247 AD-A241 I 1111 IH IiI 111III li 111III ii ~~~liiil

ELASTIC SECOND-ORDER COMPUTER ANALYSIS OF BEAM-COLUMNS AND FRAMES

APPROVED
Supervisor

_____________

Michael D. E-ngd.hardt

Dan L. Wheat

10

91-12349 I iII1' II II'11 !ll

ELASTIC SECOND-ORDER

COMPUTER ANALYSIS OF BEAM-COLUMNS AND FRAMES

".

'

APPROVED

Supe rvi sor

~~~"~~
Michael D. 1ingihardt

Dan L. Wheat

To my wife Virginia and my children Mervin and Exrica

ELASTIC SECOND-ORDER COMPUTER ANALYSIS OF BEAM-COLUMNS AND FRAMES

by

HERMOGENES REYES MUNOZ, B.S.

THESIS

Presented to the Faculty of the Graduate School of The University of Texas at Austin in Partial Fulfillment of the Requirements for the Degree of

MASTER OF SCIENCE IN ENGINEERING

THE UNIVERSITY OF TEXAS AT AUSTIN AUGUST 1991

ACKNOWLEDGMENTS
The author extends his sincere gratitude to Dr. Michael D. Engelhardt for his guidance and supervision of this thesis, and his moral support to my academic endeavors at The University of Texas at Austin. The author also wishes to express his sincere

appreciation to Dr. Dan L. Wheat for his support, and his genuine effort to read the thesis. Special thanks is also extended to Dr. James 0. Jirsa for his help in achieving my Educational Plan at the University of Texas at Austin. And most importantly, my sincere thanks to the

United States Navy for sponsoring my graduate education, and to the personnel of the Naval Reserve Officers'

Training Corps Unit at the University of Texas at Austin for their administrative support. Finally, the author is indebted to his wife and children for their love, understanding, and moral

support, and in return this thesis is dedicated to them.

TABLE OF CONTENTS Page

ACKNOWLEDGMENTS TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES iv v viii x

INTRODUCTION 1.1 1.2 1.3 General Obiective Scope

1.3-1 Methods of Analysis

1 1 2 4
4

1.3-2 1.3-3

Assumptions Organization of the Report

5 7
9

BACKGROUND ON BEAM-COLUMNS

ELASTIC SECOND-ORDER ANALYSIS 3.1 3.2 3.3 General Methodology Derivation of Beam-Column Stiffness Matrix 3.3-1 3.3-2 3.4 Translation Stability Functions Rotation Stability Functions

13 13 14 16 16 26 36

Load Functions and Maximum Moments

v

3.4-1 3.4-2

Load Functions Maximum Elastic Second-Order Moments

36 42 56 56 56

ELASTIC BUCKLING LOAD 4.1 4.2 General Computer Analysis of the Elastic Stability of Plane Frames

VERIFICATION OF COMPUTER PROGRAM (E S C A P 5.1 5.2 General Sample Problems

64

64 64 64

Example 5.2-1 Simply-Supported Beam-Column with Uniform Distributed Load Example 5.2-2 Fixed-Fixed Beam-Column with Uniform Distributed Load Example 5.2-3 Hinged-Fixed Beam-Column with Uniform Distributed load Example 5.2-4 Elastic Second-Order Analysis of a Frame Example 5.2-5 Elastic Critical Load of a Frame Example 5.2-6 Elastic Second-order Analysis of A Three-Story Frame

72

77

86 92 95

vi

CONCLUSIONS

99 102 127

APPENDIX REFERENCES

vii

LIST OF TABLES
Page TABLE TABLE TABLE TABLE TABLE TABLE TABLE TABLE TABLE TABLE TABLE TABLE I II III IV V VI VII VIII IX X XI XII STABILITY FUNCTIONS OF A BEAM-COLUMN MAXIMUM BENDING MOMENTS SIMPLY SUPPORTED BEAM-COLUMN MAXIMUM DEFLECTIONS SIMPLY-SUPPORTED BEAM-COLUMN RELATIVE DISPLACEMENTS SIMPLY SUPPORTED BEAM-COLUMN MAXIMUM BENDING MOMENTS FIXED-FIXED BEAM-COLUMN MAXIMUM DEFLECTION: FIXED-FIXED BEAM-COLUMN RELATIVE DISPLACEMENTS FIXED-FIXED BEAM-COLUMN MAXIMUM BENDING MOMENTS HINGED-FIXED BEAM-COLUMN MAXIMUM MOMENTS-MEMBER AC HINGED-FIXED BEAM-COLUMN DEFLECTIONS-MEMBER AC HINGED-FIXED BEAM-COLUMN RELATIVE DISPLACEMENTS HINGED-FIXED BEAM-COLUMN SECOND-ORDER MOMENTS OF A PLANE FRAME EXAMPLE No. 5 SECOND-ORDER DEFLECTIONS OF A PLANE FRAME EXAMPLE No. 5 34 69 70 70 75 76 76 82 83 84 85 90

TABLE

XIII

91

viii

TABLE

XIV

SECOND-ORDER MOMENTS THREE-STORY FRAME SECOND-ORDER DEFLECTIONS THREE-STORY FRAME

97

TABLE

XV

98

ix

LIST OF FIGURES
Page

Fig.

EFFECTS OF AXIAL FORCE IN A BEAM-COLUMN FLOWCHART-ELASTIC SECOND ORDER ANALYSIS STIFFNESS COEFFICIENTS OF A BEAM-COLUMN UNIT TRANSLATION - ROTATION OF A BEAM-COLUMN TYPES OF LOADING FIXED-END ACTIONS FOR A CONCENTRATED LOAD - COMPRESSION FIXED-END ACTION FOR A CONCENTRATED LOAD - TENSION FIXED-END ACTIONS FOR A UNIFORM LOAD MAXIMUM ELASTIC SECOND-ORDER MOMENTS DERIVATION OF MAXIMUM MOMENT TRANSFORMED AXIS OF A

10

Fig.

15

Fig.

17

Fig.

18

Fig. Fig.

5 6

38 39

Fig.

40

Fig.

41

Fig.

43

Fig. Fig.

10 11

45 46

BEAM-COLUMN Fig. Fig. 12 13 CHOLESKY'S DECOMPOSITION FLOWCHART-ELASTIC STABILITY

ANALYSIS Fig. Fig. 14 15 FLOWCHART-SUBROUTINE DECOMP SIMPLY-SUPPORTED BEAM-COLUMN 63 66

59 61

Fig. Fig Fig. Fig.

16 17 18 19

LOAD VS. DEFLECTION/MOMENT CURVE FIXED-FIXED BEAM-COLUMN HINGED-FIXED BEAM-COLUMN SUPERPOSITION OF LATERAL LOAD AND MOMENT ANALYSIS OF A RIGID PLANE MOMENT DIAGRAMS-FIRST/SECOND ORDER ELASTIC DUCKLING ANALYSIS OF PORTAL FRAME THREE-STORY FRAME

71 73 78 79

Fig. Fig. Fig. Fig.

20 21 22 23

88 89 94 96

yi

INTRODUCTION

1.1 GENERAL Today, many computer programs in structural analysis are readily available and each is designed for a specific use. Just to name a few are SAP90, ETABS, SAFE, M-

STRUDL, MICROFEAP, MULTIFRAME, RISA-2D, etc... and many more. These programs are generally used for structural and design. Most of these programs have

analysis

graphics capability which provides visual presentation on the behavior and response of the structure. While most of these programs are developed using sophisticated computer methods in elastic structural

analysis, few are designed to perform elastic secondorder analysis. Its use can be of considerable

importance in the design of various structural systems, particularly frame type structures. In this report, the development and implementation of computerized elastic second-order analysis of frames is presented. examples. The technique is illustrated by several

2 1.2 OBJECTIVES

The primary objective of this report is to develop and implement into a computer program the concepts of elastic second-order analysis. An additional objective is to

extend these concepts to perform computerized elastic stability analysis. In buildings the
1

current 1

LRFD

Specification

for

steel

(1st Ed.,AISC 1986),

"

Chapter C, Section

Cl states that,

second order ( P-Delta ) effects shall

"

be considered in the design of frames.

Chapter H,

Section 2a, of the Manual, which deals with beam-column design, specifies three (3) ways to determine the

required flexural strength M, a.

of a member:

M u may be determined using second-order

elastic analysis using factored loads

b. M . may be determined using plastic

analysis that satisfies the requirements of Sections C1 and C2 of Chapter C of the Manual. Section C1 deals with

the second-order analysis while Section C2 deals with frame stability. c. analysis Mu may be determined using linear-elastic combined with approximate

(first-order)

amplification factors.

3 The LRFD Specification therefore requires explicit

consideration of second-order effects and permits either exact or approximate methods of analysis. Typical current design practice uses elastic firstorder analysis with approximate amplification factors to determine M u. consuming, This procedure may be tedious and time if computers are being used. In

even

addition, the approximate amplification factors provided in the LRFD Specification, are generally applicable only to simple, rectangular framing. of determining the The approximate methods factors for complex

amplification

irregular framing are not yet well developed. It is the specific purpose of this report to develop an automated elastic-second order analysis computer

program that will meet the requirements of Section 2a, Chapter H and Chapter C of the LRFD Manual. The use of an automated elastic second-order analysis program will

remove the burden of hand calculations associated with the approximate methods. elastic second-order It will also permit accurate of planar frames of

analysis

arbitrary geometry.

1.3 SCOPE

1.3-1 Methods of Analysis The elastic second-order analysis of a beam-column which was
2

originated from the

"Buckling of Columns"

first introduced in 1729 by Van Musschenbroek 1744, Euler introduced the first accurate

In

stability
3.

analysis of a centrally loaded elastic column members that were the subject of

Other study

subsequent

included eccentrically loaded columns, columns with end moments, and laterally loaded columns. columns were treated as "beam-columns". There are two basic methods commonly used in These types of

performing computerized elastic second-order analysis of beam-columns and frames. These are called the stability

function approach and the geometric stiffness approach. The latter is an approximate approach, and will provide accurate solutions for a limited class of problems. stability function approach, on the other hand, The is

directly based on the governing differential equations of the elastic beam-column, and provides exact solutions. Further, in order to perform accurate elastic stability analysis requires the use of stability functions. approach is used in the development of the This

Elastic

Second-order Computer Analysis Program described in this report.

Timoshenko and Gere and Lui [ 5, provide
,

[ E S C A P ]

Chen and Atsuta I2 to various

Chen of

solutions

cases

elastic beam-columns.

Salmon and Johnson ( 6 1 give an subject of beam-

extensive list of references on the columns.

1.3-2 Assumptions The development of elastic second-order analysis is based on the following assumptions [2, a. b.
3, 7 ] :

The material is linearly elastic. Rotations are small, so that the following

approximations can be made :

i. sin 0 = ii. iii. cos

2

=1
=0

6 c. is The exact expression for the curvature of a line

_ P where p
*

1Eq.

(1)

[1 + (y') 21 3/ 2 radius of curvature

= curvature y = deflection

As a consequence of the small rotation assumption ( y

=

rotation ), the usual relationship between curvature

and deflection results

4 = y"

Eq. (la)

d. e. f.

