although shear deformation and distributed mass work

against each other, it is hard to tell by how much, and the result could be an overall

101
increase or decrease in maximum deflection compared to that of the SDOF system.

For this reason, Z is too simple of a parameter to use as a limit for SDOF accuracy.

As load intensifies, SDOF and MDOF solutions for pinned-pinned conditions vary in

magnitude. However, the shape of the deflected element remains fairly similar.

However, for fixed-fixed conditions, there is a point where the MDOF solution will

yield at points along its length not predicted by static analysis. This causes the rest of

the solution (post-yield) for MDOF analysis to be very different than for SDOF analysis.

This change in overall shape at first yield can occur at 5.8𝑓𝑡⁄𝑙𝑏 1⁄3 , which

approximately twice the currently accepted limit.

6.4 Future Research

Several areas of the research included in this paper have room to be expanded upon

and further developed. It is very imperative that the model being used be as accurate

as possible. Therefore, more blast load testing needs to be recorded for further

verification of the MDOF model. Damping and dynamic increase factors play a large

role in the sensitivity of the model. Therefore, it is very important that these factors

be set to values that most closely represent the testing results.

The MDOF model used for this research has limitations as well. A more accurate

model can be used to achieve more accurate results. This model could account for

102
material nonlinearity, since the model for this research used an elastic-perfectly plastic

material. Steel is to a certain extent elastic-perfectly plastic, but only if the strain in

the steel does not get high enough to cause strain hardening. Much of the analyses in

this research have large enough strains to cause strain hardening, yet this was

neglected. However, since strain hardening is also neglected in SDOF analysis, the

comparison is valid. Additionally, yielding was lumped at the nodes for MDOF analysis.

In reality, there is a region of plasticity that spreads outward along the element from

the point of initial yielding. Fiber elements may be used to better represent this

plastic flow. Additionally, fiber elements would be able to distinguish between initial

yielding of the outermost fiber and full yielding of the section, which would change the

results of an analysis.

Only steel material was used in this research, for simplicity in analysis. The result of

reinforced concrete elements under blast load is a topic that can be explored, as the

resistance curves for these elements are different than those for steel. Considering

failure modes due to bending about the strong axis only, reinforced concrete sections

can fail in concrete crushing or steel reinforcement rupture. The stress-strain curve for

concrete softens for most of the curve, whereas the stress-strain curve for steel is

linear until a definitive point of yielding (see Figure 2.5.1). Due to the different

material properties of concrete, reinforced concrete sections will have different

responses than steel.

103
For MDOF analysis, geometric effects are accounted for in this research with the use of

the geometric stiffness matrix. However, the co-rotational formulation could be

implemented in the MDOF solution to obtain a more accurate representation of

geometric effects, since this formulation takes into account the deformations from the

chord of each element, instead of only taking into account deformations from a global

coordinate system. The results of the MDOF solution with a co-rotational formulation

may be more precise at predicting response due to blast loading.

Under initial axial loading, the moment capacity of the section will be reduced, due to

initial stress within the element. There are different equations that model this

decrease in capacity, or P-M interaction. Further investigation on the best P-M

interaction to use would increase the accuracy of this research. However, SDOF and

MDOF analysis used in this research use the same P-M interaction equation, so there is

no inconsistency in solutions with regards to P-M interaction.

Lastly, the loading under blast analysis can be further investigated. The load used in

this research is converted from a charge and standoff into a peak pressure and impulse

with a ramp-down shape. This load is then applied to the element as a uniform load.

However, depending on the standoff of the charge, the load may not be uniform along

the length of the element at different time steps. However, results from this research

are consistent between SDOF and MDOF analysis, since they both represent the load

104
as uniform. Additionally, negative-phase can be introduced to the forcing function, to

better represent a real blast.

105
References

AISC (2011), Steel Construction Manual Fourteenth Edition. American Institute of Steel
Construction.

ASCE/SEI 59-11 (2011), Blast Protection of Buildings. American Society of Civil

Engineers.

Biggs, J. M. (1964), Introduction to Structural Dynamics. McGraw-Hill, Inc.

Chopra, A.K. (1995), Dynamics of Structures. Pearson Education, Inc.

FEMA 356 (2000), Prestandard and Commentary for the Seismic Rehabilitation of
Buildings. American Society of Civil Engineers.

Friedlander, F. G. (1946), “The diffraction of sound pulses. I. Diffraction by a semi-

infinite plate.” Proc. Roy. Soc. Lond. A, pp. 186, 322-344.

Gavin, Henri P (2014), Structural Element Stiffness, Mass, and Damping Matrices.
Department of Civil and Environmental Engineering, Duke University.

Logan, Daryl L (2012), “A First Course in the Finite Element Method.” Cengage
Learning.

Nassr, A.A. (2012), “Experimental Performance of Steel Beams under Blast Loading.”
Journal of Performance of Constructed Facilities. American Society of Civil Engineers.
Vol. 26, Issue 5, pp. 600-619.

Przemieniecki, J. S. (1968), Theory of Matrix Structural Analysis. McGraw-Hill, Inc.

US Army Corps of Engineers (2008), Methodology Manual for the Single-Degree-of-

Freedom Blast Effects Design Spreadsheets (SBEDS). U.S. Army Corps of Engineers
Protective Design Center Technical Report.

Yokoyama, T. (2014, November 4), “SDOF Limits on Static Deflected Shape

Assumptions.” Hinman Consulting Engineers Internal Technical Brief. Retrieved by
request from [email protected].

Yokoyama, T. (2014), “Limits to Deflected Shape Assumptions of the SDOF

Methodology for Analyzing Structural Components Subject to Blast Loading.” Journal
of Performance of Constructed Facilities. American Society of Civil Engineers. Vol. 29,
Issue 5.

106
Yokoyama, T. (2011), Verification and Expansion of Single-Degree-of-Freedom
Transformation Factors for Beams Using a Multi-Degree-of-Freedom Non-Linear
Numerical Analysis Method. Master of Science Thesis, Department of Architectural
Engineering, California Polytech State University, San Luis Obispo, California.

107
Appendix A: MATLAB Code for SBEDS Replica (SDOF Solver)

Main Program (Executable)

clc
clear

%% For SBEDs Check (Enter the result of an SBEDS analysis to compute

error

y_yield_SBED=0;
y_yield2_SBED=0;
rotation_SBED=0;
ymax_SBED=0;
%duct_SBED=0;

%% inputs

del_t =.03; % time increment for Newmark (ms)

p_giv=14.6161; % input pressure of pulse (psi)

t_giv=15.2042; %length of pulse (ms)

bc='pinpin'; %set as pinpin, fixpin, fixfix, or cantil

load_type = 'unifrm'; % set as unifrm for uniform, pointd for point

zeta=0.02; %damping ratio of element (enter as decimal)

axial=0; %axial load in pounds

L=10; %length of beam (ft)

B=5; %tributary width that gets transferred to element (ft) if

distributed load

cross_area=3.54; %cross-sectional are (in^2)

A=B*L; %tributary area that gets transferred to element (ft^2) if

point load

E=29000; %modulus of elasticity (ksi)

fy=50; %yield strength (ksi)

stat=1.05; %static increase factor for yield strength

dyn=1.19; %dynamic increase factor for yield strength

108
I=53.8; %moment of inertia (inches^4)

Z = 12.6; %Plastic section modulus (in^3)

weight=12; %weight of element in (lb/ft)

weight_supported=0; %supported weight in psf

steps = 2000; % number of steps for Newmark

u_init=0; % initial displ

v_init=0; % initial vel (ft/s)

gam=1/2; %factor for Newmark

beta=1/4; %factor for Newmark (between 1/4 and 1/6)

%% unit conversions to s, lb, in

L=L*12;

B=B*12;

A=A*144;

weight=weight/12;

weight_supported=weight_supported/144;

%%

[t p load] = Load(t_giv,p_giv,del_t,steps,load_type,B,A); % creates

linearly decreasing pressure

g=0.0003864;

weight=weight*L/A+weight_supported;

m=weight/g; %convert weight/in to mass in psi with g in in/s^2

fy=fy*dyn*stat;

rg=sqrt(I/cross_area);

if bc=='cantil'
klr_fact=2;

109
elseif bc=='pinpin'
klr_fact=1;
elseif bc=='fixpin'
klr_fact=.7;
elseif bc=='fixfix'
klr_fact=.5;
end

Klr=klr_fact*L/rg;

Cc=Klr/pi*sqrt(fy/E);

if Cc<=1.5
Pcr=(.658^(Cc^2))*fy*1000*cross_area;
else
Pcr=.877/Cc^2*fy*1000*cross_area;
end

Mpset(1)=fy*1000*Z;

Mpset(2)=(1-axial/Pcr)*fy*1000*Z;

Mpset(3)=(1-axial/(fy*1000*cross_area))*1.18*Z*fy*1000;

Mp=min(Mpset);

Mp=Mp/1000;

axial=axial/B;

if load_type=='pointd' %for P-delta effects

if bc=='cantil'
k_fact=1/L^2;
axial=axial*k_fact;
else
k_fact=4/L^2;
axial=axial*k_fact;
end

elseif load_type=='unifrm'

if bc=='cantil'
k_fact=2/L^2;
axial=axial*k_fact;
else
k_fact=8/L^2;
axial=axial*k_fact;
end

110
end

[k_m k_F k R F m k_V_RF] = Biggs(L, bc, load_type, E, I, Mp, load, m,

A); % creates Biggs Factors

%%

wn=sqrt(k/m); %natural freq

wd=wn*sqrt(1-zeta^2); %damped freq

c=zeta*2*m*wn; %damping coeff

[u, v, a, alldata, V] = const_vel(bc, load_type, u_init, v_init,

m, c, k, k_F, k_m, R, t, F, zeta, wn, wd, steps, del_t, p, A, k_V_RF,
axial);

%figure(1)
%plot(t*1000,p*1000)
%title('Load')
%xlabel('time (ms)')
%ylabel('Pressure (psi)')

%figure(2)
%hold on
%plot(t*1000,u,'r');
%plot(t*1000,u_actual);
%grid on
%legend('Newmark', 'Actual');
%title('Displacements')
%xlabel('Time (ms)')
%ylabel('Deflection (in)')
%hold off

%figure(3)
%plot(t*1000,u_residual);
%title('Residual')
%xlabel('Time (ms)')
%ylabel('Deflection (in)')

ymax=max(u)
Vmax=max(V)

111
%ymax=max(u_actual);

half=L/2;

if bc=='cantil'
rotation=atan(ymax/L);

else
rotation=atan(ymax/half);
end

rotation=rotation*180/pi;

R=R/A*1000; %converts R values to psi.

y_yield=R(1)/(k(1));

y_yield2=y_yield+(R(2)-R(1))/(k(2));

%if duct==1
% ductility=ymax/y_yield;

%elseif duct==2
% ductility=ymax/y_yield2;
%end

er(1)=Error(y_yield, y_yield_SBED); %first hinge displacement

sdf=1;

if bc=='pinpin'

sdf=0;
end

if bc=='cantil'

sdf=0;
end

if bc=='fixfix'
if load_type=='pointd'

sdf=0;
end
end

112
if sdf==0

er(2)=0;

else

er(2)=Error(y_yield2, y_yield2_SBED); %2nd hinge displacement

end

er(4)=Error(rotation, rotation_SBED); % total rotation

er(3)=Error(ymax, ymax_SBED); %max displacement
%er(5)=Error(ductility, duct_SBED) %ductility

Load Function

function [t p load] = Load(t_giv,p_giv,del_t,steps,load_type,B,A)

t(1)=0;
p(1)=p_giv;

for i=2:1:steps

if t(i-1)<t_giv-del_t

t(i)=t(i-1)+del_t;
p(i)=p(i-1)-p_giv/t_giv*del_t;

else

t(i)=t(i-1)+del_t;
p(i)=0;

end

end

if load_type == 'unifrm'
load=p.*B;
end

if load_type == 'pointd'
load=p.*A;

end

113
Biggs Function

function [k_m k_F k R F m k_V_RF] = Biggs(L, bc, load_type, E, I, Mp,

load, m, A)

if bc == 'pinpin'

if load_type == 'unifrm'

k_F=[.64 .64 .5];

k_m=[.5 .5 .33];
R=[0 8*Mp/L];
k=[384*E*I/5/L^3 384*E*I/5/L^3 2*10^-6/1000*A]/A*1000;
k_V_RF=[.39 .11; .39 .11; .38 .12];

elseif load_type == 'pointd'

k_F=[1 1 1];
k_m=[.49 .49 .33];
R=[0 4*Mp/L];
k=[48*E*I/L^3 48*E*I/L^3 2*10^-6/1000*A]/A*1000;
k_V_RF=[.78 -.28; .78 -.28; .75 -.25];

end

elseif bc == 'fixpin'

if load_type == 'unifrm'

k_F=[.58 .64 .5];

k_m=[.45 .5 .33];
R=[8*Mp/L 12*Mp/L];
k=[185*E*I/L^3 384*E*I/5/L^3 2*10^-6/1000*A]/A*1000;
k_V_RF=[.43 .19; .43 .19; .46 .12];

elseif load_type == 'pointd'

k_F=[1 1 1];
k_m=[.43 .49 .33];
R=[16*Mp/3/L 6*Mp/L];
k=[107*E*I/L^3 48*E*I/L^3 2*10^-6/1000*A]/A*1000;
k_V_RF=[.97 -.28; .97 -.28; .92 -.25];

114
 end

elseif bc == 'fixfix'

if load_type == 'unifrm'

k_F=[.53 .64 .5];

k_m=[.41 .5 .33];
R=[12*Mp/L 16*Mp/L];
k=[384*E*I/L^3 384*E*I/5/L^3 2*10^-6/1000*A]/A*1000;
k_V_RF=[.36 .14; .39 .11; .38 .12];

elseif load_type == 'pointd'

k_F=[1 1 1];
k_m=[.37 .37 .33];
R=[0 8*Mp/L];
k=[192*E*I/L^3 192*E*I/L^3 2*10^-6/1000*A]/A*1000;
k_V_RF=[.71 -.21; .71 -.21; .75 -.25];

end

elseif bc == 'cantil'

if load_type == 'unifrm'

k_F=[.4 .4 .5];
k_m=[.26 .26 .33];
R=[0 2*Mp/L];
k=[8*E*I/L^3 8*E*I/L^3 2*10^-6/1000*A]/A*1000;
k_V_RF=[.69 .31; .69 .31; .75 .25];

elseif load_type == 'pointd'

k_F=[1 1 1];
k_m=[.24 .24 .33];
R=[0 Mp/L];
k=[3*E*I/L^3 3*E*I/L^3 2*10^-6/1000*A]/A*1000;
k_V_RF=[1.74 -.74; 1.74 -.74; 1.5 -.5];

end

115
end

k_F=transpose(k_F);

if load_type == 'unifrm'

F=load*L;

elseif load_type == 'pointd'

F=load;

end

Constant Velocity Function

function [u, v, a, alldata, V] = const_vel(bc, load_type, u_init,

v_init, m, c, k, k_F, k_m, R, t, F, zeta, wn, wd, steps, del_t, p, A,
k_V_RF, axial)

R=R/A*1000; %converts R values to psi.

y_yield=R(1)/(k(1));

y_yield2=y_yield+(R(2)-R(1))/(k(2));

skip2=0;

Rmax=.95*R(2);

KLM=k_m./transpose(k_F);

c_cr=2*sqrt(k(1)*m*KLM(1));

c_use=zeta*c_cr;

c_min=.05/100*c_cr;

R_datum=0;

%% initial constant velocity calcs

i=1;

116
t(i)=0;

u(i)=u_init;

v(i)=v_init;

equiv_pdelt(i)=axial*u(i);

if u(i)<=y_yield

c=c_use;
Res(i)=u(i)*k(1);
a(i)=(p(i)+equiv_pdelt(i)-c*v(i)-Res(i))/(m*KLM(1));
stiffness(i)=k(1);
factor(i)=KLM(1);
V(i)=k_V_RF(1,1)*Res(i)+k_V_RF(1,2)*p(i);

elseif u(i)>y_yield && u(i)<y_yield2

Res(i)=R(1)+(u(i)-y_yield)*k(2);

if Res(i)<Rmax
c=c_use;
elseif Res(i)>=Rmax
c=c_min;
end

a(i)=(p(i)+equiv_pdelt(i)-c*v(i)-Res(i))/(m*KLM(2));
stiffness(i)=k(2);
factor(i)=KLM(2);
V(i)=k_V_RF(2,1)*Res(i)+k_V_RF(2,2)*p(i);

elseif u(i)>=y_yield2

Res(i)=R(2)+(u(i)-y_yield2)*k(3);
c=c_min;
a(i)=(p(i)+equiv_pdelt(i)-c*v(i)-Res(i))/(m*KLM(3));
stiffness(i)=k(3);
factor(i)=KLM(3);
V(i)=k_V_RF(3,1)*Res(i)+k_V_RF(3,2)*p(i);

end

d_u(i)=.5*a(i)*del_t^2+v(i)*del_t;

alldata(i,1)=t(i);
alldata(i,2)=p(i);
alldata(i,3)=equiv_pdelt(i);
alldata(i,4)=u(i);

117
alldata(i,5)=v(i);
alldata(i,6)=a(i);
alldata(i,7)=stiffness(i);
alldata(i,8)=Res(i);
alldata(i,9)=d_u(i);
alldata(i,10)=factor(i);
alldata(i,11)=V(i);

i=2;

%% Main Constant Velocity loop

while i<=steps %loop for Constant Velocity iterations

t(i)=t(i-1)+del_t;

u(i)=u(i-1)+d_u(i-1);

v(i)=(u(i)-u(i-1))/del_t+.5*a(i-1)*del_t;

equiv_pdelt(i)=axial*u(i);

R_check=abs(Res(i-1)-R_datum); %amount of change in resistance

from inflection point ("0 strain")

%%%%%elastic region%%%%%%

if R_check<=R(1) || skip2==1

Res(i)=Res(i-1)+(u(i)-u(i-1))*k(1);

if Res(i)>R(2) %prevents going past maximum resistance

Res(i)=R(2);

end

if Res(i)<-R(2)%prevents going past maximum resistance

(negative)

