Lehigh University

Lehigh Preserve
Fritz Laboratory Reports Civil and Environmental Engineering

1976

Design criteria for i-columns subject to biaxial

bending, Feb. 1976
D. A. Ross

W. F. Chen

Follow this and additional works at: http://preserve.lehigh.edu/engr-civil-environmental-fritz-lab-

reports

Recommended Citation
Ross, D. A. and Chen, W. F., "Design criteria for i-columns subject to biaxial bending, Feb. 1976" (1976). Fritz Laboratory Reports.
Paper 2092.
http://preserve.lehigh.edu/engr-civil-environmental-fritz-lab-reports/2092

This Technical Report is brought to you for free and open access by the Civil and Environmental Engineering at Lehigh Preserve. It has been accepted
for inclusion in Fritz Laboratory Reports by an authorized administrator of Lehigh Preserve. For more information, please contact
preserve@lehigh.edu.
j ..:
.; .. ,

by

D. A 6 1{.\)SS \.J. .F. Cllen

:(~r:SE.2TCil .t\SH'istnnt Professor of Civil Engineering
Fritz Engine~=ing L~boratory Frit~ Eoginee~ing Laboratory
Lehigh Jniversity Lehi0t University
}jsU·.tlcher;!, P;L 1.8015 Reth.Ic.:1e:1>, Pa. 1:3015

FRITZ ENGINEERING
A8r-.no···. u'\!J'{ LIBRARY
Lrt. Ut\r' l ' '

.·.
L'~ h j_gh /FL/ 393,. 3B
 -i

TABLE OF CONTENTS

. "
ABSTRACT iv

1. Introduction 1

2. Background 3

3. Previous Work q.

4. Analysis 5

4.1 Assumptions Used in the Computer Model 5

4.2 Shapes Adopted for Analysis 6
4.3 Analytical Variations Considered 6

5. Results and Discussions 8

6. Comparison with Vinnakota's Work 10

7. Summary and Conclusions 10

8. Acknowledgments 11

9. References 1.?.

10. Nomenclature 13

TABLES 14

FIGURES 17
 -ii

LIST OF TABLES

. . Table

1 Load Range Considered

2 Haximum Strength Tables for Bj_axially Loaded Beam-Columns

 -iii

•.
LIST OF FIGURES

Figure

1 A Typical Maximum Strength Interact:i.on Surface for a Parti-

cular CoJ.umn of Kaown Length Subject to Biaxial Bending

2 Maximum Strength Interaction Curves for W8x31 at p = 0.3

(From Tebedge and Chen's Results, 1974)

3 Maximum Strength Interaction Curves for Wll.~x38 at p - 0. 3

4 Haximum Strength Interaction Curves for W2lx44 at p = 0.3

5 Maximum Strength Interaction Curves for W8x31 at p = 0.5

(From Tebedge and Chen's Results, 1974)

6 Maximum Strength Interaction Curves for Wl4x38 at p ::: 0.5

7 Maximum Strength Interaction Curves for w2lx44 at p - 0.5

8 Comparison of Observed and Recommended Values of ~

9 Testing of Program in Region 1-.1here ~ == 1.0

10 Comparison of Results Predicted by Proposed Formula and ·

Vinnakota's Solution. (Solid Line is Proposed Formula, and

Dotted Line is Vinnakota's Solution.)

 DESIGN CRITERIA FOR I-COLUMNS SUBJECT TO BIAXIAL BENDING
by

D. A. Ross H. F. Chen
Research Assistant Professor of Civil Engineering
Fritz Engineering Laboratory Fritz Engineering Laboratory
Lehigh University Lehigh University
Bethlehem, Pa. 18015 Bethlehem, Pa. 18015

ABSTRACT

The design code Canadian Standard Association 516.1-1974

(limit states design) permits ultimate strength design of steel H-

columns subjected to axial load and biaxial bending moment. Hm.;ever,

the design formula recommended by the CSA Sl6 for stability is applicable

only for H-sections in which the flange width to section depth ratio

is equal to or greater than 0.8. In this paper a simple modification

to the stability design formula is proposed which enables the restl:ic-

tion on flange width to section depth ratio to be removed so that they

are also applicable to steel I-columns.

