ACI STRUCTURAL JOURNAL

Title no. 99-S41

TECHNICAL PAPER

Comparative Study of Strength-Computation Methods for Rectangular Reinforced Concrete Columns

by S. Ali Mirza and Edward A. Lacroix
Comparisons of strengths determined from 354 physical tests of rectangular reinforced concrete columns available in the literature with the strengths calculated from selected computational procedures were conducted. The computational procedures compared in this study include ACI 318-99, Eurocode 2, and a commercially available nonlinear finite element modeling software. The physical tests used for comparison were conducted on tied, reinforced concrete columns that were pinned at both ends and subjected to short-term loads, producing pure axial force or axial force combined with symmetrical single-curvature bending. The study included only those columns for which the complete information required for analysis was available from the published physical test data and for which the compressive strength of normal density concrete ranged from approximately 2500 to 8500 psi (17 to 58 MPa). No further physical tests were conducted as part of this study. Major variables include the concrete strength, the end eccentricity ratio, the slenderness ratio, the reinforcing steel index, and the transverse reinforcement (tie/hoop) volumetric ratio. The comparative study provides a critical review of the reliability of the computational methods examined. Most of these methods are affected to some degree by at least some of the major variables studied. A recommendation for improving the ACI 318-99 procedure is also presented.
Keywords: column; reinforced concrete; strength; test.

INTRODUCTION A comparative study of selected methods used for computing the strengths determined from physical tests of rectangular reinforced concrete columns was undertaken. Physical tests included in this study involve tied columns made of normal density concrete. The computational methods compared include ACI 318-99 (ACI Committee 318 1999), Eurocode 2 (European Committee for Standardization 1992), and a commercially available nonlinear finite element modeling software (ABAQUS 1994a,b). The ACI design method is strongly influenced by the column effective flexural rigidity (EI), which varies due to cracking, creep, and the nonlinearity of the concrete stress-strain curve. In an attempt to account for some of these variables, Mirza (1990) proposed a refined equation for calculating the flexural rigidity of reinforced concrete slender columns designed according to the ACI procedure. This equation is also included in the comparative study reported herein. During the past 10 to 15 years, commercial finite element modeling (FEM) software has become more readily available, and its use by design engineers has been steadily increasing. Presently, there are several FEM programs that are able to model the reinforced concrete column strength. In an attempt to examine the applicability of FEM for computing the strength of reinforced concrete columns, the results from a commercially available nonlinear FEM software (ABAQUS 1994a,b) were also compared with the physical tests. ACI Structural Journal/July-August 2002

To determine the influence of a full range of variables on the computational methods examined, 354 physical tests of rectangular reinforced concrete columns were taken from 25 studies reported in the literature. Due to practical reasons and concrete stress block modeling implications, it was decided not to include columns with concrete strengths of less than 2500 psi (17 MPa) or more than 8500 psi (58 MPa). Furthermore, only those columns for which the complete information required for strength calculations was available in the published literature were included. The columns investigated were pinned at both ends and subjected to short-term loads, producing pure axial force or axial force combined with symmetrical single-curvature bending. No new physical tests were conducted for this study. The major variables investigated in this study included the concrete strength fc , the end eccentricity ratio e/h, the slenderness ratio l/h, the reinforcing steel index rs fy /fc, and the transverse reinforcement (tie/hoop) volumetric ratio , where e = eccentricity of the axial load at column ends; h = overall depth of the flexural rigidity area taken perpendicular to the axis of bending; l = length of the column; fy = yield strength of reinforcing bars; and rs = ratio of the crosssectional area of longitudinal reinforcing bars to the gross flexural rigidity area (longitudinal reinforcement ratio). Based on statistical analyses of the ratios of tested-tocomputed strengths (strength ratios), comparisons of different computational methods used, as well as evaluations of major variables affecting the strength, were conducted. Most of these methods were affected to some degree by at least some of the major variables studied. The evaluations and comparisons provide an insight for critical review of the variability and other statistics related to the computational methods examined, and are presented and discussed in this paper. The results of the study are limited to normal-strength concrete. A similar study of high-strength concrete columns is currently in its initial stages at Lakehead University. RESEARCH SIGNIFICANCE This paper provides a critical review of procedures used for, and related statistics of strength computations of, reinforced concrete columns. The evaluations and comparisons of computational procedures presented in the paper provide guidance for the reliability of several design methods used for reinforced concrete columns. A recommendation for improving the ACI procedure is presented.
ACI Structural Journal, V. 99, No. 4, July-August 2002. MS No. 01-231 received August 9, 2001, and reviewed under Institute publication policies. Copyright 2002, American Concrete Institute. All rights reserved, including the making of copies unless permission is obtained from the copyright proprietors. Pertinent discussion will be published in the May-June 2003 ACI Structural Journal if received by January 1, 2003.

399

S. Ali Mirza, FACI, is a professor of civil engineering at Lakehead University, Thunder Bay, Ontario, Canada. He is a member and past chair of Joint ACI-ASCE Committee 441, Reinforced Concrete Columns; and is a member of ACI Committees 335, Composite and Hybrid Structures; and 340, Design Aids for ACI Building Codes. He was a recipient of the ACI Structural Research Award in 1990. Edward A. Lacroix is a project engineer, Walters Inc., Hamilton, Ontario, Canada. He received his MSc from the University of Manitoba. His research interests include reinforced concrete and composite columns.

Table 1Summary of geometric and material properties of reinforced concrete specimens used*
Properties b x h, in. x in. fc , psi e/h l/h fy, psi rs, % , %
*Number