Res(i)=-R(2);

end

118
 if sign(Res(i))==sign(v(i)) %decreases damping when close to
yielding

if Res(i)<Rmax
c=c_use;
elseif Res(i)>=Rmax
c=c_min;
end

else

c=c_use;
end

a(i)=(p(i)+equiv_pdelt(i)-c*v(i)-Res(i))/(m*KLM(1));
stiffness(i)=k(1);
factor(i)=KLM(1);
V(i)=k_V_RF(1,1)*Res(i)+k_V_RF(1,2)*p(i);

if sign(v(i))~=sign(v(i-1)) %once an inflection point occurs

(sign change of velocity), R_datum is set as the current resistance

R_datum=Res(i);

end

%with skip2, it forces stage 1 on turnaround while the net

%Resistance is still larger than yield Resistance. Resets the
%datum
if sign(Res(i))~=sign(Res(i-1))
skip2=0;
R_datum=0;

end

%%%%%elasto-plastic region%%%%%%

elseif R_check>R(1) && R_check<R(2)

Res(i)=Res(i-1)+(u(i)-u(i-1))*k(2);

if Res(i)>R(2) %prevents going past maximum resistance

Res(i)=R(2);

end

119
 if Res(i)<-R(2)%prevents going past maximum resistance
(negative)

Res(i)=-R(2);

end

if Res(i)<Rmax %decreases damping when close to yielding

c=c_use;
elseif Res(i)>=Rmax
c=c_min;
end

a(i)=(p(i)+equiv_pdelt(i)-c*v(i)-Res(i))/(m*KLM(2));
stiffness(i)=k(2);
factor(i)=KLM(2);
V(i)=k_V_RF(2,1)*Res(i)+k_V_RF(2,2)*p(i);

if sign(v(i))~=sign(v(i-1)) %once an inflection point occurs

(sign change of velocity), R_datum is set as the current resistance

R_datum=Res(i);
skip2=1;

end

% %%%%%plastic region%%%%%%

elseif R_check>=R(2)

Res(i)=Res(i-1)+(u(i)-u(i-1))*k(3);

c=c_min;
a(i)=(p(i)+equiv_pdelt(i)-c*v(i)-Res(i))/(m*KLM(3));
stiffness(i)=k(3);
factor(i)=KLM(3);
V(i)=k_V_RF(3,1)*Res(i)+k_V_RF(3,2)*p(i);

if sign(v(i))~=sign(v(i-1)) %once an inflection point occurs

(sign change of velocity), R_datum is set as the current resistance

R_datum=Res(i);
skip2=1;

end

120
 end

d_u(i)=u(i)-u(i-1)+a(i)*del_t^2;

alldata(i,1)=t(i);
alldata(i,2)=p(i);
alldata(i,3)=equiv_pdelt(i);
alldata(i,4)=u(i);
alldata(i,5)=v(i);
alldata(i,6)=a(i);
alldata(i,7)=stiffness(i);
alldata(i,8)=Res(i);
alldata(i,9)=d_u(i);
alldata(i,10)=factor(i);
alldata(i,11)=V(i);

i=i+1;

end

figure(1)
subplot(2,2,1)
plot(t,a)
title('acc')
xlabel('time')
ylabel('Acceleration')
grid on

subplot(2,2,2)
plot(t,v)
title('vel')
xlabel('time')
ylabel('Velocity')
grid on

subplot(2,2,3)
plot(t,u)
title('displ')
legend('Displacement');
xlabel('time')
ylabel('displacement')
grid on

121
Error Function

function [error] = Error(i, j)

error=(j-i)/j*100;

122
Appendix B: SDOF Solver Verifications (Time-History Plots)

All verifications use 15 ft Length and 0.5 ft Tributary Width

W10X12 Pinned-Pinned Point Load (30 psi in 17.8 ms) 0% Damping

5
Deflection (in)

3 MATLAB DISPL
SBEDS Displ
2

0
0 50 100 150
Time (ms)

20

15

10
MATLAB RESISTANCE
Internal Moment (psi)

5 SBEDs RESISTANCE

0
0 20 40 60 80 100 120 140 160
-5

-10

-15

-20 Time (ms)

