Design of Beam-Columns

IRA H O O P E R

THE DESIGN OF BEAM-COLUMNS in the 1963 AISG Specifi- Columns Without Sidesway—The classical derivation
cation 1 recognizes two important features of structural of the Euler formula for the critical load of a concen-
behavior—frame stability and moment amplification d u e trically loaded column applies to a vertical member of
to lateral deflection. uniform cross-section pinned at each end (Fig. 1 ) ; lateral
F r a m e stability m a y be defined as the ability of a displacement is prevented at each end, b u t vertical dis-
structure to resist sidesway when vertical load is applied. placement is permitted at the top. At critical load, the
I t is not usually a problem for simply connected column deflects laterally in a smooth curve between the
structures braced against sidesway. Old buildings have ends or nodes.
heavy masonry walls that provide bracing although no All other column types are referred to this classical
calculations were made for the bracing effect. M o d e r n case, so for convenience its actual length is defined as its
buildings can no longer afford the cost or wasted space of effective length, or K = 1.0. Such columns occur in ac-
massive walls. Also, rigidly connected frames have be- tual practice. T h e A I S C rules permit the use of K = 1.0
come common. Rigidly connected frames offer several for columns in frames with sidesway prevented by vertical
advantages: reduced beam deflections, greater economy bracing, by shear walls, or by attachment to other braced
and elimination of floor cracks. structures. T h e rules would be theoretically correct for
W h e n rigidly connected frames are not braced against frames with hinged connections between beams a n d col-
sidesway, the columns and beams must provide the stiff- umns; the rules are conservative for rigidly connected
ness needed to maintain frame stability in addition to braced frames.
their load resisting function. In comparing other column types with the classic
T h e interaction formula in use prior to the 1963 AISG case it is necessary to find the distance between nodes to
Specification was intended to apply to members with establish the effective length. I n some cases it m a y be
pinned ends in braced frames. I t did not account for the difficult to locate two nodes, a n d another concept may be
effects of lateral deflections in the columns. helpful. Each half of the classic column m a y be consid-
T h e new formulas include the effects of restrained ered to be a "flagpole" (Fig. 2). A "flagpole" is fixed in
ends a n d of lateral deflection. T h e range of design in-
cludes rigidly connected frames with or without sidesway
bracing. T h e new sections of the A I S C Specification for
beam-columns deal with (a) effective length and (b) in-
teraction formulas (6), (7a) a n d (7b). i
EFFECTIVE L E N G T H

Effective length is a geometrical property of the frame

incorporating the column being considered. Effective
length does not depend upon the application of moments KL
to the column, nor will it be influenced by the presence
or absence of lateral loads applied to the frame. Effective
length is a useful comparison between actual columns
in a frame a n d the isolated column of theory.
£

Ira Hooper is Associate, Seelye, Stevenson, Value and Knecht, New

York, N. Y., and is a Professional Member of AISC.
t
Fig, 7. Classic case—2 ends pinned, no sidesway. K = 1.0

A P R I L / 1967
 L ' i

L
4
\ 1
L

1 ,
L
4

4 1
1
Fig. 2. "Flagpole". K = 2.0 V777/
vertical position at its base and supports the critical ver- t
tical load at its free end. T w o "flagpoles" with joined Fig. 3. Fixed ends, no end rotation, no sidesway. K = 0.5
bases form the classic column; therefore, the effective
length of each ''flagpole" is twice its actual length.
Figure 3 shows a column braced against sidesway, but
with ends fully restrained against rotation. At critical
load, the deflection curve includes four "flagpoles" of
equal length and two inflections at the quarter points.
T h e critical load for the restrained column is equal to the
critical load for a pinned-ended column of half the length.
Therefore, the effective length of the restrained column is
half the actual length. Such columns rarely occur in ac-
tual practice, since it is not possible to provide full re-
straint against end rotation.
T h e two cases with ends pinned and ends fully re-
strained are the limiting conditions for columns with
sidesway prevented, and the it-values cannot exceed 1.0
nor be less than 0.5. It is conservative to use K equal to
1.0 for all columns with sidesway prevented; for columns
of usual proportions the extra weight will be small. For
Fig. 4. Fixed ends, free to sway. K = 7.0
tall slender columns with rigid beam connections, the
true K-values will result in worthwhile savings.

Columns With Sidesway—When sidesway is not pre-

vented, it is possible to set only a lower bound to the
effective length of a column. This will occur when both
ends are fully restrained against rotation (Fig. 4). U n d e r
critical load, the column will assume a reverse curve
consisting of two "flagpoles" with tips joined at mid-
height. T h e effective length is, therefore, equal to the
actual unsupported length.
Full restraint against rotation is a convenient concept
that can only be approached in practice (Fig. 5). T h e
beams will permit rotation at the ends of a real column,
so the curves above and below the inflection point will
be only parts of a "flagpole." T h e curves must be ex-
tended beyond the actual column ends to reach the points
where the tangents would become vertical. T h e effec- Rotation
tive length will always be larger than the actual length
and will increase as the restraint against rotation is re-
duced. Fig. 5. Semi-fixed ends, free to sway. K > 7.0

42

A I S C E N G I N E E R I NG JOURNAL
Design Methods for Efrective L e n g t h — T w o methods T h e second method uses an alignment chart based
of evaluating effective length are offered by the A I S C on stiffnesses of a column and of the beams rigidly con-
Specification. Table CI.8.2 in the Commentary 2 (Fig. 6) nected to each end of the column. T h e A I S C Com-
shows diagrams for all the cases discussed above with mentary shows only one alignment chart for side-sway
theoretical and recommended values of K. This table uninhibited, but another chart for side-sway prevented
may be used for preliminary selection of K, but it is better is available in the Column Research Council Guide 3 (Fig.
to use the second method for final iC-values, especially 7). T h e charts are intended to be used for frames with
when sidesway is not prevented by bracing. rectangular panels.
T o enter the alignment charts, it is necessary to cal-
culate the stiffness ratios at the top and at the bottom of
(a) (b)

u (c)