The beam-column is prismatic. Shear deformation is neglected. The frame members are two-dimensional and subjected to in-plane forces and deformations only.

1.3-3

Organization of the Report The first chapter

The report consists of six chapters. is the Introduction. theory of beam-columns.

Chapter 2 deals with the basic The concept of elastic second-

order is explored at some depth to give an appreciation of the subject. Its relative importance to the elastic

stability problem is briefly discussed and later covered in Chapter 4. Chapter 3 presents the methodology that is

implemented in the development of the computer program L E S C A P I. linear-elastic It explains how an ordinary first-order program using the direct-

analysis

stiffness method 1 8 1 is modified to include second-order effects through the implementation of stability

functions. matrix is

The derivation of the beam-column stiffness presented using In Section 3.4-2, second-order differential

equations.

a method for determining

the maximum elastic second-order moment within a member is presented. In Chapter 4, the reader is introduced to the

concept of elastic stability analysis of beam-columns and frames. The use of Cholesky's decomposition of the

stiffness matrix in determining the elastic buckling load of frames and beam-columns is introduced and discussed in

8 detail. Chapter 5 presents several examples of beam-columns and frames. The computer solutions are compared to

closed-form or published solutions.

Some of the computer

solutions to the sample problems are listed in Appendix B. Also, a Users Manual for the computer program is

listed in Appendix A. And finally, Chapter 6 provides a conclusion on the results of the analysis of the computer pzraa ( E S C A P ]. The importance of elastic second-order

analysis and elastic stability analysis of frames are also summarized. provided. A list of selected references is also

BACKGROUND ON BEAMCOLUMNS

The term beam-column refers to a structural member that is subjected to both axial compression tension) force and bending moment. (or

It is a structural

element simplifying to a beam or a column as special cases. A column is a structural element where axial

compression predominates, while a beam is a structural element where bending predominates. the axial force The interaction of result to a

and bending moment may

nonlinear behavior of a member in terms of : a. the axial force and the resulting second-order deflections, and ; b. the axial force and the resulting second-order moments. The elastic analysis of beam-columns must consider the effect of change in geometry of the structural

element, i.e, second-order effects.

As shown in Fig. 1,

the axial force will magnify the deflection and moment in the beam. When the axial force approaches a critical

value, the deflection and moment of the beam-

10~

(a)
2rid-.,rderdeflection

Deflection due to 0 Q
first-order deflection

(b)

Deflection amplified by axial force P

2nd-order moment

~fir,

t-order moment Mo

(c) Moment amplified

by axial force P

Fig. 1

Effects of Axial Force In a Beam-Column

11 column increase without bound. the elastic buckling load and This critical value is can be determined by

performing elastic stability analysis of the member or structure. Most structures are composed of many structural

elements, where each element serves a unique purpose, and the combination meets the functional objectives of the structure. Because these elements are interconnected, When both

they interact under a given loading pattern.

axial forces and moments are present in the members, they respond as beam-columns and the deformations resulting from the bending moments loads may be induced by the externally of

applied

amplified

by

the

presence

compressive axial loads.

However, when the axial force

is tension, the effect on the resulting deformations is opposite of the compressive axial force. Since nearly all the members of a structure are subjected to both bending moments and axial forces, the interactive effect of these forces may be significant in the analysis of beam-columns. When the magnitude of the

internal stresses and deflections produced by the axial force in a member is relatively small compared to the bending effect, it may be practical to analyze and design the member as a beam element. On the other hand, if the

12 bending effect is relatively small, analysis and design of the member as a column is appropriate. The presence of a compressive axial force reduces the force needed to cause a unit rotation or a unit

translation in the transverse direction at one end of a member. Conversely, the required forces increase when In other words, the presence

the axial force is tensile.

of axial force affects the flexural stiffness of a member or frame. The case in which the axial force since is it

compressive is

usually of greater interest,

leads to the possibility of member or frame instability. Using second-order analysis techniques, the instability can be predicted as the axial compressive force

approaches a critical value, which reduces the flexural stiffness to zero. Second-order analysis is therefore

central to the solution of stability problems.

3 ELASTIC SECOND-ORDER ANALYSIS

3.1 GENERAL

Elastic second-order analysis of members and frames can be accomplished by developing and solving the secondorder differential equations of equilibrium. closed form solutions for single members Classical and simple

frames are available

2,4,5,6] .

However, for analysis of

large frames of arbitrary geometry, closed-form solutions are difficult and cumbersome to obtain. However, with the advent of computers, automated second-order analysis for such cases is possible. The direct stiffness method is commonly used for computerized first-order structural analysis. In this

Chapter, this method will be used in extending the firstorder analysis to a second-order analysis described in the computer program [ E S C A P I using stability

functions.

13

14 3.2 METHODOLOGY

As previously noted, the presence of large axial force ( i.e, P / P , is large, where P , is the elastic buckling load ), affects the geometry and stiffness of a beamIf the axial force P is known, the

column element.

maximum deflection and bending moment can be calculated using second-order formulations. However, the member or

frame is initially indeterminate and the axial forces are unknown. The axial forces are related to the joint

displacements and must be first calculated because the axial forces influence the members' fixed-end actions, which are the elements of the load vector used in

determining the joint displacements.

Additionally, the

axial forces also influence the stiffness matrix of the member or frame. One way of determining the initial

axial forces is to perform a first-order linear-elastic analysis method. ( first cycle ) using the direct-stiffness

This will give the first-order shears, moments,

deflections and axial forces. In the second cycle, the previously computed axial forces are used to modify member stiffness and fixed-end actions. Then new values of joint deflections and

15

Elastic Analysis

Input

SAssemble & Modify

holesky's

Calculate
Deflections

s4semble Load

E Vector

deflections

__________

D(i)

D(i) old new ess; than or equal

-

Yes

Calculate Member
Forces

.005

_________

No

ICalculate Member
End forces

[2.

Print Output 1.Joint Deflections Member Forces

Fig. 2 - Flowchart Elastic Second-Order Analysis

16 member-end actions are calculated. This process is

repeated until the differences between the deflections in successive cycles are within the acceptable tolerance. This is the basic algorithm used in implementing the computer program [ E S C A P ]. The flowchart shown in

Fig. 2 outlines the basic algorithm of the computer program.

3.3

DERIVATION OF THE SECOND-ORDER STIFFNESS MATRIX

3.3 - 1

Translation Stability Functions

To begin our derivation of the stiffness matrix of a beam-column, let us consider Fig. 3 . The beam-column

element is subjected to a unit translation ( Fig. 3a ) and a unit rotation ( Fig. 3b ) under the presence of a compressive axial load P The elastic stiffness

coefficients for these two beam-column elements will be derived from second-order differential equations of

equilibrium. The degrees of freedom of a beam-column element are shown in Fig. 4a. A free-body diagram of a section of from the left

the beam-column element cut at distance x support is shown in Fig. 4b.

The sum of external

17

SM 22 SM
~52

k I~ SM 32 L (a) unit translation SM 23SM

m M 62

53

63

(b) unit rotation

Fig. 3

Stiffness Coefficients of a Beam-Column

18

SM2

SMM5

SM
(a) member degrees of freedom

SM22

X 32

.k

wx.

(b)

free-body diagram unit translation

SM 2O

(c) free-body diagram unit rotation Fig. 4 Unit Translation - Unit Rotation of a Beam-Column

19 moments at point 0 (Fig. 4b), assuming counter-clockwise moment is positive is given by

Maxt = -P(A - y) + SM
2

SM 2 2 x,

Eq. (2)

where

SM il =

is

the

member

stiffness i

coefficient induced by a

corresponding to the degree of freedom

unit displacement corresponding to the degree of freedom j while the other displacements are kept to zero.

Applying the differential equation of an elastic curve

M
=

-EIy",

Eq.(3)

then

the

interior moment must

balance

the

exterior

moments, resulting in :

M = -EI d
-P(A
Y) + SM32 -

2y

dx 2
Eq. (4)

SM

22

M=Ed2y

P(A - Y) + SM22x - SM32

Eq. (5)

20 Rearranging the terms of Eq. ( 5 ) yields

d 2Y + k 2y
dx
2

k 2A

(SM 22X - SM 2 )

Eq. (6)

El

where k

P El

Eq. (7)

The complementary solution of Eq. ( 6 ) is given by

YC

A sinkx + B coskx

Eq. (8)

and the particular solution of Eq. ( 6 ) has the form

yP = Cx + D.

Eq. (9)

Differentiating Eq. ( 9 ) twice with respect to

and

substituting the first and second derivatives into Eq. ( 6 ) yields Eq. ( 10 ). The constants C and D are

determined by comparing coefficients of the terms on the left hand side of Eq. ( 10 ) to the terms cl the right

21

hand side as shown below

0 + k 2 (Cx + D) =

kA SM2 x SM 2A + kE EI - El
where C = SM2 2 k 2EI

Eq. (10) SM3 2

The general solution of Eq. particular solution and

( 6 ) is the sum of its complementary solution.

Inserting the values of C and D into Eq. general solution has the form

( 9 ), the

y = A sinkx + B coskx + A + (SM22 x - SM 3 2) k 2EI

Eq. (11)

22 To solve for the constants A and B, apply the boundary conditions at joint k then
B Eq. (12)

of Fig.

4b

At

x = 0,

=A

k 2 EI

q3

At

x = 0, dy/dx = 0; then

A kSEI

Eq.

(13)

The stiffness coefficients

SM2 2

and

SM3 2

are determined
of Fig.

by applying the boundary conditions at joint m 3a. When x = L

,

y = 0 ; then

A sinkL + B coskL + A +(SM

2 2 L2

SM3 2 ) EI

Eq.

(14)

When

x =L , dy/dx =0 ;then

Ak coskL

-Bk

sinkL + SM 2 22

EI

Eq.

(15)

23

inserting the values of A and B into Eqs. yields Eqs. (16a) & (16b)

(14)

& (15)

SM 22

sin kL +

SM 32

cos kL + A+

EI
=-

k 2 EI Eq. (16 a)

(SM 2 2 L 2- SM 32)

k EI

S2 2 2 k EI

cos kL -

S32sin

kL

kEX

SM22 k EI

= 0

Eq.

(16 b)

Simplifying Eqs.

(16a) & (16b) yields

SM 2 2 (kL SM 32

sin kL)
-k 3 EIA&

k (cos kL
-

1)

Eq. (17) (18)

SM 22

(1

cos kL)

- SM 32

sin kL =0

Eq.

24

Solving

Eqs.

(17)

& (18)

for

SM

22

yields

SM 22

[kL sin kL
-

sin 2 kL + cos kL)] Eq.

(19)

(cos kL
=-k 3 EI sin

COS 2 kL +

kL

Simplifying Eq. (19) yields

SM 22

[2

2 cos kL

kL sin kL]

k 3 Elsin kL A

Eq. (20)

Multiply Eq. (20) by 12L 3

12L 3 and introduce the term

4= (2

2 cos kL

kL sin kL)

Eq. (21)

25

the stiffness coefficient SM

22

is

SM2 2 =2I)

( 312 6 kL3 sin # kL)

Eq. (22)

SM 2

E (6 s 1) A

Eq. (23)

where S I is

referred to as a

stability

function and is

equal to

s, = (kL)

12 ,

sin kL

Eq.