123
W10X12 Pinned-Pinned Point Load (30 psi in 17.8 ms) 5% Damping

4
MATLAB DISPL
3 SBEDS Displ

0
0 50 100 150

20

15

10

5
MATLAB RESISTANCE
0
SBEDs RESISTANCE
0 20 40 60 80 100 120 140 160
-5

-10

-15

-20

124
W10X12 Pinned-Pinned Uniform Load (40 psi in 17.8 ms) 0% Damping

3.5

2.5

2 MATLAB DISPL
SBEDS Displ
1.5

0.5

0
0 50 100 150

40

30

20

10
MATLAB RESISTANCE
0
SBEDs RESISTANCE
0 20 40 60 80 100 120 140 160
-10

-20

-30

-40

125
W10X12 Pinned-Pinned Uniform Load (40 psi in 17.8 ms) 5% Damping

3.5

2.5

2
MATLAB DISPL
1.5 SBEDS Displ

0.5

0
0 50 100 150

40

30

20

10 MATLAB RESISTANCE
SBEDs RESISTANCE
0
0 20 40 60 80 100 120 140 160
-10

-20

-30

126
W10X12 Fixed-Pinned Point Load (100 psi in 17.8 ms) 0% Damping

80

70

60

50

40 MATLAB DISPL
SBEDS Displ
30

20

10

0
0 20 40 60 80 100 120

30

20

10

MATLAB RESISTANCE
0
SBEDs RESISTANCE
0 20 40 60 80 100 120

-10

-20

-30

127
W10X12 Fixed-Pinned Point Load (100 psi in 17.8 ms) 5% Damping

80

70

60

50

40 MATLAB DISPL
SBEDS Displ
30

20

10

0
0 20 40 60 80 100 120

30
25
20
15
10
5 MATLAB RESISTANCE

0 SBEDs RESISTANCE
0 20 40 60 80 100 120
-5
-10
-15
-20
-25

128
W10X12 Fixed-Fixed Point Load (50 psi in 17.8 ms) 0% Damping

4 MATLAB DISPL
SBEDS Displ
3

0
0 10 20 30 40 50 60 70

40

30

20

10
MATLAB RESISTANCE
0
SBEDs RESISTANCE
0 10 20 30 40 50 60 70
-10

-20

-30

-40

129
W10X12 Fixed-Fixed Point Load (50 psi in 17.8 ms) 5% Damping

4
MATLAB DISPL
3 SBEDS Displ

0
0 10 20 30 40 50 60 70

40

30

20

10 MATLAB RESISTANCE
SBEDs RESISTANCE
0
0 10 20 30 40 50 60 70
-10

-20

-30

130
W10X12 Fixed-Fixed Uniform Load (100 psi in 17.8 ms) 0% Damping

4 MATLAB DISPL
SBEDS Displ
3

0
0 10 20 30 40 50 60 70

80

60

40

20 MATLAB RESISTANCE
SBEDs RESISTANCE
0
0 10 20 30 40 50 60 70
-20

-40

-60

131
W10X12 Fixed-Fixed Uniform Load (100 psi in 17.8 ms) 5% Damping

4
MATLAB DISPL
3 SBEDS Displ

0
0 10 20 30 40 50 60 70

80

60

40

20 MATLAB RESISTANCE
SBEDs RESISTANCE
0
0 10 20 30 40 50 60 70
-20

-40

-60

132
W10X12 Cantilever Point Load (15 psi in 17.8 ms) 0% Damping

30

25

20

15 MATLAB DISPL
SBEDS Displ
10

0
0 100 200 300 400 500

1
MATLAB RESISTANCE
0
SBEDs RESISTANCE
0 50 100 150 200 250 300 350 400 450
-1

-2

-3

-4

-5

133
W10X12 Cantilever Point Load (15 psi in 17.8 ms) 5% Damping

30

25

20

15 MATLAB DISPL
SBEDS Displ
10

0
0 100 200 300 400 500

1 MATLAB RESISTANCE
SBEDs RESISTANCE
0
0 50 100 150 200 250 300 350 400 450
-1

-2

-3

-4

134
W10X12 Cantilever Uniform Load (20 psi in 17.8 ms) 0% Damping

5
MATLAB DISPL
4
SBEDS Displ
3

0
0 100 200 300 400 500

10

2
MATLAB RESISTANCE
0
SBEDs RESISTANCE
0 50 100 150 200 250 300 350 400 450
-2

-4

-6

-8

-10

135
W10X12 Cantilever Uniform Load (20 psi in 17.8 ms) 5% Damping

5
MATLAB DISPL
4
SBEDS Displ
3

0
0 100 200 300 400 500

10

2 MATLAB RESISTANCE
SBEDs RESISTANCE
0
0 50 100 150 200 250 300 350 400 450
-2

-4

-6

-8

136
W10X12 Pinned-Pinned Point Load (70 psi in 17.8 ms) 0% Damping 30 psf Added

Weight

25

20

15
MATLAB DISPL

10 SBEDS Displ

0
0 50 100 150 200 250

20

15

10

5
MATLAB RESISTANCE
0
SBEDs RESISTANCE
0 50 100 150 200 250
-5

-10

-15

-20

137
W10X12 Pinned-Pinned Point Load (70 psi in 17.8 ms) 5% Damping 30 psf Added

Weight

25

20

15
MATLAB DISPL

10 SBEDS Displ

0
0 50 100 150 200 250

20

15

10

5
MATLAB RESISTANCE
0
SBEDs RESISTANCE
0 50 100 150 200 250
-5

-10

-15

-20

138
W10X12 Pinned-Pinned Uniform Load (70 psi in 17.8 ms) 0% Damping 30 psf Added

Weight

4
MATLAB DISPL
3 SBEDS Displ

0
0 50 100 150 200 250

40

30

20

10
MATLAB RESISTANCE
0
SBEDs RESISTANCE
0 50 100 150 200 250
-10

-20

-30

-40

139
W10X12 Pinned-Pinned Uniform Load (70 psi in 17.8 ms) 5% Damping 30 psf Added

Weight

4
MATLAB DISPL
3 SBEDS Displ

0
0 50 100 150 200 250

40

30

20

10
MATLAB RESISTANCE
0
SBEDs RESISTANCE
0 50 100 150 200 250
-10

-20

-30

-40

140
W10X12 Fixed-Pinned Point Load (70 psi in 17.8 ms) 0% Damping 30 psf Added Weight

16

14

12

10

8 MATLAB DISPL
SBEDS Displ
6

0
0 50 100 150

30

20

10

MATLAB RESISTANCE
0
SBEDs RESISTANCE
0 20 40 60 80 100 120 140 160

-10

-20

-30

141
W10X12 Fixed-Pinned Point Load (70 psi in 17.8 ms) 5% Damping 30 psf Added Weight

16

14

12

10

8 MATLAB DISPL
SBEDS Displ
6

0
0 50 100 150

30
25
20
15
10
5 MATLAB RESISTANCE

0 SBEDs RESISTANCE
0 20 40 60 80 100 120 140 160
-5
-10
-15
-20
-25

142
W10X12 Fixed-Pinned Uniform Load (70 psi in 17.8 ms) 0% Damping 30 psf Added

Weight

3.5

2.5

2
MATLAB DISPL
1.5 SBEDS Displ

0.5

0
0 50 100 150

60
50
40
30
20
10 MATLAB RESISTANCE

0 SBEDs RESISTANCE
0 20 40 60 80 100 120 140 160
-10
-20
-30
-40
-50

143
W10X12 Fixed-Pinned Uniform Load (70 psi in 17.8 ms) 5% Damping 30 psf Added

Weight

3.5

2.5

2
MATLAB DISPL
1.5 SBEDS Displ

0.5

0
0 50 100 150

60
50
40
30
20
10 MATLAB RESISTANCE

0 SBEDs RESISTANCE
0 20 40 60 80 100 120 140 160
-10
-20
-30
-40
-50

144
W10X12 Fixed-Fixed Point Load (70 psi in 17.8 ms) 0% Damping 30 psf Added Weight

12

10

6 MATLAB DISPL
SBEDS Displ
4

0
0 20 40 60 80 100 120

40

30

20

10
MATLAB RESISTANCE
0
SBEDs RESISTANCE
0 20 40 60 80 100 120
-10

-20

-30

-40

145
W10X12 Fixed-Fixed Point Load (70 psi in 17.8 ms) 5% Damping 30 psf Added Weight

10
9
8
7
6
5 MATLAB DISPL

4 SBEDS Displ

3
2
1
0
0 20 40 60 80 100 120

40

30

20

10
MATLAB RESISTANCE
0
SBEDs RESISTANCE
0 20 40 60 80 100 120
-10

-20

-30

-40

146
W10X12 Fixed-Fixed Uniform Load (70 psi in 17.8 ms) 0% Damping 30 psf Added

Weight

1.8

1.6

1.4

1.2

1
MATLAB DISPL
0.8
SBEDS Displ
0.6

0.4

0.2

0
0 20 40 60 80 100 120

80

60

40

20 MATLAB RESISTANCE
SBEDs RESISTANCE
0
0 20 40 60 80 100 120
-20

-40

-60

147
W10X12 Fixed-Fixed Uniform Load (70 psi in 17.8 ms) 5% Damping 30 psf Added

Weight

1.6

1.4

1.2

0.8 MATLAB DISPL

SBEDS Displ
0.6

0.4

0.2

0
0 20 40 60 80 100 120

80

60

40

20 MATLAB RESISTANCE
SBEDs RESISTANCE
0
0 20 40 60 80 100 120
-20

-40

-60

148
W10X12 Cantilever Point Load (40 psi in 17.8 ms) 0% Damping 30 psf Added Weight

90

80

70

60

50
MATLAB DISPL
40
SBEDS Displ
30

20

10

0
0 100 200 300 400 500 600 700

1
MATLAB RESISTANCE
0
SBEDs RESISTANCE
0 100 200 300 400 500 600 700
-1

-2

-3

-4

-5

149
W10X12 Cantilever Point Load (40 psi in 17.8 ms) 5% Damping 30 psf Added Weight

80

70

60

50

40 MATLAB DISPL
SBEDS Displ
30

20

10

0
0 100 200 300 400 500 600 700

1 MATLAB RESISTANCE
SBEDs RESISTANCE
0
0 100 200 300 400 500 600 700
-1

-2

-3

-4

150
W10X12 Cantilever Uniform Load (40 psi in 17.8 ms) 0% Damping 30 psf Added Weight

16

14

12

10

8 MATLAB DISPL
SBEDS Displ
6

0
0 100 200 300 400 500 600 700

10

2
MATLAB RESISTANCE
0
SBEDs RESISTANCE
0 100 200 300 400 500 600 700
-2

-4

-6

-8

-10

151
W10X12 Cantilever Uniform Load (40 psi in 17.8 ms) 5% Damping 30 psf Added Weight

14

12

10

8
MATLAB DISPL
6 SBEDS Displ

0
0 100 200 300 400 500 600 700

10

2 MATLAB RESISTANCE
SBEDs RESISTANCE
0
0 100 200 300 400 500 600 700
-2

-4

-6

-8

152
W10X12 Pinned-Pinned Point Load (40 psi in 17.8 ms) 0% Damping 10 kip Axial Load

18

16

14

12

10
MATLAB DISPL
8
SBEDS Displ
6

0
0 50 100 150

20

15

10

5 MATLAB RESISTANCE
SBEDs RESISTANCE
0
0 20 40 60 80 100 120 140 160
-5

-10

-15

153
W10X12 Pinned-Pinned Uniform Load (70 psi in 17.8 ms) 0% Damping 10 kip Axial

Load

16

14

12

10

8 MATLAB DISPL
SBEDS Displ
6

0
0 50 100 150

40

30

20

10 MATLAB RESISTANCE
SBEDs RESISTANCE
0
0 20 40 60 80 100 120 140 160
-10

-20

-30

154
W10X12 Fixed-Pinned Point Load (40 psi in 17.8 ms) 0% Damping 10 kip Axial Load

4 MATLAB DISPL
SBEDS Displ
3

0
0 20 40 60 80 100 120

30
25
20
15
10
5 MATLAB RESISTANCE

0 SBEDs RESISTANCE
0 20 40 60 80 100 120
-5
-10
-15
-20
-25

155
W10X12 Fixed-Pinned Uniform Load (70 psi in 17.8 ms) 0% Damping 10 kip Axial Load

4
MATLAB DISPL
3 SBEDS Displ

0
0 20 40 60 80 100 120

60

50

40

30

20
MATLAB RESISTANCE
10
SBEDs RESISTANCE
0
0 20 40 60 80 100 120
-10

-20

-30

-40

156
W10X12 Fixed-Fixed Point Load (40 psi in 17.8 ms) 0% Damping 10 kip Axial Load

3.5

2.5

2 MATLAB DISPL
SBEDS Displ
1.5

0.5

0
0 10 20 30 40 50 60 70

40

30

20

MATLAB RESISTANCE
10
SBEDs RESISTANCE

0
0 10 20 30 40 50 60 70

-10

-20

157
W10X12 Fixed-Fixed Uniform Load (70 psi in 17.8 ms) 0% Damping 10 kip Axial Load

2.5

1.5 MATLAB DISPL

SBEDS Displ
1

0.5

0
0 10 20 30 40 50 60 70

70

60

50

40

30
MATLAB RESISTANCE
20
SBEDs RESISTANCE
10

0
0 10 20 30 40 50 60 70
-10

-20

-30

158
W10X12 Cantilever Point Load (40 psi in 17.8 ms) 0% Damping 1 kip Axial Load

200
180
160
140
120
100 MATLAB DISPL

80 SBEDS Displ

60
40
20
0
0 100 200 300 400 500

2
MATLAB RESISTANCE
1
SBEDs RESISTANCE
0
0 50 100 150 200 250 300 350 400 450
-1

-2

-3

159
W10X12 Cantilever Uniform Load (40 psi in 17.8 ms) 0% Damping 10 kip Axial Load

50
45
40
35
30
25 MATLAB DISPL

20 SBEDS Displ

15
10
5
0
0 100 200 300 400 500

4 MATLAB RESISTANCE
SBEDs RESISTANCE
3

0
0 50 100 150 200 250 300 350 400 450

160
W40X655 Pinned-Pinned Point Load (10,000 psi in 17.8 ms) 5% Damping

70

60

50

40
MATLAB DISPL
30 SBEDS Displ

20

10

0
0 10 20 30 40 50

5000

4000

3000

2000

1000 MATLAB RESISTANCE

SBEDs RESISTANCE
0
0 5 10 15 20 25 30 35 40 45
-1000

-2000

-3000

-4000

161
W40X655 Pinned-Pinned Uniform Load (10,000 psi in 17.8 ms) 5% Damping

3.5

2.5

2 MATLAB DISPL
SBEDS Displ
1.5

0.5

0
0 10 20 30 40 50

10000

8000

6000

4000 MATLAB RESISTANCE

SBEDs RESISTANCE
2000

0
0 5 10 15 20 25 30 35 40 45
-2000

-4000

162
W40X655 Fixed-Pinned Point Load (10,000 psi in 17.8 ms) 5% Damping

25

20

15
MATLAB DISPL

10 SBEDS Displ

0
0 5 10 15 20 25

7000
6000
5000
4000
3000
2000 MATLAB RESISTANCE

1000 SBEDs RESISTANCE

0
0 5 10 15 20 25
-1000
-2000
-3000
-4000

163
W40X655 Fixed-Pinned Uniform Load (10,000 psi in 17.8 ms) 5% Damping

0.6

0.5

0.4

0.3 MATLAB DISPL

SBEDS Displ
0.2

0.1

0
0 5 10 15 20 25

14000

12000

10000

8000
MATLAB RESISTANCE
6000
SBEDs RESISTANCE
4000

2000

0
0 5 10 15 20 25
-2000

164
W40X655 Fixed-Fixed Point Load (10,000 psi in 17.8 ms) 5% Damping

3 MATLAB DISPL
SBEDS Displ
2

0
0 5 10 15 20 25

9000
8000
7000
6000
5000
4000 MATLAB RESISTANCE

3000 SBEDs RESISTANCE

2000
1000
0
0 5 10 15 20 25
-1000
-2000

165
W40X655 Fixed-Fixed Uniform Load (20,000 psi in 17.8 ms) 5% Damping

5
4.5
4
3.5
3
2.5 MATLAB DISPL

2 SBEDS Displ

1.5
1
0.5
0
0 5 10 15 20 25

18000
16000
14000
12000
10000
8000 MATLAB RESISTANCE

6000 SBEDs RESISTANCE

4000
2000
0
0 5 10 15 20 25
-2000
-4000

166
W40X655 Cantilever Point Load (2,000 psi in 17.8 ms) 5% Damping

16

14

12

10

8 MATLAB DISPL
SBEDS Displ
6

0
0 20 40 60 80 100 120

1500

1000

500
MATLAB RESISTANCE
SBEDs RESISTANCE
0
0 20 40 60 80 100 120

-500

-1000

167
W40X655 Cantilever Uniform Load (2,000 psi in 17.8 ms) 5% Damping

2
1.8
1.6
1.4
1.2
1 MATLAB DISPL

0.8 SBEDS Displ

0.6
0.4
0.2
0
0 20 40 60 80 100 120

2500

2000

1500

1000
MATLAB RESISTANCE
500
SBEDs RESISTANCE
0
0 20 40 60 80 100 120
-500

-1000

-1500

168
W10X12 Pinned-Pinned Point Load (70 psi in 8 ms) 5% Damping

10
9
8
7
6
5 MATLAB DISPL

4 SBEDS Displ

3
2
1
0
0 50 100 150

20

15

10

5
MATLAB RESISTANCE
0
SBEDs RESISTANCE
0 20 40 60 80 100 120 140 160
-5

-10

-15

-20

169
W10X12 Pinned-Pinned Uniform Load (70 psi in 8 ms) 5% Damping

3.5

2.5

2
MATLAB DISPL
1.5 SBEDS Displ

0.5

0
0 50 100 150

40

30

20
MATLAB RESISTANCE
10 SBEDs RESISTANCE

0
0 20 40 60 80 100 120 140 160
-10

-20

-30

-40

170
W10X12 Fixed-Pinned Point Load (70 psi in 8 ms) 5% Damping

4
MATLAB DISPL
3 SBEDS Displ

0
0 20 40 60 80 100 120

30
25
20
15
10
5 MATLAB RESISTANCE

0 SBEDs RESISTANCE
0 20 40 60 80 100 120
-5
-10
-15
-20
-25

171
W10X12 Fixed-Pinned Uniform Load (70 psi in 8 ms) 5% Damping

2.5

1.5

MATLAB DISPL
1
SBEDS Displ

0.5

0
0 20 40 60 80 100 120

-0.5

60
50
40
30
20
10 MATLAB RESISTANCE

0 SBEDs RESISTANCE
0 20 40 60 80 100 120
-10
-20
-30
-40
-50

172
W10X12 Fixed-Fixed Point Load (100 psi in 8 ms) 5% Damping

12

10

6 MATLAB DISPL
SBEDS Displ
4

0
0 10 20 30 40 50 60 70

40

30

20

10
MATLAB RESISTANCE
0
SBEDs RESISTANCE
0 10 20 30 40 50 60 70
-10

-20

-30

-40

173
W10X12 Fixed-Fixed Uniform Load (100 psi in 8 ms) 5% Damping

2.5

1.5
MATLAB DISPL

1 SBEDS Displ

0.5

0
0 10 20 30 40 50 60 70

80

60

40

20 MATLAB RESISTANCE
SBEDs RESISTANCE
0
0 10 20 30 40 50 60 70
-20

-40

-60

174
W10X12 Cantilever Point Load (70 psi in 8 ms) 5% Damping

120

100

80

60 MATLAB DISPL
SBEDS Displ
40

20

0
0 100 200 300 400 500

3
MATLAB RESISTANCE
2
SBEDs RESISTANCE
1

0
0 50 100 150 200 250 300 350 400 450
-1

-2

-3

-4

175
W10X12 Cantilever Uniform Load (70 psi in 8 ms) 5% Damping

20
18
16
14
12
10 MATLAB DISPL

8 SBEDS Displ

6
4
2
0
0 100 200 300 400 500

10

6
MATLAB RESISTANCE
4
SBEDs RESISTANCE
2

0
0 50 100 150 200 250 300 350 400 450
-2

-4

-6

-8

176
Appendix C: MATLAB Code for More Accurate Model (MDOF

Solver)

Main Program (Executable)

clc
clear

ii=sqrt(-1);

[Nodes Node_class ElementNodes ElementProp ElementSec BC Force

Force_distr, Axial_distr, Moment_distr analysis modes lumping
nonstr_mass gamma beta forcecurve del stopt ICu ICv TOL rayleigh axial
bc_type]=Input(); %Brings in data from input file

NumNodes=numel(Nodes(:,1)); %number of nodes

NumElem=numel(ElementNodes(:,1)); %number of elements

% preprocessing

sizBC=size(BC,1); %adds a row of 0's for BC, Force,

Axial_distr, Force_dist, Moment_distr, nonstr_mass, ICs
BC(sizBC+1,:)=zeros;

sizForce=size(Force,1);
Force(sizForce+1,:)=zeros;

sizAxial=size(Axial_distr,1);
Axial_distr(sizAxial+1,:)=zeros;

sizForce_d=size(Force_distr,1);
Force_distr(sizForce_d+1,:)=zeros;

sizMoment=size(Moment_distr,1);
Moment_distr(sizMoment+1,:)=zeros;

sizmass=size(nonstr_mass,1);
nonstr_mass(sizmass+1,:)=zeros;

sizICu=size(ICu,1);
ICu(sizICu+1,:)=zeros;

sizICv=size(ICv,1);

177
ICv(sizICv+1,:)=zeros;

siz1=size(Axial_distr,2); %number of columns of axial_distr

siz2=size(Force_distr,2); %number of columns of force_distr
siz3=size(Moment_distr,2); %number of columns of moment_distr

counter1=1; %counter to loop over Axial_distr

counter2=1; %counter to loop over Force_distr
counter3=1; %counter to loop over Moment_distr

degree1=siz1-1; %degree of polynomial for axial(x)

degree2=siz2-1; %degree of polynomial for force(x)
degree3=siz3-1; %degree of polynomial for moment(x)

for i=1:1:NumElem %preprocessing for elements (fills out full distr.

load)

if i==Axial_distr(counter1,1) %for the size of the constants

(a's), loops and stores into temporary full load vector

Axial_distrfull(i,1)=i;

for count=2:1:siz1
Axial_distrfull(i,count)=Axial_distr(counter1,count);
end

counter1=counter1+1;

else

Axial_distrfull(i,1)=i;

for count=2:1:siz1
Axial_distrfull(i,count)=0;
end

end

if i==Force_distr(counter2,1) %for the size of the constants

(a's), loops and stores into temporary full load vector

Force_distrfull(i,1)=i;

for count=2:1:siz2
Force_distrfull(i,count)=Force_distr(counter2,count);

178
 end

counter2=counter2+1;

else

Force_distrfull(i,1)=i;

for count=2:1:siz2
Force_distrfull(i,count)=0;
end

end

if i==Moment_distr(counter3,1) %for the size of the constants

(a's), loops and stores into temporary full load vector

Moment_distrfull(i,1)=i;

for count=2:1:siz3
Moment_distrfull(i,count)=Moment_distr(counter3,count);
end

counter3=counter3+1;

else

Moment_distrfull(i,1)=i;

for count=2:1:siz3
Moment_distrfull(i,count)=0;
end

end

end

if degree1==0

Axial_distrfull(:,2)=zeros;
end

if degree2==0

Force_distrfull(:,2)=zeros;
end

if degree3==0

179
 Moment_distrfull(:,2)=zeros;
end

counter=1; %counter to loop over Force

counter2=1; %counter for BC's
counter3=1; %counter for nonstr_mass
counter4=1; %counter for ICu
counter5=1; %counter for ICv

position=1; %position in Force vector

position2=1; %position in BC vector
position3=1; %position in nonstr_mass vector
position4=1; %position in ICu vector
position5=1; %position in ICv vector

for i=1:1:NumNodes %preprocessing for nodes

if i==Force(counter,1) %if the node i matches the node from Force

Forcefull(position,1)=Force(counter,2);
Forcefull(1+position,1)=Force(counter,3);
Forcefull(2+position,1)=Force(counter,4);

counter=counter+1;

else %otherwise, put a 0

Forcefull(position,1)=0;
Forcefull(1+position,1)=0;
Forcefull(2+position,1)=0;

end

position=position+3;

if i==BC(counter2,1) %if the node i matches the node from BC

BCfull(position2,1)=BC(counter2,2);
BCfull(1+position2,1)=BC(counter2,3);
BCfull(2+position2,1)=BC(counter2,4);

counter2=counter2+1;

else % otherwise put sqrt(-1) as placeholder

BCfull(position2,1)=ii;
BCfull(1+position2,1)=ii;
BCfull(2+position2,1)=ii;

180
 end

position2=position2+3;

if i==nonstr_mass(counter3,1) %if the node i matches the node

from nonstr_mass

nonstr_massfull(position3,1)=nonstr_mass(counter3,2);
nonstr_massfull(1+position3,1)=nonstr_mass(counter3,3);
nonstr_massfull(2+position3,1)=nonstr_mass(counter3,4);

counter3=counter3+1;

else % otherwise put 0 mass

nonstr_massfull(position3,1)=0;
nonstr_massfull(1+position3,1)=0;
nonstr_massfull(2+position3,1)=0;

end

position3=position3+3;

if i==ICu(counter4,1) %if the node i matches the node from ICu

ICufull(position4,1)=ICu(counter4,2);
ICufull(1+position4,1)=ICu(counter4,3);
ICufull(2+position4,1)=ICu(counter4,4);

counter4=counter4+1;

else % otherwise put 0 IC

ICufull(position4,1)=0;
ICufull(1+position4,1)=0;
ICufull(2+position4,1)=0;

end

position4=position4+3;

if i==ICv(counter5,1) %if the node i matches the node from ICv

ICvfull(position5,1)=ICv(counter5,2);
ICvfull(1+position5,1)=ICv(counter5,3);
ICvfull(2+position5,1)=ICv(counter5,4);

181
 counter5=counter5+1;

else % otherwise put 0 IC

ICvfull(position5,1)=0;
ICvfull(1+position5,1)=0;
ICvfull(2+position5,1)=0;

end

position5=position5+3;

end

BCfull_original=BCfull;

for i=1:1:NumElem %Loops over each element for lengths and material
properties

%% lengths and angles

node1=ElementNodes(i,1); %temporarily stores node 1 for element i

node2=ElementNodes(i,2); %temporarily stores node 2 for element i

x1=Nodes(node1,1); %temporarily stores x coordinate of node 1 for

element i
y1=Nodes(node1,2); %temporarily stores y coordinate of node 1 for
element i
x2=Nodes(node2,1); %temporarily stores x coordinate of node 2 for
element i
y2=Nodes(node2,2); %temporarily stores y coordinate of node 2 for
element i

[ang Length]=Lengths(x1, x2, y1, y2); %Function to calculate

lengths/angles

angle(i)=ang;
L(i)=Length;

%% Material properties

Mat_num=ElementProp(i,1);
Type_num=ElementProp(i,2);

Area=ElementSec(i,1);

182
 Inert=ElementSec(i,2);
Flange=ElementSec(i,3);
Z=ElementSec(i,4);
d_l=ElementSec(i,5);
b_l=ElementSec(i,6);
f_l=ElementSec(i,7);
w_l=ElementSec(i,8);
nodeset=ElementNodes(i,:);

[EM GM DENS Ar Ar_sh Inertia Fy Mpl]=Data(Mat_num, Type_num, Area,

Inert, Flange, Z);

E(i)=EM;
G(i)=GM;
A(i)=Ar;
As(i)=Ar_sh;
I(i)=Inertia;
rho(i)=DENS;
Fyield(i)=Fy;
Mp(i)=Mpl;
Tp_num(i)=Type_num;

%uses first element as section for axial capacity

if i==1

Py=Ar*Fy;
axial=axial*Py

if axial~=0 % P-M reduction in plastic moment (to match

SBEDS)

rg=sqrt(Inertia/Ar);

if bc_type=='cantil'
klr_fact=2;
elseif bc_type=='pinpin'
klr_fact=1;
elseif bc_type=='fixpin'
klr_fact=.7;
elseif bc_type=='fixfix'
klr_fact=.5;
end

Klr=klr_fact*Length*NumElem/rg;

Cc=Klr/pi*sqrt(Fy/EM);

if Cc<=1.5
Pcr=(.658^(Cc^2))*Fy*Ar;
else

183
 Pcr=.877/Cc^2*Fy*Ar;
end

Mpset(1)=Fy*Z;

Mpset(2)=(1-axial/Pcr)*Fy*Z;

Mpset(3)=(1-axial/(Fy*Ar))*1.18*Z*Fy;

Mp_red=min(Mpset)

end

end

if axial~=0

Mp(i)=Mp_red;
end

% if axial~=0
%
% Pboundary=w_l*(d_l-2*f_l)*Fy;
%
% if axial>=Pboundary
%
% NA=(2*b_l*f_l*Fy-axial+w_l*Fy*(d_l-2*f_l))/(2*b_l*Fy);
%
% Mp(i)=b_l*NA*Fy*(d_l-NA);
%
%
% elseif axial<Pboundary
%
% NA=-(axial-d_l*w_l*Fy)/2/w_l/Fy;
%
% Mp(i)=-Fy*(b_l*f_l^2-f_l^2*w_l+w_l*NA^2-
b_l*d_l*f_l+d_l*f_l*w_l-d_l*w_l*NA);
% Mp(i)=557.3;
% Mp(i)=511;
% Mp(i)=473;
% end
%
% end

184
 %% connectivity b's

[b_e] = connectivity(ElementNodes(i,:),NumNodes);

belement(:,:,i)=b_e;

%% Initializing internal forces as all 0 for first state

determination

Finternal(:,1,i)=[0;0;0;0;0;0];
FsElement(:,1,i)=[0;0;0;0;0;0];
FdElement_b(:,1,i)=[0;0;0;0;0;0];
FdElement_a(:,1,i)=[0;0;0;0;0;0];
FiElement(:,1,i)=[0;0;0;0;0;0];

end

[F_t t steps] = Forcetime(forcecurve, del, stopt);

timecount=1;

%first timestep outside the loop

kglobal=[]; kglobal=zeros; %clears matrices every timestep

gglobal=[]; gglobal=zeros;
mglobal=[]; mglobal=zeros;
cglobal=[]; cglobal=zeros;
reducer=[]; reducer=zeros;

for i=1:1:NumElem %Loops over each element for stiffness, mass,

damping, and equivalent nodal loads

node1=ElementNodes(i,1); %temporarily stores node 1 for element i

node2=ElementNodes(i,2); %temporarily stores node 2 for element i
nodeclass1=Node_class(node1,1);
nodeclass2=Node_class(node2,1);

moment_left=abs(FsElement(3,timecount,i));
moment_right=abs(FsElement(6,timecount,i));

if Type_num==6

order=ElementProp(i,3);

185
 [N_e B_e C_e N_e_r B_e_r hinging] = NBC_element(L(i), E(i),
A(i), As(i), G(i), I(i), Tp_num(i), order);

else

N_e=0;
B_e=0;
C_e=0;
N_e_r=0;
B_e_r=0;

end

[k_e m_e r kms kss] = k_element(L(i), E(i), A(i), As(i), G(i),

I(i), Tp_num(i), B_e_r, C_e, N_e_r, lumping, analysis, rho(i),
FsElement(:,timecount,i), Mp(i), nodeclass1, nodeclass2);

kelemento(:,:,i)=k_e;

melemento(:,:,i)=m_e;

if nodeclass1=='I' %Nonstructural masses block (adds half if

internal node, adds full if external node)

melemento(1,1,i)=melemento(1,1,i)+0.5*nonstr_massfull(node1*3-
2);
melemento(2,2,i)=melemento(2,2,i)+0.5*nonstr_massfull(node1*3-
1);

melemento(3,3,i)=melemento(3,3,i)+0.5*nonstr_massfull(node1*3);

elseif nodeclass1=='E'

melemento(1,1,i)=melemento(1,1,i)+nonstr_massfull(node1*3-2);
melemento(2,2,i)=melemento(2,2,i)+nonstr_massfull(node1*3-1);
melemento(3,3,i)=melemento(3,3,i)+nonstr_massfull(node1*3);

end

if nodeclass2=='I'

melemento(4,4,i)=melemento(4,4,i)+0.5*nonstr_massfull(node2*3-
2);
melemento(5,5,i)=melemento(5,5,i)+0.5*nonstr_massfull(node2*3-
1);

melemento(6,6,i)=melemento(6,6,i)+0.5*nonstr_massfull(node2*3);

elseif nodeclass2=='E'

