STATISTICAL ANALYSIS OF SLENDER COMPOSITE

BEAM-COLUMN STRENGTH
By S. A. Mirza, 1 Fellow, ASCE, and B. W. Skrabek2
Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

ABSTRACT: The strength of a slender composite steel-concrete beam column dif-

fers from the computed nominal value, which is based on specified nominal strengths
of constituent materials, geometric properties, and code design equations. The
variability in the strength is caused by the variations in the strengths of concrete
and steel, the cross-section dimensions of concrete and steel sections, the placement
of steel sections and reinforcing bars, and the strength model itself, among other
factors. This study was undertaken to investigate the variability of ultimate strength
of slender composite beam columns in which steel shapes are encased in cast~in-
place concrete. The results of this study indicate that the slenderness ratio, the
structural steel ratio, and the end eccentricity ratio significantly influence the prob-
ability distribution properties of slender composite beam-column strength, whereas
the effect of specified concrete strength seems to be significant only for beam
columns with a low slenderness ratio.

INTRODUCTION

The strength of a slender composite steel-concrete beam column differs

from the computed nominal value, which is based on specified nominal
strengths of constituent materials, geometric properties, and code design
equations. This is caused by the variations in the strengths of concrete and
steel, the cross-section dimensions of concrete and steel sections, the place-
ment of steel sections and reinforcing bars, and the strength model itself,
among other factors. This variability in strength is included in member design
through safety provisions of the structural codes.
The safety provisions of American Concrete Institute (ACI) Standard
318-89 ("Building Code" 1989) involve the use of load factors and resistance
(capacity reduction) factors to account for probable overloading and un-
derstrength of structural members. Computing the strength variability is an
essential component for developing probability-based safety criteria for de-
sign of structural members. An earlier investigation studied the strength
variability of short composite beam columns (Mirza and Skrabek 1991),
However, that study did not consider the composite beam columns in which
length or stability effects are significant. This study was undertaken to in-
vestigate the variability of ultimate strength of slender composite beam
columns in which steel shapes are encased in cast-in-place concrete. Square
beam columns selected for study were bent in symmetrical single curvature
in braced frames subjected to short-time loads, and are shown in Figs. 1(a)-
1(c). The moment magnifier approach specified in the ACI Standard 318-
89 ("Building Code" 1989) was developed for this type of beam column.
The results reported in this study will be employed in calculating prob-
ability-based resistance factors for slender composite steel-concrete beam

2
'Prof, of Civ. Engrg., Lakehead Univ., Thunder Bay, Ontario, Canada P7B 5E1.
Struct. Engr., C.E. Mickelson Assoc, 846 MacDonnell St., Thunder Bay, On-
tario, Canada P7B 5J1.
Note. Discussion open until October 1, 1992. To extend the closing date one
month, a written request must be filed with the ASCE Manager of Journals. The
manuscript for this paper was submitted for review and possible publication on
December 17, 1990. This paper is part of the Journal of Structural Engineering, Vol.
118, No. 5, May, 1992. ©ASCE, ISSN 0733-9445/92/0005-1312/$1.00 + $.15 per
page. Paper No. 1085.
1312

J. Struct. Eng., 1992, 118(5): 1312-1332

 P1

j
r '
M •• ^ ir 3 I.---"" I
A

\
Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

\
1 \\
\ V
\\
\ vz 112

dm

e/2 f/2

;
.l^s^ i

t-i-
PI

(a) (b)

0.5 - inch diameter

lies at 10 inches
on centres

1.5 inch 0.875 - Inch diameter

concrete cover vertical bars

FIG. 1. Type of Beam Columns Studied: (a) Model of Beam Column in Symmetrical
Single Curvature; (b) Forces on Beam Column; (c) Bending-Moment Diagram; and
(d) Details of Nominal Cross Section (Dimensions Shown Are in Inches; 1 in. =
25.4 mm; W10 x 54 = W250 x 80; W10 x 112 = W250 x 167)

columns. These results are needed, since the probability-based load and
resistance factors are currently under discussion by the structural engineer-
ing profession for possible inclusion into the ACI Standard 318-89.

DESCRIPTION OF BEAM COLUMNS USED FOR STUDY

Twenty-two hypothetical square-tied slender composite beam columns

subjected to single curvature bending with equal bending moments acting
at both ends were used for this study. The representation of the nominal
cross section used is shown in Fig. 1(d), and the specified properties of the
beam columns studied are given in Table 1.
For series B and SI (Table 1), a sequence of four numbers is used to
designate a beam column: The first number stands for the specified concrete
1313

J. Struct. Eng., 1992, 118(5): 1312-1332

 TABLE 1. Specified Properties of Slender Composite Beam Columns Studied'

Beam-column Structural Strain-hardening

designation (psi) steel / v (psi) Pa llr pJvlf'c included
(1) (2) (3) (4) (5) (6) (7)
Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

(a) Series B
4-50-4-22 4,000 50,000 0.040 22.1 0.50 No
4-50-4-33 4,000 50,000 0.040 33 0.50 No
4-50-4-66 4,000 50,000 0.040 66 0.50 No
4-50-4-100 4,000 50,000 0.040 100 0.50 No
6-50-4-22 6,000 50,000 0.040 22.1 0.33 No
6-50-4-33 6,000 50,000 0.040 33 0.33 No
6-50-4-66 6,000 50,000 0.040 66 0.33 No
6-50-4-100 6,000 50,000 0.040 100 0.33 No
4-50-8-22 4,000 50,000 0.082 22.1 1.03 No
4-50-8-33 4,000 50,000 0.082 33 1.03 No
4-50-8-66 4,000 50,000 0.082 66 1.03 No
4-50-8-100 4,000 50,000 0.082 100 1.03 No
6-50-8-22 6,000 50,000 0.082 22.1 0.68 No
6-50-8-33 6,000 50,000 0.082 33 0.68 No
6-50-8-66 6,000 50,000 0.082 66 0.68 No
6-50-8-100 6,000 50,000 0.082 100 0.68 No
(6) Series SI
4-36-4-33 4,000 36,000 0.040 33 0.36 No
4-44-4-33 4,000 44,000 0.040 33 0.44 No
6-36-4-33 6,000 36,000 0.040 33 0.24 No
6-44-4-33 6,000 44,000 0.040 33 0.29 No
(c) Series S2
4-50-4-66-STH 4,000 50,000 0.040 66 0.50 Yes
6-50-4-66-STH 6,000 50,000 0.040 66 0.33 Yes
"Each beam column listed had grade 60 (/,. = 414 MPa) reinforcing bars with p„ = 0.012, and was
studied for 16 nominal end eccentricity ratios elh of 0.05, 0.1, 0.15, 0.2, 0.25, 0.3, 0.4, 0.5, 0.6, 0.7,
0.8, 1,0, 1.5, 2.0, 4.0, and =°.
Note: 1,000 psi = 6.895 MPa; 1 in. = 25.4 mm.