Minimum values 3.0 x 3.0 2550 0.00 2 39,503 0.80 0.04

Maximum values 17.7 x 17.7 8246 1.25 40 104,500 7.06 4.73

SUMMARY OF PHYSICAL TESTS USED The experimental data used for this study were taken from 354 physical tests reported by Bresler (1960); Bresler and Gilbert (1961); Bunni (1975); Chang and Ferguson (1963); Cusson and Paultre (1992); Drysdale and Huggins (1971); Ernst, Hromadik, and Riveland (1953); Fang, Hong, and Wu (1994); Gaede (1958); Goyal and Jackson (1971); Green and Hellesland (1975); Heimdahl and Bianchini (1975); Hognestad (1951); Hudson (1965); Kim and Yang (1995); Martin and Olivieri (1965); Mehmel et al. (1969); Pfister (1964); Ramu et al. (1969); Razvi and Saatcioglu (1989); Roy and Sozen (1964); Scott, Park, and Priestley (1982); Sheikh and Uzumeri (1980); Todeschini, Bianchini, and Kesler (1964); and Viest, Elstner, and Hognestad (1956). These tests were conducted on reinforced concrete column specimens subjected to pure axial load or axial load combined with symmetrical singlecurvature bending. The failure strength of a test column was taken as the peak strength reached on the load-deflection response. The geometric and material properties of the test specimens are summarized in Table 1 and cover significantly large ranges of the column cross section size, fc , e/h, l/h, fy , rs, and . This helped to examine the effects of different variables on the strength of reinforced normal-strength concrete columns. In this study, the concrete strength fc was defined as the strength obtained from the standard (6 in. [150 mm] diameter by 12 in. [300 mm] high) cylinder tests, or as the equivalent standard cylinder strength computed from cube tests. For some of the physical tests, the cube test strengths instead of the cylinder test strengths were reported. In such cases, the reported strengths were converted to the equivalent standard cylinder strengths by employing the following procedure. Equation (1) was used to first convert the strength of a cube of a given size to the strength of a 4 in. (100 mm) cube, and then to convert the strength of the 4 in. (100 mm) cube to the strength of a 6 in. (150 mm) cube v o 1 3 f = fo 0.58 + 0.42 --- v (1)

of specimens = 354; h = depth of concrete cross section perpendicular to axis of bending; and b = width of concrete cross section parallel to axis of bending. Note: 1.0 in. = 25.4 mm; 1000 psi = 6.895 MPa.

in which fcu is the strength of a 6 in. (150 mm) cube. For SI units, replace 2840 psi with 19.6 MPa. The longitudinal reinforcing steel yield strengths fy were taken as those reported for bar sample tests in individual studies available from the literature. The transverse reinforcement yield strengths were not reported for some studies. In those instances, the yield strength of the transverse reinforcement was assumed to be equal to the yield strength of the longitudinal reinforcing bars. DESCRIPTION OF COMPUTATIONAL METHODS EXAMINED The strengths of column specimens were computed using ACI 318-99, Eurocode 2, and a nonlinear FEM procedure. In these computations, the understrength factors for ACI 318-99 and the so-called partial safety factors (material resistance factors) for Eurocode 2 were taken equal to 1.0. Similarly, no understrength factors, material resistance factors, or both, were applied for computing strengths from the nonlinear FEM procedure. The computed strengths, therefore, represent the unfactored strengths in this study. ACI 318-99 procedure The ACI unfactored axial load strengths for given (test) e/h ratios were computed from the cross section and column axial force-bending moment-strength interaction curves, similar to the ones shown in Fig. 1. These strength-interaction curves were generated for each test column used in this study and were based on the equilibrium of forces and the compatibility of strains. The assumptions given in ACI 318-99 (ACI Committee 318 1999), including the maximum usable strain of 0.003 at the extreme concrete compression fiber, the equivalent rectangular compressive stress block with a stress ordinate of 0.85fc for concrete, the elastic perfectly plastic stress-strain relationship for reinforcing steel, and the moment magnifier equations for slenderness effects, were used in computations of the ACI strengths. The following ACI limitations, however, were ignored in this study: 1. ACI 318-99 specifies that the maximum axial load acting on a tied reinforced concrete column shall be limited to 0.8Po , where Po is the column cross section pure axial load strength, computed using Po = 0.85fc (Ag Ast) + fy Ast (3)

in which fo and vo are the concrete strength and volume of a 4 in. (100 mm) cube, respectively; and f and v represent the concrete strength and volume of the cube of a given size, respectively. Equation (1), which is based on the statistical theory of brittle fracture of solids, was first presented by Bolotin (1969) and is also documented by Mirza, Hatzinikolas, and MacGregor (1979). Once the strength of an equivalent 6 in. (150 mm) concrete cube was obtained, it was converted to the equivalent standard cylinder strength by using Eq. (2), proposed by LHermite (1955) f cu - f f c = 0.76 + 0.2log10 ---------- 2840 cu 400 (2)

in which Ag = gross area of the column cross section; and Ast = total cross-sectional area of the longitudinal reinforcement. The ACI upper limit on axial load is intended primarily to account for accidental eccentricities not considered in the ACI Structural Journal/July-August 2002

design but that may exist in a real structure. As the column specimens used in this study were prepared and tested under controlled laboratory conditions, this limit was not used for computing the ACI strengths; 2. ACI 318-99 also specifies that the cross-sectional area of longitudinal reinforcing bars shall be not less that 0.01Ag. The ACI lower limit on Ast is intended primarily to guard against yielding of the longitudinal reinforcement due to creep and shrinkage of the concrete under sustained compressive stresses at service loads. Because the column specimens used were subjected to short-term loads, this limit was not applied to the ACI computations in this study; and 3. ACI 318-99 further specifies that the slenderness ratio l/h of pin-ended columns subjected to symmetrical singlecurvature bending shall be not more than 30 when such columns are designed using the moment magnifier method. The ACI upper limit on l/h is intended to represent the upper range of physical tests of slender columns used in calibration studies when the moment magnifier method was developed. As some test results for columns with l/h greater than the ACI upper limit are now available in the literature, this limit was ignored herein to investigate whether the moment magnifier approach was applicable to columns that exceeded the maximum l/h permitted by ACI. The cross section strength interaction curve (Fig. 1) for each test column was represented by 102 points. The first point was determined from Eq. (3) for pure axial load strength (e/h = 0), the last point was computed for pure bending moment strength (e/h = ), and the remaining 100 points were distributed uniformly along the interaction curve between these two points. The area of concrete displaced by reinforcing bars was included in computations. The column (member) axial load-bending moment interaction curve (Fig. 1) was developed from the cross-sectional strength interaction curve. The pure axial load strength of a column was first determined as the lesser of the two strengths computed from Eq. (3) and (4) Pc = 2EI/l2 (4)