186
 melemento(4,4,i)=melemento(4,4,i)+nonstr_massfull(node2*3-2);
melemento(5,5,i)=melemento(5,5,i)+nonstr_massfull(node2*3-1);
melemento(6,6,i)=melemento(6,6,i)+nonstr_massfull(node2*3);

end

%% Equivalent Nodal Loads

axial_nodal=Axial_distrfull(i,:);
forces_nodal=Force_distrfull(i,:);
moment_nodal=Moment_distrfull(i,:);

[g_e] = g_element(Length, degree1, degree2, degree3, axial_nodal,

forces_nodal, moment_nodal, N_e_r, Tp_num(i), r, kms, kss,
FsElement(:,timecount,i), Mp(i), nodeclass1, nodeclass2); %Equivalent
Load Function

gelemento(:,i)=g_e;

%% transformation matrices

c=cos(angle(i));
s=sin(angle(i));

transf=[c s 0 0 0 0
-s c 0 0 0 0
0 0 1 0 0 0
0 0 0 c s 0
0 0 0 -s c 0
0 0 0 0 0 1];

kelement_tr(:,:,i)=transpose(transf)*kelemento(:,:,i)*transf;
%rotates elemental stiffness to global coordinates
gelement_tr(:,1,i)=transpose(transf)*gelemento(:,i); %rotates
elemental equivalent nodal loads to global coordinates
melement_tr(:,:,i)=transpose(transf)*melemento(:,:,i)*transf;
%rotates mass matrix

%% Assembly

kglobal=kglobal+transpose(belement(:,:,i))*kelement_tr(:,:,i)*belement
(:,:,i); %assembles k matrix, k=sum(b_eT*k_e*b_e)

gglobal=gglobal+transpose(belement(:,:,i))*gelement_tr(:,1,i);
%assembles g vector, k=sum(b_eT*g_e)

mglobal=mglobal+transpose(belement(:,:,i))*melement_tr(:,:,i)*belement
(:,:,i); %assembles global mass matrix m=sum(b_eT*m_e*b_e)

187
 Fnet=gglobal+Forcefull; %calculates total load vector, which is
sum of nodal loads and equivalent nodal loads

% update BC vector for size of reducer matrix (and reduced

% stiffness, mass, damping matrices)

if nodeclass1=='E' %if a fixed end is yielding, creates different

boundary conditions

if moment_left>=Mp(i) %Hinge at left node

BCfull(node1*3,1)=ii;

elseif moment_left<Mp(i)

BCfull(node1*3,1)=BCfull_original(node1*3,1);

end

end

if nodeclass2=='E' %if a fixed end is yielding, creates different

boundary conditions

if moment_right>=Mp(i) %Hinge at left node

BCfull(node2*3,1)=ii;

elseif moment_right<Mp(i)

BCfull(node2*3,1)=BCfull_original(node2*3,1);

end

end

end

reduce_count=1;

for i=1:1:NumNodes*3 %calculates reducer matrix

if BCfull(i,1)~=0

reducero(reduce_count,i)=1;
reduce_count=reduce_count+1;

else

reducero(:,i)=0;

188
 end

end

%applying B.C.s with reducer matrix

ko=reducero*kglobal*transpose(reducero);

go=reducero*gglobal;

mo=reducero*mglobal*transpose(reducero);

%elastic frequencies

modes_solve=min(modes,size(ko,1));

[phi w2]=eigs(ko,mo,modes_solve,'SM');

w=sqrt(w2);

for freq=1:1:size(w,1)

T(freq,freq)=2*pi/w(freq,freq);

end

firstmode=rayleigh(1,1);
secondmode=rayleigh(2,1);
z1=rayleigh(1,2);
z2=rayleigh(2,2);

w1=w(firstmode,firstmode);
w2=w(secondmode,secondmode);

c_alpha=2*w1*w2/(w2^2-w1^2)*(w2*z1-w1*z2);
c_beta=2/(w2^2-w1^2)*(w2*z2-w1*z1);

c_alpha=0;
c_beta=.000002;

if axial~=0
kglobal=[]; kglobal=zeros;
end

189
for i=1:1:NumElem %Loops over each element damping

node1=ElementNodes(i,1); %temporarily stores node 1 for element i

node2=ElementNodes(i,2); %temporarily stores node 2 for element i
nodeclass1=Node_class(node1,1);
nodeclass2=Node_class(node2,1);

moment_left=abs(FsElement(3,timecount,i));
moment_right=abs(FsElement(6,timecount,i));

if Type_num==6

order=ElementProp(i,3);

[N_e B_e C_e N_e_r B_e_r hinging] = NBC_element(L(i), E(i),

A(i), As(i), G(i), I(i), Tp_num(i), order);

else

N_e=0;
B_e=0;
C_e=0;
N_e_r=0;
B_e_r=0;

end

damp_switch=1;

[celemento_a(:,:,i) celemento_b(:,:,i)]=Damping(kelemento(:,:,i),
melemento(:,:,i), c_alpha, c_beta, damp_switch);

celemento(:,:,i)=celemento_a(:,:,i)+celemento_b(:,:,i);

if axial~=0 %adds geometric stiffness to elastic stiffness AFTER

calculating stiffness-proportional damping

k_g = k_geom(L(i),axial,r);

kgelemento(:,:,i)=k_g;

kelemento(:,:,i)=kelemento(:,:,i)+k_g;
end

c=cos(angle(i));
s=sin(angle(i));

190
 transf=[c s 0 0 0 0
-s c 0 0 0 0
0 0 1 0 0 0
0 0 0 c s 0
0 0 0 -s c 0
0 0 0 0 0 1];

celement_tr(:,:,i)=transpose(transf)*celemento(:,:,i)*transf;
%rotates damping matrix

if axial~=0
kelement_tr(:,:,i)=transpose(transf)*kelemento(:,:,i)*transf;
%rotates damping matrix
end

%% Assembly

cglobal=cglobal+transpose(belement(:,:,i))*celement_tr(:,:,i)*belement
(:,:,i); %assembles global damping matrix c=sum(b_eT*c_e*b_e)

if axial~=0

kglobal=kglobal+transpose(belement(:,:,i))*kelement_tr(:,:,i)*belement
(:,:,i);
end

end

co=reducero*cglobal*transpose(reducero);

%applying axial load statically before dynamic analysis

if axial~=0

ko=reducero*kglobal*transpose(reducero);
siz=size(ko,1);
F_axial=zeros(siz,1);
F_axial(1,1)=axial;
u_axial=ko\F_axial;
u_axial_full=reducero'*u_axial;
ICufull=u_axial_full;

end

% for i=3:3:3*NumNodes

191
%
% if abs(Fnet(i,1))>Mp(1)
%
% Fnet(i,1)=Fnet(i,1)/abs(Fnet(i,1))*Mp(1);
%
% end
%
% end

Fexternal=reducero*Fnet;

% for first timecount, sets initial conditions

minvo=inv(mo);

Re(:,1)=Fexternal*F_t(1);

Ru(:,1)=Re(:,1)*0; %sets unbalanced to 0 initially

Ri(:,1)=Ru(:,1); %sets internal to 0 intially

u_t(:,1)=reducero*ICufull;

Fs(:,1)=ko*u_t(:,1);

v_t(:,1)=reducero*ICvfull;

Fd(:,1)=co*v_t(:,1);

a_t(:,1)=minvo*(Re(:,1)-Fd(:,1)-Fs(:,1));

Fi(:,1)=mo*a_t(:,1);

u_t_total(:,1)=ICufull;
v_t_total(:,1)=ICvfull;

count=1;

for i=1:1:NumNodes*3 %this loop adds back the boundary conditions to

initial a_t_total vector

if BCfull(i,1)==ii

a_t_total(i,1)=a_t(count,1);

count=count+1;

192
 elseif BCfull(i,1)==0

a_t_total(i,1)=0;

end

end

%stores the original (elastic) matrices

ao=gamma/beta*co+1/beta/del*mo;

bo=1/2/beta*mo+del*(gamma/2/beta-1)*co;

kbaro=1/del*ao;

kstaro=ko+kbaro;

invko=inv(kstaro);

for i=1:1:NumNodes % initializes

yieldnodes(i,1)=0;
yieldvalues(i,1)=0;
end

while (timecount<=steps) %Timestepping

%% intial calcs done every timesetep

Re(:,timecount+1)=Fexternal*F_t(timecount+1); %external applied

load of timestep

%Re(:,timecount+1)=Re(:,timecount+1)+F_axial;

dRe(:,timecount)=Re(:,timecount+1)-Re(:,timecount); %change in
applied load of timestep

%The remainder of the timestepping is split into 2 sets of

solutions
%The first set is the timestep solved elastically to determine
yielding
%The second set is only calculated if there is yielding (using
hinged elements where required)

193
 %% First set (elastic for yielding determination)

%yielded='n'; %determines whether the second set is calculated

% for i=1:1:NumNodes % stores all yielded nodes as unyielded

%
% yieldnodes(i,timecount+1)=yieldnodes(i,timecount);
% end

Ru(:,timecount+1)=dRe(:,timecount)+ao*v_t(:,timecount)+bo*a_t(:,timeco
unt)+Ru(:,timecount); %unbalanced load

du(:,timecount)=Ru(:,timecount+1)*0; %initializes du

norm_check=TOL+1; %initializes the norm as greater than tolerance

while (norm_check>TOL) %NR iteration

yielded='n';

du(:,timecount)=du(:,timecount)+invko*Ru(:,timecount+1);

dv(:,timecount)=gamma/beta/del*du(:,timecount)-
gamma/beta*v_t(:,timecount)+del*(1-gamma/2/beta)*a_t(:,timecount);

da(:,timecount)=1/beta/del^2*du(:,timecount)-
1/beta/del*v_t(:,timecount)-1/2/beta*a_t(:,timecount);

u_t(:,timecount+1)=u_t(:,timecount)+du(:,timecount);

v_t(:,timecount+1)=v_t(:,timecount)+dv(:,timecount);

a_t(:,timecount+1)=a_t(:,timecount)+da(:,timecount);

count=1;
for i=1:1:NumNodes*3 %this loop adds back all values to u, v,
a vectors

if BCfull(i,1)==ii

u_t_total(i,timecount+1)=u_t(count,timecount+1);
du_total(i,timecount)=du(count,timecount);

v_t_total(i,timecount+1)=v_t(count,timecount+1);
dv_total(i,timecount)=dv(count,timecount);

a_t_total(i,timecount+1)=a_t(count,timecount+1);

194
 da_total(i,timecount)=da(count,timecount);

count=count+1;

elseif BCfull(i,1)==0

u_t_total(i,timecount+1)=0;
du_total(i,timecount)=0;

v_t_total(i,timecount+1)=0;
dv_total(i,timecount)=0;

a_t_total(i,timecount+1)=0;
da_total(i,timecount)=0;

end

end

Finternal_global=[]; %initialize internal as 0

Finternal_global=zeros;

for i=1:1:NumElem %Loops over each element for internal

forces

set=size(kelemento(:,:,i),2);

node1=ElementNodes(i,1); %temporarily stores node 1 for

element i
node2=ElementNodes(i,2); %temporarily stores node 2 for
element i
nodeclass1=Node_class(node1,1);
nodeclass2=Node_class(node2,1);

du_t_element(:,timecount,i)=[du_total(node1*3-2,timecount)
%creates du for element i
du_total(node1*3-1,timecount)
du_total(node1*3,timecount)
du_total(node2*3-2,timecount)
du_total(node2*3-1,timecount)
du_total(node2*3,timecount)];

dv_t_element(:,timecount,i)=[dv_total(node1*3-2,timecount)
%creates dv for element i
dv_total(node1*3-1,timecount)
dv_total(node1*3,timecount)
dv_total(node2*3-2,timecount)
dv_total(node2*3-1,timecount)

195
 dv_total(node2*3,timecount)];

da_t_element(:,timecount,i)=[da_total(node1*3-2,timecount)
%creates da for element i
da_total(node1*3-1,timecount)
da_total(node1*3,timecount)
da_total(node2*3-2,timecount)
da_total(node2*3-1,timecount)
da_total(node2*3,timecount)];

u_t_element(:,timecount+1,i)=[u_t_total(node1*3-
2,timecount+1) %creates u for element i
u_t_total(node1*3-1,timecount+1)
u_t_total(node1*3,timecount+1)
u_t_total(node2*3-2,timecount+1)
u_t_total(node2*3-1,timecount+1)
u_t_total(node2*3,timecount+1)];

v_t_element(:,timecount+1,i)=[v_t_total(node1*3-
2,timecount+1) %creates v for element i
v_t_total(node1*3-1,timecount+1)
v_t_total(node1*3,timecount+1)
v_t_total(node2*3-2,timecount+1)
v_t_total(node2*3-1,timecount+1)
v_t_total(node2*3,timecount+1)];

a_t_element(:,timecount+1,i)=[a_t_total(node1*3-
2,timecount+1) % creates a for element i
a_t_total(node1*3-1,timecount+1)
a_t_total(node1*3,timecount+1)
a_t_total(node2*3-2,timecount+1)
a_t_total(node2*3-1,timecount+1)
a_t_total(node2*3,timecount+1)];

dFsElement(:,timecount,i)=kelemento(:,:,i)*du_t_element(:,timecount,i)
;

FsElement(:,timecount+1,i)=FsElement(:,timecount,i)+dFsElement(:,timec
ount,i); %Spring forces for element i (incremental)

FdElement_a(:,timecount+1,i)=celemento_a(:,:,i)*v_t_element(:,timecoun
t+1,i);

196
FdElement_b(:,timecount+1,i)=celemento_b(:,:,i)*v_t_element(:,timecoun
t+1,i);

FdElement(:,timecount+1,i)=FdElement_a(:,timecount+1,i)+FdElement_b(:,
timecount+1,i);

FiElement(:,timecount+1,i)=melemento(:,:,i)*a_t_element(:,timecount+1,
i);

Finternal(:,timecount+1,i)=FsElement(:,timecount+1,i)+FdElement(:,time
count+1,i)+FiElement(:,timecount+1,i); %Internal forces for element i

moment_right=abs(FsElement(6,timecount+1,i)); %right end

moment for element i
moment_left=abs(FsElement(3,timecount+1,i));

if moment_right>=Mp(i) %allows for recalculation using

end released elements after this section

yielded='y'; %this allows the second set to be

calculated
yieldnodes(node2,timecount+1)=1; %tracks which node
yielded
yieldvalues(node2,timecount+1)=moment_right;

%%%%%%%%%%%%%%%%%%%%%%decrease timestep here and get

more accuracy for other elements spring forces

% value=abs(FsElement(6,timecount,i));
%This scales the previous spring force to the yielded value so that
the second set has this value in the internal force
%
% scale=Mp(i)/value;
%
%
FsElement(:,timecount,i)=scale*FsElement(:,timecount,i);

else

yieldnodes(node2,timecount+1)=0;
yieldvalues(node2,timecount+1)=0;

end

197
 if nodeclass1=='E'

if moment_left>=Mp(i)

yielded='y';
yieldnodes(node1,timecount+1)=1;
yieldvalues(node1,timecount+1)=moment_left;

else

yieldnodes(node1,timecount+1)=0;
yieldvalues(node1,timecount+1)=0;

end

end

Finternal_global=Finternal_global+transpose(belement(:,:,i))*Finternal
(:,timecount+1,i); %assembles the internal force vector

end

Ri(:,timecount+1)=reducero*Finternal_global; %reduces the

internal force vector to the correct size

Ru(:,timecount+1)=Re(:,timecount+1)-Ri(:,timecount+1);
%calculates the unbalanced load

norm_check=norm(Ru(:,timecount+1)); %restores the norm to be

checked against tolerance

end

%% This is the second set (recalculation of all matrices and

solution for timestep if yielding occurs)

if yielded=='y' %only goes into this set if the yielded marker is

'y'

yield_values_sorted(:,timecount+1)=sort(yieldvalues(:,timecount+1),'de
scend');

loop='stay';

198
 while (loop=='stay') %iterates this whole section until the
yielded nodes are correct

a=[];
b=[];
kbar=[];
kstar=[];
invk=[];

kglobal=[]; kglobal=zeros;
gglobal=[]; gglobal=zeros;
mglobal=[]; mglobal=zeros;
cglobal=[]; cglobal=zeros;
reducer=[]; reducer=zeros;

for i=1:1:NumElem

moment_left=abs(FsElement(3,timecount+1,i)); %These
moments were calculated in the first set so the yielded values will be
above Mp(i)
moment_right=abs(FsElement(6,timecount+1,i));
node1=ElementNodes(i,1); %temporarily stores node 1
for element i
node2=ElementNodes(i,2); %temporarily stores node 2
for element i
nodeclass1=Node_class(node1,1);
nodeclass2=Node_class(node2,1);

%reassemble k

kelement(:,:,i)=kelemento(:,:,i);
melement(:,:,i)=melemento(:,:,i);

damp_switch=1;

scale=0;

if nodeclass1=='E'

if moment_left==yield_values_sorted(1,timecount+1)

[kelement(:,:,i)
Tr_k(:,:,i)]=Condenseleft(kelemento(:,:,i));%-kgelemento(:,:,i));
%creates hinged matrix
[melement(:,:,i)
Tr_m(:,:,i)]=Condenseleft(melemento(:,:,i)); %creates hinged matrix

%kelement(:,:,i)=kelement(:,:,i)+kgelemento(:,:,i);

199
 value=abs(FsElement(3,timecount,i)); %This
scales the previous spring force to the yielded value so that the
second set has this value in the internal force

scale=Mp(i)/value;

FsElement(:,timecount,i)=scale*FsElement(:,timecount,i);

if moment_right==Mp(i) %if the right end is

previously yielded, becomes bar
'error- increase mesh to avoid fully
yielded element'
break

end

elseif moment_left==Mp(i)

[kelement(:,:,i)
Tr_k(:,:,i)]=Condenseleft(kelemento(:,:,i));%-kgelemento(:,:,i));
%creates hinged matrix

[melement(:,:,i)
Tr_m(:,:,i)]=Condenseleft(melemento(:,:,i)); %creates hinged matrix

%kelement(:,:,i)=kelement(:,:,i)+kgelemento(:,:,i);

end

end

if moment_right==yield_values_sorted(1,timecount+1)
%Hinge at right node for first node yielded