strength (f'c); the second shows the specified yield strength (/,) of the struc-
tural steel; the third indicates the approximate ratio of structural steel section
area to the gross (concrete) area of the cross section (pM); and the fourth
gives an approximate slenderness ratio (llr). Thus, 6-50-8-22 designates a
beam column for which f'c = 6,000 psi (41.4 MPa), structural steel/ =
50,000 psi (345 MPa), pss = 0.082, and llr = 22.1. For beam columns of
series S2 (Table 1), a suffix consisting of the letters "STH" is added to the
end of the designation. This indicates that the effect of strain hardening of
structural and reinforcing steel was included in the theoretical analysis of
series S2 beam columns.
Four llr ratios were examined: 22.1, 33, 66, and 100. The llr ratio of 22.1
nearly represents the lower limit of the ACI Standard 318-89 ("Building
Code" 1989) for which the length effects must be included in design for the
type of beam columns studied, whereas the llr ratio of 100 is the maximum
slenderness ratio permitted for the evaluation of stability effects by the
moment magnifier method of the ACI Standard 318-89. Each beam column
was studied for specified end eccentricity ratios (elh) = 0.05, 0.1, 0.15, 0.2,
0.25, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 1.0, 1.5, 2.0, 4.0, and °o (pure bending).
These end eccentricities produced bending moments about the major axis
of the steel section. The definitions of e, h, and / are given in Figs. 1(b)
and 1(d).
1314

J. Struct. Eng., 1992, 118(5): 1312-1332

 All beam columns had .the reinforcmg-steel specified yield strength of
60,000 psi (414 MPa). The longitudinal reinforcing bars and lateral ties
conformed to the minimum requirements of the ACI Standard 318-89. The
steel sections satisfied the geometric requirements of a class 3 (noncompact)
Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

section under pure compression and those of a class 2 (compact) section

when subjected to pure flexure or combined axial force and flexure. In all
cases, concrete cross sections with dimensions shown in Fig. 1(d) were used.
This was done because an earlier parametric study on this subject (Mirza
1989) established that the gross cross-section size had only an insignificant
effect on the strength variability of composite beam columns.

STATISTICAL EVALUATION PROCEDURE

Monte Carlo Technique

The Monte Carlo technique was used in this study to simulate the vari-
ability of the ultimate strength of slender composite steel-concrete beam
columns. Two items were required: (1) A theoretical deterministic calcu-
lation procedure to express the ultimate strength of composite beam columns
as a function of all variables affecting the strength; and (2) the properties
of probability distributions of all variables involved in calculation of the
ultimate strength. By simulating a large number of ultimate strength values,
each one calculated using the deterministic strength relationships and a
randomly generated set of values of the variables affecting strength, the
statistical properties of the ultimate strength probability distribution were
estimated by analyzing the simulated sample. Details of the Monte Carlo
technique as applied to safety problems in reinforced concrete structures
are given elsewhere (Mirza 1985), and the required items for the simulations
used in this study are described in the next two sections.

Theoretical Strength Model

Theoretical Strength Analysis of Slender Beam Columns

The theoretical analysis used a strain-compatibility solution to generate
the moment-curvature curves for different levels of axial load acting on a
composite cross section. The maximum bending moment from the cross-
section moment-curvature curve for a given axial-load level defined one
point on the axial-force-bending-moment interaction diagram. This was
repeated until sufficient points were obtained to define accurately the entire
interaction diagram for the cross section.
The bending capacity of the member (slender beam column) for a given
axial force was calculated by solving for the maximum end eccentricity for
which equilibrium could be maintained between the ends and midheight of
the slender beam column. This was done by approximating the deflected
shape of the slender beam column by a fourth-order parabola, as Quast
(1970) suggested, and employing a trial-and-error procedure to solve for
the maximum eccentricity acting at the ends of the beam column.
The lateral deflection at midheight of the beam column can be computed
from (Quast 1970)
_ P(5<j>„, + 4>e) m
W
48
in which <$>e = curvature at the beam-column ends; <)>,„ = curvature at
midheight of the beam column; and / = length or height of the beam column.
1315

J. Struct. Eng., 1992, 118(5): 1312-1332

 Since the midheight eccentricity is the sum of end eccentricity and lateral
deflection at midheight, the end eccentricity can be computed from
_ Mm P{5$m + k )
Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

e (2)
~~P 48
in which Mm = bending moment at midheight; and P = axial load acting
on the beam column. The midheight curvature was incremented from a
minimum value until a maximum end bending moment was calculated. For
each value of the midheight curvature investigated, the end curvature value
was initially set equal to a minimum, then incremented until an equilibrium
combination was f6und. The largest curvature that could be attained at
midheight was equal to the curvature that corresponded to the maximum
moment on the cross-section moment-curvature diagram for the axial-force
level under consideration. Once all possible midheight curvatures had been
investigated for a given axial force, the largest calculated end bending mo-
ment established one point on the member (slender beam column) axial-
force-bending-moment interaction curve. The process was repeated for all
axial-force levels that permitted equilibrium of the member under deflected
conditions. This generated the entire axial-force-bending-moment inter-
action curve for the slender beam column that included the length effects.
The axial-force and bending-moment resistances of slender beam columns
were then calculated for specified elh through interpolation from the gen-
erated points on the member interaction diagram.