(24)

Substitute Eq. (22) into Eq. (17) and solve for SM

32

SM32_

2 LEI 3 (3L s 2 ) A

Eq. (25)

The other translational stability function S

is

s2 =

2 (1 (kL) 6 4ccos kL)

Eq. (26)

26
3.3 - 2 Rotation Stability Functions

Consider now the beam-column element in Fig. 3b which is subjected to a compressive axial force rotation at joint a distance x n
.

and a unit

A free-body diagram is taken at

from the left support as shown in Fig. 4c

Summing the external moments about point 0, gives

M=

Py-

SM23 x + SM 3

Eq.(27)

Applying the differential equation of an elastic curve {Eq. {Eq. (2)} and using equilibrium, the external moments (27)), must be equal to the internal moment.

Rearranging the terms yields

2 EI d 2 d~xy

'Y = SM 2 3 x

SM 33

Eq. (28)

27

The complementary solution of Eq. (8).

(28) is the same as Eq.

Using similar procedures described in Section 3.2-

1, the particular solution of Eq. (28) has the form

YP-

SM 2 3 k 2 EI

SM 33

k 2 EI

Eq. (29)

The general complementary

solution

of Eq. {Eq.

(28) (8))

is the and the

sum of the particular

solution

solution {Eq. (29))

, and has the form

2 x y = A sin kx + B cos kx + k 2 EI

SM 33 k 2EE

Eq.(30)

SM 23 2 Y/ = Ak cos kx - Bk sin kx +kE k EI

Eq. (31)

28 The constants A and B of Eq. (30) are determined by applying the boundary conditions at joint n of Fig. 3b:

at x =0 ; y =0

then B=

k 2 EI

SM33

Eq. (32)

at x 0

;y

then A

0
k

SM 23
k 3 EI

Eq. (33)

Inserting the values of A and B

to Eq. (31) yields

0
k

SM23 ) k cos kx-

k 3EI

SM 3 k sin kx + k 2EI

SM 22 3

k 2EI

M 3

Eq. (34)

29 Applying the boundary conditions ( joint r of Fig. 3b

)to Eq.

( 34 )at

x = L

y'

0 ; then

G9

S23

k 3 EI

k cos kL-

S3k

k2 EI

sin

kL +

(SM2 3 )

=Q

Eq. (35)

At

=L

,y

=0

;then

(0

-SM

k 3 EI

23

) si

L+S33coskL + k 2 EI

k 2 EI

SM 23 ) L

-M

Eq. (36)

NI

30 Solving Eqs. (35) and (36) simultaneously yields

SM 3 3 =-E[sin kL -kLcos kL]

*0

Eq. (37) where 4,= 2

-

2 cos kL

kL]I

Eq.

(21)

Multiplying Eq. terms

(37)

by

WL

/ 4L 3 and simplifying the

SM 33

2E1 * (2L 2 )

(S3 ) 0

Eq. (38)

where
_3

kL (sin kL

-kL

cos WL

Eq.

(39)

Substitute Eq.

(37) into Eq. (36) yields

SM2 3

-Elk

o kL ) 0 cs

Eq. (40)

31 Equations (25) and (40) are equivalent and satisfy

Betti's Law and Maxwell's Law

This indicates that

the second-order stiffness matrix

is symmetric.

The

remaining rotation stability function is calculated by summing moments at joint r of FI.g. 3b assuming counterclockwise moment is positive. By equilibrium we have

SM 2 3 L -

P(O)

SM 33

SM63

Eq. (41)

Substituting Eqs. (37) & (40) into Eq. (41) yields

SM 6

L coskL] -

EIke [sin kL - kL cos kL]

EIkO [kL - sin

kL]

Eq.(42)

32 Introducing the term to Eq. (43), 2L 3/2L 3 in Eq. (42) and simplify

which is the coefficient of stiffness at

joint r caused by a unit rotation at joint n , Fig. 3b

$IS 3

L3 L 2E12

(s 4 ) * 0
where

Eq.(43)

S4 kL =

(kL 2 - oosin kL)

Eq. (44) E. 44

For negative values of axial force

(tension)

the stability functions are derived in similar fashion applying the same concept of elementary beam analysis. To illustrate, let us investigate Eq. (4). Assuming that

axial tensile force is negative, then the homogeneous second-order differential equation has the form of Eqs. (45) & (46)

33 Ely" - Py = 0
y/ - k 2 y = 0 where k
2 =-

Eq. (45) Eq. (46)

Eq. (7)

The complimentary solution of Eq. (46) is

YC

A sinh kx + B cosh kx
=

Eq. (47) Eq.(8)

YC

A sin kx + B cos kx

Equations

(8) and (47) are similar in form except the

former contains the trigonometric functions while the latter has the hyperbolic of functions. Eq. (47), the Using the

complimentary

solution

stability

functions for an axial force in tension are calculated following the same procedures as when the axial load is in compression. The four stability functions are listed

in TABLE I for axial forces in compression, tension, and

zero [12]

TABLE -

STABILITY FUNCTIONS

34

OF A BEAM-COLUMN

Stability

Functions

for a Beam Subjected to

Axial Forces Direction of Axial Force Function s

1

Compression (kL)3 sin kL 12

Tension (kL)3 sinh kL 124t (kL)2 (cosh kl - 1)

(kL)

(1 -

cos kL)

S3

kL (sin kL kl cos kL

kL (sinh kL kl cosh kL)

S4

kL (kL
2 0

sin kL)

kL

(sinh kL - kL)

2 0t

0C

= 2 - 2 cos kL - kL sin kL

ot

= 2 - 2 cosh kL + kL sinh kL

Note: When axial force is zero

s
= S2 = S

35 The Tremb-r Ym~fnes atrix with (48), the stability the

functions is represented by Eq.

where [ K, 3 ] is

modified stiffness matrix and includes the effects of axial forces in tension or compression.

0
12 1

0 S1
I

-A
0
-

0
12 1

0 S1 -IS 2Is 4 0
6is

o
[K E L L-A 0
0

-S -

41s3 0
6I

0
A

_6s
0
121S

0
12Is 6I S

21s4
Eq.

0
(48)

7-Is2

41s3

36 When *-Ie axi ] force is zero, the stability functions are (48) is reduced to Eq. (49)

reduced to unity and Eq.

0 121 L2

0 61 L

-A

0 121 L2

0 61 L

0
[K]-~L L -A

61
0 121

41
0
_61

0
A

6
L

21
0
_

0
0

.2 L

. L
21

0 0

0 121

-LT
L

-L
41

61

61
L

Eq. (49)

Equation matrix.

(49)

is

the

first-order

elastic

stiffness

3.4

LOAD FUNCTIONS AND MAXIMUM MOMENTS

3.4

Load Functions

The

fixed-end actions

for

a beam-column produced by

applied lateral loads depends upon the magnitude of the axial force acting on the member as well as upon the

37 distribution and intensity nf the lnteral lo~ds. These

fixed-end actions or joint loads are used in determining the joint deflections and member forces in the structure. The fixed-end actions form the elements of the load

vector in the direct stiffness method. The two types of member loading ( Fig. 5 )

implemented in the computer program are: (1) uniform load distributed over the full length of the member, and (2) concentrated loads applied anywhere along the span. The

procedures used in deriving these fixed-end actions are similar to those described in Section 3.3 and will not be repeated here. A complete derivation of the fixed-end

actions for the two types of loading is presented in detail in reference [ 14 ]. Timoshenko and Gere 1 4
1

present solutions to specific cases of beam-columns with various constraints at the ends. This includes a member

simply supported on one end and built in at the other end, as shown in Example 5.2-3, Section 5.2 Ghali and

Neville (9] present a different procedure in determining the fixed-end actions of a beam-column using matrix

methods. The fixed-end actions of a beam-column for the two types of member loading mentioned above are shown in
figures Fig. 6 through 8 110, 12, 14]

38

FEA

2t (a) Concentrated Load

FEA

FEA 5

3
(b) Uniform Load

Fig. 5

Types of Loading

39 Axial Compression

FEA=

(coskL -coska +cosb + kb sinkL -1)

FFA3=
-

(sin ict - sin ka - sin kb

+

kb cos kL

kL cos kb

ka)

FEas

-k (cos kL
41C

cos kb
-

cos ka

ka sin kL

1)

FF.% ' (-sin kL +sin kb +sin ka + ka cos kL -kL cos ka +kb)

40 2
-

2 cos kL

kL sin kL

2F

Concentrated Load
rig. 6
-

Fixed-End Actions for a Concentrated Load

40 Axial Tension

FEKA 2 = -P

(cosh kL -cosh ka +coshkb

- kb sinh kL -1)

FA3=
-

(sinh kL -sinh ka-sinh kb kb cosh kL

+

kL cosh kd,

ka)

FEA 5

~(coshkL -cosh kb +cosh ka

-

ka sinh kL

1)

FF-&
+

(-sinh kL
-

sinh kb
+

sinh ka

ka cash kL
=2

kL cosh ka
+

kd,)

2 cosh kL

kL sinh kL

k-

Fig.

Fixed-End Actions for a Concentrated Load

41

Axial Compression

FE

= FEA

wL 2

wL 2 12

12 (kL) 2

6 sinkL kL (coskL - 1)

FEA 6 = - FEA 3

Axial Tension

FE% - YEAs

wL

w L 2 [12 12 (kL)2

kL (cosh kL - 1)

6 sinh kL

FEAG

-FEA 3

FEA 2
p

FEA 5
'PA

3
Uniform Loud

Fig. 8 - Fixed-End Actions for a Uniform Load

42 3.4 - 2 Maximum Elastic Second-Order Moments

It is important for a designer to know the location and magnitude of the maximum moment within a member. Some

structural analysis programs only calculate the member end actions ( i.e. shears,moments and axial forces ) and the joint deflections. In some cases, the maximum

moments occur at the joints, when the axial forces are small and no member lateral loads are present. However,

when the axial forces acting on the members become large, the maximum moment may occur within the span as shown in Fig. 9 , even if no lateral loads are present. In this section, the derivation of the maximum

moment is limited to the case of a uniform distributed load. The formulation for this case is straight forward,

since an equation for bending moment can be expressed as a continuous function along the length of the member. The bending moment at a concentrated load point can be easily obtained by introducing a node at the location of the load. The maximum moment in a member with a uniform

distributed load is calculated based on the procedure suggested by White and Hajjar 1 13. An isolated

43

p M
0 elastic
first-order

M1-0 moment

elastic/ scond-order moment

small axial force

large axial force

Fig. 9

Maximum Elastic

Second-Order Moment

44 member of a frame structure is treated as a simply

supported beam-column that has translation.