[kelement(:,:,i)
Tr_k(:,:,i)]=Condense(kelemento(:,:,i));%-kgelemento(:,:,i)); %creates
hinged matrix
if i~=NumElem
[melement(:,:,i)
Tr_m(:,:,i)]=Condense(melemento(:,:,i)); %creates hinged matrix
end

%kelement(:,:,i)=kelement(:,:,i)+kgelemento(:,:,i);

200
 value=abs(FsElement(6,timecount,i)); %This scales
the previous spring force to the yielded value so that the second set
has this value in the internal force

scale=Mp(i)/value;

FsElement(:,timecount,i)=scale*FsElement(:,timecount,i);

if nodeclass1=='E' %if it is the left end member

and the left end is also previously yielded, becomes bar

if moment_left==Mp(i)
'error- increase mesh to avoid fully
yielded element'
break

end

end

%damp_switch=10^-3;

elseif moment_right==Mp(i) %any already yielded nodes

(takes value of Mp)

[kelement(:,:,i)
Tr_k(:,:,i)]=Condense(kelemento(:,:,i));%-kgelemento(:,:,i)); %creates
hinged matrix
if i~=NumElem
[melement(:,:,i)
Tr_m(:,:,i)]=Condense(melemento(:,:,i)); %creates hinged matrix
end

%kelement(:,:,i)=kelement(:,:,i)+kgelemento(:,:,i);
%damp_switch=10^-3;

end

if moment_left>=Mp(i) %accounts for the other side of

each node yielded to decrease damping as well

%damp_switch=10^-3;

%value=abs(FsElement(3,timecount,i));

201
 %scale=Mp(i)/value;

%
FsElement(:,timecount,i)=scale*FsElement(:,timecount,i);

end

[celement_a(:,:,i)
celement_b(:,:,i)]=Damping(kelement(:,:,i), melement(:,:,i), c_alpha,
c_beta, damp_switch);
celement(:,:,i)=celement_a(:,:,i)+celement_b(:,:,i);

%
celement_a(:,:,i)=celemento_a(:,:,i);
celement_b(:,:,i)=celemento_b(:,:,i);

damp_switch_left=1;
damp_switch_right=1;

if moment_right>=Mp(i)

damp_switch_right=10^-3;

end

if i==1

moment_previous(i)=moment_left;
end

if moment_previous(i)>=Mp(i)

damp_switch_left=10^-3;

end

celement_b(6,:,i)=celement_b(6,:,i)*damp_switch_right;
%"zeros" damping for right end rotational DOF of element i
celement_b(:,6,i)=celement_b(:,6,i)*damp_switch_right;
celement_b(6,6,i)=celement_b(6,6,i)/damp_switch_right;

celement_b(3,:,i)=celement_b(3,:,i)*damp_switch_left;
%"zeros" damping for left end rotational DOF of adjacent member (same
DOF as above)
celement_b(:,3,i)=celement_b(:,3,i)*damp_switch_left;
celement_b(3,3,i)=celement_b(3,3,i)/damp_switch_left;

202
%
%
celement_b(:,:,i)=celement_b(:,:,i)*damp_switch_right;
%
% if damp_switch_right==1 && damp_switch_left~=1
%
%
celement_b(:,:,i)=celement_b(:,:,i)*damp_switch_left;
% end

moment_previous(i+1)=moment_right;

%celement_b(:,:,i)=celement_b(:,:,i)*damp_switch;
celement(:,:,i)=celement_a(:,:,i)+celement_b(:,:,i);

%% Equivalent Nodal Loads

axial_nodal=Axial_distrfull(i,:);
forces_nodal=Force_distrfull(i,:);
moment_nodal=Moment_distrfull(i,:);

[g_e] = g_element(Length, degree1, degree2, degree3,

axial_nodal, forces_nodal, moment_nodal, N_e_r, Tp_num(i), r, kms,
kss, FsElement(:,timecount+1,i), Mp(i), nodeclass1, nodeclass2);
%Equivalent Load Function

gelement(:,i)=g_e;

%% transformation matrices

c=cos(angle(i));
s=sin(angle(i));

transf=[c s 0 0 0 0
-s c 0 0 0 0
0 0 1 0 0 0
0 0 0 c s 0
0 0 0 -s c 0
0 0 0 0 0 1];

kelement_tr(:,:,i)=transpose(transf)*kelement(:,:,i)*transf; %rotates
elemental stiffness to global coordinates
gelement_tr(:,1,i)=transpose(transf)*gelement(:,i);
%rotates elemental equivalent nodal loads to global coordinates

melement_tr(:,:,i)=transpose(transf)*melement(:,:,i)*transf; %rotates
mass matrix

celement_tr(:,:,i)=transpose(transf)*celement(:,:,i)*transf; %rotates
damping matrix

203
 %% Assembly

kglobal=kglobal+transpose(belement(:,:,i))*kelement_tr(:,:,i)*belement
(:,:,i); %assembles k matrix, k=sum(b_eT*k_e*b_e)

gglobal=gglobal+transpose(belement(:,:,i))*gelement_tr(:,1,i);
%assembles g vector, k=sum(b_eT*g_e)

mglobal=mglobal+transpose(belement(:,:,i))*melement_tr(:,:,i)*belement
(:,:,i); %assembles global mass matrix m=sum(b_eT*m_e*b_e)

cglobal=cglobal+transpose(belement(:,:,i))*celement_tr(:,:,i)*belement
(:,:,i); %assembles global damping matrix c=sum(b_eT*c_e*b_e)

Fnet=gglobal+Forcefull; %calculates total load vector,

which is sum of nodal loads and equivalent nodal loads

% update BC vector for size of reducer matrix (and

reduced
% stiffness, mass, damping matrices)

end

%applying B.C.s with reducer matrix

k=reducero*kglobal*transpose(reducero);

g=reducero*gglobal;

m=reducero*mglobal*transpose(reducero);

c=reducero*cglobal*transpose(reducero);

%Newmark Beta solution

Fexternal=reducero*Fnet;

Re(:,timecount+1)=Fexternal*F_t(timecount+1); %external
applied load of timestep

%Re(:,timecount+1)=Re(:,timecount+1)+F_axial;

dRe(:,timecount)=Re(:,timecount+1)-Re(:,timecount);
%change in applied load of timestep

204
 a=gamma/beta*c+1/beta/del*m;

b=1/2/beta*m+del*(gamma/2/beta-1)*c;

kbar=1/del*a;

kstar=k+kbar;

invk=inv(kstar);

Ru(:,timecount+1)=dRe(:,timecount)+a*v_t(:,timecount)+b*a_t(:,timecoun
t)+Ru(:,timecount);

du(:,timecount)=Ru(:,timecount+1)*0;

norm_check=TOL+1;

while (norm_check>TOL) %NR iteration

du(:,timecount)=du(:,timecount)+invk*Ru(:,timecount+1);

dv(:,timecount)=gamma/beta/del*du(:,timecount)-
gamma/beta*v_t(:,timecount)+del*(1-gamma/2/beta)*a_t(:,timecount);

da(:,timecount)=1/beta/del^2*du(:,timecount)-
1/beta/del*v_t(:,timecount)-1/2/beta*a_t(:,timecount);

u_t(:,timecount+1)=u_t(:,timecount)+du(:,timecount);

v_t(:,timecount+1)=v_t(:,timecount)+dv(:,timecount);

a_t(:,timecount+1)=a_t(:,timecount)+da(:,timecount);

count=1;
for i=1:1:NumNodes*3 %this loop adds back the
boundary conditions to u vector

if BCfull(i,1)==ii

u_t_total(i,timecount+1)=u_t(count,timecount+1);
du_total(i,timecount)=du(count,timecount);

v_t_total(i,timecount+1)=v_t(count,timecount+1);
dv_total(i,timecount)=dv(count,timecount);

205
a_t_total(i,timecount+1)=a_t(count,timecount+1);
da_total(i,timecount)=da(count,timecount);

count=count+1;

elseif BCfull(i,1)==0

u_t_total(i,timecount+1)=0;
du_total(i,timecount)=0;

v_t_total(i,timecount+1)=0;
dv_total(i,timecount)=0;

a_t_total(i,timecount+1)=0;
da_total(i,timecount)=0;

end

end

Finternal_global=[];
Finternal_global=zeros;

for i=1:1:NumElem %Loops over each element for

internal forces

set=size(kelement(:,:,i),2);

node1=ElementNodes(i,1); %temporarily stores node

1 for element i
node2=ElementNodes(i,2); %temporarily stores node
2 for element i
nodeclass1=Node_class(node1,1);
nodeclass2=Node_class(node2,1);

du_t_element(:,timecount,i)=[du_total(node1*3-
2,timecount) %creates du for element i
du_total(node1*3-1,timecount)
du_total(node1*3,timecount)
du_total(node2*3-2,timecount)
du_total(node2*3-1,timecount)
du_total(node2*3,timecount)];

206
 dv_t_element(:,timecount,i)=[dv_total(node1*3-
2,timecount) %creates dv for element i
dv_total(node1*3-1,timecount)
dv_total(node1*3,timecount)
dv_total(node2*3-2,timecount)
dv_total(node2*3-1,timecount)
dv_total(node2*3,timecount)];

da_t_element(:,timecount,i)=[da_total(node1*3-
2,timecount) %creates da for element i
da_total(node1*3-1,timecount)
da_total(node1*3,timecount)
da_total(node2*3-2,timecount)
da_total(node2*3-1,timecount)
da_total(node2*3,timecount)];

u_t_element(:,timecount+1,i)=[u_t_total(node1*3-
2,timecount+1) %creates u for element i
u_t_total(node1*3-1,timecount+1)
u_t_total(node1*3,timecount+1)
u_t_total(node2*3-2,timecount+1)
u_t_total(node2*3-1,timecount+1)
u_t_total(node2*3,timecount+1)];

v_t_element(:,timecount+1,i)=[v_t_total(node1*3-
2,timecount+1) %creates v for element i
v_t_total(node1*3-1,timecount+1)
v_t_total(node1*3,timecount+1)
v_t_total(node2*3-2,timecount+1)
v_t_total(node2*3-1,timecount+1)
v_t_total(node2*3,timecount+1)];

a_t_element(:,timecount+1,i)=[a_t_total(node1*3-
2,timecount+1) % creates a for element i
a_t_total(node1*3-1,timecount+1)
a_t_total(node1*3,timecount+1)
a_t_total(node2*3-2,timecount+1)
a_t_total(node2*3-1,timecount+1)
a_t_total(node2*3,timecount+1)];

dFsElement(:,timecount,i)=kelement(:,:,i)*du_t_element(:,timecount,i);

FsElement(:,timecount+1,i)=FsElement(:,timecount,i)+dFsElement(:,timec
ount,i); %Spring forces for element i (incremental)

207
 moment_left=abs(FsElement(3,timecount+1,i));
%These moments were calculated in the first set so the yielded values
will be above Mp(i)
moment_right=abs(FsElement(6,timecount+1,i));

% if moment_left>=Mp(i)
%
%
FsElement(3,timecount+1,i)=abs(FsElement(3,timecount+1,i))/FsElement(3
,timecount+1,i)*Mp(i);
% end

FdElement_a(:,timecount+1,i)=celement_a(:,:,i)*v_t_element(:,timecount
+1,i);

FdElement_b(:,timecount+1,i)=celement_b(:,:,i)*v_t_element(:,timecount
+1,i);

FdElement(:,timecount+1,i)=FdElement_a(:,timecount+1,i)+FdElement_b(:,
timecount+1,i);

FiElement(:,timecount+1,i)=melement(:,:,i)*a_t_element(:,timecount+1,i
);

Finternal(:,timecount+1,i)=FsElement(:,timecount+1,i)+FdElement(:,time
count+1,i)+FiElement(:,timecount+1,i); %Internal forces for element i

Finternal_global=Finternal_global+transpose(belement(:,:,i))*Finternal
(:,timecount+1,i);

moment_right=abs(FsElement(6,timecount+1,i));
%again this value will be over Mp(i) if yielded
moment_left=abs(FsElement(3,timecount+1,i));

if moment_right>=Mp(i)

yieldnodes(node2,timecount+1)=1;
yieldvalues(node2,timecount+1)=moment_right;

else

yieldnodes(node2,timecount+1)=0;
yieldvalues(node2,timecount+1)=0;

end

if nodeclass1=='E' %leftmost node

208
 if moment_left>=Mp(i)

yieldnodes(node1,timecount+1)=1;

yieldvalues(node1,timecount+1)=moment_left;

else

yieldnodes(node1,timecount+1)=0;
yieldvalues(node1,timecount+1)=0;

end

end

end

Ri(:,timecount+1)=reducero*Finternal_global;

Ru(:,timecount+1)=Re(:,timecount+1)-Ri(:,timecount+1);

norm_check=norm(Ru(:,timecount+1));

end

yield_values_sorted(:,timecount+1)=sort(yieldvalues(:,timecount+1),'de
scend');

loop='exit';

for i=1:1:NumNodes %if there is a value that isn't scaled

to Mp, it stays in the loop

j=i; %moment check is associated with the left node of

the element

if i==NumNodes

j=i-1; %moment check is associated with right

moment for last node

end

if yield_values_sorted(i,timecount+1)~=0 &&
yield_values_sorted(i,timecount+1)~=Mp(j)

loop='stay';

209
 end
end

end %end of yieldcheck convergence (converges to the correct

number of yielded nodes

end %end of yielded='y' section

%% increase timesetep for next timestep

timecount=timecount+1;

end

DOFs=NumNodes*3;

midpoint=(DOFs+1)/2;

figure(1)

subplot(4,1,1)
plot(t,F_t)
title('Force function')
ylabel('Forcing shape with max=1 for scaling')
grid on

t(timecount+1)=[];

subplot(4,1,2)
plot(t,u_t_total(midpoint,:))
title('midspan deflection')
grid on

subplot(4,1,3)
plot(t,v_t_total(midpoint,:))
title('midspan velocity')
grid on

subplot(4,1,4)
plot(t,a_t_total(midpoint,:))
title('midspan acceleration')
grid on

F1=Finternal(:,:,1);
F2=Finternal(:,:,2);
F3=Finternal(:,:,3);
F4=Finternal(:,:,4);

210
Fs1=FsElement(:,:,1);
Fs2=FsElement(:,:,2);
Fs3=FsElement(:,:,3);
Fs4=FsElement(:,:,4);

Fd1=FdElement(:,:,1);
Fd2=FdElement(:,:,2);
Fd3=FdElement(:,:,3);
Fd4=FdElement(:,:,4);

Fi1=FiElement(:,:,1);
Fi2=FiElement(:,:,2);
Fi3=FiElement(:,:,3);
Fi4=FiElement(:,:,4);

u1=u_t_element(:,:,1);
u2=u_t_element(:,:,2);
u3=u_t_element(:,:,3);
u4=u_t_element(:,:,4);

v1=v_t_element(:,:,1);
v2=v_t_element(:,:,2);
v3=v_t_element(:,:,3);
v4=v_t_element(:,:,4);

a1=a_t_element(:,:,1);
a2=a_t_element(:,:,2);
a3=a_t_element(:,:,3);
a4=a_t_element(:,:,4);

text1='Element ';

text3=' left moment';

text4=' right moment';

%figure(2)

% counter=1;
%
% for i=1:1:NumElem
%
% text2=num2str(i);
%
% lefttext=strcat(text1,text2,text3);
% righttext=strcat(text1,text2,text4);
%
% subplot(NumElem,2,counter)
% plot(t,Finternal(3,:,i));
% title(lefttext)
% grid on
%

211
%
% subplot(NumElem,2,counter+1)
% plot(t,Finternal(6,:,i));
% title(righttext)
% grid on
%
%
%
% counter=counter+2;
%
% end

spring='spring';

figure(3)

counter=1;

for i=1:1:NumElem

text2=num2str(i);

lefttext=strcat(text1,text2,spring,text3);
righttext=strcat(text1,text2,spring,text4);

subplot(NumElem,2,counter)
plot(t,FsElement(3,:,i));
title(lefttext)
grid on

subplot(NumElem,2,counter+1)
plot(t,FsElement(6,:,i));
title(righttext)
grid on

counter=counter+2;

end

%
%
%
%
%
% spring='damping';
%
% figure(4)
%
% counter=1;
%
% for i=1:1:NumElem
%

212
% text2=num2str(i);
%
% lefttext=strcat(text1,text2,spring,text3);
% righttext=strcat(text1,text2,spring,text4);
%
% subplot(NumElem,2,counter)
% plot(t,FdElement(3,:,i));
% title(lefttext)
% grid on
%
%
% subplot(NumElem,2,counter+1)
% plot(t,FdElement(6,:,i));
% title(righttext)
% grid on
%
%
%
% counter=counter+2;
%
% end
%
%
%
%
% spring='inertial';
%
% figure(5)
%
% counter=1;
%
% for i=1:1:NumElem
%
% text2=num2str(i);
%
% lefttext=strcat(text1,text2,spring,text3);
% righttext=strcat(text1,text2,spring,text4);
%
% subplot(NumElem,2,counter)
% plot(t,FiElement(3,:,i));
% title(lefttext)
% grid on
%
%
% subplot(NumElem,2,counter+1)
% plot(t,FiElement(6,:,i));
% title(righttext)
% grid on
%
%
%
% counter=counter+2;
%
% end
%

213
%
% text3=' left';
%
% text4=' right';
%
% spring='displ';
%
% figure(6)
%
% counter=1;
%
% for i=1:1:NumElem
%
% text2=num2str(i);
%
% lefttext=strcat(text1,text2,spring,text3);
% righttext=strcat(text1,text2,spring,text4);
%
% subplot(NumElem,2,counter)
% plot(t,u_t_element(3,:,i));
% title(lefttext)
% grid on
%
%
% subplot(NumElem,2,counter+1)
% plot(t,u_t_element(6,:,i));
% title(righttext)
% grid on
%
%
%
% counter=counter+2;
%
% end
%
%
% spring='vel';
%
% figure(7)
%
% counter=1;
%
% for i=1:1:NumElem
%
% text2=num2str(i);
%
% lefttext=strcat(text1,text2,spring,text3);
% righttext=strcat(text1,text2,spring,text4);
%
% subplot(NumElem,2,counter)
% plot(t,v_t_element(3,:,i));
% title(lefttext)
% grid on
%
%