Assumptions and Approximations

The theoretical analysis used the following assumptions: (1) The strain
in the cross section was proportional to the distance from the neutral axis
at all levels of axial load; (2) there was no slip between the structural steel
or reinforcing bars and the surrounding concrete; (3) the concrete and steel
stresses were calculated as functions of the strains; (4) the effects of residual
stresses in the structural steel section were included; and (5) the concrete
confinement provided by the lateral ties and the component plates of the
structural steel section was considered. These are described in more detail
in the following.
The composite beam-column cross section consists of three materials:
concrete, structural steel, and reinforcing steel. The concrete was subdivided
into three types: unconfined concrete outside the lateral ties, highly confined
concrete in between the steel section web and flanges, and partially confined
concrete inside the lateral ties but outside the influence of the steel section.
The steel section was subdivided into two parts: the web and flanges. Hence,
the cross section was comprised of materials with six different stress-strain
curves. The six parts of the cross section are identified in Fig. 2(a).
The distribution of residual stresses was assumed to be linear for the steel
section web and flanges. The residual stress at the flange tips was taken as
a function of the steel section geometry and was computed using the expres-
sion suggested by Young (1971). The residual stress at the flange-web junc-
tures was considered a function of the residual stress at the flange tips and
the geometry of the steel section, as Galambos (1963) suggested. The dis-
tribution of residual stresses in the flanges was determined first. The dis-
tribution of residual stresses in the web was then computed using equilib-
rium. A comparison of computed residual stresses with the test data on
North American steel sections reported by Beedle and Tall (1960) suggested
a reasonable correlation between the computed and tested values. A detailed
1316

J. Struct. Eng., 1992, 118(5): 1312-1332

 Limit of Partially
Confined Concrete
Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

Highly
Confined
Concrete

Partially
Confined
Concrete
Unconfined
Concrete
Vertical Reinforcing Bar

Lateral Tie
FIG. 2(a). Material Types in Composite Cross Section

0.2 f - -

8u = 0.004

FIG. 2(b). Stress-Strain Curve for Unconfined Concrete

analysis of the residual stresses is presented elsewhere (Skrabek and Mirza

1990).
A modified Kent and Park stress-strain relationship was used for concrete
in compression (Park et al. 1982) with the first part of the curve up to the
maximum strength described by a second-order parabola and the part of
the curve after the maximum strength by a straight line. The slope of the
straight line depended on concrete strength for the unconfined concrete as
well as on the degree of concrete confinement provided by lateral ties for
the partially confined concrete. For the highly confined concrete, the slope
of the straight line was arbitrarily assumed to be zero due to the confinement
provided by the partially confined concrete on one side and by the com-
1317

J. Struct. Eng., 1992, 118(5): 1312-1332

 ponent plates of the structural steel section on the remaining three sides.
The assumed zones of concrete confinement are defined in Fig. 2(a), and
the corresponding schematic stress-strain curves are given in Figs. 2(b) and
2(c). The factor K in Fig. 2(c) represents the degree of concrete confinement
Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

provided by lateral ties. For the concrete in tension, a linear brittle stress-
strain relationship with the rupture tensile strain equal to fJEc was used,
where fr stands for the modulus of rupture and Ec for the modulus of
elasticity of concrete.
An elastic-plastic stress-strain relationship was assumed for both steels
for series B and SI beam columns, where the effect of strain hardening of
steel was neglected (Table 1). The actual stress-strain curves of steel were
used for series S2 beam columns for which the theoretical strength analysis
included the effect of strain hardening (Table 1), as indicated by. Fig. 2(d).
Computation of Model Error
The bias and variability of the theoretical strength model itself were
computed from test data available in the literature. The ratio of tested to

0.2Kf

e0 = 2 K f o / E c

FIG. 2(c). Stress-Strain Curves for Partially Confined and Highly Confined Con-
cretes

ys sh us

FIG. 2(d). Stress-Strain Curve for Steel

1318

J. Struct. Eng., 1992, 118(5): 1312-1332

 computed strength is affected by in-batch variations in material strengths
and geometric properties in addition to variations caused by inaccuracies in
the theoretical strength model and the beam-column testing procedures,
i.e.
Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

' t/c 'in-batch "^ 'model ' 'test W/

in which V„c = coefficient of variation of the ratio of tested to computed

strengths; Fmodei = coefficient of variation of beam-column strength due to
inaccuracies in the theoretical model; Vtest = coefficient of variation of
beam-column strength due to testing procedures; and Vin.billch = coefficient
of variation of beam-column strength due to in-batch variabilities of all
variables affecting the beam-column strength. The variability of the model
error then can be calculated by rearranging (3).

^model = W t / C — ^in-batch ~~ 'test (4)

V„c computed for 63 physical tests from Johnson and May (1978), Morino
et al. (1984), Procter (1967) and Suzuki et al. (1984) was significantly af-
fected by the elh ratio. Hence, VmoM was taken as a function of the elh
ratio. Vtest was assumed to be equal to 0.02 for pure bending and 0.04 for
pure compression. Similar values were suggested earlier (Mirza and MacGregor
1989). Vin.batch w a s computed from Monte Carlo simulations assuming the
in-batch coefficients of variation of the compressive and tensile strengths
of concrete to be 0.05, the modulus of elasticity of concrete 0.035, the yield
strength of steel 0.02, and the modulus of elasticity of steel 0.01, as suggested
in earlier studies (Mirza et al. 1979; Mirza and MacGregor 1979a, b; Skrabek
and Mirza 1990). V^^ was then calculated from (4) using the estimated
values of Vt/C, ^in-bateh, and Vtest.
The probability distribution of the model error in this study was described
by a normal curve with a mean value of 1.00 and a coefficient of variation
that was a function of the end eccentricity and varied from 0.025 for pure
bending to 0.11 for pure compression. This model error was included in the
subsequent Monte Carlo simulations by introducing a random variable rep-
resenting the ratio of tested to theoretical strengths.