 -1

Canadian structural engineers have been quick to recognize

.
the usefulness of design criteria ~tich enable columns subjected to

bending moment about both axes, as well as axial load, to be designed

by the ultimate strength design method. In the design code Canadian

Standard Association 816.1-1974 (limit states design), a new method for

the stability design of H-sections under such loading conditions was

presented.

For wide-flange cross-sections in which the flange width, B,

to section depth, D, ratio exceeds the value,

B/D _:: 0.8 (1)

the code allows the maximum bending moment capacities of the column

to be calculated by the relationship

~
,,M X \
!--) 1.0 (2)
\M 1
ux

where

M ,M - a[lplied end moments about the x- and y-axes, respectively

X y
H
ux
)H
uy
= maximum end moment capacity about the x- or y-axis, respec-

tively, that can be resisted by the ~ember in the plane of

bending when the

. ,
ax~a1. loal is included, but in the absence

of the other moment.

The values of M ,M may be computed with good ac~uracy from the for-
ux uy
mulae given also in AISC specifications, 1969:

. p \/ p \
H ::: M 1 - - ·- il l - -~- l (3a)
ux m( P 1\ p J
cr ex
 -2

p \(1 p \
-
p I I\ p--; {3b)
cr ey

in which

P
cr
= axial load producing failure in the absence of bending

moment

Pex' P ey == Euler buckling load of the column about the x- or y-axis,

respectively

Mm maximum moment that can be resisted by the member in strong

axis bending in the absence of axial load and \vea.k axis

bending moment, i.e., plastic moment capacity about the

x-axis, reduced for the presence of lateral torsional

buckling if necessary.

In lieu of a more precise calculation, M may be taken as

m
(AISC Specification, 1969)

(t/r )/;-,
r
M == I 1.07 - y YJM <M (4)
m 3160 J px px

\<7here

~1 M
px' PY
= plastic bending moment capacity of section about x- or

y-axis, respectively in the absence of axial load

ay = material yield stress

t == column effective length

r
y
= radius of gyration irr y-direction.

The code recommends the exponent a in Eq. (2) to be taken as

~ = 1.40 + p (5)
 -3

p
where p = p (6)
y

in which P is axial load at full yield condition.

y

However, the section shape restriction given by Eq. 1 limits

the applicability of this method to I-sections. Two pertinent qaestions

might then be asked: Does the exponent 13 have any shape dependence; and,

if so, how should Eq. 5 be modified to account for this? This paper

investigates the question of shape dependence and suggests a simple

modification to Eq. 5 which then extends the formula (2) to include

both H- and !-sections.

2. Background

An examination of Fig. 1 clarifies many of the premises of

this paper. Figure 1 represents, in tHo dimensions, what is essentially

a three-dimensional surface describing the maximum strength of columns

subject to axial load and biaxial bending moment. Thus it shaHs a

typical maximum strength interaction surface for a particular beam-

column length.

If the s6lid lines on the mutually perpendicular planes of

Fig. 1 represent the actual failure curves under the relevant restricted

loading conditions, then the dotted lines represent previously established

approximations. In particular, the dotted lines on planes P-H anr~ P-M

X y
are reorganized versions of equations 3(a) and 3(b). Both the AISC

design expressions and CRC Eq. (6.19) of the Second Edition (1966) of

the Column Research Council's Guide to Design Criteria for Metal Com-

pression Members are straight line interaction equations of moments

 -4

on planes P = constant. Recent research has shown that the interaction

of moments about the orthogonal axes is not linear; on the contrary,

the interaction curve resembles more closely the quadrant of a circle

(shaded horizontal plane in Fig. 1). This research reported herein

uses the two dotted curves on planes P-M and P-M (with equations as
X y
given) as end points. From these, a nonlinear expression is derived

to describe the circle-iike curve Hhich corresponds to a particular

value of applied axial load. This nonlinear expression is taken to

be adequately approximated by Eq. 2. By deriving the correct exponent,

B, an interaction curve can be defined which fits the actual strength

curve of a biaxially loaded beam-column of a particular cross section,

such as the H-shape given previously by Eq. 5. Studies of !-shapes made

using the same computer model are presented herein.