214
% subplot(NumElem,2,counter+1)
% plot(t,v_t_element(6,:,i));
% title(righttext)
% grid on
%
% counter=counter+2;
%
% end
%
%
% spring='acc';
%
% figure(8)
%
% counter=1;
%
% for i=1:1:NumElem
%
% text2=num2str(i);
%
% lefttext=strcat(text1,text2,spring,text3);
% righttext=strcat(text1,text2,spring,text4);
%
% subplot(NumElem,2,counter)
% plot(t,a_t_element(3,:,i));
% title(lefttext)
% grid on
%
%
% subplot(NumElem,2,counter+1)
% plot(t,a_t_element(6,:,i));
% title(righttext)
% grid on
%
%
%
% counter=counter+2;
%
% end
%

% text3=' left shear';

%
% text4=' right shear';
%
% figure(3)
%
% counter=1;
%
% for i=1:1:NumElem
%
% text2=num2str(i);
%

215
% lefttext=strcat(text1,text2,text3);
% righttext=strcat(text1,text2,text4);
%
% subplot(NumElem,2,counter)
% plot(t,Finternal(2,:,i));
% title(lefttext)
% grid on
%
%
% subplot(NumElem,2,counter+1)
% plot(t,Finternal(5,:,i));
% title(righttext)
% grid on
%
%
%
% counter=counter+2;
%
% end

t=transpose(t);
u_t=transpose(u_t);
Fs2=-transpose(Fs2);
Fs3=transpose(Fs3);
Fs4=transpose(Fs4);

a_output(:,1)=t;
a_output(:,2)=u_t(:,7);
a_output(:,3)=Fs2(:,6);
a_output(:,4)=-Fs4(:,6);

%if BC has left fixed end

%
% a_output(:,1)=t;
% a_output(:,2)=u_t(:,6);
% a_output(:,3)=Fs2(:,6);
% a_output(:,4)=-Fs4(:,6);

% %
% % % if cantilever
%
% a_output(:,1)=t;
% a_output(:,2)=u_t(:,2);
% a_output(:,3)=-Fs4(:,6);
% a_output(:,4)=-Fs4(:,6);

% Fs3=transpose(Fs3);
%
%
% Fs1=transpose(-Fs1);
%
% a_outputspr(:,1)=t;

216
% a_outputspr(:,2)=Fs1(:,3);
%
% mcheck=melemento(:,:,1);

aaimpose(1)=c_alpha;
aaimpose(2)=c_beta;
aaimpose(3)=axial;
aaimpose(4)=axial*L(1)*NumElem/A(1)/E(1);

min(u_t_total(8,:));

Input Function

%% This is the input file. Enter in units of inches, seconds, and kips

function [Nodes Node_class ElementNodes ElementProp ElementSec BC

Force Force_distr, Axial_distr, Moment_distr, analysis, modes,
lumping, nonstr_mass gamma beta forcecurve del stopt ICu ICv TOL
rayleigh axial bc_type] = Input

%% Input Section

i=sqrt(-1);

% enter type of analysis

% 1 for static analysis
% 2 for modal/dynamics analysis

analysis=2;

%NR iteration unbalanced load tolerance

TOL=10^-5;

% If modal analysis: enter number of modes to calculate, otherwise

enter
% any. Enter type of mass matrix. otherwise, enter any.
% 1: Consistent mass matrix
% 2: Particle Lumping (SAP)
% 3: Row Summing
% 4: HRZ Lumping

modes=1000;
lumping=2;

217
type=5; %mbeam type (4 bernoulli, 5 timoshenko)

% If modal analysis: gamma and beta values for newmark time stepping

beta=1/4;
gamma=1/2;

% If modal analysis, list linear endpoints of force time curve

% Scale so that largest point is 1 since you will be entering in the
values
% of the scaling later

% example: triangle function with 0 to 15k in 0.3 sec and back down
is:
% [0 0
% .3 1
% .6 0]

%enter delta t for timestepping (must be dividable into each

%segment)
%Also, enter time to stop analysis (delta t must be dividable)

del=0.0001;

ts=.0178;

% forcecurve=[0 0
% .01-del 0
% .01 1
% .01+ts 0
% .1 0 ];

forcecurve=[0 0
.0099 0
.01 1
.01+ts 0
.1 0 ];

% forcecurve=[0 1
% ts 0];

stopt=.08;

% enter pinpin, fixpin, fixfix, or cantil

218
bc_type='pinpin';

midforce=-40; %this gets put into the force vector at midspan (see
line 211)
midforce=0;

distr_load=-.45; % k/in this gets put into the distr load vector
(see line 258)
%distr_load=0;

axial=0; %enter axial load as proportion of axial cap

% Enter initial conditions in node order

% for nodes with nonzero ICs, list conditions
% example ICu=[1 5 4 0
% 6 -2 0 0];
% ICu is the displacement initial conditions (node, ux0, uy0, r0)
% ICv is the velocity initial conditions (node, vx0, vy0, rdot0,

ICu=[];

ICv=[];

%enter 2 modes followed by damping value as proportion

rayleigh=[1 .02
2 .02];

%List total length of beam (inches), and number of elements (must be

even)

ltotal=120;
nElem=4; %even number
nNodes=nElem+1;

lElem=ltotal/nElem;

219
 %{
for below,

material:
1: 50 ksi steel with 1.05 static increase, 1.19 dynamic
increase
2. Unit Material(E=1, G=1, density=1, fy=1

type:
1: bar (1d or 2d) (Can only have constant distributed
loads)

4: Bernoulli Beam with decoupled axial (Can only have

constant distributed loads)
5: Timoshenko Beam with decoupled axial (Can only have
constant distributed loads)
6: Polynomial Finite Beam (Can have any polynomial
distributed load

% order (for type 6 elements only, if not type 6, enter 0 here):

1: linear
2: quadratic
3: cubic
etc...

%}

matrl=1;
tp=type;
ord=0;

%Section info

Area=3.54;
Inertia=53.8;
shear_area=9.87*.19;
Z=12.6;
d_l=9.87;
b_l=3.96;
f_l=0.21;
w_l=0.19;

220
%% Node Looping

for n=1:1:nNodes

Nodes(n,1)=(n-1)*lElem; %Creates coordinates of each node

(constrained to a beam)
Nodes(n,2)=0;

if n==1 || n==nNodes %Labels node as exterior or interior

Node_class(n,1)='E';

else

Node_class(n,1)='I';
end

end

%% Element Looping

for n=1:1:nElem

ElementNodes(n,1)=n; %Connectivity info (orders nodes left to

right)
ElementNodes(n,2)=n+1;

ElementProp(n,1)=matrl;
ElementProp(n,2)=tp;
ElementProp(n,3)=ord;

ElementSec(n,1)=Area;
ElementSec(n,2)=Inertia;
ElementSec(n,3)=shear_area;
ElementSec(n,4)=Z;
ElementSec(n,5)=d_l;
ElementSec(n,6)=b_l;
ElementSec(n,7)=f_l;
ElementSec(n,8)=w_l;

end

221
%% More Direct Inputs

%list boundary conditions in node order (node, BCx, BCy, BCrot)

%put 0 for constrained, i for free,

if bc_type=='pinpin'

BC=[1 i 0 i
nNodes 0 0 i];

elseif bc_type=='fixpin'

BC=[1 i 0 i
nNodes 0 0 0];

elseif bc_type=='fixfix'

BC=[1 i 0 0
nNodes 0 0 0];

elseif bc_type=='cantil'

BC=[1 i i i
nNodes 0 0 0];

end

%list applied nodal forces in order by node (node, x force, y force,

moment)

Force=[

];

Force= [(1+nNodes)/2 0 midforce 0 %for point load at midspan

];

if bc_type=='cantil'

Force=[];

Force=[1 0 midforce 0];

222
 end

%If modal analysis: Enter non-structural masses to nodes by listing

nodes in order, then mass in x, y, and rotation
% (node, x-direction mass, y-direction mass, rotational inertia)

nonstr_mass= [1 1.619171e-4 1.619171e-4 0

2 3.238342e-4 3.238342e-4 0
3 3.238342e-4 3.238342e-4 0
4 3.238342e-4 3.238342e-4 0
5 1.619171e-4 1.619171e-4 0

];

nonstr_mass=[];

%list all distributed forces on elements in order by element(element,

values of constants)
%for values of constants list a's , ao + a1x + a2x^2 + .....)

Axial_distr= [

];

%list all distributed forces on elements in order by element(element,

values of constants)
%for values of constants list a's , ao + a1x + a2x^2 + .....)

Force_distr= [

];

for nn=1:1:nElem

Force_distr(nn,1)=nn;
Force_distr(nn,2)=distr_load;

end

223
%list all distributed forces on elements in order by element(element,
values of constants)
%for values of constants list a's , ao + a1x + a2x^2 + .....)

Moment_distr= [

];

end

Lengths Function

%% This function calculates the length and angle for each element sent
to it

function [ang Length] = Lengths(x1, x2, y1, y2)

Lx=x2-x1; % x component of length

Ly=y2-y1; % y component of length

ang=atan(Ly/Lx); %computes angle

Length=sqrt(Lx^2+Ly^2); %computes length

end

Force Time History Function

%% This function creates force vs time lines for dynamic analysis

function [F_t t steps] = Forcetime(forcecurve, del, stopt)

ii=sqrt(-1);

steps=stopt/del;

F_t(1)=forcecurve(1,2);

t(1)=forcecurve(1,1);

curvecount=1.0;

224
siz=size(forcecurve,1);

forcecurve(siz+1,:)=ii; %adds row of zeros at end

for i=1:1:steps+1

t(i+1)=t(i)+del;

if forcecurve(curvecount+1,1)~=ii

if t(i+1)>=real(forcecurve(curvecount+1,1))+10^-13

curvecount=curvecount+1;

end

slope=(forcecurve(curvecount+1,2)-
forcecurve(curvecount,2))/(forcecurve(curvecount+1,1)-
forcecurve(curvecount,1));
F_t(i+1)=F_t(i)+del*slope;
end

if forcecurve(curvecount+1,1)==ii

F_t(i+1)=0;

end

end

t=real(t);
F_t=real(F_t);

end

Connectivity Function

%% this function calculates the b matrix for each element

(connectivity)

function [b_e] = connectivity(Nodes,NumNodes)

225
 b_e=zeros(6,NumNodes*3); %first fills with zeros of size, 6 by
Nodes*3

for i=1:1:NumNodes %loops over the nodes

if Nodes(1,1)==i %if element node 1 is global node of

iteration

b_e(1,3*i-2)=1;
b_e(2,3*i-1)=1;
b_e(3,3*i)=1;

end

if Nodes(1,2)==i %if element node 2 is global node of

iteration

b_e(4,3*i-2)=1;
b_e(5,3*i-1)=1;
b_e(6,3*i)=1;

end

end

Data Function

%% This function contains all the data for materials

function [EM GM DENS Ar Ar_sh Inertia Fy Mpl] = Data(Mat_num,

Type_num, Area, Inert, Flange, Z)

% list of materials E, G, density, fy

materials=[29000 29000/2/(1+.29) .284/386.0886/1000 50*1.05*1.19 %ksi,

ksi, kip*s^2/in^3, ksi
1 1 1 1];

EM=materials(Mat_num,1); %assigns E to element sent to function

GM=materials(Mat_num,2); %assigns G to element sent to function
DENS=materials(Mat_num,3); %assigns density to element sent to
function
Fy=materials(Mat_num,4);

226
Ar=Area;

Ar_sh=Flange; %Shear area

Inertia=Inert;

Mpl=Fy*Z;

end

Elemental NBC Function

%% This function calculates N, B, and C matrices for each element sent

to it

function [N_e B_e C_e N_e_r B_e_r] = NBC_element(Length, EM, Ar,

Ar_sh, GM, Inertia, Type_num, order)

E=EM;
L=Length;
A=Ar;
As=Ar_sh;
G=GM;
I=Inertia;

syms x
z=2/L*x-1; %zeta for shape functions

%{

if Type_num==1

N=[.5*(1-z)
.5*(1+z)];

N_e=[N(1,1) 0 0 N(2,1) 0 0
0 0 0 0 0 0
0 0 0 0 0 0];

B_e=diff(N_e);

C_e=[E*A 0 0
0 As*G 0
0 0 E*I];

227
elseif Type_num==4

N=[1/4*(1-z)^2*(2+z)
1/4*(1-z)^2*(z+1)
1/4*(1+z)^2*(2-z)
1/4*(1+z)^2*(z-1)];

Nx=[.5*(1-z)
.5*(1+z)];

N_e=[Nx(1,1) 0 0 Nx(2,1) 0 0
0 N(1,1) L/2*N(2,1) 0 N(3,1) L/2*N(4,1)
0 diff(N(1,1),x) L/2*diff(N(2,1),x) 0 diff(N(3,1),x)
L/2*diff(N(4,1),x)];

B_e=diff(N_e,x);

B_e(2,:)=zeros;

C_e=[E*A 0 0
0 0 0
0 0 E*I];

elseif Type_num==5

N=[.5*(1-z)
.5*(1+z)];

N_e=[N(1,1) 0 0 N(2,1) 0 0
0 N(1,1) 0 0 N(2,1) 0
0 0 N(1,1) 0 0 N(2,1)];

B_e=diff(N_e);
B_e(2,3)=-N(1,1);
B_e(2,6)=-N(2,1);

C_e=[E*A 0 0
0 As*G 0
0 0 E*I];

%}

n=order+1;

space=2/order;

228
zee(1)=-1;

for i=2:1:n

zee(i)=zee(i-1)+space;

end

for k=1:1:n %calculates lagrange functions

num1=1;

for i=1:1:k-1

num1=num1*(zee(i)-z);

end

num2=1;

for i=k+1:1:n

num2=num2*(zee(i)-z);

end

den1=1;

for i=1:1:k-1

den1=den1*(zee(i)-zee(k));

end

den2=1;

for i=k+1:1:n

den2=den2*(zee(i)-zee(k));

end

N(k,1)=num1*num2/(den1*den2);

end

229
count=1;

for i=1:3:order*3+1

N_e(1,i)=N(count,1);
N_e(2,i+1)=N(count,1);
N_e(3,i+2)=N(count,1);

count=count+1;

end

%reorders N matrix so that external nodes are first 2 nodes

for i=1:1:3

N_e_r(i,i)=N_e(i,i);

end

for i=1:1:3

N_e_r(i,3+i)=N_e(i, 3*(order)+i);

end

for i=4:3:order*3

N_e_r(1,i+3)=N_e(1,i);
N_e_r(2,i+4)=N_e(2,i+1);
N_e_r(3,i+5)=N_e(3,i+2);

end

B_e=diff(N_e,x);

B_e_r=diff(N_e_r,x);

230
 for i=1:3:order*3+1 %adds the -1 part for B matrices

B_e(2,i+2)=-N_e(1,i);
B_e_r(2,i+2)=-N_e_r(1,i);

end

C_e=[E*A 0 0
0 As*G 0
0 0 E*I];

end

K_Element Function (Elemental Stiffness and Mass Matrices)

%% This calculates k matrix (and mass if modal) for each element sent
to it

function [k_e m_e r kms kss] = k_element(Length, EM, Ar, Ar_sh, GM,
Inertia, Type_num, B_e_r, C_e, N_e_r, lumping, analysis, DENS,
Finternal, Mp, nodeclass1, nodeclass2) %Stiffness Matrix Function

L=Length;
E=EM;
A=Ar;
As=Ar_sh;
G=GM;
I=Inertia;
type=Type_num;
density=DENS;

if type~=6 %If not polynomial finite element

if analysis==1 %if not modal, set mass equal to 0

m_e=0;
end

r=12*E*I/L^2/G/As; %rho for timoshenko

beta=r/12; %beta for timoshenko mass

231
%block that defines the type of stiffness matrix
if type==1 %if bar element

I=0;
r=0;

elseif type==4 %if bernoulli element

r=0;

end

k_e1=E*A/L*[1 -1;-1,1]; %element stiffness for bar

if analysis==2 %if modal, element mass for bar

mass_e=density*L*A;

m_e1=mass_e/6*[2 1; 1 2];

if lumping==2

m_e_new=zeros;

for i=[1 2]

m_e_new(i,i)=density*L*A/2;

end

m_e1=m_e_new;

end

if type==1 %if bar, zeros mass so following calcs are zero for
bending

mass_e=0;
end

end

moment_left=Finternal(3,1);
moment_right=Finternal(6,1);

%initially sets element stiffness as having no hinges

232
 k_e2=E*I/L^3/(1+r)*[12 6*L -12 6*L %element stiffness
for beam
6*L L^2*(4+r) -6*L L^2*(2-r)
-12 -6*L 12 -6*L
6*L L^2*(2-r) -6*L L^2*(4+r)];

if nodeclass2=='I' %any member except rightmost

member

if abs(moment_right)>=Mp %Hinge at right node

k_e2=Condense(k_e2);

end

end

if analysis==2 %if modal, element mass for beam

if type==5 %if timoshenko element

comp1=[13/35+7/10*r+1/3*r^2
(11/210+11/120*r+1/24*r^2)*L 9/70+3/10*r+1/6*r^2 -
(13/420+3/40*r+1/24*r^2)*L
(11/210+11/120*r+1/24*r^2)*L
(1/105+1/60*r+1/120*r^2)*L^2 (13/420+3/40*r+1/24*r^2)*L -
(1/140+1/60*r+1/120*r^2)*L^2
9/70+3/10*r+1/6*r^2
(13/420+3/40*r+1/24*r^2)*L 13/35+7/10*r+1/3*r^2 -
(11/210+11/120*r+1/24*r^2)*L
-(13/420+3/40*r+1/24*r^2)*L -
(1/140+1/60*r+1/120*r^2)*L^2 -(11/210+11/120*r+1/24*r^2)*L
(1/105+1/60*r+1/120*r^2)*L^2];

radius=sqrt(I/A);

comp2=(radius/L)^2*[6/5 (1/10+1/2*r)*L
-6/5 (1/10-1/2*r)*L
(1/10+1/2*r)*L
(2/15+1/6*r+1/3*r^2)*L^2 (-1/10+1/2*r)*L (-1/30-
1/6*r+1/6*r^2)*L^2