77771

T
(d)
t
(e)

1 I
? P
(f)
the column under investigation (Fig. 8). T h e stiffness
ratio is equal to the sum of the column stiffnesses at a
i joint, divided by the sum of the restraining girder stiff-
Buckled shape of column I
is shown by dashed line nesses at the joint:

2IjLc
G=
'////) ////// ?IS/LS
Theoretical K value 0.5 0.7 1.0 1.0 2.0
In general, one or two columns will be present at a
Recommended design
value when ideal condi- 0.65 0.80 1.2 1.0 2.10 2 0 joint, restrained by one or two girders. For column ends
tions are approximated bearing on a footing but not rigidly connected, the
Rotation fixed and translation fixed stiffness ratio is theoretically infinite, but may be taken
Rotation free and translation fixed equal to 10 unless actually designed as a true friction-
End condition code
Rotation fixed and translation free free pin. If the column is rigidly attached to the footing,
Rotation free and translation free the stiffness ratio may be taken as 1.0.
The appropriate chart is entered with the stiffness
Fig. 6. Table C1.8.2 of Commentary on AISC Specification ratios at top and bottom to find the K-value.

oAL K GB G„ K GB

00 J
50.0^ -1.0 rso.o 100.0 -
~1
2mF20.0
r— OO

- 100.0
10.0- rio.o 50.0 -
f 10.0
r 50.0
5.0- ~- 5.0 30.0- - 30.0
4.0- - 5.0
3.0- -0.9 - 3.0 2.00- - 20.0
- 4.0
2.0- - 2.0
10.0- - 3.0 - 10.0
-0.8 8.0- - 8.0
1.0- - 1.0 7.0- - 7.0
- 0.8 6.0- - 6.0
0.8-
0.7- - 0.7 5.0- - 5.0
0.6- - 0.6 4.0- - 2.0 - 4.0
-0.7
0.5- - 0.5
3.0- - 3.0
0.4- - 0.4

0.3- - 0.3 2.0- - 2.0

- 1.5 -
-0.6
0.2- - 0.2
1.0- - 1.0
0.1- - 0.1

0- -0.5 - 0 0- - 1.0 0
(a) (b)
Sidesway prevented Sidesway not prevented

Fig. 7. Alignment charts for effective length of column in continuous frames (from Column Research Council Guide*)

43

A P R I L / 1967
 T h e alignment chart is not a direct design method;
it analyzes assumed sizes. T h e girder sizes are easily
estimated (Fig. 9) by assuming the maximum moment
will not exceed the fixed end moment due to uniformly
distributed load. T h e assumption covers a wide range of
end-fixities.
Estimates of column sizes are more difficult. T o find
it, one must choose a column section for strength, but
the choice depends on the effective length. This vicious
circle can be broken by the following procedure:

1. Assume a trial it-value. Design aids are described £k col

k = G=
below. L I k gird
2. Choose a column section for strength based on the
k
trial iC-value. l + k2
3. Calculate K and compare with procedure 1 above; 'top' k
3+k4
adjust the column size if required.
k
l+k7
Design Aids for Trial K-Values—Two design aids are 'bof k
5 + k6
available for assistance in assuming the trial K-value.
Table CI.8.2 of the AISC Commentary has already been Figure 8
mentioned. A more detailed table is included in the
Bethlehem Steel Corp. publication, Beam-Column Tables
for Structural Shapes4 (Fig. 10).
T h e best design aid for estimating ^-values is knowl- wL= W
edge of previous values used elsewhere in a structure. M
(
T h e greatest difficulty lies in guessing the values for the
first beam-column. In multi-story frames, the K-values
vary little between stories. After the calculations have
been m a d e for the top story, close estimates can be made 100% 0 . 0 8 3 wlf
for the succeeding stories below. Fixed 0.042 wLT
In similar fashion, calculations for the first few bays
of a large single-story building give excellent clues for
estimating the it-values for the remaining bays.
T h e alignment charts were originally developed for 50% 0 . 0 4 2 wl?
use with rigidly-connected frames. It is possible to use Fixed
0 . 0 8 3 wlf
the charts with partially-rigid connections. A method
has been developed to reduce the girder stiffness to
Figure 9
account for the moment-rotation characteristics of the
connections. 5 T h e method has been elaborated and
practical design information has been made available.
T h e method is particularly useful for one- or two-story Connections
Single
Multistory
Bottom Story
Story of Multistory
buildings with ordinary framed connections and without Case
Bracing
Condition
Ky &
sway bracing; in a low building, the column effective Top Bottom GB
Kx
Ky Kx Ky Kx
length with framed connections has been shown to be
A Braced NR NR 1.00 1.00 1.00 1.00 1.00
only slightly greater than with rigid connections. 6 B Braced NR R 1.00 .87 .95 .95 .87 _£Z
C Braced R NR 10.0 .80 .95 .95 .90 .90
D Braced R R 1.00 .75 .90 .90 .80 .80
INTERACTION FORMULAS
E Unbraced R R 1.00 1.25 1.80 2.25 1.60 1.60

After the effective length has been established or as- F Unbraced R NR 10.0 1.80 - - 2.0 2.50

sumed, beam-column design is controlled by Formulas Legend NR —Non-rigid connection free to rotate under load (i.e., simply supported girders,
(6), (7a) and (7b) of the AISC Specification. hinged column connections, etc.)
R —Rigid connection capable of holding the original angle between members virtually
unchanged.
W h e n / a / F f l < 0.15: Underlined items are final values and need not be modified for the actual member stiffnesses.

fa h = 1 Formula (6) Fig. 70. Suggested first trial values of K (from Bethlehem Beam-
Fa Column Tables4)