Probability Models of Variables

The statistics of the variables used in this study are shown in Table 2.
These statistics were based on the probability distribution data of basic
variables summarized by Beedle and Tall (1960), Galambos and Ravindra
(1978), Kennedy and Aly (1980), Mirza et al. (1979), and Mirza and
MacGregor (1979a, b), or were developed by Skrabek and Mirza (1990) as
part of this project from the data available in the literature.
In developing the statistics for concrete properties reported in Table 2,
the quality control of 4,000-psi (27.6-MPa) concrete was assumed to be
average (coefficient of variation of the test cylinder strength ycyl = 0.15),
and that of 6,000-psi (41.4-MPa) concrete was taken to be better than
average or excellent (Vcyl = 0.10), reflecting somewhat extra care taken for
manufacturing concrete with/c > 4,000 psi (27.6 MPa). Similar values of
Vcy, were suggested in previous studies, including an extensive statistical
study of concrete strength by Rusch et al. (1969).
The static yield and ultimate strengths of reinforcing bars were assumed
to follow beta probability distributions, whereas the static yield strength of
structural steel web, the ratio of actual to specified flange width, and the
1319

J. Struct. Eng., 1992, 118(5): 1312-1332

 TABLE 2. Overall Statistics of Basic Variables
Standard Coefficient of
Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

Variable Mean value deviation variation

d) (2) (3) (4)
(a) Concrete in Structure Loaded to Failure in 1 Hour
Compressive strength fc (psi)
Average quality control, f'c = 4,000 psi 3,388 596 0.176
Excellent quality control, f'c = 6,000 psi 4,640 630 0.136
Modulus of rupture /, (psi)
Average quality control, f'c — 4,000 psi 445 97 0.218
Excellent quality control, f'c = 6,000 psi 523 110 0.211
Modulus of elasticity Ec (ksi)
Average quality control, f'c — 4,000 psi 3,260 388 0.119
Excellent quality control, f'c = 6,000 psi 3,795 398 0.105
(b) Structural Steel
Modulus of elasticity E, (ksi) 29,000 580 . 0.020
Static yield strength of web /„„, (psi)
/,. = 36,000 psi 39,240 3,375 0.086
/,. = 44,000 psi 47,960 - 4,125 0.086
/,. = 50,000 psi 54,500 4,687 0.086
Static yield strength of flange f„f 0.95 /„,, —a —a
Static ultimate strength of web /,„„, 1.5/,,,,. —aa —aa
Static ultimate strength of flange fmj 1.5/,/ — —
Strain at start of strain hardening E,(, 0.017 0.0045 0.266
Initial tangent modulus of strain-hardening curve
E,i, (ksi) 600 150 0.250
Residual stresses (psi)
W10 X 54 (W250 x 80)
At flange tips -18,576" 2,786 0.150
At flange-web juncture 12,089 8,825 0.730
W10 x 112 (W250 x 167)
At flange tips -19,311" 2,897 0.150
At flange-web juncture 12,240 8,935 0.730
Ratio of actual to specified dimensions
Section depth d 1.000 0.0 0.0
Flange width b 1.005 0.014 0.014
Flange thickness t 0.976 0.041 0.042
Web thickness w 1.017 0.039 0.038
(c) Reinforcing Steel
Modulus of elasticity Es (ksi) 29,000 957 0.033
Static yield strength/, (psi), / . = 60,000 psi 66,800 6,200 0.093
Static ultimate strength /,„ 1.55/„ —a —a
Strain at start of strain-hardening ES/, 0.015 0.004 0.266
Ultimate strain em 0.150 0.030 0.200
(d) Deviation of Overall Beam-Column Dimensions from Specified Values"
Length (in.) 0.0 0.670 —
Cross-section depth (in.) + 0.06 0.250 —
Cross-section width (in.) + 0.06 0.250 —
Concrete cover to lateral hoops (in.) + 0.33 0.166 —
Spacing of lateral hoops (in.) 0.0 0.530 —
a
The value of this variable was assumed to be dependent on the value of another variable, and no
further variation was applied.
"The negative sign indicates compressive stress.
"These statistics represent cast-in-place construction of average quality.
Note: 1,000 psi = 1 ksi = 6.895 MPa; 1 in. = 25.4 mm.

1320

J. Struct. Eng., 1992, 118(5): 1312-1332

 ratio of actual to specified web thickness of the steel section were repre-
sented by modified lognormal probability distributions with lower limits of
0.75 times the specified yield stress, 0.88, and 0.80, respectively. The static
yield strength of flange and static ultimate strengths of web and flange also
Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

followed the modified lognormal probability distributions, because these

variables were functions of the static yield strength of web. All other var-
iables were assumed to follow normal probability distributions. This is doc-
umented in detail by Skrabek and Mirza (1990).

SIMULATION AND ANALYSIS OF STRENGTH DATA

Using the theoretical strength model and the probability distributions of

the variables given in Table 2, the ultimate strengths were simulated 500
times by a Monte Carlo technique for each of the 16 elh ratios studied for
each of the 22 slender beam columns listed in Table 1. These strengths were
divided by the corresponding ultimate strengths predicted by the ACI Stan-
dard 318-89 ("Building Code" 1989) using the specified nominal properties
of the variables. This produced 352 simulated samples of the ratios of the
theoretical to ACI strength referred to as the strength ratios (RJR„) in this
study. An R,/R„ ratio was taken as the ratio of the bending-moment capac-
ities for elh = o° and the ratio of axial-load capacities for all other elh values.
The simulated samples of RJR„ ratios were analyzed statistically to describe
their probability distribution properties.
All computations for the ACI Standard 318-89 strengths were carried out
as they would be in a design office with the exception of the understrength
factors <$>• The <j) factors for the cross-sectional strength and those for the
critical buckling strength Pc of the beam column were taken equal to 1.0 in
this study. The calculated strength ratios presented in this paper are, there-
fore, based on the unfactored ACI strength. For computing Pc, the flexural
stiffness EI was calculated from the ACI Standard 318-89, and the modulus
of elasticity of concrete was taken as 57,000 Vfc psi (4,733 Vfc MPa).

