ACI STRUCTURAL JOURNAL TECHNICAL PAPER

Title No. 119-S95

Strength Reduction Factor for Circular Reinforced

Concrete Columns
by Tomasz A. Lutomirski, Andrzej S. Nowak, and Marta Lutomirska

A reliability-based calibration for new strength reduction factors is performed for loads and load-combination factors spec-
for eccentrically loaded short circular reinforced concrete columns ified by ASCE 7-16 and available statistical data.1,7,8 The
is presented. The statistical parameters of resistance used in the procedure for the adjustment of strength reduction factors
calibration procedure are based on a developed procedure for the to the target reliability index, as well as their variability with
determination of the capacity of columns and Monte Carlo simu-
respect to strain level and dead load to live load plus dead
lations for representative design cases. The reliability indexes
load ratios, are presented.
were calculated for predetermined resistance factors considering
various dead load to live load ratios and all possible failure
modes represented by the steel strain. Then, they were compared INVESTIGATION PROCEDURE
to the target reliability index, which was assumed as βT = 4.0. The calibration procedure for the strength reduction factor
The determined strength reduction factors correspond to load and for circular concrete columns consisted of the selection of
load-combination factors specified by ASCE 7-16 and adopted by representative design cases, the determination of a load
ACI 318-19. They are compatible with the existing provisions for model and resistance model, the selection of a limit state
resistance factors for rectangular columns. function and target reliability index, and, finally, the adop-
tion of the reduction factors, which assure that the reliability
Keywords: calibration; circular columns; reinforced concrete; reliability
model; resistance factor; strength reduction factor.
indexes are the closest to the target value.

INTRODUCTION Selected design cases

Columns perform a fundamental role in structures by The study was performed for selected circular concrete
supporting and transmitting loads from the upper level to the columns, including five different column sizes for each
lower one or through the foundations to the soil. The failure of the five reinforcement ratios (1, 2, 3, 4, and 5%). The
of one column in a critical location may cause the progres- dimensions of the column and the reinforcing bars were
sive collapse of a whole structure, resulting in significant chosen according to Wight and MacGregor.9 Four of the
economic as well as human losses. Current knowledge allows most common concrete compressive strengths were taken
for the adjustment of the resistance and the load factors to into consideration: 27.6, 41.4, 55.2, and 82.7 MPa (4, 6, 8,
the desired level of safety, corresponding to the importance and 12 ksi). In total, 100 cases were analyzed. To narrow
of the structure and consequences of failure. Therefore, it down the scope of the study and make it more practical, the
is important to provide structural engineers with adequate analysis was limited to not-slender columns, because they
design criteria based on reliability analysis, which is a constitute the majority of erected columns. The effect of
recognized rational approach in a design process involving the slenderness of columns is limited by ACI 318-19 using
uncertainties.1-4 A strength reduction factor for rectangular Eq. (1) and (2)
reinforced concrete columns has already been investigated5
and implemented in ACI 318-19.6 Therefore, this study is (klu)/r ≤ 34 – 12(M1/M2) (1)
focused on the strength reduction factor for circular columns
as they are very eagerly used by architects. The reason for (klu)/r ≤ 40 (2)
the common use of circular columns is, first, their aesthetic
appeal, but also better stress distribution in the cross section, where k is the effective length factor for compression
which compensates for more difficult construction. It is also members; lu is the unsupported length of the column or wall;
known that circular columns with spiral transverse rein- r is the radius of gyration of the cross section; and M1/M2
forcement exhibit more ductile behavior when compared to is the ratio of lesser-to-greater factored end moments on a
columns with a rectangular cross section. compression member.