44

AISC ENGINEERING JOURNAL

When/B/Fa>0.15: W h e n the axial load exceeds 15 percent of the allow-
able axial load capacity, Formula (7a) or (7b) may
/. Cmfb govern the design.
= 1 Formula (7 a)
+ (1 -fJF.')Fh T h e middle of a column may deflect horizontally
when end-moments are applied, creating eccentricity
fa , fb _ ,
Formula (7 b) of the compressive load and additional moment (Fig.
0.6F,
11). Such amplified moment is accounted for by
Formula (6) is the familiar unity-ratio formula of pre- Formula (7a) and it is understandable that Formula (7a)
vious A I S C rules, but now applies only when the axial should be affected by the column slenderness.
load is less than 15 percent of the member's allowable When the end moments are applied in the same
axial load capacity. sense (Fig. 12), the beam-column will be bent into a
reverse curve with smaller moments due to deflection.
U n d e r these conditions, Formula (7a) may not govern
and Formula (7b) will tend to establish the design.
Since no eccentric effects due to deflection can occur at
the ends, Formula (7b) is independent of the column
slenderness.

Interaction Curves—Formulas (7a) and (7b) are modi-

fications of Formula (6). For understanding of the formu-
las it will be helpful to plot them as graphs. I n Fig. 13,
the vertical coordinate is the ratio of computed axial
stresses to 0.6Fy and the horizontal coordinate is the
ratio of actual to allowable bending stresses.
With these coordinates, Formula (7b) is a straight
line between values + 1 . 0 on the horizontal axis and
+ 1 . 0 on the vertical axis.
Formula (6) is a series of straight lines from various
intercepts on the vertical axis converging to + 1 . 0 on the
horizontal axis. T h e F-axis intercepts are determined by
the column slenderness ratio; note that Formula (6) for
M
the case of a column of zero length is identical with
total = M + p
y Formula (7b).

(7^) All cases

(6) Kl/r*0

Figure 72 Fig. 73. Curves for Formulas (6) and (7b) at indicated Kl/r values

45

A P R I L / 1967
 Figure 14 shows the graphs for Formula (7a). Formula
(7 a) is similar to formula (6) except that the second term
(6)and(7 Q )
is multiplied by a fraction. T h e denominator of the
entire fraction is always less t h a n 1; it accounts for the
added moment due to lateral deflection of the beam- Cm-I.O
column. T h e reciprocal of the denominator is called
the "amplification factor55 and gives correct results when
the applied moment neglecting lateral deflection is
uniform for the full height of the column. T h e graphical
significance of the amplification factor is shown in Fig.
14. For a given value of slenderness ratio and axial stress
ratio, the distances a and b represent the bending stress
ratios in the second terms of Formulas (6) and (7a) when
Cm = 1. T h e ratio a/b is equal to the amplification factor:

1
1 - fa/F.'
Also,
Formula (7 a )
Formula ( 6 )
-(-£)• Fig. 74. Curves for Formulas (6) and (7a) at indicated Kl/r values

and
C m = 0.6 + 0.4 r r - ^ 0 . 4

T h e A I S C Manual 7 includes tables for calculating the

i M •M

amplification factor. T h e Bethlehem Beam-Column

Tables 4 yield quick results when calculating the ampli-
fication factor.
I n Fig. 14, the value of the numerator, Cm, is 1.0. Any
M
other arrangement of moments causes less lateral deflec-
tion, resulting in less amplified moment, so that the value
'3
of Cm becomes less than 1.0; hence, it is called the ''reduc- Cm=I.O 0.6 0.4 0.4
tion factor 55 . T h e A I S C Specification gives rules for Fig. 75. Reduction factor, Cm, for columns in braced frames
evaluating the reduction factor. For beam-columns in
braced frames, Cm may vary between 0.4 and 1.0 (Fig. Formula (7b) for most of the range, so Formula (7b)
15); for unbraced frames, Cm is specified as 0.85. governs as would be expected. Note that slender columns
W h e n the Cm-value is 1.0, the intercepts on the two have a region where Formula (7a) will govern; in this re-
axes for Formula (7a) are the same as for Formula (6). gion the axial stress is high and bending stress is low.
Between the intercepts for any given column slenderness, An important value of Cm is 0.85, which is specified by
the amplification factor depresses the curve for Formula A I S C to be used with unbraced frames. Figure 17 shows
(7a) below the straight line for Formula (6). that Formula (7a) will govern for slenderness ratios over
W h e n the C m -value is less t h a n 1.0 (Fig. 16), the inter- 80, except for a small region of very high moment. For
cepts on the vertical axis are the same as for Formula (6), slenderness ratios less than 60, each of the three formulas
but the point of convergence on the horizontal axis shifts controls a part of the range. T h e combined curve is very
to the right to a value equal to the reciprocal of Cm. complicated, but the straight line for Formula (6) is a
T h e shift causes the curves for Formula (7 a) to cross the close approximation, generally within 5 percent.
line for Formula (7 b ) ; where this occurs, Formula (7 a) no Columns with slenderness ratio less t h a n 60 occur fre-
longer governs and (7b) will determine the design. quently. For practical purposes, it is suggested that in de-
T h e lowest value of Cm permitted by AISC rules is signing A36 columns with sidesway permitted, Formula
0.4; this value applies to beam-columns in braced frames (6) will give satisfactory results when the slenderness ratio
with moments applied at the ends causing a reverse is less than 60, and that Formula (7a) will govern when
curve. I n Fig. 16 the curves for Formula (7a) lie above the ratio is above 80.