joint rotation and no

To illustrate this procedure, Fig. 10 shows a rigid frame with its deformed shape when subjected to the two types of loading described in Section 3.4-1. Member 1

is isolated and the free-body diagram is shown in Fig. 10b. Let us assume the member end actions are known from the results of the elastic second-order analysis. we connect joints 1 and 2 ( Fig. 11a ), the When

member's chord location is established and calculated using the new joint coordinates If the member end

forces and any other loads applied directly to the member are transformed to local axes oriented along the member's chord, White and Hajjar suggest that there is no

difference between the behavior of this isolated member ( Fig. 10 and 11 ) and the behavior of a simply supported beam-column with the Example 5.2-1 ). To derive the location of the maximum elastic second-order moment, we shall use { Eq. the differential same applied loads ( Fig. 15,

equation of an

elastic curve

(3) } about the

transformed local axes of Fig. 11b

pt

45

2
W f

w13

(a) Frame Geometry Loading, and Deflected Shape

pg
2

VI

(b) Member 1 Loading and Deflected Shape

chord ' w

VI
1+ .7 M1 =0

Fig. 10

Derivation of Maximum

Elastic Second-Order Moment

46

2' b

(a) E-lefleOt end totces chord's 8aOS oriented at the

0

ma a

(b) tresbody

diagr~fn

igI

Transformed

Axis

of a 3eafn.Column

47 The second-order differential equation of a beam-column subjected to a uniformly distributed load with member end forces known is given by

2~YtM
d;C2

-Vx+

(wX

Eq. (50)

The general solution of Eq.

(50) is given by

Y=A sinkx +B coskx +

kEl

+ V x-

-M. a

Eq. (51)

48 Differentiating differential yields Eq. (51) twice and applying the (3)

equation of an elastic curve

{ Eq.

Y"

2 -A k 2 sin kx- Bk2 coskx+ k E

Eq. (52)
Mx

-Ely"l

=Elk 2 [A sin

kx+ B cos kx Eq. (53)

w~

When x then B
=

0; Mx

Ma

1[M. + P
When x

wE.(4 k2
=

E.(4 Mb Eq. (55)

L; Mx
+

then A-M

, ok PsinkL

wtan kL 2 pk2

where P = k 2 El

49 Differentiating Eq. (53) once with respect to x gives

the expression of shear force anywhere along the span and has the form

V, = MI

[A coskx - B sinkx] EIk2

Eq. (56)

When the shear force is equal to zero, the moment is at local maximum. Inserting the values of A and B into Eq.

(56), the location of the maximum moment anywhere along the span of a beam-column with a uniform distributed is determined using Eq. (57):

tankL + Mb - Ma coskL

tan kx

k2

M.+

w sinkL

Eq. (57)

where

is measured from the left support.

All the

applied loads must be transformed in the local axes when using Eq. (57). When the end moments in Fig. 11, are equal to zero, the constants A and B are given by Eq. (58). The moment

anywhere along the member is then given by Eq. (59)

50 B= ; A-wtank'2 pk Eq. (58)

pk 2

. =w (tan kL sin kx + cos kx - 1]

k2 kx

-2

Eq. (59)

The maximum moment for a simply-supported beam-column with uniform distributed load is located at the midspan. Thus

tan kx = tan-S; where x =2 2

Eq. (60)

Inserting the value of

x into Eq. (59) yields

[tan kLsin kL + cos kL

kc
2

2 Eq. (61)

51 Introducing the term u

,

Eq. (61) becomes

M. :

w2 k

[tanu sinu + cosu - I]

Eq. (62)

where u 2

Applying trigonometric identities and simplifying the terms in Eq. (62) yields

wL

[ 2 (sec u-I)

Eq. (63)

U2

Equation (63) is the general expression of the maximum moment at midspan of a simply-supported beam-column with uniform distributed load. / 8 ) The first-order moment ( wL 2

is amplified by the terms inside the brackets. called the moment amplification factor

This term is (MAP).

Let us now investigate what happens when the lateral load w = 0 and M
,

M b are not equal to zero.

52 Applying the boundary conditions at x = 0 and x = L, from Eq. (51), the constants A and B are determined.
= 0 , y = 0 , and w = 0 ; then

When

-R , P

Eq. (64)

When

x = L,

y = 0 , and w = 0 ; then

A
P

L2

Ma PsinkL
Mb

Eq. (65)

where V8 L =

L (Ma +Mb) L

Eq. (66)

Inserting the value of the trigonometric identity { Eq. (67)

} into Eq.

(65),

tan kL 2

(1 sinkL - coskL)

Eq. (67)

53 Then the constant A is given by Eq. (68).

A =_ (M coskL + Mb)
PsinkL Eq. (68)

Inserting the values of A and B into Eq. (53) when

equal 0 , yields

w is

tan icx

A _ B

(Mb + Macos

kL)

Eq.

(69)

MasinkL

M,

(Mb + MCoskL) I sifkx + Macoskx

sinkL

Eq. (70)

Thus, Eq.

(70) is the moment at any location along a

Mb

beam-column when subjected to end moments M,, axial force P. Inserting the value of kx

and an

from Eq. (69)

into Eq. (70), the maximum moment is given by Eq. (71). The minus sign that appears in Eq. (71) simply indicates that the moment M . causes tension on the top fiber of the cross section [
I

54

Mm~rax = -Mb[

_M M

(M

+2 (M /Mb) coskL + i] i n 2 + S(Ma/Mb) kL

Eq. (71)

When end moments are present in the member, there is always a possibility that the maximum moment will occur at the end of the span. Therefore, the value of x in Eq. When x is greater than

(69) must always be evaluated.

the span length or less than zero, the maximum moment occurs at the end of the member and it is equal to the larger of the two end moments. Equations (51), (53), (54), (55) and (57) are used

in the computer program to calculate the location of the maximum moment of a member with uniform load and

subjected to end-moments. developed for the For case

These equations have been axial force is in

where the force in

compression.

axial

tension,

similar

equations are used, except hyperbolic functions replace the corresponding trigonometric functions. moment is then compared with member, and the larger of the The maximum

the end moments of the three controls. The

program also calculates the deflection at the point of maximum moment. If the maximum moment occurs at one

55 end of the member, then the computer program prints out the deflection at that joint. member, the computer For deflections within a only calculates the

program

displacements relative to the chord member. (See Example 5.2-3)

ELASTIC BUCKLING LOAD

4.1

GENERAL

This chapter deals with the elastic stability analysis of beam-columns and frames. The same stability functions

and stiffness matrix developed in Chapter 3 are used in determining the elastic critical loads. A load factor

increment is applied successively to the external Loads until the stiffness matrix of the structure has become singular. The decomposition of the stiffness matrix Several

using Cholesky's method will also be presented.

examples are examined and presented in Chapter 5.

4.2

COMPUTER ANALYSIS OF THE ELASTIC STABILITY OF PLANE FRAMES

In Section 3.3, the column were derived

stability functions of the beamand used to modify the member

stiffness matrix (Eq. (48)) axial forces.

to include the effects of

The joint displacements and member end

forces are calculated iteratively.

56

57 The computer program was modified to calculate the elastic critical load using load increments. critical load is determined by performing The elastic successive

computations until the stiffness matrix of the structure becomes singular. In the computer program, the structure's stiffness matrix is decomposed using Cholesky's method The stiffness matrix [ K, ]
8, 14

{Eq. (48)) is factored into

an upper and lower triangular matrices. In general, structural systems generate stiffness matrices that are symmetrical, square, and with positive diagonal elements. They display the unique property of If this is the case, then

positive-definiteness.

Cholesky's method is valid and the stiffness matrix can be expressed as

{ K, ) = {Ut)
the upper

{ U ). The matrix { U
elements } of of the the lower

represents

diagonal

structure's decomposition

stiffness process.

matrix The

{K m elements

after the

triangular matrix are represented by the matrix { Ut ), which is the transpose of the matrix { U }. Eq. (72)

shows the elements of the structure's stiffness matrix { K, ) and the elements of the upper and lower triangular matrices after the decomposition process.

58

kil
k2l k3 1

k12 k13 k14

k2 2
k32

. . kln
. . . .

k2 3
k
33

k24
k
34

n
k..
-3

k., k2

k,

U1 1
U1 2
U13

0
U22 U23

0
0 U33

...
...
. . .

0
0 0

U11
0 0

U1 2
U22 0

U 13
U23 U33

U1 4
U 24 U 34

...
... ...

Uln
U 2n U3n

Uln

U2 n

U3n

. .

Un

..

Ur

Eq. (72)

Performing the indicated multiplication of { U ) and { U t } and equating the product to the elements in { K, yields Eqs. (73) & (74) shown in Fig. 12. If during }

decomposition one of the diagonal elements u i.approaches zero or is less than zero, the structure has reached the elastic buckling load.

59

k1

2k 11

12
U11 U1
2

k 1 2 = U1 1 U 1 2 k 13 =
U1I U 3

OKI kU11

"

U 3

k13
U1

k2
kc2 3

2 2 =U k2 = U12u2
=

22

U22

k22

-u12 2

U12 U1 3

+ U22 U23

123 U

(k23 -

U12 U1 3 )
U22

k2
=

23 +/

2U+
+

U U32

-U33 = k33 (U13+

U2 ) 23

3 + U2 3

k 34

U1 3 U14

+ U2 3 U24 + U k
-

U33 U34 (u
13

34

14

+ U23 U24)

U33

a~Ia~

E-1(3

where u.,

Eq. (73)

are diagonal elements of [U]

kii = diagonal elements of [K]

a=i -I kii Ua i Uaj

8=1 Uii

Eq.

(74)

i z j (lower diagonal elements) i < j (upper diagonal elements)

Fig. -

12

DECOMPOSITION OF MATRIX [ K,

60 To better understand the concept, a flowchart is shown in Fig. 13. In the flowchart, the process starts

with the calculation of the first-order axial forces. Then these axial forces are used ]. to modify the

structure' s stiffness matrix [ K ,

The subroutine

DECOMP decomposes the structure's stiffness matrix using the procedure described in Fig. 12. The program then

checks the diagonal elements u ii for value less than or equal to zero. If all the diagonal elements are greater is calculated and

than zero, a new load factor ( LF ) the process continues.

If any one of the diagonal elements is less than zero, the program stores the load factor LF, calculates the joint displacements, prints the output and then

terminates the process.

The output contains the load To obtain an

factor LF and the joint displacements.