Fig. 1Schematic cross section and column axial loadbending moment interaction diagrams. EI = 0.4Ec Ig EI = 0.2Ec Ig + Es Ise (6) (7)

in which Ec = modulus of elasticity of concrete taken equal to 57,000 f c psi (4733 f c MPa); Es = modulus of elasticity of reinforcing steel bars taken as 29,000,000 psi (199,955 MPa); and Ig and Ise = moment of inertia of the gross cross section and of the reinforcing bars, respectively, taken about the centroidal axis of the cross section. In an attempt to take into account the cracking and nonlinearity of the concrete stressstain curve in determining EI, Mirza (1990) proposed the following equation for calculating the flexural rigidity of reinforced concrete columns subjected to short-term loads e EI = 0.3 1 -- E I + E s I se E s I se h c g (8)

in which Pc = critical load of a pin-ended column. This established the maximum axial load that could be applied on the column. The cross-sectional bending moment resistance Mcs was then divided by the moment magnification factor ns to obtain the column bending moment resistance Mcol for each point on the cross-sectional interaction curve with axial load Pu less than the column pure axial load strength, as indicated in Fig. 1. The ns factor used in this study was developed for pin-ended columns subjected to symmetrical single-curvature bending from the moment magnifier equations of ACI 318-99, and was taken as 1 - 1.0 ns = -------------------------- Pu l2 - 1 ----------- 2 EI (5)

Note that the stiffness reduction factor of 0.75 used in the ACI ns expression was not included in Eq. (5) for reasons described in a previous part of this section. ACI 318-99 permits two alternative EI expressions for use in Eq. (4) and (5). For columns subjected to short-term loads and used in this study, the ACI EI equations simplify to ACI Structural Journal/July-August 2002

This equation was also included in the comparative study reported herein. Hence, the ACI column strengths were computed in three different ways, using Eq. (6), (7), and (8). For short columns (l/h 6.6 for the type of columns studied), the column strengths were taken equal to the cross-sectional strengths as permitted by ACI 318-99. Hence, the axial load strength Pu of a short column was computed through interpolation from the points generated on the cross-sectional interaction curve, using the e/h ratio evaluated from the test data. Similarly, the axial load strength Pu of a slender column (l/h > 6.6 for the type of columns studied) was determined from the column interaction curve using the test e/h ratio, as indicated in Fig. 1. Software was developed for computing the ACI unfactored strengths based on the ACI 318-99 assumptions and procedure given in this section. Further details are documented elsewhere (Lacroix and Mirza 1998). Eurocode 2 procedure The Eurocode 2 (1992) unfactored axial load strengths were computed in a similar manner as that described in the previous section. The cross-section and column strength 401

indicated in Fig. 2. For the purpose of analysis, Eq. (9) was rearranged as e = etot (ea + e2) (10)

Fig. 2Model column used by Eurocode 2 for determining second-order eccentricities. interaction curves were generated for each test specimen and were based on the assumptions and requirements of Eurocode 2. However, the following exceptions were applied to these computations. Eurocode 2 places an upper limit of 7300 psi (50 MPa) on fc , limits the smaller cross section dimension to a minimum of 8 in. (200 mm), and requires a minimum end eccentricity ratio e/h of 0.05 and 0.10 for design of short columns (l/h < 7.5 for the type of columns studied) and slender columns (l/h 7.5 for the type of columns studied), respectively. As the column specimens used in this study were prepared and tested under laboratory conditions, these requirements were ignored for computing the Eurocode 2 strengths. The Eurocode 2 cross section strength interaction curve (Fig. 1) for each test column was defined also by 102 points generated using a similar procedure as that described for ACI. The Eurocode 2 assumptions for computing the crosssectional strengths are identical to those of ACI 318-99, with the exception that Eurocode 2 uses the maximum strain at the extreme compression fiber of concrete equal to 0.0035, and the depth of concrete stress block equal to 0.8 times the distance from the extreme compression fiber to the neutral axis. The Eurocode 2 column (member) strength interaction curves (Fig. 1) were developed also from the cross-sectional strength interaction curves. The Eurocode 2 method for computing second order effects in columns, however, is different from that of ACI 318-99. Eurocode 2 uses a model column, which is an isolated cantilever fixed at the base and free at the top, as illustrated in Fig. 2. The model column is assumed to be bent in single curvature under the applied axial load and bending moment, producing the maximum bending moment at the base. The column is designed for a total eccentricity etot, which consists of: a) the equivalent first order eccentricity that includes the effect of moment gradient; b) the eccentricity caused by initial imperfections; and c) the second order eccentricity acting on the failure cross section. For the type of columns studied, etot was taken as etot = e + ea + e2 (9)

Equation (10) represents the relationship between the bending moment strength of a column and that of its cross section in the sense that the column and cross-sectional bending moment strengths (Mcol and Mcs in Fig. 1) can be computed as Pu e and Pu etot , respectively. Hence, for each point on the cross section interaction curve, etot was computed as Mcs /Pu, and the sum of ea and e2 was subtracted from etot to obtain the column end eccentricity e. The column bending moment strength was then computed as Pue. Note that any points on the cross section interaction curve that generated negative values of e were not considered for developing the column interaction curve. The Eurocode 2 equations for ea and e2 were simplified to Eq. (11) and (12) for the pin-ended columns used in this study l l - -------e a = --------31.9 400 l 2 k 2 y ----- -----------e 2 = k1 10 0.9 d
2

(11)

(12)

where y = yield strain of reinforcing steel; d = effective depth of the cross section in the direction of stability failure; k1 = coefficient that applies correction to e2 for columns with low slenderness ratio = (l 4.5h)/6h 1.0; and k2 = coefficient that takes into account the decrease in curvature with increasing axial force = (Po Pu)/(Po Pbal) 1.0; Po = pure axial load strength of the cross section computed from Eq. (3); and Pbal = axial load that maximizes the bending moment strength of the cross section computed from 0.4fc (Ag Ast), as permitted by Eurocode 2. For l /31.9 in Eq. (11), ea and l are in in. For SI conversion (ea and l in m), replace 31.9 by 200. Eurocode 2 does not provide a specific equation for determining the pure axial load strength of a column (member). In this study, the Eurocode 2 column pure axial load strength was defined as the condition when the column end eccentricity e, which is computed from Eq. (10), equals zero, and it was determined by interpolating between two adjacent points on both sides of the column pure axial load strength (Fig. 1). For a short column (l/h < 7.5), second order effects were neglected, and the axial load strength was determined for the test e/h ratio from the cross section interaction curve, as permitted by Eurocode 2. Similarly, the axial load strength of a slender column (l/h 7.5) was computed for the test e/h ratio from the column interaction curve, as indicated in Fig. 1. A computer program was developed for computing the Eurocode 2 unfactored strengths based on the procedure and assumptions outlined previously. Further details are documented by Lacroix and Mirza (1998). Finite element modeling (FEM) method The FEM of the strengths and load-deflection responses of reinforced concrete test columns was carried out by using a commercially available nonlinear FEM software (ABAQUS 1994a,b). The software was capable of performing a static second order strength analysis involving both the material and geometric nonlinearities. It included an extensive library of predefined three-dimensional (space) beam sections that ACI Structural Journal/July-August 2002