46

AISC ENGINEERING JOURNAL

 — — — — Formula (7 D )
Formula (7 a ) for Cm= 1.0
— . — Formula (7 a ) for C m = 0.4
40
0.
\
80 40
\
\ \
\ . ' 8 0 ..
120
~I20

L^v
(2.5)

Fig. 16. Effect of reduced Cm value on Formula (7a)

Formula (7b) for all cases

Formulas (6) and (7a) for i l l = 0

Formula (7a) for indicated values of

Kl
—

Formula (6) for indicated values of —

0.6 R

Fig, 77. Interaction curves for Cm — 0.85, Fy — 36 ksi

47

A P R I L / 1967
 GENERAL INTERACTION CHART Dividing the numerator and the denominator by 0.6Fy:
It is possible to construct a design chart of the inter- 3.4/0.6F^
action formulas that includes all values of Cm (Fig. 18). Column weight = • „ ^ ^ X P
fa/0.6Fy
T h e coordinates are the same as for the previous interac-
tion charts, except that the horizontal coordinate is multi- For Fv — 36 ksi:
plied by Cm.
0.157
I n Fig. 18, the curves for Formula (7a) remain the Column weight (lbs per ft) = X P (kips)
same as in Fig. 14 with Cm equal to 1.0. Formula (7b) is fJolF,
now represented by a series of straight lines intersecting When the slope, s, is greater than 1.0, it is convenient to
the vertical axis at 1.0 and the horizontal axis at values use the reciprocal for interpolation; these values are in-
equal to Cm. Formula (6) can be plotted as a series of lines dicated on the general interaction chart.
intersecting the 7-axis at the usual points, converging to In addition to giving a close estimate of column
an X-axis intercept at the Cm-value. weight, the general chart shows which formula governs.
T h e slope of a line through the origin is: Furthermore, the chart shows the effect of changing any
variable, which is often of interest to the designer b u t is
not evident from the formulas or from tables.
Cmfb/Fb
Substituting
Use of the General Interaction Chart—
P __ M M A
fa = B = - 1. K n o w n items: Fy, P, Mtop, Mbot, sway bracing
; h
A ~ ~S S condition, Cm, effective length.
2. Choose overall column dimensions; find cor-
= — ( — ) responding B and r in the table on the chart.
CmeB\0.6FyJ 3. Evaluate Kl/r; locate the corresponding curve for
Formula (7 a).
This expression holds the key to rapid trial design, since 4. Draw Formula (7b) as straight line between 1.0 on
all factors can be readily evaluated: the vertical axis and Cm on the horizontal axis.
5. Evaluate CmeB or \/CmeB; draw the corresponding
1. Cm is determined by A I S C rules.
line through the origin; find points of intersection
2. e is the eccentricity obtained by dividing the known
with the curves for Formulas (7 a) and (7 b).
moment by the known axial load.
6. Using the vertical coordinate of the lower intersec-
3. B is the bending factor listed in the AISC M a n u a l .
tion point, calculate the estimated section weight
I t varies little for column sections of equal depth.
and choose a trial column size.
4. T h e quotient Fb/0.6Fy can be estimated equal to
7. Check the trial column size. A check is necessary be-
1.0 for almost all cases. Examination of the column
cause average values of B and r were used. T h e
design tables in the A I S C M a n u a l will show that
check will be simple, since the governing formula
most members with slenderness ratio less than 120
has been determined.
have an allowable bending stress of 0.6Fy; columns
with greater slenderness ratios are neither usual nor
Design Aids for Checking the Trial Column Size—The
economical. Compact sections with slenderness ra-
AISC M a n u a l can be used to find the ratios of axial
tios less than 50 m a y use 0.667%, in which case the
stresses in the interaction formulas, b u t it requires much
use of 0.6 Fy is slightly conservative.
calculation to find the ratio of bending stresses and
T h e intersection of the 5-line with the appropriate amplification.
Formulas (7a) and (7b) represent solutions. Values of B T h e Beam-Column Tables published by Bethlehem
and r must be known in order to find the intersections. Steel Corporation, 3 available for Fy values of 36 and 50
Fortunately, B a n d r are almost constant for steel mem- ksi, are very helpful for all parts of the final check. T h e
bers of the same dimension. Recommended average val- Tables show data for every rolled section. Each pair of
ues are given in the table on the general interaction facing pages deals with the same sections; the left page
chart (Fig. 18). gives the allowable axial load for all practical lengths as
T h e vertical coordinate of the point of intersection is well as the Euler load, P / , to be used in the amplification
used to find the required weight per foot of the column factor. T h e right page shows allowable moments for all
section: practical lengths, including allowances for flanges un-
braced laterally.
3.4P
Column weight = 3.4 X A = A quick b u t less accurate check of the results from
7T the general interaction chart can be made by inspecting

AISC ENGINEERING JOURNAL

d B
x
r
x b B r
y y
nom av av nom av av
16 0.47 4.0
14 0.185 6.5 141 0.52 3.7
12 0.22 5.5 12 0.64 3.0
10 0.265 4.5 10 0.77 2.5
8 0.33 3.5 8 1.00 2.0

ei 1.25 1.6
0.157
COL.WT/FT = P
f a / 0 . 6 F y X kip

i-c pBra6Fy'
_ c
s meB ^ ^ T " .
(0.5) (1.0)