3. Previous Work

Santathadaporn and Chen (1973), reported the development of

a computer model for the analysis of biaxially loaded columns of wide

flange cross section with inelastic stability problems. The load-

displacement curves for such colurr,ns could b2 derived, and by taking

the maximum strength from each of these curves, maximum strength inter-

action curves could be derived for use in design. The computer model

considered material yielding, residual stresses, end warping restraint,

end bending restraint, and initial imperfections. A tangent stiffness

method of analysis was used, using an incremental iterative procedure

to derive points on the load-displacement curves. Tebedge and Chen

(1974), then developed and proposed 2 design ~ethcd for the biaxial

bending of steel H-columns using this computer model. In addition they

 -5

produced design curves which could also be used once the applied axial

load and bending moments about each axis were known. On the basis of

this work Eqs. 2 and 5 Here included in the provisions of Canadian

Standard Association Sl6.1-1974.

The investigation into the biaxial bending behavior of I-

columns for which B/D :::_ 0.8 required additional numerical results from

thE: computer model, and simple modifications of the exponent, ~' to

enable formula (2) to be applicable to all wide flange shapes. As

discussed by Tebedge and Chen (1974): the computer program has been

developed for a symmetrical loading condition, i.e., applied loads at

one end of the column are resisted by equal and opposite loads at the

other end of the column. This is not a limitation to the extent that

the "equivalent" or "C ·-metho:i" may be used to transform unsymmetrical

m
loading conditions b.to symmetrical condition~: appropriate to this

ma~hematical model. An up-to-date summary of this new development is

given in a recent paper by Springfield (1975).

The reader is referred to the paper of Santathadaporn and

Chen (1973), for the complete formulation of the analytical procadure

and assumptions. As stated therein, the assumpU.ons include:

1. Failure due to local buckling does not occur,

2. Slopes are sufficiently small such that curvatures can be

approximated by the second derivative of the relevant dis-

placement with respect to distance along the column,

 -6

• 3. The rotation or twist of the column is small, but terms in-

eluding products with one term either a second derivative or

a rotation are not negligible,

4.. The initial out-of-straightness may be approximated by a

half.:sine Have,

5. Bending restraint but no twisting restraint may exist at

the column ends, and,

6. The material is elastic-perfectly plastic.

For derivation of the biaxial moment interaction curves, Tebedge and

Chen (1974) assumed that there was no mo~ent restraint at the ends,

thus giving a pin-ended condition. This assumption is also followed

for the analysis herein.

Two narrower !-shapes in the light-~eight range were adopted!

one with B/D ~ 0.5, and the other with B/D ~ 0.3.

Wl4x38 for >·7h:i.ch. B/D 0.48

W2lxl~4 for v7hich B/D = 0.32

In addition, the results repo~ted previously by Tebedge and Chen (1974)

for the nearly square shape W8x31 for which B/D ~ 1.00 were also included

for comparison.

Tebedge. and Che:1 (1974), found that the ratio of length, t,

to radius of gyration about the x-axis,

.
r X , was no~ a significant

variable iu the results and proposed Eq. (5) such that it was conserva-

tive for all practical t/r X . Thl! ratio t/r

' y
i.s more commonly used
 -7

• for design purposes, although the t/r ratio can reaJily be deduced
y
from the t/r ratio for a particular column section. The variation of
X

t/r vias selected such that a reasonable mo:nent capacity of the column
X

remained after application of the axial load. The axial load cases

considered are p ~ 0.1, 0.3, 0.5 and 0.7, and are sur.~arized in Table 1.