RESEARCH SIGNIFICANCE
The aim of the presented research is the calibration of
ACI Structural Journal, V. 119, No. 4, July 2022.
strength reduction factors for eccentrically loaded short MS No. S-2021-372, doi: 10.14359/51734653, received September 17, 2021, and
reviewed under Institute publication policies. Copyright © 2022, American Concrete
circular reinforced concrete columns, which would be in Institute. All rights reserved, including the making of copies unless permission is
accordance with existing provisions for rectangular columns obtained from the copyright proprietors. Pertinent discussion including author’s
closure, if any, will be published ten months from this journal’s date if the discussion
and could possibly be implemented in ACI 318. The study is received within four months of the paper’s print publication.

ACI Structural Journal/July 2022 303

Fig. 1—Example of interaction diagrams from Monte Carlo simulations.
Load model the approximations involved in the structural analysis and
The assumed statistical parameters for dead load are idealized models. It was based on values recommended by
based on Ellingwood et al.,1 and for cast-in-place concrete Ellingwood et al.1 for spiral columns: bias factor λ = 1.05
elements are λ = 1.05 and V = 0.10. The statistical parame- and coefficient of variation V = 0.06.
ters of live load depend on the type of occupancy and influ- The development of statistical parameters of resistance
ence area. For the scope of this study, the maximum 50-year for short circular concrete columns consisted of the formu-
live load for office buildings with an influence area of 400 ft2 lation of an analytical model for the representation of their
(37.2 m2) was chosen: λ = 1.00 and V = 0.18. behavior and the determination of the variability of nominal
resistance based on force-moment interaction diagrams and
Resistance model Monte Carlo simulations. The interaction diagrams were
The actual resistance of reinforced columns was consid- divided into compression control and tension control zones
ered as a random variable that is a product of nominal resis- and included the characteristic points: axial load, the point
tance Rn and three factors: the material factor (M), fabrica- corresponding to zero strain in the bottom reinforcement,
tion factor (F), and professional factor (P) the balanced failure point, and the point representing the end
of the transition zone (Fig. 1). The developed incremental
R = RnMFP (3) procedure for the calculation of resistance of columns
considered the changing number of bars and their localiza-
The material factor represents the variability of mate- tion (Fig. 2). The position of each bar in the cross section
rial properties such as strength and modulus of elasticity. was calculated with respect to the top of the cross section
Its value was assumed as recommended by Nowak et al.7,8 using Eq. (4)
Bias factors and coefficients of variation for each concrete
compressive strength were the following: for 27.6 MPa Di = R – cos(ψ + β ∙ (i – 1)) ∙ (R – cover – d/2) (4)
(4 ksi), λ = 1.24 and V = 0.15; for 41.4 MPa (6 ksi), λ =
1.15 and V = 0.125; for 55.2 MPa (8 ksi), λ = 1.11 and V = where R is the radius of the cross section; ψ is the angle of
0.11; for 68.9 MPa (10 ksi), λ = 1.09 and V = 0.11; and for initial rotation of the reinforcement; β is the angle between
yield strength of reinforcing steel Grade 60 ksi, λ = 1.13 and the consecutive reinforcement bars; i is the number of the
V = 0.03. The fabrication factor represents variation in the i-th reinforcement bar in the cross section; and d is the diam-
geometry of a component. It was based on values recom- eter of a single reinforcement bar.
mended by Mirza and MacGregor10 and Ellingwood et al.1 The possible eccentricities in the column were described
for the dimensions of the concrete column: λ = 1.005 and as decreasing depths of the compression blocks of concrete,
V = 0.04, and for the dimensions of reinforcing steel bars: from the initial value aI up to the point when the compression
λ = 1.000 and V = 0.015. The professional factor represents block does not exist. Failure zones were related to strains in
the most tensile reinforcement. For each reinforcement bar