Fig. 18. General column interaction chart, Fy = 36 ksi

49

APRIL/ 1967
the difference between the average values assumed and Kl
For — = 80: ^80
the actual values of the steel section selected, then adjust- r V e80/
ing the size if required.
Kl
For = 60: ^60
B e n d i n g About the Non-critical Axis—The discussion \re&0/
to this point has applied to columns being bent about the 149,000,000
axis critical for concentric load. The general interaction F' = (as defined in A I S C Specification)
(Kl/r)2
chart can be used for size estimates when bending occurs
in the non-critical direction. It is also possible to use the '60V
chart to estimate sections with bending in both directions. ^80 .80/
Formula (7b) is not affected by bending about the
non-critical axis, but Formula (7a) requires adjustment. These equations show that the ^-distances measured
To understand the effect on Formula (7a), visualize a col- horizontally at a given value offa/0.6Fy will vary directly
u m n with the same slenderness ratio in both directions, as the square of the slenderness ratio of the non-critical
resisting moment in only one direction. T h e interaction axis.
curve for Formula (7a) in the general chart will apply For each value of critical slenderness ratio, a "fam-
for the column described. T h e lowest curve in Fig. 19 ily" of curves similar to Fig. 19 could be drawn for values
represents the interaction equation for such a column. of slenderness ratio in the non-critical direction. Each
"family" of curves would have the same intercepts on
Allow the slenderness ratio resisting the moment to
the chart axes and would be closely spaced for low values
decrease; the amplification due to lateral deflection will
of non-critical slenderness so that straight-line interpola-
also decrease and the curve for Formula (7a) in Fig. 19
tion between the limiting curves of a "family" will be
will tend to straighten. When the slenderness ratio has re-
conservative. For legibility, the additional curves have
duced to zero, no amplification will take place and the in-
not been drawn on the general chart. T h e curves in Fig.
teraction curve will be a straight line between the intersec-
18 are the lower "family" limits for the indicated slender-
tion on the vertical axis and 1.0 on the horizontal axis.
ness ratio; the upper "family" limits are easily obtained
The intersection on the vertical axis of the chart is deter-
by placing a straightedge between the intercepts on the
mined by the slenderness ratio in the direction critical for
two axes.
concentric load; it is not affected by changes in slender-
ness ratio for the other direction. Solutions of Formula (7a) for bending about the non-
Applying to Fig. 19 the expressions derived in explain- critical axis are obtained by finding the intersection of
the 5-line through the origin with the appropriate inter-
ing the graphical significance of the amplification factor
polated member curve of a "family" (Fig. 20).
in Fig. 14:

(7a) Non-crit. — • l/S ^

(7a) Crit. &1 -, 1.0

"80 / 80. 0
80 20 fq ^(Upper family 1
/ / '
80 40 0.6 R, / limit)
\y / / 1

0.6 R,
/y//— so, 60 r-(lo) Non-crit.

%>yyC/ so, 80 X ILL

/\^ r
^s^y
(7 b)

/ ( 7 a ) Crit. * r
/ (Lower family limit) ^""^^W.
c I.O
'm mfp
F
b
Figure 19 Figure 20

50

AISC E N G I N E E R I NG JOURNAL
Wind and Seismic Stresses—The AISC Specification When fa and fb are reduced one-quarter in Formula
allows a 33 percent increase in allowable stresses for mem- (7a), the two terms are not reduced in the same propor-
bers carrying wind or seismic loads, acting alone or in tion, since fa appears in the denominator of the second
combination with dead and live loads. term. T h e graphical significance oijjFj was explained
Designers usually reduce the combined loads by one- for Fig. 14; the reduction of/a by one-quarter will reduce
quarter and design at the usual allowable stresses. This the ^-distance also by one-quarter.
procedure applies directly to general chart solutions for Figure 21 illustrates the recommended procedure for
Formula (7b); fa and fb and both terms of the formula are chart solution of Formula (7a) with wind loads and bend-
reduced by the same proportion. ing about the axis critical for concentric load:

1. Reduce the axial load, P, and the bending moment,

M, by 25 percent.
2. Plot the j-line.
3. Locate the curve for Formula (7 a).
4. Find the point on the s-line where cw equals 0.75^
(or interpolate 0.75z as shown in Fig. 21 as a good
approximation).
5. Calculate the column weight and check.

Figure 22 illustrates the same procedure for solving

Formula (7a) when bending occurs about the non-critical
axis.
Biaxial Bending—The general interaction chart can be
of greatest help in estimating the size of a beam-column
with bending in both directions.
T o understand the use of the chart for Formula (7a),
visualize a column with equal slenderness ratios about
both axes, supporting an axial load and moments about
both axes. T h e eccentricities can be computed and the
Cm*b
.y-lines through the origin can be plotted on the chart
(Fig. 23). T h e bending in each direction must use only a
portion of the strength available for single axis bending;
Figure 27
the solution must lie below points g and h. T h e solution
may be found by adding \/sx to \/sv to obtain the "slope-
i/s reciprocal", \/sx + l/sy, of a line through the origin
that intersects the curve for Formula (7a) at point j .
T h e proof follows:

Upper family Ja , ^mxjbx ^"myjby

limit Ftt (1 - fJFex')Fbx (1 - fa/Fev')Fbi
0.6 F„ Formula (7 a)
(7a) Non-crii
K_l but
r
jXxi<x ^-y^y

therefore,

ex — " ev " e

C m fb Factoring,
r
b
Figure 22 H ^mxjbx ^myjby

/I -fa/Fe'

51

A P R I L / 1967
From Fig. 23,

^mxflbx , j ^myjby ,

F,bx bx and
J^by
= bv l/Sx+l/Sy

b — bx + by
Substituting,

— + (b) - = 1
fa/F.'
T h e last equation is Formula (7a) for single-axis bend-
ing, using terms explained in Fig. 14; it shows that biaxial
bending of a column with the same slenderness in both
directions can be reduced to an equivalent form of Form-
ula (7a) for single-axis bending and that point j repre-
sents the solution.
T h e assumption of equal slenderness in two directions
is not intended for practical use. Columns usually have
different slenderness ratios for the two directions; the use
of the higher ratio for both directions will yield too con-
servative an estimate.
T h e recommended method is to add the two "slope-
reciprocals" to find the line through the origin as before,
but to interpolate as shown in Fig. 24. It has been found
that straight-line interpolation yields good estimates of
the size required by Formula (7a).
Formula (7b) for biaxial bending is independent of
slenderness ratios. In Fig. 25, the alines have been shown
for both axes and for the sum. T h e solution lies between