In each case a linear variation of residual stress was

assumed with a maximum value of 0.3 c . Tebedge and Che::1 (1974),

y
showed that the actual stress magnitudes do not affect the shape of

the interaction curves, and thus this assumption is not critical.

The relationship of the present interaction curves to the

multiple column curves, such as those presented in the revised CRC

Guide (Johnston, 1975), is also of interest. These multiple column

curves nt,rovide an accurate estimate of the ultimate axial load P

Cl"

which an axially loaded column may carry, given a certco.in slenderness

ratio. As such it corresponds to the end point P of the interaction

surface being derived herein. Hence, in using the proposed interaction

equation (2), the determination of the axial lnad capacity, P of

cr'
the column should be made as realistic as possible, such as using the

multiple column curves.

As a check on the lower limit of formula appli~ability, one

interaction curve was derived for a column of a \\T2lxMJ. section \vith a

t/r
X
ratio of 30 for t/r
' y
= 190)
'
and an axial load ratio of D. = 0.1.
 -8

5. Results and Discussions

The curves of maximum strength under axial load and biaxial

bending moments \vere plotted on axes of M /M vs. M /N , and are also

y PY x px
tabulated in Table 2. If the end points are considered to be M where
uy
the curve crosses the M /H axis, and M 1·1here the curve crosses the
Y PY ux
H /M axis, then the curves were replotted on axes of M /H vs.
X ~ y ey
H /M . Some of these curves are shmvn in Figs. 2 through 7. In vie\v
X UX

of the results obtained, a modification is suggested to Eq. (5), such

that the expression for the exponent ~ be adopted as

.f3 = 0. 4 + p + B/D ~ 1 for B/D _2: 0.3

and (7)
~ :·:: 1.0 for B/D < 0.3

Using Eq. 7, the nonlinear interaction of moments calculated from Eq. 2

ic also plotted in each of Figs. 2 through 7, dashed curves.

Figures 2 through 7 illustrate the applicability of the

~xpression (7). When B/D ~ 1.0, this expressio~ becomes identical

to Eq. (5), as proposed previously by Tebedge and Chen (1974). The

equation vms seen to give remarkably accurate predictions of maximum

strength interaction of moments for sections with B/D ~ 1.0 (Figs. 2

and 5). For sections which deviate markedly from being nearly square

(i.e., B/D I 1.0), Eq. (7) is found to be adequate, but to give more

conservative predictions (Figs. 3,4,6,7)~

Some attempts were made to find a more accurate estimate of B

for sections \,•ith BiD << 1..0, and the results of these "true" ~ curves

are plotted in dashed lines, together with the proposed expression (7)

in Fig. 8. The more accurate values of S -..1ere derived by a trial-and-

error process to fit the actual strength curve of computer solution.

For simplicity, it is proposed that Eq. (7) be adopted for general use.
 -9

The results presented are derived for columns in the light

to intermediate \oleight range. However, Tebedge and Chen (1974), investi-

gated sections of varying weight with the same B/D ratio, and found

little or no variation in the curves derived. Thus, it is assumed that

this result is similarly applicable to the present results.

It is noted also that Eq. (7) proposed that ~ = 1.0 be

adopted as a lower limit. The curve ~ = 1.0 is the straight line

representing the design expressions assumed in traditional code requ1re-

ments. There is evidence that for B/D:::; 0.3 the true value of ~ is

close to B= 1.0. As a check on the viability of this lower limit, an

interaction curve was derived for the W2lx44 section with an axial load

ratio of p = 0.1. Thus, by calculation, the corresponding value of

B "" 0.82. But we have stipulated a B . of 1.0. It can be seen from

:n1.n
Fig. 9 that there does not have any concavity in the predicted strength

curve. The adoption of the lower limit of: ~::: 1.0 i.s recom:1.1ended. I:1

design practice, rolled sections with B/D < 0.3 are very rare. Further,

the formula is not recommended for use with non-standard sections.