233
 -6/5 (-1/10+1/2*r)*L
6/5 (-1/10+1/2*r)*L
(1/10-1/2*r)*L (-1/30-
1/6*r+1/6*r^2)*L^2 (-1/10+1/2*r)*L (2/15+1/6*r+1/3*r^2)*L^2];

m_e2=density*A*L/(1+r)^2*(comp1+comp2);

elseif type==4 %if bernoulli element

beta=0;

m_e2=mass_e*[48*beta^2+42/5*beta+13/35 -(-6*beta^2-
11/10*beta-11/210)*L 24*beta^2+18/5*beta+9/70 -
(6*beta^2+9/10*beta+13/420)*L
-(-6*beta^2-11/10*beta-11/210)*L
(6/5*beta^2+1/5*beta+1/105)*L^2 -(-6*beta^2-9/10*beta-13/420)*L (-
6/5*beta^2-1/5*beta-1/140)*L^2
24*beta^2+18/5*beta+9/70 -(-6*beta^2-
9/10*beta-13/420)*L 48*beta^2+42/5*beta+13/35 -
(6*beta^2+11/10*beta+11/210)*L
-(6*beta^2+9/10*beta+13/420)*L (-
6/5*beta^2-1/5*beta-1/140)*L^2 -(6*beta^2+11/10*beta+11/210)*L
(6/5*beta^2+1/5*beta+1/105)*L^2];

end

if lumping==2

for i=[1 3] %fills translationals

m_e_new(i,i)=density*L*A/2;

end

for i=[2 4] %fills rotationals

m_e_new(i,i)=density*L*A*L^2/24;

end

m_e2=m_e_new;

end

if nodeclass2=='I' %any member except rightmost member

234
 if abs(moment_right)>=Mp %Hinge at right node

m_e2=Condense(m_e2);

end

end

end

k_e(2,:)=k_e2(1,:); %shifting the k_e2 matrix with zeros in

axial spots
k_e(3,:)=k_e2(2,:);
k_e(5,:)=k_e2(3,:);
k_e(6,:)=k_e2(4,:);
k_e(4,:)=zeros;
k_e(1,:)=zeros;

k_e(:,6)=k_e(:,4);
k_e(:,5)=k_e(:,3);
k_e(:,4)=zeros;
k_e(:,3)=k_e(:,2);
k_e(:,2)=k_e(:,1);
k_e(:,1)=zeros;

if analysis==2 %if modal, shifting the m_e2 matrix with zeros

in axial spots

m_e(2,:)=m_e2(1,:);
m_e(3,:)=m_e2(2,:);
m_e(5,:)=m_e2(3,:);
m_e(6,:)=m_e2(4,:);
m_e(4,:)=zeros;
m_e(1,:)=zeros;

m_e(:,6)=m_e(:,4);
m_e(:,5)=m_e(:,3);
m_e(:,4)=zeros;
m_e(:,3)=m_e(:,2);
m_e(:,2)=m_e(:,1);
m_e(:,1)=zeros;

end

235
 for i=[1,4] %fills stiffness matrix for axial components

for j=[1,4]

k_e(i,j)=k_e1(sqrt(i),sqrt(j));

if analysis==2 %if modal

m_e(i,j)=m_e1(sqrt(i),sqrt(j));

end

end

end

kms=0;
kss=0;

elseif type==6 %if polynomial finite element

if analysis==1 %if not modal, set mass equal to 0

m_e=0;
end

r=0;
syms x

indef=int(transpose(B_e_r)*C_e*B_e_r);

smaller=subs(indef,x,0);

larger=subs(indef,x,L);

k_e=larger-smaller;

k_e=double(k_e);

if analysis==2 %if modal

sizeM=size(N_e_r,2);

236
 for i=1:3:sizeM-2

N_e_m(i,1)=N_e_r(1,i);
N_e_m(i+1,1)=N_e_r(2,i+1);
N_e_m(i+2,1)=N_e_r(3,i+2);

end

indefM=int(N_e_m*transpose(N_e_m));

smallerM=subs(indefM,x,0);

largerM=subs(indefM,x,L);

m_e=DENS*A*(largerM-smallerM)
end

sizek=size(k_e,1);

if sizek>6

%static condensation to get rid of internal nodes

for i=1:1:6 % loop for kmm matrix

for j=1:1:6

kmm(i,j)=k_e(i,j);

end

end

for i=1:1:sizek-6 %loop for kss matrix

for j=1:1:sizek-6

kss(i,j)=k_e(i+6,j+6);

if analysis==2

mss(i,j)=m_e(i+6,j+6);

end

end

237
end

for i=1:1:sizek-6 %loop for ksm matrix

for j=1:1:6

ksm(i,j)=k_e(i+6,j);

if analysis==2

msm(i,j)=m_e(i+6,j);

end

end

end

kms=transpose(ksm);

k_e=kmm-kms*inv(kss)*ksm;

if analysis==2

mbot=-inv(mss)*msm;

sizembot=size(mbot,2);

Ident=eye(sizembot);

TransfM=vertcat(Ident,mbot);

rows0=size(TransfM,1);

cols0=size(mbot,2);

for i=cols0+1:1:rows0

TransfM(:,i)=0;

end

m_e=transpose(TransfM)*m_e*TransfM;

end

238
 else

kms=0;
kss=0;

end

end

end

K_geometric Function (Geometric Stiffnes Matrices)

%% This calculates k matrix (and mass if modal) for each element sent
to it

function [k_g] = k_geom(L,P,r) %Stiffness Matrix Function

k_g=-P/L/(1+r)^2*[1 0 0 -1 0
0 %geometric stiffness
0 6/5+2*r+r^2 L/10 0 -6/5-
2*r-r^2 L/10
0 L/10 2*L^2/15+L^2*r/6+L^2*r^2/12 0 -L/10
-L^2/30-L^2*r/6-L^2*r^2/12
-1 0 0 1 0
0
0 -6/5-2*r-r^2 -L/10 0
6/5+2*r+r^2 -L/10
0 L/10 -L^2/30-L^2*r/6-L^2*r^2/12 0 -L/10
2*L^2/15+L^2*r/6+L^2*r^2/12];

end

239
G_Element Function (Equivalent Nodal Loads)

%% This function calculates g for each element

function [g_e] = g_element(Length, degree1, degree2, degree3,

axial_nodal, forces_nodal, moment_nodal, N_e_r, Type_num, r, kms, kss,
Finternal, Mp, nodeclass1, nodeclass2) %Equivalent Load Function

L=Length;

Pbeam=forces_nodal(1,2);
Pbar=axial_nodal(1,2);
Pmoment=moment_nodal(1,2);

Loadvect=[Pbar
Pbeam
Pmoment];

if Type_num==6

g=int(transpose(N_e_r)*Loadvect,x);

smaller=subs(g,x,0);

larger=subs(g,x,L);

g_e=larger-smaller;

g_e=double(g_e);

sizeg=size(g_e,1);

if sizeg>6

%static condensation to get rid of internal nodes

for i=1:1:6 %loop for gm vector

gm(i,1)=g_e(i,1);

end

for i=1:1:sizeg-6 %loop for gs vector

240
 gs(i,1)=g_e(i+6,1);

end

g_e=gm-kms*inv(kss)*gs;

end

else

moment_left=Finternal(3,1);
moment_right=Finternal(6,1);

%%initially sets equivalent load vector as having no hinges

a=[0 0
-L/2 1/(1+r)
-L^2/12 -L*r/2/(1+r)
0 0
-L/2 -1/(1+r)
L^2/12 -L*r/2/(1+r)];

b=[Pbeam
Pmoment];

g_e=-a*b;

if abs(moment_right)>=Mp %Hinge at right node

a=[0 0
-5*L/8 1/(1+r)
-L^2/8 -L*r/2/(1+r)
0 0
-3*L/8 -1/(1+r)
0 -L*r/2/(1+r)];
%

a=[0 0
-7*L/12 1/(1+r)
-L^2/12 -L*r/2/(1+r)
0 0
-5*L/12 -1/(1+r)
0 -L*r/2/(1+r)];
% %

b=[Pbeam

241
 Pmoment];

g_e=-a*b;

end

if nodeclass1=='E'

if abs(moment_left)>=Mp %hinge at leftmost node

a=[0 0
-3*L/8 1/(1+r)
0 -L*r/2/(1+r)
0 0
-5*L/8 -1/(1+r)
L^2/8 -L*r/2/(1+r)];

a=[0 0
-5*L/12 1/(1+r)
0 -L*r/2/(1+r)
0 0
-7*L/12 -1/(1+r)
L^2/12 -L*r/2/(1+r)];
% %
%
%
%
% %
%

b=[Pbeam
Pmoment];

g_e=-a*b;

end
end

g_e(1,1)=Pbar*L/2;

g_e(4,1)=Pbar*L/2;

end

end

Damping Function

242
%% This function creates Rayleigh Damping Matrix

function [c_matrix_alpha c_matrix_beta] = Damping(kglobal_BC2,

mglobal_BC2, c_alpha, c_beta, damp_switch)

m=mglobal_BC2;

k=kglobal_BC2;

a=c_alpha;

b=c_beta;

c_matrix_alpha=m*a;
c_matrix_beta=k*b*damp_switch;

end

Condense Function (Static Condensation for Right Rotation)

%% This function condenses the right rotation out of the matrix

function [ksend Tr] = Condense(k)

for i=1:1:5

kab(i,1)=k(i,6);
end

kbb=k(6,6);

kinput=-kbb^-1*transpose(kab);

Id=[1 0 0 0 0;0 1 0 0 0; 0 0 1 0 0; 0 0 0 1 0; 0 0 0 0 1];

col1=vertcat(Id,kinput);

col2=[0;0;0;0;0;0];

T=horzcat(col1,col2);

ksend=transpose(T)*k*T;

Tr=kinput;

243
end

Condense Left Function (Static Condensation for Left Rotation)

%% This function condenses the right rotation out of the matrix

function [ksend Tr] = Condenseleft(k)

for i=1:1:5

kab(i,1)=k(i,6);
end

kbb=k(6,6);

kinput=-kbb^-1*transpose(kab);

Id=[1 0 0 0 0;0 1 0 0 0; 0 0 1 0 0; 0 0 0 1 0; 0 0 0 0 1];

col1=vertcat(Id,kinput);

col2=[0;0;0;0;0;0];

T=horzcat(col1,col2);

kcond=transpose(T)*k*T;

kint=kcond;

kint(:,3)=kcond(:,6);
kint(:,6)=kcond(:,3);

kint2=kint;

kint2(3,:)=kint(6,:);
kint2(6,:)=kint(3,:);

ksend=kint2;

244
Tr=kinput;

end

245
Appendix D: MDOF Solver Verifications (Time-History Plots)

All verifications use 15 ft Length, 4 Elements, W10X12, 2% damping in first two modes

unless specified otherwise

Pinned-Pinned Point Load (17 k in 17.8 ms) 40% Axial Capacity Applied

246
Pinned-Pinned Point Load (30 k in 17.8 ms) 20% Axial Capacity Applied

247
Pinned-Pinned Uniform Load (0.25 k/in in 17.8 ms) 40% Axial Capacity Applied

248
Pinned-Pinned Uniform Load (0.4 k/in in 17.8 ms) 20% Axial Capacity Applied

Large deviation in deflection curve (7%) likely due to poorly adjusted yield stress in SAP2000,
causing a smaller reduced plastic moment capacity

249
Fixed-Pinned Point Load (25 k in 17.8 ms) 40% Axial Capacity Applied

250
Fixed-Pinned Point Load (37.5 k in 17.8 ms) 20% Axial Capacity Applied

251
Fixed-Pinned Uniform Load (0.35 k/in in 17.8 ms) 40% Axial Capacity Applied

252
Fixed-Pinned Uniform Load (0.6 k/in in 17.8 ms) 20% Axial Capacity Applied

253
Fixed-Fixed Point Load (30 k in 17.8 ms) 40% Axial Capacity Applied

254
Fixed-Fixed Point Load (40 k in 17.8 ms) 20% Axial Capacity Applied

255
Fixed-Fixed Uniform Load (0.45 k/in in 17.8 ms) 40% Axial Capacity Applied

256
Fixed-Fixed Uniform Load (0.7 k/in in 17.8 ms) 20% Axial Capacity Applied

257
Cantilever Point Load (3 k in 17.8 ms) 40% Axial Capacity Applied

258
Cantilever Point Load (6 k in 17.8 ms) 20% Axial Capacity Applied

259
Cantilever Uniform Load (0.06 k/in in 17.8 ms) 40% Axial Capacity Applied

260
Cantilever Uniform Load (0.125 k/in in 17.8 ms) 20% Axial Capacity Applied

261
Pinned-Pinned Uniform Load (0.45 k/in in 17.8 ms) with Added Mass 10X Self Weight,

α=0 β=0.0000002 (damping)

262
Cantilever Point Load (20 k in 17.8 ms) with Added Mass 10X Self Weight

α=0 β=0.0000002 (damping)

263
Appendix E: First Data Set Plots
Various Sections With 2% Damping in First and Second Modes

W14X48 Section
KL factor

Pinpin KL factor
200
15psf
150
KL deviation (%)

40psf
81.25psf
100

50

-50
0 5 10 15 20 25
Z (ft/lb(1/3))
Fixpin KL factor
200
15psf
150
KL deviation (%)

40psf
81.25psf
100

50

-50
4 6 8 10 12 14 16 18 20 22 24
Z (ft/lb(1/3))
Fixfix KL factor
200
15psf
150
KL deviation (%)

40psf
81.25psf
100

50

-50
4 6 8 10 12 14 16 18 20 22 24
Z (ft/lb(1/3))

264
SDOF-Bernoulli

Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection
0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

265
SDOF Timoshenko

Max deflection deviation (%)

Pinpin SDOF-Timoshenko Deflection
50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

266
Bernoulli Timoshenko

Max deflection deviation (%)

Pinpin Timoshenko-Bernoulli Deflection
100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

267
W14X53:
KL factor

Pinpin KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
Z (ft/lb(1/3))

Fixpin KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

4 6 8 10 12 14 16 18 20 22 24
Z (ft/lb(1/3))

Fixfix KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

4 6 8 10 12 14 16 18 20 22 24
Z (ft/lb(1/3))

268
SDOF Bernoulli

Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection
0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

269
SDOF Timoshenko

Max deflection deviation (%)

Pinpin SDOF-Timoshenko Deflection
50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

270
Bernoulli Timoshenko

Max deflection deviation (%)

Pinpin Timoshenko-Bernoulli Deflection
100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

271
W14X61
KL Factor

Pinpin KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
Z (ft/lb(1/3))

Fixpin KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

4 6 8 10 12 14 16 18 20 22 24
Z (ft/lb(1/3))

Fixfix KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

4 6 8 10 12 14 16 18 20 22 24
Z (ft/lb(1/3))

272
SDOF Bernoulli

Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection
0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

273
SDOF Timoshenko

Max deflection deviation (%)

Pinpin SDOF-Timoshenko Deflection
50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

274
Bernoulli Timoshenko

Max deflection deviation (%)

Pinpin Timoshenko-Bernoulli Deflection
100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

275
W14X82
KL Factor

Pinpin KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
Z (ft/lb(1/3))

Fixpin KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

4 6 8 10 12 14 16 18 20 22 24
Z (ft/lb(1/3))

Fixfix KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

4 6 8 10 12 14 16 18 20 22 24
Z (ft/lb(1/3))

276
SDOF Bernoulli

Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection
0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

277
SDOF Timoshenko

Max deflection deviation (%)

Pinpin SDOF-Timoshenko Deflection
50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

278
Bernoulli Timoshenko

Max deflection deviation (%)

Pinpin Timoshenko-Bernoulli Deflection
100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

279
W14X109
KL Factor

Pinpin KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
Z (ft/lb(1/3))

Fixpin KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

4 6 8 10 12 14 16 18 20 22 24
Z (ft/lb(1/3))

Fixfix KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

4 6 8 10 12 14 16 18 20 22 24
Z (ft/lb(1/3))

280
SDOF Bernoulli

Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection
0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

281
SDOF Timoshenko

Max deflection deviation (%)

Pinpin SDOF-Timoshenko Deflection
50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

282
Bernoulli Timoshenko

Max deflection deviation (%)

Pinpin Timoshenko-Bernoulli Deflection
100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

283
W14X132
KL Factor

Pinpin KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
Z (ft/lb(1/3))

Fixpin KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
Z (ft/lb(1/3))

Fixfix KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
Z (ft/lb(1/3))

284
SDOF Bernoulli

Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection
0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

285
SDOF Timoshenko

Max deflection deviation (%)

Pinpin SDOF-Timoshenko Deflection
50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

286
Bernoulli Timoshenko

Max deflection deviation (%)

Pinpin Timoshenko-Bernoulli Deflection
100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

287
W14X132
KL Factor

Pinpin KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
Z (ft/lb(1/3))

Fixpin KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
Z (ft/lb(1/3))

Fixfix KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
Z (ft/lb(1/3))

288
SDOF Bernoulli

Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection
0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

289
SDOF Timoshenko

Max deflection deviation (%)

Pinpin SDOF-Timoshenko Deflection
50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

290
Bernoulli Timoshenko

Max deflection deviation (%)

Pinpin Timoshenko-Bernoulli Deflection
100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

291
W14X145
KL Factor

Pinpin KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
Z (ft/lb(1/3))

Fixpin KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
Z (ft/lb(1/3))

Fixfix KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
Z (ft/lb(1/3))

292
SDOF Bernoulli

Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection
0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

293
SDOF Timoshenko

Max deflection deviation (%)

Pinpin SDOF-Timoshenko Deflection
50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

294
Bernoulli Timoshenko

Max deflection deviation (%)

Pinpin Timoshenko-Bernoulli Deflection
100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

295
W14X257
KL Factor

Pinpin KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
Z (ft/lb(1/3))

Fixpin KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
Z (ft/lb(1/3))

Fixfix KL factor
200
KL deviation (%)

15psf
100 40psf
81.25psf
0

0 5 10 15 20 25
Z (ft/lb(1/3))

296
SDOF Bernoulli

Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection
0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Bernoulli Deflection

0
0 axial 15 psf
0 axial 40 psf
-50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-100
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

297
SDOF Timoshenko

Max deflection deviation (%)

Pinpin SDOF-Timoshenko Deflection
50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix SDOF-Timoshenko Deflection

50
0 axial 15 psf
0 axial 40 psf
0 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
-50
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

298
Bernoulli Timoshenko

Max deflection deviation (%)

Pinpin Timoshenko-Bernoulli Deflection
100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixpin Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
Z (ft/lb(1/3)
) 0.4 axial 15 psf
0.4 axial 40 psf
0.4 axial 81.25 psf
Max deflection deviation (%)