l/S X + l/Sy l/S X + l/Sy

c
mfb c
mfb

Figure 23 Figure 25

52

AISC ENGINEERI NG JOURNAL

points m and n ; point m would give too conservative a Then,
solution; point n would be unconservative. When m and
n are close together, interpolation can be done by inspec- U:
tion. When m and n are not close together, the following ax (1 - fa/0.6Fy)Fbx
procedure should be used when Cmy > Cmx (Fig. 26): and
1. Find the intersection point p of the sx-line with the
graph of Formula (7b) for the X-axis.
/.by
av (1 -fa/0.6Fy)Ftby
2. Find the horizontal distance q' between p and the
graph of Formula (7 b) for the T-axis.
3. Lay off distance q' to the right of point u ; draw line
vo. Intersection of line vo with the graph of Formula Therefore,
(7b) for the 7-axis indicates the solution at point t.
1
When Cmx > Cmy the same procedure may be used for + -*
interpolation except the X a n d Y terms are interchanged.
It is now necessary to find a geometrical construction
T h e proof of the construction follows:
to locate the horizontal line at the appropriate level of
I fax _|_ fan fa/0.6Fy to satisfy the equation above:
= 1 Formula (7b)
0.6F, by
&X = T T " \ay)
By transposing and dividing, ^my

fax , fay Substituting in the equation above,

= 1
(1 - fa/0.6Fy)Fbx (1 - fa/0.6Fy)Fby
\(-"my/L>mx)\bx)
= l
In Fig. 25, +
, ^mxjbx , j ^mytby ^"m

ox = — — and by = —— *y - W~ Q>x) + v
b = Cp - l) (*.) + b* + b„
Fbx Fby ^m.x

At point p in Fig. 26, a convenient relationship exists:

1= and av = Cm
^ 0 - oik) k " oik) q = bx I — I - bx = I 1 16,
\^mx/ S^mx /
l/S x +l/Sy
If distance q' is drawn to the right of point u and line
vo is drawn, then at point t, by proportionality,

\^mx I
a = q + (bx + by)

1 ) bx + bx + by = ay

which indicates the solution.

In using the chart for biaxial bending, inspection of
the intersection points of the (\/sx + l/j-J-line with the
graphs for Formulas (7a) and (7b) will usually indicate
which formula governs the design.

DESIGN EXAMPLES
Six design examples are provided in the Appendix to this
paper. These are arranged in three groups:

Examples 1 and 2—Bending about the axis critical for

concentric load
Examples 3 and 4—Bending about the axis not critical
for concentric load
Figure 26 Examples 5 and 6—Biaxial bending

53

A P R I L / 1967
 ACKNOWLEDGMENTS 2. Commentary on the Specification for the Design, Fabrica-
tion and Erection of Structural Steel for Buildings, American
This paper contains material originally presented by the
Institute of Steel Construction, New York, N. Y., 7963.
author to the Steel Structures Symposium at the Uni-
versity of Illinois, U r b a n a Campus, October 25-26, 3. Guide to Design Criteria for Metal Compression Members,
1966. T h e symposium was sponsored by the Mississippi Column Research Council of the Engineering Foundation, New York,
N. Y.
Valley Structural Steel Company, the American Insti-
tute of Steel Construction and the University of Illinois. 4. Beam-Column Tables for Structural Shapes, Bethlehem Steel
Corporation, Bethlehem, Pa.
The work leading to this paper was sponsored by
Bethlehem Steel Corporation. The author wishes to thank 5. Tall, Lambert, Editor-in-Chief Structural Steel Design,
Dr. William C. Hansell and M r . William A. Milek for Ronald Press, New York, N. Y., 7964, p. 720.
their review and helpful suggestions.
6. De Falco, Fred and Marino, Frank J. Column Stability in
REFERENCES Type 2 Construction, AISC Engineering Journal, Vol. 3, No. 2,
Apr., 7966.
1. Specification for the Design, Fabrication and Erection of
Structural Steel for Buildings, American Institute of Steel Con- 7. Manual of Steel Construction, American Institute of Steel Con-
struction, New York, N. Y., 7963. struction, New York, N. Y., 7963.

APPENDIX

Example 1—Bottom story of multi-story building.

f(0.Z4) 5
1.0 (1.0
Y X k£/r(7a)
-40
42 k-ft -55
Nss( -60
0.65
Fy - 36 ksi
fa \s-C7b)
P = 395 kip 10 ft 10 ft 0.6Fy
USE I2 1AF

>
(6) 1
BRACED UNBRACED CO)
k = I.O k=2.5 Cm fb 0.85 1.0
Fb
Given:

Item Unit X Remarks

ksi
P kips 395
Size 12\AF
Sway Br. Unbr
L ft 10 10
K 1.0 2.5 (Fig. 10, Case F for Z-axis)
Mtop k-ft 0 42
Mhot k-ft 0 0

54

AISC E N G I N E E R I NG J O U R N A L
Solution:
3.0 5.5 Fig. 18
Kl/r (40) 55 X-axis governs
ksi 36 Assumed
0.85 Unbraced
e in. 1.28 Mm X 12/P = 42 X 12/395
•Bav in." 0.22 Fig. 18
0.24 CmeBav(0.6Fy/Fb) = 0. 85 X 1.28 X 0.22 X (22/22)
A/o. 6FV 0.65 Fig. 18, chart intersection (7a)
Col. wt. lbs/ft 95.5 0.157P/(fa/0.6Fy) = 0.157 X 395/0.65

Try 12V\F99: Chart (Fig. 18) shows Formula (7a) governs.

Item Actual Assumed Remarks
rx 5.43 5.5 A I S C M a n u a l p . 1-15
Bx 0.216 0.22 A I S C M a n u a l p. 3-20
Fb 24 22 Compact, L < Lc; A I S C M a n u a l p. 3-20
Kl/rx 55 55 —

\/s = CmeB(0.6F„/Fb) = 0.85 X 1.28 X 0.216 X (22/24) = 0 . 2 2 (was 0.24)

fa/0.6Fyfrom Fig. 18 = 0.67 (was 0.65)
This is a small, conservative change; therefore, continue check for 12\AF99.