Experimental verification of the proposed formula is not

undertaken herein. · Extensive investigation Has conducted on the

experimental justification of the formula proposed by Tebedge and Chen,

197l~. This propasal was merely a restricted version of the p"Lesent

proposal, and it was found (Springfield and Hegan, 1973) to be a good

prediction of experimental results.

 ··10

6. .f~arison with Vinnakota' s Hork

Hork by Vinnakota (1975), has now advanced to the stage \vhere

some comparisons are possible. This work has been undertaken indepen-

dently and with a different approach, and so the comparisons made are

quite valuable. Figure 10 is a comparison of Vinnakota's results and

those of Tebedge and Chen (1974), for a wide-flange shape, i.e., W8x31

at an axial load ratio, p, given by

p = 0.3
It can be seen that the results of Tebedge and Chen are generally

a conservative prediction of Vinnakota's results (dashed lines). Since

Eqs. 7 and 8 are a conservative approximation of Tebedge and Chen's

results it is considered they are also a satisfactory approximation

to Vinnakota's results.

7. -~~_IP_11_!.ary_ and Concll~S ipn~.

The basis of the design equations permitted in the Canadian

Standard Association 816.1-1974 (Limit States Design) for steel H- and

!-columns subjected to axial load and biaxial bending moment has been

smmrt-'1-rized. A simple modification to these equations was presented

in Eqs. 7 and 8, which enabled these equations to be used for columns

with a wider range of flange width to section depth ratio. It is

recommended for general use.

An investigation was undertaken in the region where S = 1. 0

to determine lvhether there could be any bas is for allowing ~ < 1. 0.

The results are shown in Fig. 9~ and on this basis retention of a

minimum value of e given by ~ ·- 1.0 is reconmended.

 -11

Further interpretation of these results has been recommended

for the proposed revision to the Commentary of AISC Specification (1969).

An up-to-date summary on this subject along Hith design examples is

given in a recent paper by Springfield (1975). Furthermore, the for-

mulae are included in the forthcoming revised CRC Guide (Johnson, 1975),

for the use of designers.

8. A~lmovJlejgmen~-~

This work is a part of the general investigation on 11

Biaxially

Loaded Columns 11 , currently being carried out at the Fritz Engineering

Laboratory, with funds furnished by the National Science Foundation under

Grant Eng 73-08045 AOl (GK-35886) to Lehigh University.

This p<).rticnlar- phase of r.vork v7as initiated through the

Coltm1n R~search Council, and sponsored by the Canadian Steel InduJtries

Ccustruction Council. Mr. J. Springfield was project coordinator.

 -12

' 9. References

1. Canadian Standard S16 .1-1974, "Steel Structures for Buildings--

Limit States Design."

2. "Commentary to the Specification for the Design, Fabrication ar:.d

Erection of Structural Steel for Buildings," AISC, Ne~·l York,
N.Y., 1969.

3. Johnston, B., ed., "Guide to Design Criteria for Metal Compression

Members," Column Research Council, 2nd ed., 1966, John \.Jiley
and Sons, Inc., New York, N.Y., 3rd ed., 1975 (to appear).

lf. Santathadaporn, S. and Chen, W. F., "Analyses of Biaxially Loaded

Steel H-Sections," Journal of the Structural Division, li.SCE,
Vol. 99, No. ST3, Proc. l1 aper 9621, Harch 1973, pp. /f91.-509.

5. "Specifice.tion for the Design, Fabrication and Erection of St1~uctural

Steel for Buildings," AISC, Ne\v York, N.Y., February· 1969.

6. Springfield, J., "Design of Columns Subject to Biaxial Ile.ndings,"

AISC Engineering Journal, Third Quarter, 1975, pp. 73-81.