accurate buckling load, small load increments must be used. The buckling mode shape is determined by converting the calculated joint displacements into relative joint displacements. The joint displacements at buckling are

expressed in terms of relative displacements. This approach of determining the elastic buckling load is simple to implement in the computer program,

ILozd Factor

LF =1.0

O l rd Foe cesF Axi

[Modify
Do i Km For Second-Order Effects

lo wc h art

Global Stiffness

Choleskys Dcoiion7
Check Diagoni Element

U 1 1
Less or Equal 0t

Yes

oa
Vector

NO
0WMT

PDeflection

icu late Joint

Loa

vecor

iculate Memberr Deflections Calculate New Load Factor LF= + STEP

tOiu 1. Load Factor 2. EJolnt DI placements

Fig. 13

Elastic Stability

Analysis of a Frame and Beam-Column

62 since the decomposition of the matrix is routinely

performed during the first and second-order analyses. A flowchart showing the subroutine DECOMP t shown in Fig. 14. matrix is also is

The decomposed structure stiffness used in solving for the unknown

displacements.

In the flowchart, there are two returns.

One occurs when the value of I becomes greater than N and the processing is complete. decomposition fails. The second is when the

The latter serves as a safeguard

against an undefined arithmetic operation and it also indicates that the critical load of the structure has been reached. The undefined arithmetic operation will

result to an error in determining the critical load if at the beginning of the linear-elastic (first-order)

analysis, the structure stiffness matrix is not positivedefinite. In order to safeguard against this error, one If the error

should perform a first-order analysis.

message "WARNING !!! DECOMPOSITION FAILED" shows, then the stiffness matrix is "singular." The input file

should be checked for incorrect data of joint coordinates and member connectivity, which are the common sources of error.

63

Return

Yes

No

Fig 14

Flowchart

<Yes

Subroutine Decomposition

J>N No

Ye sY es

A ~ ,J ) =

u rn Te m p ,J

VERIFICATION OF COMPUTER PROGRAM

5.1 GENERAL

In order to verify the program

[ E S C A P ], the

following Sections will compare closed-form or published solutions to the computer solutions. The second-order

moments and deflections are calculated first using the closed-form solution of the differential equation. a computer solution is generated and results Then are

compared.

5.2

SAMPLE PROBLEMS

OF BEAM-COLUMNS

EXAMPLE 5.2 - 1

Simply Supported Beam-Column with Uniform Distributed Load.

The simply supported beam-column in Fig. 15 has a uniform load w of 1 kip/ft acting downward. = 500 in. and I = 1000 in . The span L

Calculate the maximum

deflection and moment at midspan based on second-order alatzc analysis. Determine the elastic buckling load

64

65 and the buckled shape of the member. Assume shear strain

is negligible and neglect the weight of the beam. The maximum deflection and moment occur at midspan. The exact closed-form solutions are given by Timoshenko and Gere K
1,

and Chen and Lui

5]

The general solutions

are given by Eqs.

(75) a (76).

Ymax

Yo L 12 (2secu - 2
5u

Eq

(75)

where ukL

;k
2

P E=

-5

wL

384EI

MO

2 (secu
U2

1)

Eq. (76)

where M,
0

wL 2 8

66

Example 5.2-1
-W

(a)

Loading

Diagram

ly max
(b) Deflection Diagram

M M

max

(C)

Moment Diagram

Fig. 15

Simply Supported

Beam-Column with uniform Load

67

The computer results are compared with these solutions in TABLES II and III. These tables show the computer The

solution agrees very closely to the exact solution.

amplification factors (AF ) reported in these tables are the ratio of the second-order to the first-order moments or deflections. Note that second-order effects are quite significant at high values of axial compressive force. The theoretical elastic critical load for the beamcolumn in Fig. 15 is given by ;

2 n EI-L2 n2 * 30000 Pe -5002

1000

1184.35 kips.

Eq. (77)

The symbol P , is used to represent the elastic critical load. The computer solution reports a load factor LF = 2.37, with an applied axial force of 500 kips. The

computed elastic critical load is therefore = 1185 kips, versus the exact 1184.35 kips.

500 * 2.37 The computer

program used a load factor increment of .005 The computed relative joint displacements for the buckled member are listed in TABLE IV. The buckled shape

68 is plotted in Fig. 15b. The buckled shape from the

computer solution agrees with that predicted by the exact solution. The values of P

P , versus

the

moment

and

deflection amplification factors (MAF,DAF) are plotted in Fig. 16. are made : - the moment and deflection amplification factors are non-linear with respect to P, - the amplification factors are very small for small values of P / P 0 ,on the order of .05 to .10,
-

From this figure, the following observations

AF increases without bound as P approaches P,.

These observations are typical for second-order elastic response.

69 Table IIMAXIMUM BENDING MOMENTS FOR A SIMPLY SUPPORTED BEAM-COLUN

Maximum Bending Moments (at midspan Axial Load kips ) 100 200 300
400 500

Exact solution

Computer solution

MAF

(kip-in
2851.0 3148.3 3513.1
3971.0 4563.0

(kip-in
2851.1 3148.5 3513.2
3971.2 4563.2

1.09 1.21 1.35

1.53 1.75

Table III - MAXIMUM DEFLECTIONS FOR A SIMPLY SUPPORTED BEAM-COLUMN Maximum Deflections ( at midspan ) Axial Load Exact solution Computer solution DAF

70

(kips )
100
200 300 400 500
-

( in. )
2.46965 2.72135 3.02999 3.41735 3.91793
-

( in. )
2.4698 2.7215 3.0301 3.4175 3.9181 1.09 1.20 1.34 1.51 1.73

Table - IV

RELATIVE DISPLACEMENTS AT INSTABILITY OF SIMPLY SUPPORTED BEAM-COLUMN RELATIVE DISPLACEMENTS

JOINT

D(X)

D(Y)

ROTATION -1 0

1 2 3

0 0 0 -1

71

Fig.

16

Load vs. Deflection/Moment Curve 1.0-

0.8U-

0.6M AIF -s--PIPe 0.41DAF P/Pe

0.2DAF, 0 10 Simply-Supported MAF 20 Beam-Column

72
EXAMPLE 5.2 - 2 FIXED-FIXED BEFA-COLUM WITH UNIFORM

DISTRIBUTED LOAD The beam-column shown in Fig. 17 is similar to 5.2-1 except the the end supports moment, are fixed. Example We will

determine

maximum

deflection,

elastic

critical load and the buckled shape by following the same steps described in Example 5.2-1. The exact solution to the differential equation of this particular beam-column is found in Timoshenko and
Gere [ , and Chen and Lui [ 5 The general solutions

are given by Eqs. (78) and (79):

Y.'. =

You12
y.

(2 - 2cosu - u sinu 2sinu

L occurs at x = 2 ;

Eq.(78)
4

where m

3 84 El Eq. (79)

wL

_ wL 12

[ 3 (tanu u 2 tan -u u)

Using w = 1 kip/ft, L = 500 in, and I = 1000 in 4 as

in Example 5.2-1, the deflection at midspan and moment at the end supports are calculated and compared to TABLES V and VI. Again, the the

computer

solutions in

results are in agreement.

73

Example 5.2-2 -w

(a) Loading Diagram y

(b) Deflection Diagram

MM X

(c) Moment Diagram

Fig. 17

Fixed-Fixed Beam-Column with Uniform Load

74 The theoretical elastic critical load for a fixedfixed beam-column is given by :

47r2 EI

Eq.

(80)

7*

30000
5002

1000 = 4737 .41 kips

The computed load factor LF = 9.4752 with an applied axial load of 500 kips.
load is

The computed elastic critical

The load factor

9.4752 * 500 = 4737.6 kips.

increment was

.005.

The elastic critical load for a

fixed-fixed beam-column is four (4x) times larger than the simply-supported beam. The computed relative joint displacements for the buckled member are listed in TABLE VII. shape is plotted in Fig. 17b. The buckled

The buckled shape from the

computer solution agrees with that predicted by the exact solution. The relationships between P / P
*

and

the

amplification factors are similar to those in Example 5.2-1.

75

Table - V

MAXIMUM BENDING MOMENTS FIXED-FIXED BEAM-COLUMN

Maximum Bending Moments ( at end supports ) Axial Load kips )

100

Exact Solution ( kip-in )

1760.6

Computer Solution ( kip-in

1760.7

MAF

1.01

200 300 400

500

1786.3 1813.0 1841.0

1870.0

1786.3 1813.0 1841.0

1870.0

1.03 1.04 1.06

1.08

Table - VI

MAXIMUM DEFLECTIONS

76

FIXED-FIXED BEAM-COLUMN Maximum Deflections ( at midspan ) Axial Load Exact Solution Computer Solution DAF

kips )
100 200 300 400 500
-

( in. )
.46172 .47176 .48226 .49324 .50474
-

( in.
.46174 .47178 .48228 .49326 .50476 1.02 1.04 1.07 1.09 1.12

Table - VII

RELATIVE DISPLACEMENTS AT INSTABILITY

OF FIXED-FIXED BEAM-COLUMN RELATIVE DISPLACEMENTS JOINT 1 2 3 D(X) 0 0 0 -1 0 D(Y) 0 ROTATION 0 0 0

77 EXAMPLE 5.2 - 3 HINGED-FIXED BEAM-COLUMN WITH UNIFORM DISTRIBUTED LOAD

This problem is similar to the previous two examples, except this beam-column is restrained against rotation at joint B and free to rotate at joints A and C ( Fig. 18 ) The maximum moment, deflection at the location of the maximum moment, elastic critical load, and the buckled shape of the member will be determined by following the same procedures described in the previous examples. The exact solution is determined by superposition. The general solution is a combination of a simply-

supported beam-column with a uniform load ( Example 5.2-1 and Fig. 19a ) and one with an end moment acting at joint B without transverse load ( Fig. 19b ). For the latter

case the general solutions are given by Timoshenko and Gere 14 :

y = P

sinx s sinkL L

Eq.

(81)

M= -M

sinkx sinkL

Eq. (82)

78

Example 5.2-3 P

A (a) Load Diagram y

x
~max

(b) Deflection Diagram M M 1

(c) Moment Diagram

max

Fig. 18

Hinged-Fixed Beam-Column with Uniform Load

79

Example 5.2-3

(a) uniform load

PNb

(b) end moment

Fig. 19 Superposition of Lateral Load and Moment

80
The rotation at joint B for the case in Fig. 19b is

bL3 .3EI2u

2u

tanu

Eq.

(83)

For the casein Fig. 19a, the rotation at joint B is

1b -24EI

wL

3 (tan u - u) 3
u

Eq.

(84)

By imposing the rotational boundary condition at B, such

that

lb

= 0, the moment Mb is obtained.

The moments and the deflections at the location of the maximum moments for members AC and CB are calculated and compared to the computer solution in TABLES VIII thru X.

81

The solutions are in agreement. The theoretical elastic critical load is [4

20.191EI (L) 2

20.191*30000*1000 (500 ) 2 Eq. (85)

P, = 2423 kips

The computed load factor LF = 4.8501 for an applied load of 500 kips. The computed elastic critical load is

4.8501 x 500 = 2425.05 kips. was .005.

The load factor increment

The resulting load increment was .005 x 500 The computed elastic critical load is Closer agreement is possible by

kips = 2.5 kips.

within this tolerance.