Check 12V\F99 using AISC Manual, Formula (7a):

Item Unit Value Remarks
A in.2 29.09 —
rx in. 5.43 —
Sx in.3 134.7 —
Kl/rx — 55 —
Fa ksi 17.9 Table 1-36, p. 5-68
Fe' ksi 49.3 Table 2-36, p. 5-69
Fb ksi 24
fa ksi 13.55 P/A = 395/29.09
h ksi 3.74 Mtop/S = (42 X 12)/134.7
fa t"mjb
< 1.0 Formula (7a)
+
13.55
(-£) 0.85 X 3.74
= 0.76 + 0.18 = 0.94 < 1.0 OK
17.9 + (1 - 13.55/49.3) X 24

Check 12\AF99 using Bethlehem Beam-Column Tables, Formula (7a):

rx/rv — 1.76 p. 106 (1.76 < K = 2.1
Eff. KyLy ft 14.2 2.5 X 10/1.76
Pa kips 520 p. 106, for KyLy = 14.2
Pe' kips 1,420 p. 106, for KXLX = 25
Ma k-ft 267 p. 107, for L = 10

CmM
< 1.0 (Modified Formula (7a), see p. 5)
+ (1 - P/P/)Ma
395 0.85 X 42
= 0.76 + 0.185 = 0.945 < 1.0 OK
520 + (1 - 395/1,420) X 267

55
APRIL/ 1967
Example 2—Multi-story building; weak axis braced at mid-height.

120 k-ft fa
10 ft
Fy = 36 ksi
P = 300 kcp < zo
USE 12 VsF 10 ft

BRACED BRACED
ksto k=i.o
?b

Given:
Item Unit y X Remarks
Fy ksi 36 —
P kips 300 —
Size — 12V\F —
Sway — Br. Br, —
L ft 10 20 —
K — 1.0 1.0 —
Mtov k-ft 0 + 120 Clockwise
Mhot k-ft 0 + 120 Clockwise

Solution:
rav in. 3.0 5.5 Fig. 18
Kl/r — (40) 44 X-axis governs
^m — — 0.40 0.6 - 0.4 (120/120) = 0.2 (useO.
e in. — 4.8 120 X 12/300
•t>av in." 1 — 0.22 Fig. 18
1A — — 0.42 0.40 X 4.8 X 0.22
fa/0.6Fy — — 0.48 Fig. 18, chart intersection (7b)
Col. wt. lbs/ft — 98 0.157 X 300/0.48

56

AISC EN G I N E E R I N G JOURNAL
Try 12V\F99:
Chart (Fig. 18) shows Formula (7b) governs.
Item Actual Assumed Remarks
Bx 0.216 0.22 AISC Manual p. 3-20
Fb 22 22 Lc < 20 < Lu; AISC Manual p. 3-20
rx — — Not required for (7b)
Close agreement between actual and assumed. Continue check of 12Wr99.

Check 12\AF99 using AISC Manual, Formula (7b):

Item Unit Value Remarks
A in.2 29.09 —
Sx in. 3 134.7 —
fa ksi 10.3 P/A = 300/29.09
fb ksi 10.7 M/S = (120 X 12)/134.7
Fb ksi 22 —

„ ° + — < 1.0 Formula (7b)

0.6Fy Fb ~
10.3 10.7
— - + — = 0.47 + 0.49 = 0.96 < 1.0 OK

Check 12\AF99 using Bethlehem Beam-Column Tables, Formula (7b):

Po kips 628 p. 106, for KyLv = 0
Ma k-ft 242 p. 107, for KyLv = 20

P M
1 < 1.0 (Modified Formula (7b), see p. 5)
Po Ma
300 120
1 = 0.48 + 0.49 = 0.97 < 1.0 OK
62.8 242

57
 Example 3—Multi-story building; column offset in <¥-direction from stories above and below.

(0.96)-)
y X 1.0

\ . C7t>)x

i 1 80 k-ft / ki/rK
(la)
Fy = 36 ksi , 60,0
P = 265 kip
USE 10 ¥F
12 ft 12 \ft
0.43
\\X/ r~ 60,32
y- 60,60

1 1 30 k-ft
•
\p ;—
(6)
BRACED BRACED
k'/.O k'1.0 1.0

Given:

Item Unit y X Remarks

Fy ksi 36 —
P kip 265 —
Size — 10V\F —
Sway — Br. Br. —
L ft 12 12 —
K — 1.0 1.0 —
Mtov k-ft 0 80 Clockwise
Mhot k-ft 0 80 Counterclockwise

Solution:
r
av in. 2.5 4.5 Fig. 18
Kl/r — 58 32 F-axis governs for axial
L"m — — 1.0 0 . 6 + 0 . 4 (80/80)
e in. — 3.63 M/P = 80 X 12/265
Bav in." 1 — 0.265 Fig. 18
\/s — — 0.96 1.0 X 3 . 6 3 X 0.265 X (22/22)
fa/0.6Fy — — 0.43 Fig. 18, chart intersection (7a)
Col. w t lbs/ft — 97 0.157 X 2 6 5 / 0 . 4 3

Try 10V\F100:
Chart (Fig. 18) shows Formula (7a) governs.
tern Actual Assumed Remarks
'x 4.61 4.5 —
r
y 2.65 2.5 —
Bx 0.262 0.265 A I S C Mc
Pi 22 22 Lcc<< LL<< Lu; AISC Manual p. 3-21
Close agreement. Continue check of 10V\F100.