7. Springfield, J. and Regan, B., 11 Comparison of Test Results vlith

Design Equations for Biaxially Loaded Steel Wide Flange Beam-
Columns, 11 ist.:uetl uncier the auspices of C. D. Carruthers and
\\lallace, Consultants, Ltd., as the result of a joint :Lnvesti-
gation by Canadian Steel Industries Construction Council and
C. D. Carruthers and Wallace, Consultants, Ltd., December 1973.

8. Tebedge, N. and Chen, H. F., "Design Criteria for H-Columns under

Biaxial Loading," Joe1rnal of the Structural Division, ASCE,
Vol. 100, No. ST3, Proc. Paper 10400, March 1974, pp. 579-598.

9. Vin11akota, S., 11 'J'he Influence of Initial Deformations end Residual

Stresses on the Maximum Strength of Biaxially Bent Columns,"
paper presented to Task Group 3 of Column Research Council,
Toronto, May 1975.
 -13

' 10. Nomenclature

=- flange w·id th

D -· section depth
.
'{_, ·- column length

M ,M -· applied bending moment about x- and y-axes, respectively

X y
l1 = maximum moment that can be resisted by the member in
m
strong axis bending in the absence of axial load and

weak axis bending moment

M M - plastic moment about x- and y-axes respectively

px' PY
M M - maximum end moment that can be resisted by the member
ux' ny
in the plane of bending including the axial load, but

in the absence of the other moment

p ·- axial load
)) p - elastic buckling load of the column about the x~ and
ex' ey
y-axes, respectively

P - axial load producing failure in the absence of bending

C!'

moment

- P/Py

r r
x' y
= radius of gyration in x- and y-directions, respectively

= exponent in Eq. (2)

0' = yield stress

y
., Table l Load Range Considered

.{,
-·-----------
V.i8x31 Wl4x38 W21x44
p = P/P
r (B/D=1.0Q2 (B/D-=0. 48_) (B/D=O. 32)

Hinimum t/r X 30 30 5

Minimum t/r, 50 114 32

0.1
Maximurn t/r X 60 50 30

Maximum t/ r y 100 190 190

---------------------·
Minimum .{,II .... J. 10 20 10
X

0. 3 Minimum t/r
_______ 17_ _ _ _ _ _.__________________________
J__ _ _ _..:__ 76 · 63 _

Maximum t/r 60 30 20
X

Haximum t/r 100 114 127

y
·-----· -------·----··------------
1-'I:i.nimum t/r 10 10 8
X

Ninimum -f./r 17 38 50
0 .::>,- -·---~-·

Haximum t/r !i-0 20 12

X

Haximum t/r 69 76 76
y

Hinimum t/r 10 5 5
X

Minimum t/r 17 19 32
0.7 . -· y -----------------------
Haximum t/r 30 107
X

Maximum ~,fr 50 40
y
 ~15

• Table 2 Ha~imum_Strength Tables for Biaxial1:.Y.__!,oaded Beam-Columns

a) Section Wl4:d8 (B/D = 0.48)

(Maximum Values of N /H .)
--,-----------------'--_EL____· · - · - - - - - - -
p M Slenderness Ratio t/r
X X
p M
y px 5 10 20 25 30 40 50

0.1 0.0 0.648 0.282

0.10 0.243
0.15 0.181
0.20 O.st~l 0.092
0.22 0.070
0.40 0.276
0.50 0.105
·-·--·0.53
0.3 0.0
__
.:..._ ,