where e = first order eccentricity of a column subjected to equal and opposite end moments; ea = additional eccentricity that accounts for initial imperfections such as dimensional inaccuracies and the uncertainty in the position of the line of action of the axial load; and e2 = second order eccentricity as 402

were usedthis greatly simplified the data input required to model the reinforced concrete column cross section. A reinforced concrete column cross section was assumed to consist of three different materials, each represented by a different stress-strain curve. These materials included the unconfined concrete outside the transverse tie reinforcement, partially confined concrete within the transverse ties, and longitudinal reinforcing steel bars, as indicated in Fig. 3(a), (b), and (c), respectively. The unconfined concrete was modeled by using a thin-walled box beam section with the inner wall of the box beam section coinciding with the centerline of the transverse tie reinforcement. The partially confined concrete was modeled by using a rectangular beam section with the outer edge of the rectangular beam section coinciding with the centerline of the transverse tie reinforcement and the inner edge of the box beam section. The longitudinal reinforcing bars were modeled by superimposing bar elements within the rectangular beam section mesh. The two beam sections and bar elements were superimposed at common node points to fully define the column cross section, as illustrated in Fig. 3(d). The FEM software accounted for the area of concrete displaced by longitudinal reinforcing bars and numerically integrated the cross section to obtain the generalized force/moment-strain/curvature relationships. As the integration points are used to discretize a beam section and to define its mesh for numerical computations, an increase in integration points leads to a denser mesh, resulting in a more accurate solution but a higher computational time. For this study, the number of integration points was increased substantially from the software default condition. The locations and number of integration points used are shown in Fig. 3(a) through (c) for the three materials noted previously. The modeling of a reinforced concrete column (member) was accomplished by using a predefined, three-dimensional, three-node beam (space) element. The length of the column was divided into a number of segments, each representing a beam element. Each beam element was connected to adjacent elements at the two outer (common) node points. The central node point of the element was generated and used for integration purposes by the software. A typically discretized column subjected to symmetrical single-curvature bending is illustrated in Fig. 4(a). For the purpose of analysis, however, the symmetry about the midlength permitted the use of an equivalent cantilever column that was 1/2 the length of the original column, as illustrated in Fig. 4(b), leading to substantial savings in computational time. The boundary conditions at the top node of the column restrain movements along the x- and z- axes, and the rotation about the y-axis, whereas those at the bottom node of the column restrain movements along the x-, y-, and z-axes and the rotation about the y axis (Fig. 4(a)). The node at the column midlength is restrained from movements along and rotations about the y- and z-axes (Fig. 4(b)). These restraints modeled the boundary conditions used in physical tests of column specimens analyzed in this study. To prevent the localization of the element curvature, the finite element length was taken equal to, or slightly greater than, the overall depth of the cross section in the plane of bending, as indicated in Fig. 4(c). A column was loaded by introducing a set of small applied axial load and bending moment at the top node (Fig. 4(b)), reflecting the end eccentricity used in the physical test of the column specimen. The applied axial load and bending moment were then increased in increments of constant proportions, using a second order analysis procedure, until the failure ACI Structural Journal/July-August 2002

Fig. 3FEM of reinforced concrete cross section: (a) unconfined concrete; (b) partially confined concrete; (c) longitudinal reinforcing steel; and (d) entire cross section.

Fig. 4FEM of reinforced concrete column: (a) column in symmetrical single-curvature bending; (b) 1/2 length of column used for analysis; and (c) finite element segment. occurred. The FEM failure strength of a column was defined as the peak strength reached on the load-deflection response curve. For the analysis of columns subjected to pure axial load, an imperfection was added to the initially straight element model to ensure a smooth transition from column stability to column instability. This is due to the fact that a perfectly straight column subjected to pure axial load remains straight until the critical load is reached and then buckles suddenly. The large deflections associated with this sudden buckling could not be properly captured using the FEM software. In this study, the initial imperfection in pure compression members was introduced by applying a total transverse load equal to 1% of the self-weight of the column distributed uniformly over the entire column length. For the FEM of cross sections, the elastic perfectly plastic stress-strain curves defined by measured values of yield strength fy and modulus of elasticity Es were used for reinforcing steel bars. These fy and Es values were reported in individual studies taken from the literature. The local buckling of the longitudinal reinforcing steel bars was not considered in the FEM analysis because it was assumed to be resisted by the transverse tie reinforcement and the concrete inside the transverse ties. The descriptions of both the unconfined and partially confined concretes in compression outside and 403

stress-strain curves in Fig. 5 were computed from Eq. (13) through (15) fc = 0.97fc [0.89(1 + 0.08log10(0.97fc /t))] Ec = 60,400 0.97 f c [1.16 0.08log10(t)] (13) (14)