304 ACI Structural Journal/July 2022

Fig. 2—Positioning of bars in incremental procedure: (a) ψ = 30 degrees; and (b) ψ = 0 degrees.
and for each depth of the compression block, the strains in Another observation is that the bias and coefficient of vari-
reinforcement are computed using Eq. (5) ation of resistance for concrete circular columns are more
consistent for the higher compressive strengths of concrete
 R  cos(  (i  1))  ( R  cover  d    due to better statistical parameters of higher-strength
 s (i, a )   m  1   concrete. The statistical parameters adopted in the verifica-
 c(a) 
tion of the proposed strength reduction factors were average
(5) values for each given reinforcement ratio within the failure
zone of the column. This assumption came from the obser-
where εm is the extreme compressive strain in concrete; R vation that the statistical parameters of resistance were more
is the radius of the cross section; ψ is the angle of initial associated with the change of reinforcement ratio rather
rotation of the reinforcement; β is the angle between the than the compressive strength of concrete, which was also
consecutive reinforcement bars; i is the number of the i-th confirmed by research conducted by Szerszen et al.5 More
reinforcement bar in the cross section; d is the diameter of details regarding the development of the statistical param-
a single reinforcement bar; and c(a) is the position of the eters of resistance for reinforced concrete columns with
neutral axis due to changing depth of the compression block. circular cross sections and the study on their sensitivity are
The resistance force of the cross section is expressed as described by Lutomirski11 and Lutomirski and Lutomirska.12
the sum of all forces acting in the cross section by Eq. (6),
and the bending moment resistance by Eq. (7) Calibration of strength reduction factor
The limit state function, g, is defined as the boundary
PTotal(a) = ΣiP(i, a) + 0.85 ∙ A(a) ∙ fc′ (6) between the desired performance of a structure (reliability,
safety margin) and the undesired performance of a structure,
  d  which was equal to the probability of failure (Pf)
M Total ( a ) = ∑ i  P (i, a ) ⋅ cos ( ψ + β ⋅ (i − 1)) ⋅  R − cover −  
  2 
g(R, Q) = R – Q (8)
+0.85⋅ A ( a ) ⋅ f c′⋅ Yc ( a )
(7) where g(R, Q) is the limit state function; R is the resistance;
and Q is the load effect.
where i is the number of the i-th reinforcement bar in the The factored load effect was defined using load combina-
cross section; a is the depth of the compression block of tions formulated in ACI 318-19 as the maximum of Eq. (3)
concrete; A(a) is the area of the concrete compressive block; and (4), and it is smaller than the resistance multiplied by the
fc′ is concrete compressive strength; ψ is the angle of initial strength reduction factor
rotation of the reinforcement; β is the angle between the
consecutive reinforcement bars; R is the radius of the cross 1.4D < ϕR (9)
section; d is the diameter of a single reinforcement bar; and
Yc(a) is the distance between the centroid of the compressive 1.2D + 1.6L < ϕR (10)
block and the centroid of the cross section.
Full interaction diagrams from axial compression to where D is the dead load; L is the live load; ϕ is the reduction
pure bending were generated for selected design cases, and factor; and R is the resistance.
for each of them, 10,000 Monte Carlo simulations were The next step in the calibration is the calculation and
performed. An example of simulated interaction diagrams selection of resistance factors. The approach to the estima-
for one design case (diameter 36 cm, 7 ϕ 19 mm, 27.6 MPa tion of the strength reduction factor was based on the target
[diameter 14 in., seven No. 6, 4.0 ksi]) is presented in Fig. 2. reliability index, which was selected as βT = 4.0. It corre-
The derived statistical parameters of resistance are presented sponded to the probability of failure Pf = 3.17 × 10–5 and it
in Table 1. was in accordance with the target reliability index assumed
It was concluded that the coefficient of variation of for calibration of the strength reduction factor of rectangular
resistance changes due to failure zones, and its variation columns by Szerszen et al.5 Such a level of safety margin
decreases with the increase of the reinforcement ratio. took into account potential consequences of failure and the