using a smaller load factor increment. The computed relative joint displacements for the buckled member are listed in TABLE XI. is plotted in Fig. 18b. The buckled shape from the

The buckled shape

computer solution agrees with that predicted by the exact solution.

82

Table - VIII

MAXIMUM BENDING MOMENTS

HINGED-FIXED BEAM-COLUMN Maximum Bending Moments

( at Joint B of member CB )
Axial Load kips ) 100
200 300 400 500
-

Exact Solution ( kip-in )

2679.4 2761.0 2849.9 2947.0 3053.7
-

Computer Solution ( kip-in

2679.4 2761.0 2849.9 2947.0 3053.7

MAF

1.03 1.06 1.09 1.13 1.17

The deflections at the support are 0.

83

Table - IX

MAXIMUM MOMENTS FOR MEMBER HINGED - FIXED BEAM-COLUMN Maximum Bending Moments

AC

( for member AC at distance x from left support

Axial Distance Load X kips ) ( in. ) 100 200
300 400 500

Exact Solution ( kip-in ) 1533.2 1608.0

1690.0 1780.2 1880.0

Computer Solution ( kip-in 1533.3 1608.1

1690.0 780.3 1880.1

187.11 186.72
186.32 185.91 185.49

84

Table

-X

DEFLECTIONS AT MEMBER AC HINGED-FIXED BEAM-COLUMN Deflections

(for member AC at distance x from left support)

Axial Distance Load X (kips) (in.) Exact Solution (in ) Computer Solution ( in) DAF

100 200 300 400 500

187.11 186.72 186.32 185.91 185.49

-.

96546

.96636 1.0092 1.0561 1.1076 1.1644

1.04 1.09 1.14 1.19 1.25

-1.0083

-1.0552

-1.1073

-1.1646

85

Table - XI

RELATIVE DISPLACEMENTS AT INSTABILITY HINGED-FIXED BEAM-COLUMN RELATIVE DISPLACEMENTS

JOINT 1 2 3

D(X) 0 0 0

D(Y) 0 -1 0

ROTATION 1

.33428 0

86
EXAMPLE 5.2-4 SECOND-ORDER ANALYSIS OF A FRAME

In previous examples, the given structures consisted of a single-element. In this example, a frame composed of

several elements will be analyzed. Since the closed-form solutions for frames are

difficult to derive, the published solution for the given frame is used to verify the accuracy of the computer solution. The effect of axial forces on the behavior of

a rigid frame are illustrated by means moment diagrams.

The rigid jointed plane frame is loaded as shown in Fig. 20 ( Beaufait et al. )[141 members. The deflections El is constant for all at the joints A

and moments

considering the effects of axial for-es are required.

summary of the published and computer solutions is shown in TABLES XII & XIII. are in The tables show that the solutions The first and second-order

good agreement.

moments are plotted in Fig. 21. following observations are made : -

From this figure, the

the largest amplification factors occur at the

support joints 1 and 4 , and are on the order of 23% to

25%
-

for each member of the frame, the maximum moment at the end joint other than the support

is

located

87

joints.

The

amplification

factors

at

the

points

of

maximum moments are less than 10% In both examples, Example 5.2-3 and this example, the maximum moment do occur at the member ends.

88

Example 5.2-4
200 klps w =-2 k/ft 200 klps

ilY i .1ii

.LI

12 kips

El
EI

kps

24 ft

in I = 250

E = 30,000 ksi

Fig. 20 Geometry

Rigid Jointed Frame and Loading Conditio

91

89

Example 5.2-4
857

1186

Legnds
61
_______

115stode3ki-n

rer

(i-n

Fi.26Mmn9Darm Firs an6ScndOre

90

Table - XII

SECOND-ORDEP MOMENTS

OF

A PLANE FRAME Second-Order Moments Plane Frame

Joint

Published Solution kip-in ) 480 110 -610 1153 425 653

Computer Solution ( kip-in 482 110 -609 1152 425 652

1 2 3 4 5 6

91

Table

MIT

SECOND-ORDER DEFLECTIONS OF A PLANE FRAME Second-Order Deflections Plane Frame

Joint

Published Solution Rotation D (X) 0 .8483 .8473 0

-.

Computer Solution Rotation D(X) 0 .8526 .8495 0

-.

1 3 4 6

0 007 .0035 0

0 007

.0035 0

92

EXAMPLE 5.2 - 5 ELASTIC BUCKLING LOAD CF A FRAME In this example, the elastic critical load of a frame will be computed using the method described Chapter 4. in

The results are compared to the published

9
1.

solution taken from Ghali and Neville

The frame

shown in Fig. 22 has a load Q and Q/2 applied at joints 2 and 3, respectively. equal to EI Member 2 has a flexural rigidity 1 and 3 have a flexural

, while members
.

rigidity equal to .75EI

Assuming that the loads are

applied simultaneously, calculate the elastic critical load Q. which causes buckling of the frame. Assume also

that the members are made of structiral steel with the following cross-sectional and material properties:

I x = 533 in.

4;

A = 19.10 in. 2 ; E = 29,000 ksi.

The elastic critical load is determined by implementing the procedures described in Section 4.3. Using an

applied load Q = 100 kips, the computer solution gives a load factor LF = 3.94. load is therefore solution between is the The computed elastic critical The published The error (.36 % ).

3.94 x 100 = 394 kips.

2

.254 EI / b two

392.6 kips. very small

solutions

is

The computed elastic critical load approaches the

93 exact solution when the load factor increment is reduced. The buckled shape from the computer solution agrees with that predicted by the exact solution. For solutions solutions. the are five in examples presented, the computer exact

excellent agreement with the

94

Example 5.2-5 0 0/2

2El

3/4 El 3/4E1

1
6b 50Oft.ow

Ocr

=.254

El/b
.2

1= 533 in

A = 19.10 in

=l100 kips

Fig. 22 Geometry

Portal Frame and Loading Condition

95 EXAMPLE 5.2-6 ELASTIC SECOND-ORDER ANALYSIS OF A THREESTORY FRAME

In this example, the capability of the computer program [ E S C A P ] The analyzing elastic larger structures is

demonstrated.

second-order

moments,

deflections, and the amplification factors of a threestory structural steel frame will be determined and

compared to the published solution.

The three-story frame in Fig. 23
15],

is subjected

to a combined gravity and lateral loading as shown.

The

computed elastic second-order moments and deflections are shown in TABLES IV and XV. solutions are in agreement. The maximum lateral displacement located at the top of the structure ( joint 7 ), of 1.03 This design The is amplified about a factor a very taller joint 7 TABLE IV shows that the two

lateral displacement becomes consideration lateral in cases at of

important buildings.

displacement

represents a story drift of .003H, where A is the story height of the building. The acceptable lateral

displacements for tall buildings are in the range of 0.002 to 0.004 of the height of the building
16]

96

Example 5.2-6

k 7
2.980.,_
_ __ _ _ _

.148 k /in
__ _ _ _ __ _ _ _ _

8
LO

W 14x 30 k
____

X .148 k/ in
____ ____ ___ ___ ____ ___ ____ ___

~13 ft
6

W21 x44

c')

13 ft

5.6k

.148 k /in3: W 21 x 44

00

13 ft

25 ft

Fig. 23 Three-Story Frame Geometry and Loading

Table-XIV MAXIMUM SECOND-ORDER MOMENTS

-EXAMPLE

5.2-6

MEMBER

Published Solution

Computer Solution

MAE

( kip-in )(kip-in)
1 2
----

442
----

1 .04 0.72 1.00 1.02 1.01 1.01 1.02 1.01 1.01

26

3
4 5 6 7 8 9
-------

----

333 1518

----

1199 579 846 695 579

845 695 579

98

Table-XV FIRST and SECOND-ORDER LATERAL DISPLACEMENTS EXAMPLE 5.2-6

JOINT

First-Order Solution (in) 0

0 0.5419 0.5406

2nd-Order Solution (in) 0

0 0.5603 0.5590

flAP

1
2 3 4

0
0 1.03 1.03

5 6 7 8

1.1093 1.1080 1.3755 1.3674

1.1470 1.1460 1.4190 1.4100

1.03 1.03 1.03 1.03

CONCLUSIONS

In the analysis of structures, neglecting the secondorder effects may overestimate the strength and stiffness of a member or frame. The elastic forces generated

within a member or frame can be more accurately predicted with the use of an elastic second-order analysis. The second-order effects ( amplified moments and

deflections ) are of increasing importance as lighter, and more flexible structures are constructed. The use of higher strength materials and less rigid non-structural elements is producing more flexible structures where

second-order effects are of greater importance. The behavior of multistory frames subjected to

either gravity loads only or under combined gravity and lateral loading cannot be accurately predicted by a

first-order elastic analysis, when the compressive axial load level is significant
17

The

lateral

displacements may be rather large and may not be within the acceptable range of lateral displacements for tall buildings. An elastic second-order analysis such as

[ E S C A P ] can be used to provide an accurate

99

100 estimate of the lateral displacements and the elastic internal forces generated in the structure. If the

lateral displacements become too large, they may become noticeable and disturbing to the building occupants
[16

When

second-order

effects

are

included

in

the

analysis, a more rational design of individual members may be possible. The beam-column interaction equation in LRFD is based on the use of elastic second-order moments. LRFD permits direct application of computer generated elastic second-order solutions, such as those generated by the computer program [ E S C A P ]. Structures with very slender members may fail by elastic instability. analysis is For such structures, elastic when

stability

directly

applicable

determining the elastic buckling load. Typical members or frames in structural steel

buildings will fail by inelastic instability, rather than elastic instability. analysis is In not such cases, elastic applicable frame for

stability

directly

predicting the structural strength.

However, the LRFD

beam-column interaction equations, which predict frame instability, require the use of effective length factors. An elastic frame stability analysis provides a powerful

101 tool to determine the effective length factors. The program [ E S C A P ] was written to provide exact elastic second-order analysis based on the use of stability functions. The program was verified by

comparison with published closed-form solutions. The program stability [ E S C A P ] of members also permits elastic and plane frames.

analysis

Instability is determined by increasing the applied load increments matrix until the elastic second-order stiffness the

becomes

singular.

Rather than

checking

singularity by computing the determinant, singularity was detected by checking for a zero or negative diagonal element using Cholesky's decomposition of the stiffness matrix. This provides a simple technique to incorporate

stability analysis into an elastic second-order program. The accuracy computing comparison elastic with of the program stability known [ E S C A P was ] in by The

loads

verified

closed-form

solutions.

solutions are in excellent agreement. The future of structural design practice will

include second-order effects in the computer analysis and design of structures. Structural analysis programs like

[ E S C A P ] will become routine.

AP PEND IX

102

APPENDIX

A
USERS MANUAL ELASTIC SECOND-ORDER COMPUTER ANALYSIS PROGRAM

Al

INTRODUCTION

The Elastic Second-Order Computer Analysis Program [ E S C A P ] was developed to provide designers a tool in the analysis program and design of beam-columns and frames. The

[ E S C A P ] has the capability of performing

analysis as well as elastic

first-order linear elastic second-order analysis. perform elastic columns.