fr = 8.3 0.97 f c [0.96(1 + 0.11log10(8.3 0.97 f c /t))] (15) in which fc = compressive strength of standard cylinders (in psi) as tested or computed from Eq. (2); and t = time (in s) taken by a column specimen to reach its failure load. In Eq. (13) through (15), the factor 0.97 represents the effect of in-place casting, and the term inside the square brackets represents the effect of rate of loading, as documented by Mirza, Hatzinikolas, and MacGregor (1979). For SI (MPa) conversion, t should be multiplied by 0.0069 in Eq. (13); 60,400 should be replaced by 5015 in Eq. (14); and 8.3 should be replaced by 0.69 and t multiplied by 0.0069 in Eq. (15). Further details are documented elsewhere (Lacroix and Mirza 1998). STRENGTH ANALYSES AND COMPARISONS The axial load strengths of 354 test specimens were calculated using the five different computational procedures. The tested strengths were divided by the computed strengths to obtain the nondimensionalized strength ratios. These strength ratios were used for statistical analyses and strength comparisons. Comparisons of tested and computed strengths The tested strengths are plotted against the strengths computed from the ACI method using Eq. (6), the ACI method using Eq. (7), Eurocode 2, the ACI method using Eq. (8), and the FEM in Fig. 6(a) through (e), respectively. The strength ratio statistics for each computational procedure are also shown on the figure, as is the histogram of the strength ratios in the inset of the figure for that procedure. The diagonal (45-degree) lines in these figures represent the lines of equality. The data for the FEM plotted in Fig. 6(e) indicate the least scatter about the line of equality, whereas the data for the ACI method involving Eq. 6 plotted in Fig. 6(a) indicate the highest scatter about the line of equality. The data for the remaining three computational procedures plotted in Fig. 6(b) through (d) show less scatter than that in Fig. 6(a) and more scatter than that in Fig. 6(e). These observations are evident also by the degree of tightness exhibited by the strength ratio histograms plotted in the insets of Figs. 6(a) through (e). The mathematical confirmation of these observations is provided by the strength ratio statistics, particularly by the coefficients of variation given in Fig. 6(a) through (e). The coefficient of variation of the strength ratios ranges from the highest value of 0.22 obtained with the ACI method using Eq. (6), to the lowest value of 0.13 obtained for the FEM, with the coefficients of variation for the other three computational methods falling between these two values as indicated in Fig. 6(a) through (e). It should be noted herein that a lower coefficient of variation signifies a more reliable computational method as long as the average value does not deviate significantly lower than unity. The variabilities associated with the different computational procedures studied can be seen more clearly by comparing the cumulative frequency curves of strength ratios plotted on a normal probability scale in Fig. 7. Figure 6 and 7 were prepared from the data for all 354 columns used, ACI Structural Journal/July-August 2002

Fig. 5Schematic concrete stress-strain curves used for FEM: (a) unconfined concrete in compression; (b) partially confined concrete in compression; and (c) reinforced concrete in tension. inside the lateral ties, respectively, were taken from Park, Priestley, and Gill (1982), as illustrated schematically by the stress-strain curves in Fig. 5(a) and (b). The ascending parts, up to the peak stress, and the descending parts, beyond the peak stress, in these curves are represented by second order parabolas and straight lines, respectively. It should be noted that u in Fig. 5(a) is the ultimate compressive strain of unconfined concrete; and K in Fig. 5(b) is a factor that represents the degree of concrete confinement provided by lateral ties. The concrete tensile stress-strain relationship used for the FEM of cross sections was represented by a bilinear curve shown schematically in Fig. 5(c). The ascending branch of this curve, up to the peak stress, was taken from Mirza and MacGregor (1989), and the descending branch representing tension softening after the cracking of concrete was taken from Baant and Oh (1984). In Fig. 5(c), fo is the strain where the tensile strength of concrete equals zero; and Et is the tension softening modulus of concrete taken equal to 70Ec /(57+ fr) psi (0.48Ec /[0.39 + fr ] MPa). The crushing and cracking of concrete were, therefore, modeled using the ultimate compressive strain u and the descending part of the stress-strain curve in Fig. 5(a), and the ultimate tensile strain fo and the descending part of the stress-strain curve in Fig. 5(c). In an attempt to construct a finite element model as complete as is feasible, the physical properties of concrete were modified to account for in-place conditions and rate-of-loading effects. Hence, the compressive strength fc , the modulus of elasticity Ec , and the modulus of rupture fr used for concrete 404

Fig. 6Comparison of tested strengths with strengths computed from: (a) ACI using Eq. (6); (b) ACI using Eq. (7); (c) Eurocode 2; (d) ACI using Eq. (8); and (e) FEM. (Note: 1 kip = 4.448 kN.) demonstrate the overall performance of computational methods examined, and lead to the following conclusions: 1. With an average strength ratio of 0.98 and a coefficient of variation of 0.13, the FEM computes the column strengths more accurately than all other computational procedures studied. This was expected because of the accuracy of the concrete stress-strain curves (Fig. 5) used for the FEM as well as the inherent accuracy of the FEM itself; 2. With an average strength ratio equal to 1.10 and a coefficient of variation equal to 0.22, the ACI method involving Eq. (6) computes the column strengths least accurately of all computational procedures studied; 3. The accuracy of the ACI method using Eq. (7) and that of the Eurocode 2 method are approximately the same, as these procedures produced an average strength ratio of 1.08 and 1.12, respectively, with a coefficient of variation of 0.17 in both cases; and 4. With an average strength ratio of 1.07 and a coefficient of variation of 0.14, the accuracy of the ACI method is improved when Eq. (8) is used in place of Eq. (6) or (7). This is expected because Eq. (8) computes the flexural rigidity (EI) of reinforced concrete columns more accurately than does Eq. (6) or (7) (Mirza 1990). Effects of major variables The effects of the end eccentricity ratio e/h, transverse reinforcement (tie/hoop) volumetric ratio , slenderness ratio l/h, reinforcing steel index rs fy /fc , and concrete strength fc on strength ratios obtained from different computational methods are shown in Fig. 8, 9, 10, 11, and 12, respectively. ACI Structural Journal/July-August 2002

Fig. 7Comparison of probability distributions of strength ratios computed from data for all columns studied (n = 354). Each of these figures was plotted from the data for all 354 column tests and illustrates the strength ratios computed from: a) the ACI procedure using Eq. (6); b) the ACI method involving Eq. (7); c) Eurocode 2; d) the ACI procedure using Eq. (8); and e) FEM. As indicated by Fig. 8(e), 9(e), 10(e), 11(e), and 12(e), the strength ratios obtained from FEM do not appear to be affected by any of the variables investigated. In fact, the data plotted for FEM are clustered almost equally on both sides of the line of equality (horizontal line of unity) over the entire 405

Fig. 8Effect of end eccentricity ratio e/h on strength ratios obtained from different computational methods: (a) ACI using Eq. (6); (b) ACI using Eq. (7); (c) Eurocode 2; (d) ACI using Eq. (8); and (e) FEM (n = 354).