ACI Structural Journal/July 2022 305

Table 1—Statistical parameters of resistance for circular columns*
Strain in steel reinforcement ε
0.00207 to 0 0 to –0.00207 –0.00207 to –0.005 –0.005 to –0.01
fc′ ρt VR λR VR λR VR λR VR λR
1% 0.15 1.23 0.12 1.23 0.11 1.22 0.12 1.24
2% 0.13 1.22 0.11 1.21 0.11 1.21 0.11 1.21
27.6 MPa
3% 0.12 1.20 0.10 1.20 0.10 1.20 0.11 1.20
(4 ksi)
4% 0.11 1.19 0.09 1.19 0.11 1.18 0.10 1.18
5% 0.10 1.18 0.09 1.17 0.10 1.17 — —
1% 0.13 1.16 0.10 1.15 0.09 1.15 0.10 1.15
2% 0.12 1.15 0.09 1.15 0.09 1.14 0.11 1.13
41.4 MPa
3% 0.10 1.15 0.08 1.14 0.09 1.14 0.10 1.13
(6 ksi)
4% 0.10 1.14 0.08 1.14 0.09 1.13 0.11 1.12
5% 0.09 1.15 0.08 1.14 0.09 1.13 0.09 1.12
1% 0.12 1.11 0.09 1.11 0.09 1.10 0.10 1.10
2% 0.11 1.12 0.08 1.11 0.08 1.11 0.10 1.09
55.2 MPa
3% 0.10 1.12 0.08 1.11 0.08 1.10 0.08 1.10
(8 ksi)
4% 0.09 1.12 0.07 1.11 0.08 1.09 0.09 1.09
5% 0.09 1.12 0.07 1.11 0.08 1.10 0.09 1.08
1% 0.12 1.09 0.09 1.09 0.08 1.09 0.09 1.08
2% 0.11 1.09 0.08 1.09 0.08 1.09 0.10 1.07
82.7 MPa
3% 0.11 1.10 0.08 1.09 0.08 1.08 0.09 1.07
(10 ksi)
4% 0.10 1.10 0.08 1.09 0.08 1.08 0.11 1.02
5% 0.10 1.10 0.07 1.09 0.08 1.08 0.09 1.06
*
Lutomirski.11

Fig. 3—Resistance factors for ρt = 1% versus: (a) steel strain; and (b) load ratio.
cost of increasing (or decreasing) the reliability of a column. The strength reduction factors were calculated for the
The definition of reliability index introduced by Cornell13 target reliability index for all selected design cases and
was used different dead load to dead load plus live load ratios, as
shown in Fig. 3 to 7. The values of the resistance factors
 R  Q vary between 0.75 and 0.95. Analyzing the sensitivity of the
 (11) resistance factor on the load ratio, the change of formula for
 2R  Q2
load combination was clearly visible for the dead load to
where β is the reliability index; μR is the mean value of the dead load plus live load equal to 0.90. Taking into account
resistance; μQ is the mean value of the load; σR is the stan- that the dead load constitutes 20 to 80% of the total load
dard deviation of the resistance; and σQ is the standard devi- carried by the column, the D/(D+L) ratios from 0.20 to 0.80
ation of the load.

306 ACI Structural Journal/July 2022

Fig. 4—Resistance factors for ρt= 2% versus: (a) steel strain; and (b) load ratio.

Fig. 5—Resistance factors for ρt = 3% versus: (a) steel strain; and (b) load ratio.

Fig. 6—Resistance factors for ρt = 4% versus: (a) steel strain; and (b) load ratio.

Fig. 7—Resistance factors for ρt = 5% versus: (a) steel strain; and (b) load ratio.

ACI Structural Journal/July 2022 307

Fig. 8—Reliability index for ρt = 1%: (a) ϕ = 0.75; (b) ϕ = 0.80; and (c) ϕ = 0.85.

Fig. 9—Reliability index for ρt = 2%: (a) ϕ = 0.75; (b) ϕ = 0.80; and (c) ϕ = 0.85.