Fixfix Timoshenko-Bernoulli Deflection

100
0 axial 15 psf
0 axial 40 psf
50 0 axial 81.25 psf
0.2 axial 15 psf
0.2 axial 40 psf
0
0 5 10 15 200.2 axial 81.25 psf 25
0.4 axial 15 psf
Z (ft/lb(1/3))
0.4 axial 40 psf
0.4 axial 81.25 psf

299
Appendix F: Second Data Set Plots
W14X109 With 2% Damping in First and Second Modes
KL factor

Pinpin KL factor
200
15 psf 25 lb charge
KL deviation (%)

150
15 psf 200 lb charge
100 40 psf 25 lb charge
50 40 psf 200 lb charge
81.25 psf 25 lb charge
0 81.25 psf 200 lb charge
-50
2 4 6 8 10 12 14 16
(1/3)
Z (ft/lb )

Fixpin KL factor
200
15 psf 25 lb charge
KL deviation (%)

150
15 psf 200 lb charge
100 40 psf 25 lb charge
40 psf 200 lb charge
50
81.25 psf 25 lb charge
0 81.25 psf 200 lb charge
-50
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

Fixfix KL factor
200
15 psf 25 lb charge
KL deviation (%)

150
15 psf 200 lb charge
100 40 psf 25 lb charge
40 psf 200 lb charge
50
81.25 psf 25 lb charge
0 81.25 psf 200 lb charge
-50
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

300
Pinned-Pinned SDOF Bernoulli

Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection 0 Axial
0
15 psf 25 lb charge
15 psf 200 lb charge
-50 40 psf 25 lb charge
40 psf 200 lb charge
81.25 psf 25 lb charge
-100
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection 0.2 Axial

0
15 psf 25 lb charge
15 psf 200 lb charge
-50 40 psf 25 lb charge
40 psf 200 lb charge
81.25 psf 25 lb charge
-100
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection 0.4 Axial

0
15 psf 25 lb charge
15 psf 200 lb charge
-50 40 psf 25 lb charge
40 psf 200 lb charge
81.25 psf 25 lb charge
-100
2 4 6 8 10 12 81.25 psf14
200 lb charge 16

Z (ft/lb(1/3))

301
Pinned-Pinned SDOF Timoshenko

Max deflection deviation (%)

Pinpin SDOF-Timoshenko Deflection 0 Axial
50
15 psf 25 lb charge
15 psf 200 lb charge
0 40 psf 25 lb charge
40 psf 200 lb charge
81.25 psf 25 lb charge
-50
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Pinpin SDOF-Timoshenko Deflection 0.2 Axial

50
15 psf 25 lb charge
15 psf 200 lb charge
0 40 psf 25 lb charge
40 psf 200 lb charge
81.25 psf 25 lb charge
-50
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Pinpin SDOF-Timoshenko Deflection 0.4 Axial

50
15 psf 25 lb charge
15 psf 200 lb charge
0 40 psf 25 lb charge
40 psf 200 lb charge
81.25 psf 25 lb charge
-50
2 4 6 8 10 12 81.25 psf14
200 lb charge 16

Z (ft/lb(1/3))

302
Pinned-Pinned Bernoulli Timoshenko

Max deflection deviation (%)

Pinpin Bernoulli-Timoshenko Deflection 0 Axial
150
15 psf 25 lb charge
100 15 psf 200 lb charge
40 psf 25 lb charge
50
40 psf 200 lb charge
81.25 psf 25 lb charge
0
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Pinpin Bernoulli-Timoshenko Deflection 0.2 Axial

150
15 psf 25 lb charge
100 15 psf 200 lb charge
40 psf 25 lb charge
50
40 psf 200 lb charge
81.25 psf 25 lb charge
0
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Pinpin Bernoulli-Timoshenko Deflection 0.4 Axial

150
15 psf 25 lb charge
100 15 psf 200 lb charge
40 psf 25 lb charge
50
40 psf 200 lb charge
81.25 psf 25 lb charge
0
2 4 6 8 10 12 81.25 psf14
200 lb charge 16

Z (ft/lb(1/3))

303
Fixed-Pinned SDOF Bernoulli

Max deflection deviation (%)

Fixpin SDOF-Bernoulli Deflection 0 Axial

0 15 psf 25 lb charge
-20 15 psf 200 lb charge
40 psf 25 lb charge
-40 40 psf 200 lb charge
81.25 psf 25 lb charge
-60
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Fixpin SDOF-Bernoulli Deflection 0.2 Axial

0 15 psf 25 lb charge
-20 15 psf 200 lb charge
40 psf 25 lb charge
-40 40 psf 200 lb charge
81.25 psf 25 lb charge
-60
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Fixpin SDOF-Bernoulli Deflection 0.4 Axial

0 15 psf 25 lb charge
-20 15 psf 200 lb charge
40 psf 25 lb charge
-40 40 psf 200 lb charge
81.25 psf 25 lb charge
-60
2 4 6 8 10 12 81.25 psf14
200 lb charge 16

Z (ft/lb(1/3))

304
Fixed-Pinned SDOF Timoshenko

Max deflection deviation (%)

Fixpin SDOF-Timoshenko Deflection 0 Axial
40
15 psf 25 lb charge
20
15 psf 200 lb charge
0 40 psf 25 lb charge
40 psf 200 lb charge
-20
81.25 psf 25 lb charge
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Fixpin SDOF-Timoshenko Deflection 0.2 Axial

40
15 psf 25 lb charge
20
15 psf 200 lb charge
0 40 psf 25 lb charge
40 psf 200 lb charge
-20
81.25 psf 25 lb charge
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Fixpin SDOF-Timoshenko Deflection 0.4 Axial

40
15 psf 25 lb charge
20
15 psf 200 lb charge
0 40 psf 25 lb charge
40 psf 200 lb charge
-20
81.25 psf 25 lb charge
2 4 6 8 10 12 81.25 psf14
200 lb charge 16

Z (ft/lb(1/3))

305
Fixed-Pinned Bernoulli Timoshenko

Max deflection deviation (%)

Fixpin Bernoulli-Timoshenko Deflection 0 Axial
100
15 psf 25 lb charge
15 psf 200 lb charge
50 40 psf 25 lb charge
40 psf 200 lb charge
81.25 psf 25 lb charge
0
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Fixpin Bernoulli-Timoshenko Deflection 0.2 Axial

100
15 psf 25 lb charge
15 psf 200 lb charge
50 40 psf 25 lb charge
40 psf 200 lb charge
81.25 psf 25 lb charge
0
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Fixpin Bernoulli-Timoshenko Deflection 0.4 Axial

100
15 psf 25 lb charge
15 psf 200 lb charge
50 40 psf 25 lb charge
40 psf 200 lb charge
81.25 psf 25 lb charge
0
2 4 6 8 10 12 81.25 psf14
200 lb charge 16

Z (ft/lb(1/3))

306
Fixed-Fixed SDOF Bernoulli

Max deflection deviation (%)

Fixfix SDOF-Bernoulli Deflection 0 Axial
20
15 psf 25 lb charge
0 15 psf 200 lb charge
40 psf 25 lb charge
-20
40 psf 200 lb charge
81.25 psf 25 lb charge
-40
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Fixfix SDOF-Bernoulli Deflection 0.2 Axial

20
15 psf 25 lb charge
0 15 psf 200 lb charge
40 psf 25 lb charge
-20
40 psf 200 lb charge
81.25 psf 25 lb charge
-40
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Fixfix SDOF-Bernoulli Deflection 0.4 Axial

20
15 psf 25 lb charge
0 15 psf 200 lb charge
40 psf 25 lb charge
-20
40 psf 200 lb charge
81.25 psf 25 lb charge
-40
4 6 8 10 12 81.25 psf
14 200 lb charge 16
Z (ft/lb(1/3))

307
Fixed-Fixed SDOF Timoshenko

Max deflection deviation (%)

Fixfix SDOF-Timoshenko Deflection 0 Axial
60
40 15 psf 25 lb charge
15 psf 200 lb charge
20
40 psf 25 lb charge
0
40 psf 200 lb charge
-20 81.25 psf 25 lb charge
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Fixfix SDOF-Timoshenko Deflection 0.2 Axial

60
40 15 psf 25 lb charge
15 psf 200 lb charge
20
40 psf 25 lb charge
0
40 psf 200 lb charge
-20 81.25 psf 25 lb charge
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Fixfix SDOF-Timoshenko Deflection 0.4 Axial

60
40 15 psf 25 lb charge
15 psf 200 lb charge
20
40 psf 25 lb charge
0
40 psf 200 lb charge
-20 81.25 psf 25 lb charge
2 4 6 8 10 12 81.25 psf14
200 lb charge 16

Z (ft/lb(1/3))

308
Fixed-Fixed Bernoulli Timoshenko

Max deflection deviation (%)

Fixfix Bernoulli-Timoshenko Deflection 0 Axial
60
15 psf 25 lb charge
40 15 psf 200 lb charge
40 psf 25 lb charge
20
40 psf 200 lb charge
81.25 psf 25 lb charge
0
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Fixfix Bernoulli-Timoshenko Deflection 0.2 Axial

60
15 psf 25 lb charge
40 15 psf 200 lb charge
40 psf 25 lb charge
20
40 psf 200 lb charge
81.25 psf 25 lb charge
0
2 4 6 8 10 12 81.25 psf14
200 lb charge 16
(1/3)
Z (ft/lb )
Max deflection deviation (%)

Fixfix Bernoulli-Timoshenko Deflection 0.4 Axial

60
15 psf 25 lb charge
40 15 psf 200 lb charge
40 psf 25 lb charge
20
40 psf 200 lb charge
81.25 psf 25 lb charge
0
4 6 8 10 12 81.25 psf
14 200 lb charge 16
Z (ft/lb(1/3))

309
Appendix G: Third Data Set Plots
W14X109 With 2% Damping in First and Last Modes
KL factor

Pinpin KL factor
200
KL deviation (%)

15 psf 25 lb charge
15 psf 200 lb charge
100
40 psf 25 lb charge
40 psf 200 lb charge
0 81.25 psf 25 lb charge
81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

Fixpin KL factor
200
KL deviation (%)

15 psf 25 lb charge
15 psf 200 lb charge
100
40 psf 25 lb charge
40 psf 200 lb charge
0 81.25 psf 25 lb charge
81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

Fixfix KL factor
200
KL deviation (%)

15 psf 25 lb charge
15 psf 200 lb charge
100
40 psf 25 lb charge
40 psf 200 lb charge
0 81.25 psf 25 lb charge
81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

310
Pinned-Pinned SDOF Bernoulli

Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection 0 Axial
40
15 psf 25 lb charge
20 15 psf 200 lb charge
40 psf 25 lb charge
0 40 psf 200 lb charge
81.25 psf 25 lb charge
-20 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection 0.2 Axial

40
15 psf 25 lb charge
20 15 psf 200 lb charge
40 psf 25 lb charge
0 40 psf 200 lb charge
81.25 psf 25 lb charge
-20 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Pinpin SDOF-Bernoulli Deflection 0.4 Axial

40
15 psf 25 lb charge
20 15 psf 200 lb charge
40 psf 25 lb charge
0 40 psf 200 lb charge
81.25 psf 25 lb charge
-20 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

311
Pinned-Pinned SDOF Timoshenko

Max deflection deviation (%)

Pinpin SDOF-Timoshenko Deflection 0 Axial
100
15 psf 25 lb charge
15 psf 200 lb charge
50
40 psf 25 lb charge
40 psf 200 lb charge
0 81.25 psf 25 lb charge
81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Pinpin SDOF-Timoshenko Deflection 0.2 Axial

100
15 psf 25 lb charge
15 psf 200 lb charge
50
40 psf 25 lb charge
40 psf 200 lb charge
0 81.25 psf 25 lb charge
81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Pinpin SDOF-Timoshenko Deflection 0.4 Axial

100
15 psf 25 lb charge
15 psf 200 lb charge
50
40 psf 25 lb charge
40 psf 200 lb charge
0 81.25 psf 25 lb charge
81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

312
Pinned-Pinned Bernoulli Timoshenko

Max deflection deviation (%)

Pinpin Bernoulli-Timoshenko Deflection 0 Axial
60
15 psf 25 lb charge
40
15 psf 200 lb charge
20 40 psf 25 lb charge
40 psf 200 lb charge
0
81.25 psf 25 lb charge
-20 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Pinpin Bernoulli-Timoshenko Deflection 0.2 Axial

60
15 psf 25 lb charge
40
15 psf 200 lb charge
20 40 psf 25 lb charge
40 psf 200 lb charge
0
81.25 psf 25 lb charge
-20 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Pinpin Bernoulli-Timoshenko Deflection 0.4 Axial

60
15 psf 25 lb charge
40
15 psf 200 lb charge
20 40 psf 25 lb charge
40 psf 200 lb charge
0
81.25 psf 25 lb charge
-20 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

313
Fixed-Pinned SDOF Bernoulli

Max deflection deviation (%)

Fixpin SDOF-Bernoulli Deflection 0 Axial
40
15 psf 25 lb charge
20 15 psf 200 lb charge
40 psf 25 lb charge
0
40 psf 200 lb charge
-20 81.25 psf 25 lb charge
81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Fixpin SDOF-Bernoulli Deflection 0.2 Axial

40
15 psf 25 lb charge
20 15 psf 200 lb charge
40 psf 25 lb charge
0
40 psf 200 lb charge
-20 81.25 psf 25 lb charge
81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Fixpin SDOF-Bernoulli Deflection 0.4 Axial

40
15 psf 25 lb charge
20 15 psf 200 lb charge
40 psf 25 lb charge
0
40 psf 200 lb charge
-20 81.25 psf 25 lb charge
81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

314
Fixed-Pinned SDOF Timoshenko

Max deflection deviation (%)

Fixpin SDOF-Timoshenko Deflection 0 Axial
100
15 psf 25 lb charge
15 psf 200 lb charge
50
40 psf 25 lb charge
40 psf 200 lb charge
0 81.25 psf 25 lb charge
81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Fixpin SDOF-Timoshenko Deflection 0.2 Axial

100
15 psf 25 lb charge
15 psf 200 lb charge
50
40 psf 25 lb charge
40 psf 200 lb charge
0 81.25 psf 25 lb charge
81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Fixpin SDOF-Timoshenko Deflection 0.4 Axial

100
15 psf 25 lb charge
15 psf 200 lb charge
50
40 psf 25 lb charge
40 psf 200 lb charge
0 81.25 psf 25 lb charge
81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

315
Fixed-Pinned Bernoulli Timoshenko

Max deflection deviation (%)

Fixpin Bernoulli-Timoshenko Deflection 0 Axial
80
60 15 psf 25 lb charge
15 psf 200 lb charge
40
40 psf 25 lb charge
20
40 psf 200 lb charge
0 81.25 psf 25 lb charge
-20 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Fixpin Bernoulli-Timoshenko Deflection 0.2 Axial

80
60 15 psf 25 lb charge
15 psf 200 lb charge
40
40 psf 25 lb charge
20
40 psf 200 lb charge
0 81.25 psf 25 lb charge
-20 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Fixpin Bernoulli-Timoshenko Deflection 0.4 Axial

80
60 15 psf 25 lb charge
15 psf 200 lb charge
40
40 psf 25 lb charge
20
40 psf 200 lb charge
0 81.25 psf 25 lb charge
-20 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

316
Fixed-Fixed SDOF Bernoulli

Max deflection deviation (%)

Fixfix SDOF-Bernoulli Deflection 0 Axial

20 15 psf 25 lb charge
15 psf 200 lb charge
40 psf 25 lb charge
0
40 psf 200 lb charge
81.25 psf 25 lb charge
-20 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Fixfix SDOF-Bernoulli Deflection 0.2 Axial

20 15 psf 25 lb charge
15 psf 200 lb charge
40 psf 25 lb charge
0
40 psf 200 lb charge
81.25 psf 25 lb charge
-20 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Fixfix SDOF-Bernoulli Deflection 0.4 Axial

20 15 psf 25 lb charge
15 psf 200 lb charge
40 psf 25 lb charge
0
40 psf 200 lb charge
81.25 psf 25 lb charge
-20 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

317
Fixed-Fixed SDOF Timoshenko

Max deflection deviation (%)

Fixfix SDOF-Timoshenko Deflection 0 Axial

40 15 psf 25 lb charge
15 psf 200 lb charge
40 psf 25 lb charge
20
40 psf 200 lb charge
81.25 psf 25 lb charge
0 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Fixfix SDOF-Timoshenko Deflection 0.2 Axial

40 15 psf 25 lb charge
15 psf 200 lb charge
40 psf 25 lb charge
20
40 psf 200 lb charge
81.25 psf 25 lb charge
0 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Fixfix SDOF-Timoshenko Deflection 0.4 Axial

40 15 psf 25 lb charge
15 psf 200 lb charge
40 psf 25 lb charge
20
40 psf 200 lb charge
81.25 psf 25 lb charge
0 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

318
Fixed-Fixed Bernoulli Timoshenko

Max deflection deviation (%)

Fixfix Bernoulli-Timoshenko Deflection 0 Axial
60
15 psf 25 lb charge
40 15 psf 200 lb charge
40 psf 25 lb charge
20 40 psf 200 lb charge
81.25 psf 25 lb charge
0 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Fixfix Bernoulli-Timoshenko Deflection 0.2 Axial

60
15 psf 25 lb charge
40 15 psf 200 lb charge
40 psf 25 lb charge
20 40 psf 200 lb charge
81.25 psf 25 lb charge
0 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))
Max deflection deviation (%)

Fixfix Bernoulli-Timoshenko Deflection 0.4 Axial

60
15 psf 25 lb charge
40 15 psf 200 lb charge
40 psf 25 lb charge
20 40 psf 200 lb charge
81.25 psf 25 lb charge
0 81.25 psf 200 lb charge
2 4 6 8 10 12 14 16
Z (ft/lb(1/3))

319
Vita

Evan Mullen was born in on February 15, 1992 in Philadelphia, Pennsylvania. He is the

son of Mark and Lauren Mullen. Evan graduated from Lehigh University in Bethlehem,

Pennsylvania in May of 2014 with a Bachelor of Science in Civil Engineering. In January

2016, Evan will receive his Master of Science in Structural Engineering from Lehigh

University. Evan works as a structural engineer for Thornton Tomasetti in Austin,

Texas.

320