The program was also designed to analysis of frames and beam-

stability

The program can be used to perform structural

analysis on any two-dimensional structures subjected to inplane bending. The program can be loaded in any IBM PC or compatible with a hard-disk or floppy. precision Compiler. FORTRAN with the It IBM is written in singleProfessional Fortran

103

A2

CAPABILITY

The program E S C A P has the following capabilities a. Performs first-order linear analysis of plane frames, beams, columns, plane trusses and beam-columns. b. c. d. Performs elastic second-order analysis. Performs elastic stability analysis. Determines the location and magnitude of the maximum elastic second-order moment within each member. e. Has the capability of analyzing multiple load patterns.

A3

PROGRAM PARAMETERS

The following parameters were used in the computer program [E S C A P ] : PID - Problem Identification. This is the name of the It

input file with a brief description of the problem.

consists of 12 characters per line with a maximum of three (3) lines provided in the input file. TITLE is a heading that describes the input

parameters in a read statement. 104

105 NJ - number of joints; Maximum of 40. The user has an

option to increase the size by increasing the dimension of the joint coordinates X ( ) and Y ( ). NDELTA analysis. 0 : 1st order analysis 1 : 2nd order analysis NCYCLE it is the maximum number of iterations. The
-

is

code

that

determines

the

type

of

program will terminate if after NCYCLE the second-order solution has not converged, and will priat the output of the last computed values of member forces and joint

.eflections.

It normally takes between 4 to 10 iterations

before the solution converges. EPS1 is the tolerance .005 ) increment used in the factor used to test

convergence.( recommend STEP - is the

load factor

elastic stability analysis.( recommend LB -

.005 )

is the code that determines if buckling load

analysis is desired. 0 : NO 1 : YES

X ( J )
-

is the x

coordinate of a joint J, in feet.

Y (K)

is the y - coordinate of a joint K in feet.

106
MCL
-

is the joint release code.

0 : joint is free 1 : joint is fixed. NDOF - is the total number of degrees of freedom. ( equal to 3 * NJ )
NRJ
-

number of restrained joints.

-

S ( I,J ) or a member

is the stiffness matrix of the structure

NRL - number of restrained degrees of freedom (DOF's) that are fixed and have 0 displacements.
MNDOF = NDOF - NRL , is the total number of degrees of

freedom that are free to rotate and translate. NM - is the number of members. Maximum is 80.

The user has the option to increase the size by changing the dimension of the other parameters in the main program. MPC - Matrix Print Code for the stiffness matrix. 0 : suppress output 1 : activate output

GX

- is the unit weight of the material; 490 pcf for

structural steel. IHINGE ( I )

-

is a joint code that describes the It is used to

type of restraints imposed at the joints.

107 modify the stiffness of a member that is pinned at one end. For members without supports : 0 : pinned at start joint and rigid at the end joint 1 : rigid joints at both ends 2 : joints pinned at both ends For members with support joints (hinged or fixed): 1 : if the beginning joint is a support joint 0 : if the beginning joint is pinned and not a support joint JFIX modify the joint. 1 : for a fixed support or rigid joint 0 : for a hinged support or pinned joint JM ( I ) - beginning joint of a member KM ( I ) - end joint of a member AM ( I ) - area of cross-section, sq. in. ZM (I )moment of inertia; in' ( J )
,

KFIX ( K ) = are joint codes used to force vector for a pin

load vector and the

EM ( I ) - modulus of elasticity of the material Example : Enter 29 (ksi) for steel

SPAN - member length in inches

SM ( I ) - member length in feet

108 CX ( I )
CY WT
-

cosine
sine

( I )

total weight of structure in tons.

-

INODE ( ) of a member

is a degree of freedom ( DOF ) at the ends

(1 )

DOF at joint JM ( I ) along the x - axis

( 2 ) - DOF at joint JM ( I ) along the y - axis

( 3 )
( 4 ) ( 5 )

rotation at joint JM ( I )
DOF at joint KM ( I ) along the x - axis DOF at joint KM ( I ) along the y - axis

( 6 )

rotation at joint K, ( I )

Q ( I,J )

stiffness matrix for the restrained DOF's

NLP - number of loading patterns LP - load pattern NLJ - number of loaded joints WX - joint load in x - direction ( kips ) WY - joint load in y - direction ( kips ) WM - joint moment W ( kip-ft )

( )

sum of joint loads in x , y or moments

-

FEA ( )

fixed-end actions or forces in members due

to externally applied loads. MEA ( )

-

member end actions after the analysis

109 NMC NCL

-

number of members with concentrated loads number of concentrated loads on a member

limited to 4).

The user can increase the size by changing The location of

the dimension of LCN in the main program.

loading is expressed as a fraction of the span measured from the left support.
FX - concentrated load in x - direction ( kips ) FY - concentrated load in y - direction ( kips ) MO - moment in z - direction ( kip-ft )

NMU - number of members with uniform loads NUL - number of uniform load3 on a member SUMWY ( )
-

sum of transformed

uniform distributed

loads on the member along the MWC

-

y - axis

member weight code.

0 1
R ( )
-

exclude membets weight in the analysis include member weight in t e analysis

reaction vector

G ( )

modified fixed end actions

-

RAL ( )
D '

identifies the restrained DOF

deflection vector

XNEW ( ) - new x-coordinates YNEW ( ) - new y- coordinates

Dl, D2, D3, D4, D5, D6 - deflections at member's local axis

110 FIX ( ) order

-

calculated fixed-end actions during second-

iteration analysis ST ( )
-

modified stiffness matrix for second-order

effects. LF
-

load factor use in load buckling analysis.

It is

equal to 1 at the start of the analysis. ITER counts number of times the load factor

increment STEP was added to LF before the buckling load is achieved. QT - restrained stiffness matrix for the second-order analysis
P - is the axial load or force

XKL = sqrt ( P / EI ) VA, VB effeCts MA, MB effects COMP - stability functions for axial compression TENS - stability functions for axial tension FLEXMAX - maximum moment anywhere in the span
YMAX - deflection of FLEXMAX. (Note: this may or may

vertical reactions considering axial load

moment reactions

considering axial load

not be the maximum deflection )

XMAX - location of FLEXMAX and YMAX

111 A4 LIST OF SUBROUTINES

a.

E S C A P

is the main program.

It reads the

input file and performs the first-order analysis. b. NODE is a subroutine that generates the degrees of

freedom for each node. c. STIFF calculates and assembles the global first-

order stiffness matrix. d. e. program. MATPRT prints the global stiffness matrix. DECOMP is the CHOLESKY's matrix decomposition

It decomposes the global stiffness matrix and

stores the matrix in S ( I,J ) or ST ( I,J ). f. LOAD modifies the fixed-end actions FEA for a

member that is pinned at the beginning node or both. g. SOLVER solves the equations for unknown

deflections and stores them at vector D ( ). h. MFORCE calculates the first-order member end

actions in the Z.ocal axis and stores them at vector MEA i. PNDELTA performs the second-order analysis.

j. XDELTA generates the global second-order stiffness matrix and stores the matrix at vector ST k. DFORCE solves the members end actions (MEA) to

include second-order effects.

1. CALMAX determines the location of the maximum

112 moment for a uniformly loaded member with end moments. also calculates the elastic second-order moments It and

deflections.

A5 INPUT DATA PREPARATION

An editor (such as PE FORTRAN ) input file.

is used to prepare the

The parameters of every read statements are A sample

entered in separate lines and are format free. copy of the input file is shown in Appendix B. PROCEDURES:

A. READ PID - in alphanumeric characters and consists of 12 characters per line with a maximum of 3 lines.
B. READ TITLE Enter > number of

joints,NDELTA,NCYCLE, EPS1,STEP,LB C. READ NJ, NDELTA, NCYCLE, EPS1 D. READ STEP, LB E. READ TITLE
-

Enter> Joint Coordinates Y ( J )

F. READ J, X ( J ),

J : is the joint or node G. READ TITLE - Enter> Number of restrained joints H. READ NRJ I. READ TITLE Enter> Boundary Conditions (Note:

113 these are the boundery conditions of the restrained joints in procedure G & H) J. READ J, ,MCL ( 1 ), J MCL ( 1 ) MCL ( 2 ) MCL ( 3 ) K. READ TITLE
-

MCL ( 2 ) , MCL ( 3 )

is the joint or node DOF in DOF in X - DIRECTION Y - DIRECTION

DOF in ROTATION Enter> Number of members,print code

L. READ NM,MPC
M. READ TITLE
-

Enter> Member Data

N. READ I,J,K,A,Z,E,GX 0. READ IHINGE (I),JFIX (J), KFIX (K) I : is the member

J : beginning node K : ending node A : area of cross-section ( in 2 Z : moment of inertia ( in ' ) E : section modulus ( KSI/1000 ) GX unit weight of member ( LBS/FT 3 )

P. READ TITLE - Enter> Number of Load Patterns Q. READ LP R. READ TITLE - Enter> Number of loaded joints S. READ NLJ ; if NLJ is equal to 0 go to

114 step V

T. READ TITLE - Enter> Loaded joints

U. READ J,WX,WY,WM J
V.

is the joint or node

READ TITLE - Enter>Number of members with

concentrated loads W. READ NMC ; if NC is step equal to AA 0 , go to

X. READ TITLE - Enter> Members concentrated loads Y. READ Z. READ K 0 1 A M, NCL K, A, FX, FY, MO load orientation member axis structure or global axis location of loading ( fraction of the span ) AA. READ TITLE Enter> Number of members with

distributed loads BB. READ NMU ; if NMU is equal to to step FF 0 go

CC. READ TITLE - Enter> Members uniform loads DD. READ EE. READ M,NUL K,FX,FY,MO

M : name of member

115 K : load orientation 0 : local or member axis 1 : global or structure axis MO is expressed in kip-ft / ft

FF. READ TITLE - Enter> Member weight code GG. READ MWC

0 : exclude member weight 1 : include member weight

A6 OUTPUT DATA

The computer output shows the joint coordinates, members data, joint loading, and the member loading in a load

pattern. actions maximum shown.

It also prints the joint deflections, member end and the moment, The reaction forces. The location of the are also in

its magnitude factor LF

and deflection is printed and

load

used

determining the actual buckling load, and is obtained by multiplying the load factor LF with the member's load or joint load. A sample output data is shown in Appendix B.

116 A7 RUNNING THE PROGRAM

The following steps describe how to run the Elastic SecondOrder Computer Analysis Program [ E S C A P ]: 1. Name the input file described in Section A6 above. The input file is named when saving the input data. 2. 3. Go to the Fortran Directory. To run the program ENTER > XFRAME RETURN 4. The program will prompt> What is the name of the input file? 5. ENTER > Name of input file. Make sure the input

file is saved in the Fortran directory. 6. 7. RETURN Execution of the program is completed when the

message Execution Terminated : 0 shows in the screen. 8. The output file is automatically sent to the file

named SOLOUT in the Fortran directory.