Fig. 10—Reliability index for ρt = 3%: (a) ϕ= 0.75; (b) ϕ = 0.80; and (c) ϕ = 0.85.
were considered relevant. It is clearly visible that the resis- than 2%, the resistance factor ϕ = 0.75 is recommended, and
tance factors increase for higher reinforcement ratios and for for reinforcement ratios higher or equal to 2%, the recom-
larger eccentricities. mended resistance factor is ϕ = 0.80. As a consequence,
Analysis of the figures led to the selection of three resis- axially loaded columns can have a higher reliability index
tance factors, ϕ = 0.75, 0.80, and 0.85, for which the reli- than eccentrically loaded columns. However, in most design
ability indexes were calculated. Then they were compared cases, columns are eccentrically loaded, and it is reasonable
with the target values to confirm the most appropriate choice. to provide them with the resistance factor, which results in
The plots of reliability indexes with respect to reinforcement the reliability indexes closest to the target value. Recom-
ratios and failure zones were plotted in Fig. 8 to 12. mended strength reduction factors are presented in Table 2.
It can be observed that the reliability indexes calculated
for eccentrically loaded columns are slightly lower than for SUMMARY AND CONCLUSIONS
axially loaded columns. However, it would be inconvenient In this paper, a reliability-based calibration for new strength
for designers to use different resistance factors for axially reduction factors for eccentrically loaded short circular rein-
and eccentrically loaded columns. Therefore, the recom- forced concrete columns was presented. The purpose of
mended resistance factor depends only on the reinforcement the research was to derive strength reduction factors corre-
ratio. For circular columns with reinforcement ratios less sponding to loads and load-combination factors specified by

308 ACI Structural Journal/July 2022

Fig. 11—Reliability index for ρt = 4%: (a) ϕ = 0.75; (b) ϕ= 0.80; and (c) ϕ = 0.85.

Fig. 12—Reliability index for ρt = 5%: (a) ϕ = 0.75; (b) ϕ = 0.80; and (c) ϕ = 0.85.

Table 2—Recommended resistance factors for AUTHOR BIOS

circular reinforced concrete columns Tomasz A. Lutomirski is an Engineer at Gaz-System Company in Poland.
He received his MS from Warsaw University of Technology, Warsaw,
Structural element and limit state Recommended resistance factor Poland, in 2004; and his PhD from the University of Nebraska–Lincoln,
Lincoln, NE, in 2009. He directed his doctoral research to reliability models
Eccentrically loaded circular for circular reinforced concrete columns. His research interests include the
columns, 0.75 reliability of structures and the assessment and repair of existing structures.
reinforcement ratio < 2%
Andrzej S. Nowak, FACI, is a Professor and Department Chair of Civil
Eccentrically loaded circular and Environmental Engineering at Auburn University, Auburn, AL. He
columns, 0.80 received his MS and PhD from Warsaw University of Technology. He is a
reinforcement ratio ≥ 2% member of ACI Committee 348, Structural Reliability and Safety, and Joint
ACI-ASCE Committee 343, Concrete Bridge Design. His research inter-
ests include structural reliability and bridge engineering, and his major
ASCE 7-16 and ACI 318-19, which would be compatible accomplishments include the development of a reliability-based calibration
with current provisions for strength reduction factors for procedure for the calculation of load and resistance factors.
rectangular columns. In the study, a variety of design cases
ACI member Marta Lutomirska is an Assistant Professor in the Faculty
were taken into consideration, including different column of Civil Engineering at Warsaw University of Technology. She received her
sizes, reinforcement ratios, concrete compressive strengths, MS from Warsaw University of Technology in 2003, and her PhD from the
and failure modes represented by steel strain. First, the University of Nebraska–Lincoln in 2009. Her doctoral research referred
to live load models for long-span bridges. Her research interests include
statistical parameters of resistance were calculated based structural reliability and concrete structures.
on the developed procedure for the determination of the
capacity of circular columns and Monte Carlo simulations. ACKNOWLEDGMENTS
The strength reduction factors were calculated assuming the The authors wish to express their gratitude and sincere appreciation to
the late M. M. Szerszen for her support and guidance in conducting this
target reliability index as βT = 4.0. After analysis of their research.
variability, three values were selected: ϕ = 0.75, 0.80, and
0.85, for which reliability indexes were calculated consid- REFERENCES
ering various dead load and live load ratios and all possible 1. Ellingwood, B.; Galambos, T. V.; MacGregor, J. G.; and Cornell,
failure modes. By comparison of the reliability indexes with C. A., “Development of a Probability Based Load Criterion for American
National Standard A58: Building Code Requirements for Minimum Design
the target reliability index, and accounting for the practical Loads in Buildings and Other Structures,” NBS Special Publication 577,
aspect, the most suitable resistance factors were selected. National Bureau of Standards, Washington, DC, 1980, 222 pp.
For the tied reinforced concrete circular columns with a rein- 2. Galambos, T. V., and Ravindra, M. K., “Load and Resistance Factor
Design,” Journal of the Structural Division, ASCE, ST9, Proc. Paper
forcement ratio less than 2%, the resistance factor ϕ = 0.75 is 14008, Sept. 1978, pp. 1335-1335 .
recommended, and for reinforcement ratios higher or equal 3. Nowak, A. S., and Collins, K. R., Reliability of Structures, CRC Press,
to 2%, the recommended resistance factor is ϕ = 0.8. Boca Raton, FL, 2013, 407 pp.