Dispersion of Slender Column Axial-Load-Bending-Moment

Interaction Curves
Figs. 3(a) and 3(b) show the plots of the simulated axial-load-bending-
moment interaction diagrams for two slender beam columns (6-50-4-66 and
4-50-8-33) taken from Table 1. These beam columns represent nearly the
extremes of the structural steel index (pssfylf'c = 0.33 and 1.03) for the beam
columns studied. The maximum, mean, one-percentile, and minimum strength
curves shown in Figs. 3(a) and 3(b) were obtained from an analysis of 500
axial-load-bending-moment interaction diagrams generated for each of the
two slender beam columns using the theoretical strength model and statistics
of basic variables (Table 2). Also plotted for the purpose of comparison are
the mean cross-sectional strength curves. The ACI ultimate strength curves
are the slender beam-column capacities computed using the ACI Standard
318-89 ("Building Code" 1989) and assuming a value of 1.0 for all under-
strength factors.
Fig. 3(a) plotted for beam column 6-50-4-66 having pssfy/f'c = 0.33 and
llr = 66 shows the ACI ultimate strength prediction significantly underes-
timating the mean theoretical strength for elh < 0.3. For elh > 0.3, the ACI
procedure predicts the strength slightly higher than the mean theoretical
strength. At the pure bending condition, the ACPultimate strength predic-
tion is nearly identical to the mean theoretical strength. In Fig. 3(b), which
1321

J. Struct. Eng., 1992, 118(5): 1312-1332

Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

Bending Moment (Ft-Kips)

FIG. 3. Dispersion of Axial-Load-Bending-Moment Strength Interaction Curves

of Randomly Generated Samples of 500 Slender Composite Beam Columns: (a)
Beam Column 6-50-4-66; and (b) Beam Column 4-50-8-33 (100 kips = 445 kN; 100
ft-kips = 135.6 kNm)

is plotted for beam column 4-50-8-33 having pssfy/f'c ~ 1.03 and IIr - 33,
the ACI ultimate strength prediction somewhat overestimates the mean
theoretical strength for e/h < 2.0. For e/h between 2.0 and °°, the mean
theoretical strength is somewhat underestimated by the ACI ultimate strength.
The differences between the ACI ultimate strength and the mean theoretical
strength are much less in Fig. 3(b) than those displayed in Fig. 3(a), and
perhaps reflect the effects of lower slenderness ratio and higher structural
steel index associated with the beam-column data plotted in Fig. 3(b).
Histograms of the simulated theoretical strengths for e/h of 0.1, 0.2, and
1.0 plotted in Figs. 3(a) and 3(b) are nearly symmetric. This indicates that
the positively skewed shape of the structural steel strength probability dis-
tribution does not influence the overall slender beam-column strength. A
smaller coefficient of variation at high e/h ratio is apparent in both figures.
1322

J. Struct. Eng., 1992, 118(5): 1312-1332

 Similar behavior was observed for slender reinforced concrete beam columns
(Mirza and MacGregor 1989).
Effects of Slenderness, Specified Concrete Strength, and
Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

Structural Steel
The Series B beam columns shown in Table 1 were divided into four sets
of four beam columns each to investigate the effect of slenderness ratio llr
on the strength ratio. Each set had one beam column with slenderness ratio
of 22.1, 33, 66, or 100. All other properties in each set of beam columns
were identical. The one-percentile strength ratios for beam columns with
llr of 22.1 and 33 were the lowest, followed by those for beam columns with
llr of 66. The highest one-percentile strength ratios were obtained for beam
columns with llr = 100. This is expected because the cracks in a longer
column are likely to be widely spaced, with more concrete in between the
cracks contributing to the flexural stiffness that, in turn, reduces the second-
order effects. In all cases, the differences in one-percentile strength ratios
of beam columns with different llr decreased as elh increased from 0.05 to
nearly 1.2, as indicated by Figs. 4(a) and 4(b) plotted for two of the sets.
For e/h>1.2, there were no significant differences in one-percentile strength
ratios of beam columns with different llr ratio regardless of elh value. This
is expected because the length does not affect the strength of a beam column
at or near a pure bending condition. The trends as noted were consistent
with the mean values of the strength ratios as well, as shown in Figs. 4(a)
and 4(b). From this it is reasonable to conclude that the strength ratios for
beam columns with llr = 22.1 to 33 are more critical than those for beam
columns with other llr ratios studied.
The series B beam columns listed in Table 1 provided eight sets for
investigating the effect of specified concrete strength f'c on the strength
ratio. Each set contained one beam column having f'c = 4,000 psi (27.6
MPa) and one having/c = 6,000 psi (41.4 MPa), with all other properties
of both beam columns in the set being identical. Comparisons for two of
these sets are shown in Figs. 5(a) and 5(b). These comparisons for beam
columns having llr = 22.1 showed lower one-percentile and mean strength
ratios when 6,000-psi (41.4-MPa) concrete was used. The differences in
strength ratios were especially apparent for the beam columns with the lower
percentage of structural steel, as indicated by Fig. 5(a). This is because the
ratio of the mean in situ strength to specified strength is 0.85 for 4,000-psi
(27.6-MPa) concrete and 0.77 for 6,000-psi (41.4-MPa) concrete. As the
slenderness ratio was increased to 33, the effect of specified concrete strength
on the one-percentile strength ratios tended to disappear. However, the
higher-strength concrete produced lower mean strength ratios for beam
columns with llr = 33 as well, as shown in Fig. 5(b). For beam columns
having slenderness ratios of 66 and 100, the effect of f'c on strength ratios
was noticeable only at elh < 0.15. This effect dissipated rapidly with in-
creasing elh due to the failure mode changing from compression to tension
caused by increasing second-order bending moment acting on the beam
column. From this discussion, it is concluded that for beam columns having
slenderness ratios greater than or equal to 33, the specified concrete strength
does not significantly affect the strength ratio. For slender beam columns
with llr < 33, the effect of f'c seems to be significant.
Eight comparisons were made to investigate the effect of structural steel
ratio pss on the strength ratios of the series B beam columns listed in Table
1. Each comparison involved two beam columns, one having pM = 0.040
1323

J. Struct. Eng., 1992, 118(5): 1312-1332

 ', a (Yr=22.1 (4-50-4-22)