Fig. 9Effect of transverse reinforcement (tie/hoop) volumetric ratio on strength ratios obtained from different computational methods: (a) ACI using Eq. (6); (b) ACI using Eq. (7); (c) Eurocode 2; (d) ACI using Eq. (8); and (e) FEM (n = 354). range of each of the variables examined. This is expected, as the variables studied are explicitly or implicitly included in the FEM analysis. The remaining four computational procedures are affected at least to some degree by some of the variables examined, as indicated by Fig. 8 through 12. The strength ratios for the ACI and Eurocode 2 procedures tend to decrease as e/h increases, with the highest variations in strength ratios taking place for columns subjected to pure compression (e/h = 0), as indicated by Fig. 8(a) through (d). This was expected due to inherent inaccuracies of the code equations in modeling the pure compression failures. It should be noted that the effect of e/h on the strength ratios computed from the ACI procedure using Eq. (8) is minimal 406 when e/h > 0 (Fig. 8(d)). This was also expected, as e/h is included in Eq. (8) used for computing the column EI. Figure 9(a) through (d) indicate an increase in strength ratios computed from the ACI and Eurocode 2 methods as increases. A similar trend can be seen in Fig. 10(a) through (d) with regard to l/h, although to a lesser degree. Again, these trends were expected because the ACI and Eurocode 2 methods do not account for the increase in column stiffness due to the slenderness effect (Mirza 1990) or the increase in column strength due to the concrete confinement provided by lateral reinforcement. As expected, the effects of e/h, , and l/h are the most pronounced for the strength ratios computed from the ACI method using Eq. (6) (Fig. 8(a), 9(a), and 10(a)), ACI Structural Journal/July-August 2002

Fig. 10Effect of slenderness ratio l/ h on strength ratios obtained from different computational methods: (a) ACI using Eq. (6); (b) ACI using Eq. (7); (c) Eurocode 2; (d) ACI using Eq. (8); and (e) FEM (n = 354). Table 2Summary of strength ratio statistics for short columns (l /h 6.6)
Group no. 1 2 3 4
*

Slenderness ratio l/h 3 l/h 3 3 < l/h 6.6 3 < l/h 6.6

End eccentricity ratio e/h = 0 e/h > 0 e/h = 0 e/h > 0

Strength ratio statistics* No. of specimens 10 5 79 19 ACI 1.07 (0.084) 1.10 (0.087) 1.14 (0.105) 0.94 (0.133) Eurocode 2 1.07 (0.084) 1.10 (0.085) 1.14 (0.105) 0.90 (0.140) FEM 0.94 (0.064) 1.06 (0.073) 0.99 (0.110) 0.92 (0.143)

Tie/hoop volumetric ratio = 0.39 2.5% = 0.39 1.9% = 0.04 4.7% = 0.39 1.1%

Statistics shown are average values (and coefficients of variation in parentheses) of ratios of tested-to-computed strengths.

followed by those calculated from both the ACI procedure involving Eq. (7), and the Eurocode 2 method (Fig. 8(b) through (c), 9(b) through (c), and 10(b) through (c)), and become even less significant for the strength ratios computed from the ACI method using Eq. (8) (Fig. 8(d), 9(d), and 10(d)). No clear effects of rs fy /fc and fc can be established from the strength ratios plotted in Fig. 11(a) through (d) and 12(a) through (d), respectively, for the ACI and Eurocode 2 procedures. Figure 8 through 12 and related discussions lead to the following conclusions: 1. The FEM strengths are not affected by any of the variables examined in this study; 2. Whereas the ACI and Eurocode 2 strengths are affected to varying degrees by e/h, , and l/h, no apparent effects of rs fy /fc and fc on the ACI and Eurocode 2 strengths are evident within the examined ranges of these variables; and 3. The effect of e/h on strengths computed from the ACI method involving Eq. (8) is minimal. Strength ratio statistics for columns of different slenderness For the type of columns used in this study, the ACI Code (ACI Committee 318 1999) permits the slenderness effects to be neglected when l/h 6.6. The Eurocode 2 (European Committee for Standardization 1992) raises this limit to nearly 7.5. In this study, the columns were categorized according to the ACI Code and are defined as short columns when l/h 6.6, and as slender columns when l/h > 6.6. The strength ACI Structural Journal/July-August 2002

ratio statistics (average values and coefficients of variation) for the so-defined short and slender columns are given in Table 2 and 3, respectively. The short and slender columns are further categorized into four groups each according to their l/h and e/h ratios, as indicated in Table 2 and 3. It should be noted that Table 3 shows three sets, whereas Table 2 shows only one set of ACI strength ratio statistics, as Eq. (6) through (8) do not affect the strength of short columns. Table 2 and 3 lead to the following conclusions: 1. The ACI and Eurocode 2 methods produce almost identical strength ratio statistics within each of the four groups of the short columns (Table 2). This was expected because ACI and Eurocode 2 use very similar analyses for computing the strengths of column cross sections; 2. The FEM produces a somewhat lower average value but a similar coefficient of variation of the strength ratios than do the ACI and Eurocode 2 procedures within each of the four groups of the short columns (Table 2). This was also expected because the FEM analysis is more comprehensive than the ACI or Eurocode 2 analyses, as described previously; 3. The FEM produces the least and the ACI Code involving Eq. (6) produces the greatest variation in strength ratio statistics (average values and coefficients of variation) among the four groups of the slender columns (Table 3). This reinforces a previous conclusion; 4. When Eq. (8) is used instead of Eq. (6) or (7), the coefficient of variation of the ACI strength ratios reduces, albeit to varying degrees, within each of the four groups of the slender 407

Fig. 11Effect of reinforcing steel index rsfy /fc on strength ratios obtained from different computational methods: (a) ACI using Eq. (6); (b) ACI using Eq. (7); (c) Eurocode 2; (d) ACI using Eq. (8); and (e) FEM (n = 354).

Fig. 12Effect of concrete strength fc on strength ratios obtained from different computational methods: (a) ACI using Eq. (6); (b) ACI using Eq. (7); (c) Eurocode 2; (d) ACI using Eq. (8); and (e) FEM (n = 354). (Note: 1000 psi = 6.895 MPa.) columns (Table 3). Again, this is expected for the reasons described previously; and 5. The ACI and Eurocode 2 methods are very conservative in computing the pure axial load capacity of the two test columns studied with l/h > 30, as indicated by the strength ratio statistics of Group 3 columns in Table 3. The strength ratio statistics are somewhat improved for the 21 similar columns subjected to combined axial load and bending moment (Group 4, with l/h > 30 and e/h > 0, in Table 3). This is particularly valid for Eurocode 2 and the ACI procedure involving Eq. (8). It should be noted that the FEM produces a somewhat low average value of the strength ratios for the Group 4 columns (Table 3). 408 CONCLUSIONS Comparisons of physical test strengths of 354 rectangular reinforced normal-strength concrete columns taken from the literature with the strengths of the same columns calculated from selected computational procedures are presented in this paper. The columns examined were tied, pin-ended, and subjected to short-term loads, producing pure axial force or axial force combined with equal and opposite end bending moments. These comparisons indicate that: a) the strength ratios calculated by the ACI and Eurocode 2 procedures are affected to a varying degree by the column end eccentricity ratio, transverse reinforcement volumetric ratio, and slenderness ratio, whereas the strength ratios computed by the FEM are ACI Structural Journal/July-August 2002