ACI Structural Journal/July 2022 309

 4. Iatsko, O., and Nowak, A. S., “Load and Resistance Factors for 9. Wight, J. K., and MacGregor, J. G., Reinforced Concrete: Mechanics
Concrete Bridges and Buildings,” Dennis Mertz Symposium on Design and and Design, fifth edition, Pearson Prentice Hall, Hoboken, NJ, 2008,
Evaluation of Concrete Bridges, SP-340, A. S. Nowak and H. Nassif, eds., 1126 pp.
American Concrete Institute, Farmington Hills, MI, 2020, pp. 19-31. 10. Mirza, S. A., and MacGregor, J. G., “Variability of Mechanical
5. Szerszen, M. M.; Szwed, A.; and Nowak, A. S., “Reliability Analysis Properties of Reinforcing Bars,” Journal of the Structural Division, ASCE,
for Eccentrically Loaded Columns,” ACI Structural Journal, V. 102, No. 5, V. 105, No. 5, May 1979, pp. 921-937. doi: 10.1061/JSDEAG.0005146
Sept.-Oct. 2005, pp. 676-688. 11. Lutomirski, T. A., “Reliability Models for Circular Concrete
6. ACI Committee 318, “Building Code Requirements for Structural Columns,” doctoral dissertation, University of Nebraska–Lincoln, Lincoln,
Concrete (ACI 318-19) and Commentary (ACI 318R-19),” American NE, 2009.
Concrete Institute, Farmington Hills, MI, 2019, 624 pp. 12. Lutomirski, T. A., and Lutomirska, M., “Variability of Statistical
7. Nowak, A. S.; Szerszen, M. M.; Szeliga, E. K.; Szwed, A.; and Parameters of Resistance for Reinforced Concrete Columns with Circular
Podhorecki, P. J., “Reliability-Based Calibration for Structural Concrete, Cross-Section,” Archives of Civil Engineering, V. LXVII, No. 2, 2021,
Phase 3,” SN2849, Portland Cement Association, Skokie, IL, 2008, 110 pp. pp. 165-180.
8. Nowak, A. S.; Rakoczy, A. M.; and Szeliga, E. K., “Revised Statistical 13. Cornell, C. A., “Bounds on the Reliability of Structural Systems,”
Resistance Models for R/C Structural Components,” Andy Scanlon Sympo- Journal of the Structural Division, V. 93, No. 1, Feb. 1967, pp. 171-200.
sium on Serviceability and Safety of Concrete Structures: From Research doi: 10.1061/JSDEAG.0001577
to Practice, SP-284, P. H. Bischoff and E. Musselman, eds., American
Concrete Institute, Farmington Hills, MI, 2011, 16 pp.

310 ACI Structural Journal/July 2022

Reproduced with permission of copyright owner. Further reproduction
prohibited without permission.