C/r-33 (4-50-4-33)
Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

1.5 f/r=66 (4-50-4-66)

f/r=100 (4-50-4-100)

1, \ Mean Values

0.9

One-Percentile Values

01
1.4
en lb) ftr=22.1 (6-50-8-22)
S 1.3
<7r=33 (6-50-8-33)
55
1.2 I/r=66 (6-50-8-66)

f/r=100 (6-50-8-100)

0 0.2 0.4 0.6 0.8 1.0 1.5 2.0 4.0 m

End Eccentricity Ratio (e/h)

FIG. 4. Effect of llr on Strength Ratio of Slender Composite Beam Columns

and the other having pM = 0.082, with all other properties being identical.
For slender beam columns with llr of 22.1 and 33, the smaller structural
steel ratio (pM = 0.040) produced lower one-percentile strength ratios than
did the larger structural steel ratio (pM = 0.082). At mean value levels,
however, the differences in strength ratios for beam columns with pss —
0.040 and 0.082 were somewhat less significant, as indicated by Fig. 6(a).
This behavior is similar to that described elsewhere for short composite
beam columns (Mirza and Skrabek 1991). When the slenderness ratio was
increased to 66, the beam columns with pss = 0.082 demonstrated signifi-
cantly lower one-percentile and mean strength ratios than those for beam
columns with pM = 0.040; the differences in strength ratios being very
significant for e/h < 0.4. This is shown in Fig. 6(b). The effect of pM on
strength ratios became even more significant when llr ratio was increased
to 100. This is probably because a very slender column must withstand high
second-order bending moments and, therefore, depends on the structural
steel to provide stiffness after the concrete has severely cracked. It is evident

1324

J. Struct. Eng., 1992, 118(5): 1312-1332

Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

Mean Values

One-Percentile Values

f; = 4000 psi (27.6 MPa) (4-50-4-22)

f; = 6000 psi (41.4 MPa) (6-50-4-22)

Mean Values

One-Percentile Values

r„ = 4000 psi (27.6 MPa) (4-50-4-33)

f; = 6000 psi (41.4 MPa) (6-50-4-33)

0.5
0 0.2 0.4 0.6 0.8 1.0 1.5 2.0 4.0 m

End Eccentricity Ratio (e/h)

FIG. 5. Effect of f'c on Strength Ratio of Slender Composite Beam Columns

that pM must be considered as a parameter for reliability analysis of slender

composite beam columns.

Effect of Structural Steel Grade

To study the influence of structural steel grade on the strength ratios,
two slender composite series B beam columns (4-50-4-33 and 6-50-4-33)
from Table 1 were compared to the four series SI beam columns (4-44-4-
33, 4-36-4-33, 6-44-4-33, and 6-36-4-33) listed. This provided two sets of
three slender beam columns, each beam column with the structural steel fy
= 50, 44, or 36 ksi (345, 303, or 248 MPa). All other properties of the three
beam columns in a set were identical. Again, the one-percentile and mean
strength ratios were compared for the beam columns in each set. These
comparisons indicated an insignificant effect of structural steel grade on
mean strength ratios. However, the beam columns with structural steel fy
= 50 ksi (345 MPa) exhibited lower one-percentile strength ratios than did
the companion beam columns with structural steel fy = 44 or 36 ksi (303
1325

J. Struct. Eng., 1992, 118(5): 1312-1332

Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

£ t = 0.040 (4-50-4-33)

£ s = 0.082 (4-50-8-33)
—i
e 0.6

b) ^=0.040(4-50-4-66)

£, = 0.082(4-50-8-66)

R./R =1.0

Mean Values

0.8

One-Percentile Values

0 0.2 0.4 0.6 0.8 1.0 1.5 2.0 4.0 Co

End Eccentricity Ratio (e/h)

FIG. 6. Effect of p„ on Strength Ratio of Slender Composite Beam Columns

or 248 MPa). This is indicated by Fig. 1(a) plotted for one of the two sets
examined. Hence, it is recommended that the 50-ksi (345-MPa) structural
steel be used for reliability studies of slender composite beam columns. This
is the highest structural steel grade currently permitted for composite beam
columns by the ACI Standard 318-89 ("Building Code" 1989).

Effect of End Eccentricity

Each beam column from Table 1 was investigated for 16 end eccentricity
ratios (e/h = 0.05, 0.1, 0.15, 0.2, 0.25, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 1.0, 1.5,
2.0, 4.0, and =0). This study showed that the one-percentile strength ratios
generally dropped as e/h increased from 0.05 to 0.2. This effect was very
significant for beam columns with llr = 100 and somewhat significant for
beam columns having llr = 66, while the beam columns with llr = 22.1 and
33 were least affected as e/h increased from 0.05 to 0.2. The one-percentile
strength ratios then tended to increase as e/h increased from 0.2 to 00 (pure
bending). This is evident from Figs. 4-7(a). A somewhat similar behavior
1326

J. Struct. Eng., 1992, 118(5): 1312-1332

Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

Mean Values

One-Percentile Values

Structural Steel t y = 36 ksi (248 MPa)

(4-36-4-33)
Structural Steel f = 44 ksi (303 MPa)
= 0.6.
(4-44-4-33)
Structural Steel f = 50 ksi (345 MPa)
(4-50-4-33)
0.5 .
1.4

Strain Hardening Neglected

(6-50-4-66)

Strain Hardening Considered

1.2
(6-50-4-66-STH)

Mean Values

One-Percentile Values

0.7-
0 0.2 0.4 0.6 0.8 1.0 1.5 2.0 4.0 a>
End Eccentricity Ratio (e/h)

FIG. 7. (a) Effect of Structural Steel Grade; and (b) Effect of Strain Hardening of
Steel

was observed for mean strength ratios as well, as indicated by these figures.
It is obvious that the elh ratio is the most important variable for reliability
analysis of slender composite steel-concrete beam columns.