Table 3 Summary of strength ratio statistics for slender columns (l /h > 6.6)
Group no. 1 2 3 4
*

Strength ratio statistics* Slenderness End eccentricity No. of Tie/hoop ratio ratio specimens ACI using Eq. (6) ACI using Eq. (7) Eurocode 2 ACI using Eq. (8) FEM volumetric ratio e/h = 0 6.6 < l/h 30 72 1.14 (0.339) 1.05 (0.192) 1.17 (0.218) 1.04 (0.172) 0.97 (0.125) = 0.04 1.6% 6.6 < l/h 30 l/h > 30 l/h > 30 e/h > 0 e/h = 0 e/h > 0 146 2 21 1.06 (0.129) 2.34 (0.042) 1.12 (0.292) 1.06 (0.143) 2.20 (0.081) 1.09 (0.203) 1.10 (0.145) 1.90 (0.071) 1.19 (0.114) 1.05 (0.118) 1.78 (0.074) 1.05 (0.126) 0.97 (0.044) 1.01 (0.127) = 0.07 2.9% = 0.36%

0.88 (0.146) = 0.12 2.1%

Statistics shown are average values (and coefficients of variation in parentheses) of ratios of tested-to-computed strengths.

not affected by any of the variables investigated; b) the ACI method involving Eq. (6) computes the column strength ratios least accurately, and the FEM computes the column strength ratios more accurately than all other computational procedures examined; and c) the computational accuracies of the ACI procedure using Eq. (7) and the Eurocode 2 method are similar. The comparisons also show that the accuracy of the ACI method is improved when Eq. (8) is used instead of Eq. (6) or (7). NOTATION
Ag Ast At b b d d Ec EI Es Et e e2 ea etot e/h fc , ft fc fc fcu fo fr fy h Ig Ise K l l/h Mcol Mcs n Pbal Pc Po Pu sh t vo ns c fo o, ot = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = gross area of column cross section total area of longitudinal reinforcement area of cross section of tie/hoop bar overall width of column cross section taken parallel to axis of bending outside width of ties/hoops effective depth of cross section in direction of stability failure outside depth of ties/hoops modulus of elasticity of concrete effective flexural rigidity of reinforced concrete column modulus of elasticity of reinforcing steel tension-softening modulus of concrete eccentricity of axial load at column ends (first order eccentricity) second order eccentricity acting on failure cross section additional eccentricity that accounts for initial imperfections total eccentricity of axial load Pu acting on failure cross section end eccentricity ratio compressive, tensile stress of concrete that corresponds to given value of strain concrete strength from standard cylinder tests or equivalent standard cylinder strength from cube tests compressive strength of unconfined concrete used for FEM strength of 6 in. (150 mm) concrete cube strength of 4 in. (100 mm) concrete cube modulus of rupture of concrete yield strength of reinforcing steel bars overall depth of column cross section taken perpendicular to axis of bending moment of inertia of gross concrete cross section taken about centroidal axis moment of inertia of reinforcing steel taken about centroidal axis of cross section factor representing degree of concrete confinement provided by lateral ties used for FEM column length slenderness ratio bending moment strength of column (member) at axial load equal to Pu bending moment strength of column cross section at axial load equal to Pu number of test specimens axial load that maximizes bending moment strength of cross section critical load strength of pin-ended column pure axial load strength of cross section axial load strength spacing of transverse ties/hoops testing time to failure, s volume of 4 in. (100 mm) concrete cube moment magnification factor strain in concrete strain where tensile strength of concrete equals zero compressive, tensile strain in concrete at peak stress, respectively

u y rs

ultimate compressive strain of unconfined concrete yield strain of reinforcing steel understrength (strength reduction) factor tie/hoop volumetric ratio taken equal to 2(b + d )At /bd sh ratio of cross-sectional area of longitudinal reinforcing bars to gross area of cross section (reinforcing steel ratio) rs fy /fc= reinforcing steel index

= = = = =

REFERENCES
Version 5.4 Theory Manual, 1994a, Hibbitt, Karlsson, and Sorensen, Inc., Pawtucket, R.I. ABAQUS Version 5.4 Users Manual, 1994b, Hibbitt, Karlsson, and Sorensen, Inc., Pawtucket, R.I. ACI Committee 318, 1999, Building Code Requirements for Structural Concrete (ACI 318-99) and Commentary (318R-99), American Concrete Institute, Farmington Hills, Mich., 391 pp. Baant, Z. P., and Oh, B. H., 1984, Deformation of Progressively Cracking Reinforced Concrete Beams, ACI JOURNAL, Proceedings V. 81, No. 3, May-June, pp. 268-278. Bolotin, V. V., 1969, Statistical Methods in Structural Mechanics (translated by S. Aroni), Holden-Day Inc., San Francisco, Calif. Bresler, B., 1960, Design Criteria for Reinforced Columns under Axial Load and Biaxial Bending, ACI JOURNAL, Proceedings V. 57, No. 5, Nov., pp. 481-490. Bresler, B., and Gilbert, P. H., 1961, Tie Requirements for Reinforced Concrete Columns, ACI JOURNAL, Proceedings V. 58, No. 5, Nov., pp. 555-569. Bunni, N. G., 1975, Rectangular Ties in Reinforced Concrete Columns, Reinforced Concrete Columns, SP-50, M. W. Huggins, ed., American Concrete Institute, Farmington Hills, Mich., pp. 193-210. Chang, W. F., and Ferguson, P. M., 1963, Long-Hinged Reinforced Concrete Columns, ACI JOURNAL, Proceedings V. 60, No. 1, Jan., pp. 1-25. Cusson, D., and Paultre, P., 1992, Behavior of High-Strength Concrete Columns Confined by Rectangular Ties under Concentric Loading, Department de gnie civil, Facult des sciences appliques, Universit de Sherbrooke, Aug., 39 pp. Drysdale, R. G., and Huggins, M. W., 1971, Sustained Biaxial Load on Slender Concrete Columns, Journal of the Structural Division, ASCE, V. 97, No. ST5, pp. 1423-1443. Ernst, G. C.; Hromadik, J. J.; and Riveland, A. R., 1953, Inelastic Buckling of Plain and Reinforced Concrete Columns, Plates, and Shells, Bulletin No. 3, University of Nebraska, 65 pp. European Committee for Standardization, 1992, Design of Concrete Structures Part 1: General Rules and Rules for Buildings (Eurocode 2), Brussels, Belgium, 253 pp. Fang, I. K.; Hong, K. C.; and Wu, J. Y., 1994, Strength and Ductility of High-Strength Tied Columns, Proceedings of the National Science Council, Part A: Physical Sciences and Engineering, Republic of China, V. 18, No. 1, pp. 63-74. Gaede, K., 1958, Knicken von Stahlbetonstben unter Kurz- und Langzeitbelastung, Versuche durchgefhrt im Institut fr Materialprfung und Forschung des Bauwesens der Technischen Hochschule Hannover, Deutscher Ausschuss fr Stahlbeton, Heft 129, Vertrieb durch Verlag von Wilhelm Ernst & Sohn, Berlin, 77 pp. Goyal, B. B., and Jackson, N., 1971, Slender Concrete Columns under Sustained Load, Journal of the Structural Division, ASCE, V. 97, No. ST11, pp. 2729-2750. Green, R., and Hellesland, J., 1975, Repeated Loading Tests of Reinforced Concrete Columns, Reinforced Concrete Columns, SP-50, M. W. Huggins, ed., American Concrete Institute, Farmington Hills, Mich., pp. 67-91. Heimdahl, P. D., and Bianchini, A. C., 1975, Ultimate Strength of Biaxially Eccentrically Loaded Concrete Columns Reinforced with High-Strength Steel, Reinforced Concrete Columns, SP-50, M. W. Huggins, ed., American Concrete Institute, Farmington Hills, Mich., pp. 93-117.
ABAQUS