APPENDIX

COMPUTER INPUT AND OUTPUT DATA For Example 5.2-4

117

118
INPUT FILE - SAMPLE.IN

RIGID FRAME EXAMPLE 5.2-4 NUMBER OF JOINTS,NDELTA,NCYCLE, EPSI, STEP, LB 6,1,1000, .005 .005,0 JOINT COORDINATES 1,0,0 2,0,12 3,0,20 4,24,20 5,24,6 6,24,0 NUMBER OF RESTRAINED JOINTS 2 BOUNDERY CONDITIONS 1,1,1,1 6,1,1,1 NM, MPC, MCODE 5,0,1 MEMBER DATA 1,1,2,13.3,250,30,490 1,1,1 2,2,3,13.3,250,30,490 1,1,1 3, 3, 4,13.3,250,30,490 1,1,1 4,4,5, 13.3,250, 30,490 1,1,1 5, 5, 6, 13.3,250,30,490 1,1,1 LOAD PATTERNS 1 NUMBER OF LOADED JOINTS 4 LOADED JOINTS 2,12,0,0 3,0,-200,0 4,0,-200,0 5,-6,0,0 NUMBER OF MEMBERS W/ CONCENTRATED LOADS 0 NUMBER OF MEMBERS WITH DISTRIBUTED LOADS 1 LOADED MEMBER 3,1 1,0, -2,0

119 MEMBERS WEIGHT CODE 0

120
"

OUTPUT DATA

"

ELASTIC SECOND-ORDER COMPUTER ANALYSIS OF BEAM-COLUMNS AND FRAMES PLANE FRAME ANALYSIS BY STIFFNESS METHOD

GENE

MUNOZ

** **

* **

** *

* *

** ***

** *

*
*

JOINT

DATA*

JOINT

COORDINATES X Y (FEET) (FEET)

RESTRAINT CONDITION X Y R

0.000

0.000

R R R

2 3 4 5 6

0.000 0.000 24.000 24.000 24.000

12.000 20.000 20.000 6.000 0.000

R R R

**

**

**

**

*MEMBER

DATA*

MEMBER

LENGTH (FEET)

CROSS-SECTION PROPERTIES I(ZZ) AREA (INS**4) (INS**2) 13.300 13.300 13.300 13.300 13.300 250.00 250.00 250.00 250.00 250.00

MODULUS OF ELASTICITY (KSI/1000) 30.000 30.000 30.000 30.000 30.000

1 2 3 4 5

12.000 8.000 24.000 14.000 6.000

121

TOTAL WEIGHT OF FRAME:

* ***** FIRST ******

1.448 TONS
*** *****
**
*

*** ****
* *

ORDER
** ** * **

ANALYSIS
*** **

***** **

***

* *

LOADING PATTERN

1 OF

1 LOADING PATTERNS

INT

L 0 A D I N *

OINT F (X) (KIPS) 12.000 0.000 0.000 -6.000

LOADING F (Y) (KIPS) 0.000 -200.000 -200.000 0.000

COUPLE (KIP-FEET) 0.000 0.000 0.000 0.000

2 3 4 5

**

**

**

***

*MEMBER

LOADING*

NO MEMBERS WITH CONCENTRATED LOADS IN PATTERN

MEMBER

UNIFORMLY DISTRIBUTED LOADING ORIENTATION W(X) W(Y) M(O) (AXES) (KIPS/FT) (KIPS/FT) (K-FT/FT) S

0.000

-2.000

0.000

122

*MEMBER

WEIGHTS NOT INCLUDED IN ANALYSIS*

LOADING PATTERN

1 OF

1 LOADING PATTERNS

JOINT

DE D(X) (INCHES)

FL

E C T10N D(Y) (INCHES)

S ROTATION (RADS)

1 2 3 4 5 6

O.OOOOE-O1 3.0288E-01 6.4644E-01 6.4031E-01 1.6280E-01 O.OOOOE-O1 LOADING PATTERN

O.OOOOE-O1 O.OOOOE-O1 -8..0369E--02 -2.5911E-03 -1.3395E-01 -6.306CE-03 -1.3553E-01 3.8776E-03 -4.0658E-02 -4.2348E-03 O.OOOOE-O1 O.OOOOE-O1 1 OF 1 LOADING PATTERNS

*MEMBER

END

ACTIONS*

MEMBER 1

END JOINTS 1 2 2 3 3 4

X-FORCE (KIPS) 2.2269E+02 -2.2269E+02 2.2269E+02 -2.2269E+02 8.4941E+00 -8.4941E+00

Y-FORCE (KIPS) 3.5060E+00 -3.5060E+00 -8.4940E+00 8.4940E+00

MOMENT (K-IN) 3.8739E+02 1.1748E+02 -1.1748E+02 -6.9794E+02

6.9794E+02 2.2688E+01 2.5312E+01 -1.0757E+03

123 4 4 5 5 5 2.2531E+02 -2.2531E+02 2.2531E+02 6-2.2531E+02 8.404!~E+00 -8.4941E+00 2.4941E+00 -2.4941E+00 1.0757E+03 3.5134E+02 -3.5134E+02 5.3091E+02

LOADING PATTERN

1 OF

1 LOADING rATTERNS

R EA

CT

I VE

FO0RC E S

JOINT

R EA R(X) (KIPS)

N S CTIO R(Y) (KIPS)

MOMENT (KIP-INCH)

-3.5060E+00OO29EO 3. 8739E+02

-2.4941E+002.51E0 5
.3091E+02

124

S SECO0N D

O RD

E R

LOADING PATTERN

1 OF

1 LOADING PATTERNS

J JOI NT

D E FL

EC

TIO

N S

JOINT

DE FL D(X) (INCHES) O.OOOOE-O1

EC

TIO N S D(Y) ROTATION (INCHES) (RADS) O.OOOOE-O1

1 2 3 4 5 6

O.0OOOE-01

4.0855E-01 -8.0162E-02 -3.7614E-03 8.5261E-01 -1.3360E-01 -7.0316E-03 8.4646E-01 -1.3587E-01 3.5363E-03

2.0170E-01 -4.0761E-02 -5.2386E-03 O.OOOOE-O1 O.OOOOE-01 O.OOOOE-O1

COW~kZZCE OCC-t;:PZD AT ITER

125 LOADING PATTERN 1 OF 1 LOADING PATTERNS

*MEMBER

END

ACTIONS*

MEMBER

END JOINTS 1 2 2 3 3 4 4 5 5 6

X-FORCE (KIPS) 2.2212E+02 -2.2212E+02 2.2212E+02 -2.2212E+02 8.5212E+00 -8.5212E+00 2.2588E+02 -2.2588E+02 2.2588E+02 -2.2588E+02 1 OF

Y-FORCE (KIPS) 3.4786E+00 -3.4786E+00 -8.5214E+00 8.5214E+00 2.2115E+01 2.5885E+01 8.5213E+00 -8.5213E+00 2.5213E+00 -2.5213E+00

MOMENT (K-IN) 4.8161E+02 1.1006E+02 -1.1006E+02 -6.0936E+02 6.0936E+02 -1.1522E+03 1.1522E+03 4.2502E+02 -4.2502E+02 6.5212E+02

1 2 3 4 5

LOADING PATTERN

1 LOADING PATTERNS

R EA

CT

I VE

FO0R CE

JOINT

REA R(X) (KIPS) -3.4786E+00

C TIO N S R(Y) (KIPS)

MOMENT (KIP-INCH)

2.2212E+02 4. 8161E+02 6 -2.5213E+00 2.258E+02 6. 5211E+02

126

LOADING PATTER~N

1 OF

1 LOADING PATTERNS

*MAXIMUM

MOMENTS*

MEMBER XNAX (INCH) 1 2 3 4 5

MOM

E NT YMAX (INCH)

S MOMENT (KIP-INCH)

O.OOOOE-O1 O.OOOOE-O1 4. 8161E+02 9.5948E+01 -1. 3360E-01 -6.0936E+02 2.8799E+02 _1.3587E-01 -1. 1522E+03 O.OOOOE-O1 -1. 3587E-01 1. 1522E+03 7.1960E+01 O.OOOOE-O1 6. 5212E+02

ANALYSIS COMPLETED

REFERENCES

[1.]

T.oad and Resistance Factor Deslyn Specification for Structural Steel, First Edition, American Institute Of Steel Construction, Inc., 400 North Michigan Avenue, Chicago, Illinois 60611, 1986 Chen, Wai-Fah and Atsuta, Toshio, Theory of BeamColumns, Volume 1 : In-Plane Behavior and Design, McGraw-Hill Book Company, Inc., N'w York, 1976, p. 1-170 Chajes, Alexander, Principles of Structural Stability, Prentice-Hall, Inc., Englewood Cliffs New Jersey, 1974, p.1-74, 133-143, 145-194 Stephen P. Timoshenko and James M. Gere, Theory of Elastic Stability, McGraw-Hill Book Company, Inc., New York, 1961, p. 1-45 Chen, W.F, and Lui, E.M, Structural Stability, Theory and Implementation, Elsevier Science Publishing Co., Inc., 52 Vanderbilt Avenue, New York, NY 10017, 1987, p.147-192 Salmon, Charles G. and Johnson, John E, Steel Structures, Design and Behavior Emphasizing Load and Resistance Factor Design, Third Edition Harper & Row, Publishers, Inc., 10 East 53rd Street,New York, NY 10022, 1990, p.829-832 Iyengar, N.G.R, Structural Stability of Columns and Plates, Ellis Horwood Limited, Halsted Press John Wiley & Sons, New York, 1988, p.1-65, 110137 Crawley, Stanley W., Dillon, Robert M., and Carter, Winfred 0, Steel Buildings , Analysis and Design, Third Edition, John Wiley & Sons, New York, 1984 p.590-598, 615-637 Ghali, A. and Neville, A.M, Structural Analysis A Unified Classical and Matrix Approach, Chapman and Hall Ltd.,f1 New Fetter Lane, London ECP4P 4EE, 1979, p. 356- 397 Harrison, H.B, Computer Methods in Structural Analysis, Prentice-Hall, Inc.,Englewood Cliffs, New Jersey, 1973, p. 235-276 127

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

128 [11] West, Harry, Analysis of Structures, An Integration of Classical and Modern Methods, John Wiley & Sons, Inc.. New York, 1980, p.64 Weaver, William Jr., and Gere, James M., Matrix Analysis of Framed Structures, Third Edition, Van Nostrand Reinhold, New York, 1990, p.428-433 White, Donald W., and Hajjar, Jerome F, Proceedings: "Application of Second-Order Elastic Analysis in LRFD", AISC National Steel Construction Conference, Kansas City, 1990, p. 11- 3, 11-22 Beaufait, Fred W., Rowan, William Jr. H, Hoadley, Peter G., and Hackett, Robert M, Computer Methods of Structural Analysis, Prentice-Hall, Inc., Englewood, New Jersey, 1970, p.36-41, 342-386 Lui, Eric M., A Practical P-Delta Analysis Method for Type FR and PR Frames, Engineering Journal, American Institute of Steel Construction, Third
Quarter, 1988, p. 85 - 98.

[12]

[13]

[14]

[15]

[16]

Fleming, John F., Computer Analysis of Structural Structural Systems, McGraw-Hill Book Company, New York, 1989, pp. 528-541 Wood, Brian R., Beaulieu, Dennis, and Adams, Peter F., Column Design By P-DELTA Method, Journal of the Structural Division, February 1976, pp. 411-427

[17]

..

m n

mm

unnn