Effect of Strain Hardening of Steel

To examine the effect of strain hardening of the structural steel and
longitudinal reinforcing bars on the strength ratios of slender composite
beam columns, the data from two series B beam columns (4-50-4-66 and
6-50-4-66) were compared to the data from the corresponding series S2
beam columns (4-50-4-66-STH and 6-50-4-66-STH). This provided two sets
of beam columns, each set having one beam column in which strain hard-
ening of both steels was included in the theoretical strength analysis and
one beam column for which the steel strain hardening was not permitted.
All other properties were identical for both beam-columns in a set.
These comparisons for slender composite beam columns demonstrated
1327

J. Struct. Eng., 1992, 118(5): 1312-1332

 insignificant effects of strain hardening of steel on the one-percentile and
mean strength ratios over almost the entire range of elh, as indicated by
Fig. 1(b) plotted for one set. The only exceptions were the strength ratios
close to the pure bending condition where the increase in strength ratios
Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

due to the strain hardening of steel was substantial. The beneficial effect
of strain hardening of steel at pure bending is expected because the steel
must strain to a level approximately 10 times the yield strain before the
effect of strain hardening can be obtained. Such a strain occurs at and near
pure bending.
An examination of the moment-curvature data for the series S2 beam
columns listed in Table 1 indicated that even at low axial loads, the second-
order moments were more than enough to offset the beneficial effects of
strain hardening. It is, therefore, reasonable to neglect the effect of strain
hardening of steel for reliability analysis of slender composite beam columns.

Range of Strength Ratios and Coefficients of Variation

The data from the 16 series B beam-columns listed in Table 1 were
grouped into three sets according to the slenderness ratio. The first set
included beam columns with llr = 22.1 and 33, which were grouped together
because of the similar values of strength ratios obtained for these beam
columns. The second set contained beam columns with IIr = 66, while the
third set included beam columns with llr = 100. For each set, the ranges
of the mean and one-percentile strength ratios are plotted against elh in
Fig. 8(a) and 8(b), respectively. The trends for the effect of elh ratio stated
in a preceding section are also valid for the plots shown in Fig. 8(a) and
8(b). Additionally, for reasons described earlier, it is apparent from these
figures that, for elh s 1.2, the upper boundary of the range of strength
ratios is defined by the beam columns having llr = 100 and the lower
boundary by the beam columns with llr = 22.1 and 33. For elh values greater
than 1.2, ranges of strength ratios for all three sets overlap and remain
nearly constant. The maximum spread in one-percentile and mean strength
ratios occurs at elh = 0.05, whereas the minimum spread in these values
takes place at elh around 1.2 as indicated by Fig. 8(a) and 8(b).
Fig. 8(c) shows the ranges of strength coefficient of variation (VK) for the
three sets of composite beam columns varying significantly with respect to
end eccentricity ratio. Furthermore, beam columns with lower slenderness
ratios correspond to higher coefficients of variation of strength and vice
versa. This is especially apparent at elh > 0.2. The decrease in VR with
increasing slenderness ratio of the beam columns is likely caused by the
increase in dependence on the structural steel to provide resistance against
the increased probability of tension failure due to secondary moment effects.
Since the coefficients of variation associated with the steel properties are
less than those for concrete, the columns deriving more of their strength
capacity from the structural steel will be subjected to lower-strength vari-
ations. The strength coefficient of variation initially decreases rapidly, then
decreases at a reduced rate as elh increases from 0.2 to <*>, as indicated by
Fig. 8(c). This is due to the mode of failure changing from compression to
tension as elh increases from 0.2 to °°, making VR more dependent on the
properties of structural steel than those for concrete.
From the data presented in Fig. 8(a) and 8(c), the following strength ratio
statistics are suggested for computing the understrength factors for slender
composite beam columns: (1) Use a mean strength ratio of 0.95 for elh <
0.2, increasing linearly to 1.04 as elh increases from 0.2 to oo; and (2) use
1328

J. Struct. Eng., 1992, 118(5): 1312-1332

Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

End Eccentricity Ratio (elh)

FIG. 8. Range of Mean Strength Ratios, One-Percentile Strength Ratios, and

Strength Coefficients of Variation Computed from Simulated Data for Slender Com-
posite Series B Beam Columns Listed in Table 1

a value of VR = 0.16 fore/h <0.2, decreasing linearly to 0.08 as elh increases

from 0.2 to 1.0. For elh values greater than 1.0, VR remains constant at
0.08. Adjust the mean values to include the effect of cross-sectional capacity
reduction factors.

PRACTICAL APPLICATION

The slender beam-column stability resistance factors § specified in the

ACI Standard 318-89 ("Building Code" 1989) were based on one-percentile
strengths and are assigned currently a value of 0.7 for compression failures,
increasing linearly to 0.9 as the mode of failure changes from compression
to pure bending. In this study, one-percentile strength ratios lower than 0.7
were obtained at elh < 0.4 for almost all beam columns having llr = 22.1
and 33 and for some beam columns having llr = 66, as indicated by Fig.
1329

J. Struct. Eng., 1992, 118(5): 1312-1332

 8(5). This behavior is similar to that demonstrated by composite cross sec-
tions (Mirza and Skrabek 1991) and is a consequence of two factors.
First, the cross-section capacity reduction factors applied by the ACI
Standard 318-89 to cross-section interaction diagrams were not included in
Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

this study. The ACI capacity reduction factors for composite cross sections
are currently assigned values that also vary from 0.7 to 0.9, but were set
equal to 1.0 in this study. Second, this study did not consider the ACI 318-
89 requirement that effectively cuts off the peak value of beam-column
compression capacities to 85% of pure compressive resistance of the cross
section. This requirement will not affect the strength of beam columns of
high slenderness, but could be a factor for beam columns with IIr = 22.1
and 33. These two factors will offset the understrength indicated by Fig.
8(b). A complete, probabilistic procedure will be used in future reevaluation
of the 4> factors for slender composite steel-concrete beam columns.