0.493
0.050
0.179 0.070
0.10 0.168
0.20 0.£1.48 0.110
0.30 0.037
0.40 0.290
0.45 0.250
0.52 0.160
0.56 0.096
-(). s -o:o-------- 0. 6~~0 0.165
---------------·-
0.078
o.lo 0.158
0.15 0.156
0.20 0.590 0.151
0.30 0.105
0.35 0. 06~~
0.40 0.436
0.50 0.333
0.52 0.307
0.56 0.255
0.60 No Solution
·----------~·-
t/r =10. 7
X
-Or X =15
0.7 0.0 0.579 0.369 0.105
0.10 0.339
0.20 0.512 0.287
0.30 0.430 0.207
0.32 0.4.11 0.186
0.34. 0.390 No Solution
------0.36 No Solutjon
--·-·----- --------·---
 -17

Muy P Mux
----=---
Mpy(I-P/Pey)
=I .0 --+
PeR
·-~
Mm ( 1- P/Pex)

(M
Mx-){3 + ::.:..:L)fS = I, 0
(-
Mux Muy

Mx

Fig. 1 A Typical Maximum Strength Interaction Surface for a Particular

Column of Known Length Subject to Biaxial Bending
 -18

'

R
-=60
r:t
My 10
IVIuy

W8 X31

0.6-- ( B;D ~ 1.0)

p= 0.3

( Tebedge 8 Chen )

0.2

0.2 0.6 1.0

Mx
Mux

. Fig. 2 Maximum Strength Interaction Curves for W8x31 at p = 0.3

(From Tebedge and Chen's Results, 1974)
 -19

'

1.0~,~=~-­
,
' ' ......
'' t
'' rx =20
'' 30
Wl4x38
''
0.6 B/0=0.48 ''
p=0.3

0.2

0.2 0.6 1.0

Fig. 3 Maximum Strength Interaction Curves for Wl4x38 at p = 0.3

 -20

"

My
Muy

0.6
W21 X 44

( B/ D = 0.32)
p = 0.3

0.2

0.2 0.6 Mx 1.0

Mux

Fig. 4: Maximum Strength Interaction Curves for W21x44 at p = 0.3

 -21

JL
-- = 10
rx

W8x31
(B/D~ 1.00)
p=0.5
0.6 (Tebedge 8t Chen)
My
Muy

;8=0.4+ p+ 8/D

0.2

0.2 0.6 1.0

Fig. 5 Maximum Strength Interaction Curves for W8x31 at p = 0.5

(From Tebedge and Chen's Results, 1974)
 -22

1.0 ~::::.-~~---
1 ...........
............
......
'' -=10
i .
'' rx
'' 20

Wl4x38 ''
(8/D =0.48) ''
My
0.6 p=0.5 ',
Muy '''
. /' ,
,8=0.4+p+B/D/- \
\
\
0.2 \
\
\
\

0.2 0.6 1.0

Fig. 6 Maximum Strength Interaction Curves for Wl4x38 at p = 0.5

 -23

..
•

W21x44
0.6
(B/D =0.32)
My
p=0.5
Muy .

~=0.4+p+B/D

0.2

0.2 0.6 1.0

Fig. 7 Maximum Strength Interaction Curves for W2lx44 at p = 0.5

 .....

(3 = 0.4 + p + 8/D
2.0 p= 0.7

p=0.5

p=0.3
r p=O.I

~\
11 11
\_ true {3

0 0.2 0.4 0.6 0.8 1.0

B/D

Fig. 8 Comparison. of Observed and Recommended Values of a I

N
~
 -25

. I

I .0 ~ .....

0.8

0.6 -
My
W21 x44
Muy (8/D= 0.32)
0.4- p=O.I

0.2

0.2 0.4 0.6 0.8 1.0

Fig. 9 Testing of Program in Region where S = 1. 0

 0.7~
I

0.5

0.4
X
0.
~
,..._ wax 31
)(
0.3
(8/D = 1.0) ~~
~
i.lrx =30

0.2
p = 0.3 \
\
0.1
\
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
My I Mpy I
N
a-
Fig. 10 Comparison of Results Predicted by Proposed Formula and Vinnakota's Solution. (Solid
Line :i.s Proposed l''or.mula, and Dotted Line i.s Vinnakota's Solution).