Check 10V\F100 by Bethlehem Beam-Column Tables, Formula (7a):

P CmM 265 .
265 1.0 X 80
+ = 0.50 + 0.42 = 0.92 < 1.0 OK
K (T P/P/)Ma 529 (T 265/4,500)202

58

AISC E N G I N E E R I NG J O U R N A L
Example 4—Same as Example 3, except use 12V\F and note any weight saved.

C0.80)
Y_ X_

V
1 80 k-ft

Fy = 36 ksi
P = 265 kip 12 ft IZ
USE 12 VF
V 80 k-ft
A
\P
BRACED BRACED
K-I.O

Solution:
Item Unit Y X Remarks
Yav in. 3.0 5.5 —
Kl/r — 48 26 —
^m — — 1.0 See Example 3
e in. — 3.63 See Example 3
•Bav in." 1 — 0.22 —
\/s — — 0.80 1.0 X 3.63 X 0.22
fa/0.6Fv — — 0.49 Fig. 18, chart intersection (7a)
Col. wt. lbs/ft — 85 0.157 X 2 6 5 / 0 . 4 9

T r y 12\AF85, Formula (7a):

Item Actual Assumed Remarks

rv 3.07 3.0 —
rx 5.38 5.5 —
Bx 0.216 0.22 —
Fb 24 22 Compact, L < L AISC Manual p. 3-20

Check 12V\F85 by Bethlehem Beam-Column Tables, Formula (7a):

CmM 265 1.0 X 80
= 0.57 + 0.37 = 0.94 < 1.0 OK
Pa + (1 - P/Per)Ma 465 +
(T 265/5,200)229
Next smaller column (12V\F79) is too small: (79 lbs/ft/85 lbs/ft) < 0.94
Note that the 12V\F85 carries the same load and moment as the 10V\F100 of Example 3.

59
Example 5—Single story building.

(0.96)
X 1.0
V ki/r (7a)
P
N\ /—- /0°, °
lp 48 k-ft 1i
i
1
90 k-ft V V\ / — ! 0 0 ' 6 0
N N y \ V r 100,100/
Fy - 36 ksi
P = 440 Kip 13 ft 18 ft
0 33 N^XX ,— f7b)x
USE 14 VF x^v\/ (7b)y

f
A
IP
(6)
UNBRACED UNBRACED •
k=L8 o 035 1.0

Given:
Item Unit Y X Remarks
Fy ksi 36 —
P kips 440 —
Size — 14V\F —
Sway — Unbr Unbr. —
L ft 19 18 —
K — 1.8 1.8 (Fig. 10, Case F)
Mtov k-ft 48 90 —
Mbot k-ft 0 0 —

Solution:
T av in. 3.7 6.5 —
Kl/r — 105 60 —
^m — 0.85 0.85 Unbraced
e in. 1.30 2.45 M/P
•E>av in.-1 0.52 0.185 —
1/s — 0.575 0.385 CmeB
21A — 0 96 0.575 + 0.385
fa/0.6Fv — 0 33 Fig. 18, chart intersection (7a)
Col. wt. lbs/ft 210 0.157 X 4 4 0 / 0 . 3 3

Try 14V\F158. Assumptions for r, B and Fb fairly close.

Check by Bethlehem Beam-Column Tables, Formula (7a):

P ^mx^-x ^my-tv*- y
< 1.0 (Modified Formula (7a))
+
Pa ' (1 - P/Pex')Max + (1 - P/Pey')May
440 0.85 X 90 0.85 X 48
+ = 0.53 + 0.15 + 0.28 = 0.96 < 1.0 OK
838 (1 - 440/2,620)611 (1 - 440/1,020)258

60

AISC ENGINEERING JOURNAL

Example 6—Bottom story of multi-story building.

25/c-ftj 43 k-ft
(0.7O)
Fy = 36 ksL
P = 147 kip
USE 12 W

0.24

BRACED UNBRACED
K=I.O k = 2.75 085 1.0

Given:
Item Unit Y X Remarks
Fy ksi 36 —
P kips 147 —
Size — 12V\F —
Sway — Br. Unbr. —
L ft 20 20 —
K — 1.0 2.75 (Fig. 10 shows 2 . 5 for Case F, X-axis)
Mt0V k-ft 25 43 Both clockwise
Mbot k-ft 25 0 Clockwise

Solution:
fav 3.0 5.5 —
80 120 —
Kl/r 0.6 0.85 —
in. 2.05 3.5 M/P
e 0.64 0.22 —
Bav 0.79 0.65
Us 1.44 0.79 + 0.65
21A 0.70 1/1.44
s 0.24 Fig. 18, chart intersection (7a)
fa/0.6Fy
lbs/ft 96 0.157 X 147/0.24
Col. wt.

Try 12V\F99. Assumptions for rx, ry, B, and Fb are close.

Check by Bethlehem Beam-Column Tables, Formula (7a):

^m.T^J- r ^myM-y
< 1.0 (Modified Formula (7a))
Pa +' (1 - P/Pex')Max + (1 - P/Pey')Ml
147 0.85 X 43 0.6 X 25
0.49 + 0.30 + 0.21 = 1.00 OK
298 + (1 - 147/302)242 + (1 - 147/730)90

61

APRIL/ 1967
 Errata

Design of Beam-Columns Page 56

In the abscissa Cmfb/Fb in the diagram, 0.04 should read

Paper presented by IRA H O O P E R (April, 1967, Issue)
0.4.

Page 60
Page 51
Under "Given", the Y dimension of L should read 18 ft
I n Step 4 of the recommended procedure for chart solu- instead of 19 ft. T h e beam size should be 14NAP211
tion of Formula (7a), the term 0.15e should read 0.75^. instead of 14V\F158.

Page 55 Page 61

In the 3rd line of the solution, Fb should be 22 ksi instead U n d e r "Given", Mhot under Y should read 0, and the
of 36ksi. term "clockwise" under " R e m a r k s " should be deleted.

128
AISC ENGINEERING JOURNAL