CONCLUSIONS

The results of this study indicate that the ratio of theoretical strength to
nominal strength (strength ratio) of slender composite beam columns (llr
> 22 for the type of beam columns studied) was influenced most significantly
by the slenderness ratio, the ratio of area of structural steel section to gross
area of the cross section, and the end eccentricity ratio. Slender beam
columns with IIr < 33 produced lower strength ratios than did the beam
columns with larger slenderness ratio. The effect of specified concrete strength
was significant only for beam columns with IIr < 33. The results indicate
that the effects of structural steel grade and strain hardening of steel can
be neglected.
Dispersions in the axial-load-bending-moment strength interactions are
shown in Figs. 3(a) and 3(b) for two slender beam columns. The strength
statistics suggested in this paper will be useful for future evaluation of
probability-based resistance factors for slender composite steel-concrete beam
columns.

ACKNOWLEDGMENT

The financial support provided by the Natural Sciences and Engineering

Research Council of Canada, Ottawa, Ontario, is gratefully acknowledged.

APPENDIX I. REFERENCES

Beedle, L. S., and Tall, L. (1960). "Basic column strength." J, Struct. Div., ASCE,
86(7), 139-173.
"Building code requirements for reinforced concrete." (1989). ACI Standard 318-
89, Amr. Concr. Inst., Detroit, Mich.
Galambos, T. V. (1963). "Inelastic lateral buckling of beams."/. Struct. Div., ASCE,
89(5), 217-242.
Galambos, T. V., and Ravindra, M. K. (1978). "Properties of steel for use in LRFD."
/. Struct. Div., ASCE, 104(9), 1459-1468.
lohnson, R. P., and May, I. M. (1978). "Tests on restrained composite columns."
Struct. Engr,, 56B(2), 21-27.
Kennedy, D. J. L., and Aly, M. G. (1980). "Limit states design of steel structures-
performance factors." Can. J. Civ. Engrg., 7(1), 45-77.
Mirza, S. A. (1985). "Application of Monte Carlo simulation to structural engi-
neering problems." Proc. 2nd Int. Conf. on Comput. in Civ. Engrg., Science Press,
Beijing/Elsevier, Amsterdam, the Netherlands, 622-635.
1330

J. Struct. Eng., 1992, 118(5): 1312-1332

 Mirza, S. A. (1989). "Parametric study of composite column strength variability."
J. Constr. Steel Res., 14(2), 121-137.
Mirza, S. A., and MacGregor, J. G. (1989). "Slenderness and strength reliability of
reinforced concrete columns." ACIStruct. J., 86(4), 428-438.
Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

Mirza, S. A., and MacGregor, J. G. (1979a). "Variability of mechanical properties

of reinforcing bars." /. Struct. Div., ASCE, 105(5), 921-937.
Mirza, S. A., and MacGregor, J. G. (1979b). "Variations in dimensions of reinforced
concrete members." J. Struct. Div., ASCE, 105(4), 751-766.
Mirza, S. A., and Skrabek, B. W. (1991). "Reliability of short composite beam-
column strength interaction." /. Struct. Engrg., ASCE, 117(8), 2320-2339.
Mirza, S. A., Hatzinikolas, M., and MacGregor, J. G. (1979). "Statistical descrip-
tions of strength of concrete." /. Struct. Div., ASCE, 105(6), 1021-1037.
Morino, S., Matsui, C , and Watanabe, H. (1984). "Strength of biaxially loaded
SRC columns." Proc. U.S./Japan Joint Seminar on Composite and Mixed Constr.,
ASCE, 185-194.
Park, R., Priestley, M. J. N., and Gill, W. D. (1982). "Ductility of square-confined
concrete columns." /. Struct. Div., ASCE, 108(4), 929-950.
Procter, A. N. (1967). "Full size tests facilitate derivation of reliable design meth-
ods." Consulting Engr., 31(8), 54-60.
Quast, U. (1970). "Geeignete vereinfachungen fur die losung des traglastproblems
der ausmittig gedriickten prismatischen stahlbetonstutze mit rechteckquerschnitt,"
Dr. Ing. thesis, Fakultat fur Bauwesen, Technischen Universitat Carolo-Wilhel-
mina, Braunschweig, West Germany (in German).
Riisch, H., Sell, R., and Rackwitz, R. (1969). "Statistische analyse der betonfestig-
keit." Deutscher Ausschuss fur Stahlbeton, Heft 206, Wilhelm Ernst, Berlin, West
Germany (in German).
Skrabek, B. W., and Mirza, S. A. (1990). "Strength reliability of short and slender
composite steel-concrete columns." Report No. CE-90-1, Lakehead Univ., Thun-
der Bay, Ontario, Canada.
Suzuki, T., Takiguchi, K., Ichinose, T., and Okamoto, T. (1984). "Effects of hoop
reinforcement in steel and reinforced concrete composite sections." Bull. New
Zealand Nat. Soc. for Earthquake Engrg., 17(3), 198-214.
Young, B. W. (1971). "Residual stresses in hot-rolled sections." Report No. CUED/C-
Struct/TR.8, Univ. of Cambridge, London, England.

APPENDIX II. NOTATION

The following symbols are used in this paper:

Ec, Es = modulus of elasticity of concrete and steel;

Esh = initial tangent modulus of strain-hardening curve;
e = eccentricity of axial load at beam-column ends;
fc, f'c = compressive strength of concrete in structure and specified com-
pressive strength of concrete;
fy = specified yield strength of structural or reinforcing steel;
fys, fus = static yield strength and static ultimate strength of steel;
h = overall depth of composite cross section;
K = factor representing degree of concrete confinement provided by
lateral ties;
I = beam-column length or height;
Rn, R, = nominal resistance and theoretical resistance of beam column;
r = radius of gyration of composite cross section;

1331

J. Struct. Eng., 1992, 118(5): 1312-1332

 VR = overall coefficient of variation of beam column strength;
A„, = lateral deflection at midheight of beam column;
E
J/I> 8m = strain at start of strain hardening and ultimate strain of steel;
(j> = understrength (capacity reduction) factor; and
Downloaded from ascelibrary.org by NIRMA INSTITUTE OF TECHNOLOGY on 11/16/17. Copyright ASCE. For personal use only; all rights reserved.

Pn> ?ss = area of longitudinal reinforcing bars and area of structural steel
section, both divided by gross area of cross section.

1332

J. Struct. Eng., 1992, 118(5): 1312-1332