ACI Structural Journal/July-August 2002

409

Hognestad, E., 1951, A Study of Combined Bending and Axial Load in Reinforced Concrete Members, University of Illinois Engineering Experiment Station, Bulletin Series No. 399, 128 pp. Hudson, F. M., 1965, Reinforced Concrete Columns: Effects of Lateral Tie Spacing on Ultimate Strength, Symposium on Reinforced Concrete Columns, SP-13, American Concrete Institute, Farmington Hills, Mich., pp. 235-244. Kim, J.-K., and Yang, J.-K., 1995, Buckling Behavior of Slender HighStrength Concrete Columns, Engineering Structures, V. 17, No. 1, pp. 39-51. Lacroix, E. A., and Mirza, S. A., 1998, Comparative Study of Strength Design Methods for Rectangular Reinforced Concrete and Composite SteelConcrete Columns, Research Series Report No. CE-98-2, Lakehead University Department of Civil Engineering, Thunder Bay, Ontario, Canada, 322 pp. LHermite, R., 1955, Ides Actuelles sur la Technologie du Bton, La Documentation Technique du Batiment et des Travaux Publics, Paris, 242 pp. Martin, I., and Olivieri, E., 1965, Test of Slender Reinforced Concrete Columns Bent in Double Curvature, Symposium on Reinforced Concrete Columns, SP-13, American Concrete Institute, Farmington Hills, Mich., pp. 121-138. Mehmel, A.; Schwarz, H.; Kasparek, K. H.; and Makovi, J., 1969, Tragverhalten ausmittig Beanspruchter Stahlbetondruckglieder, Deutscher Ausschuss fr Stahlbeton, Heft 204, Vertrieb durch Verlag von Wilhelm Ernst & Sohn, Berlin, 54 pp. Mirza, S. A., 1990, Flexural Stiffness of Rectangular Reinforced Concrete Columns, ACI Structural Journal, V. 87, No. 4, July-Aug., pp. 425-435. Mirza, S.A.; Hatzinikolas, M.; and MacGregor, J. G., 1979, Statistical Descriptions of Strength of Concrete, Journal of the Structural Division, ASCE, V. 105, No. ST6, pp. 1021-1037. Mirza, S. A., and MacGregor J. G., 1989, Slenderness and Strength

Reliability of Reinforced Concrete Columns, ACI Structural Journal, V. 86, No. 4, July-Aug., pp. 428-438. Park, R.; Priestley, M. J. N.; and Gill, W. D., 1982, Ductility of SquareConfined Concrete Columns, Journal of the Structural Division, ASCE, V. 108, No. ST4, pp. 929-950. Pfister, J. F., 1964, Influence of Ties on Behavior of Reinforced Concrete Columns, ACI JOURNAL, Proceedings V. 61, No. 5, pp. 521-537. Ramu, P.; Grenacher, M.; Baumann, M.; and Thrlimann, B., 1969, Versuche an gelenkig gelagerten Stahlbetonsttzen unter Dauerlast, Institut fr Baustatik, ETH, Zrich, 86 pp. Razvi, S. R., and Saatcioglu, M., 1989, Confinement of Reinforced Concrete Columns with Welded Wire Fabric, ACI Structural Journal, V. 86, No. 5, Sept.-Oct., pp. 615-623. Roy, H. E. H., and Sozen, M. A., 1964, Ductility of Concrete, Proceedings of the International Symposium on the Flexural Mechanics of Reinforced Concrete, ASCE-ACI, Miami, Fla., Nov., pp. 213-235. Scott, B. D.; Park, R.; and Priestley, M. J. N., 1982, Stress-Strain Behavior of Concrete Confined by Overlapping Hoops at Low and High Strain Rates, ACI JOURNAL, Proceedings V. 79, No. 2, Mar.-Apr., pp. 13-27. Sheikh, S.A., and Uzumeri, S. M., 1980, Strength and Ductility of Tied Concrete Columns, Journal of the Structural Division, ASCE, V. 106, No. ST5, pp. 1079-1102. Todeschini, C. E.; Bianchini, A. C.; and Kesler, C. E., 1964, Behavior of Concrete Columns Reinforced with High-Strength Steels, ACI JOURNAL, V. 61, No. 6, June, pp. 701-716. Viest, I. M.; Elstner, R. C.; and Hognestad, E., 1956, Sustained Load Strength of Eccentrically Loaded Short Reinforced Concrete Columns, ACI JOURNAL, Proceedings V. 52, No. 7, Mar., pp. 727-755.

410

ACI Structural Journal/July-August 2002