THE STRUCTURAL DESIGN OF TALL AND SPECIAL BUILDINGS Struct. Design Tall Spec. Build.

20, 247256 (2011) Published online 16 September 2009 in Wiley Online Library (wileyonlinelibrary.com/journal/tal). DOI: 10.1002/tal.534

Experimental and numerical evaluation of the strength and ductility of regular and cross spirally circular reinforced concrete columns for tall buildings under eccentric loading
Gh. Reza Havaei*, and Abolghassem Keramati
Department of Civil and Environmental Engineering, Amirkabir University of Technology No. 424, Hafez St., Enghelab Av., Tehran, Iran

SUMMARY This research investigates the behaviour of reinforced concrete (RC) columns subjected to eccentric loading (combination of axial loading and bending moment) for tall buildings with experimental and numerical methods. The columns are reinforced and conned with regular spiral and cross spiral (new technique). The new connement technique uses two opposing spirals (cross spirals) to conne RC circular columns in order to enhance their strength and ductility. Fourteen RC circular columns subjected to different load eccentricities with two different grades of spirals steel are experimentally tested. The force, axial and lateral displacement and concrete strains in different locations are measured during the testing. The columns are 1000-mm long with two hunched heads of 500-mm height. The experimental and numerical results are compared and utilized to develop design guidelines of RC columns using the new connement technique. Copyright 2009 John Wiley & Sons, Ltd.

1. INTRODUCTION There have been many studies showing the signicance of lateral reinforcement on the ductility and strength of reinforced concrete (RC) columns in tall buildings. As a column is subjected to compressive loads, the concrete expands laterally. This expansion of concrete is resisted by the lateral connement, which imposes conning stresses on the concrete. The effectiveness of this resistance to expansion is based on the method used to conne the concrete in the column. Currently, the use of spirals is the most common method of connement. Through countless research studies, it was shown that spirals were the most effective means of lateral connement for a column when large deformations (ductility) need to be achieved. Spiral reinforcement acts to resist the lateral expansion by applying a uniform pressure on the concrete core surrounded by the spiral. This uniform pressure signicantly reduces the lateral expansion, which leads to signicant strength and ductility enhancement. However, some limitations have been placed on lateral connement to aid constructability and provide certain minimum levels of ductility. The ACI 318-08 concrete building code (ACI Committee 318, 2008) recommends using spiral reinforcement to conne columns, especially in earthquake resistant structures where ductility is a very important issue. The ACI Committee 318 (2008) species a minimum allowable clear spiral spacing of 25 mm for constructability reasons, which may not be considered conservative and may show difculty in the construction. The mentioned code also species a maximum allowable clear spiral spacing of 75 mm, which may still show difculty in the construction of large columns, as well as long and slender piles. The construction of long and slender RC piles, where no visual inspection can be performed, can be very challenging if the spiral is closely spaced. Hindi and Turechek (2006) tested 12 cantilever circular columns with two different lengths and several spiral spacing and patterns. The columns were subjected to constant axial load and reversed
* Correspondence to: Gh. Reza Havaei, Department of Civil and Environmental Engineering, Amirkabir University of Technology, No. 424, Hafez St., Enghelab Av., Tehran, Iran E-mail: havaei@aut.ac.ir Copyright 2009 John Wiley & Sons, Ltd.

248

GH. R. HAVAEI AND A. KERAMATI

cyclic lateral displacement to study the inuence of the new connement technique on the lateral strength and ductility of circular columns compared to columns conned with conventional single spiral. Six of the columns (1000-mm high and 200 mm in diameter) were designed to study the exural behaviour, and the other six columns (500-mm high and 200 mm in diameter) were designed to study the shear behaviour. Hindi et al. (2005) tested the cross spiral idea in pure axial compression. The results gathered from those tests showed that when compared by volumetric ratios of lateral reinforcement, the regular and the cross spiral connement method showed comparable results in strength and ductility. Also, when a column that had two spirals, each with a spacing of S, was compared to a column with a single spiral with spacing S, the strength was increased slightly while the ductility tended to increase dramatically. The proposed cross spiral connement technique offers the same benets as spiral connement, but it also allows more exibility to better t the needs of the designer or the contractor. Columns with several congurations of lateral connement have been tested by others (i.e., Kim and Park, 1998; Kunnath et al., 1997; Saatcioglu and Grira, 1999; Lambert-Aikhionbare and Tabsh, 2001; Budek et al., 2002; Turechek, 2006), including interlocking spirals, welded wire grids, high-strength and prestressed spirals, and bre wrapping. The objective of this paper is to evaluate experimentally and numerically the strength and ductility of regular and cross spiral circular RC columns under eccentric loading using the new connement technique. Fourteen reduced-scale RC circular columns with different spirals (regular and cross) and two different steel grades for spirals under three different eccentricities of the combination load were tested. The columns were subjected to constant eccentric loading to study the inuence of the new connement technique on the strength and ductility of RC circular columns compared to columns conned with conventional regular spiral.

2. EXPERIMENTAL PROGRAM Fourteen reduced-scale RC circular columns with different spirals (regular and cross) and two different steel grades for spirals were tested. The columns were subjected to eccentric loading to study the inuence of the new connement technique on the strength and ductility of RC circular columns compared to columns conned with conventional regular spiral. The columns were constructed at Amirkabir University of Technology of Tehran, Iran, and tested at the structural laboratory of the Department of Civil Engineering. The tested portion of the columns had a clear height of 1000 mm with hunched heads of 500 mm in height. The specimens were prepared and tested under compression eccentric loading up to failure. The 14 columns were cast in two sets of six each and one set of two each because of various eccentricities, spiral steel grades and spiral types. 2.1. Specimen details The columns considered in this research were built to a scale of 1:3, and only physical dimensions of the column and reinforcing steel were scaled. In order to investigate the proposed cross spiral connement conguration, six of the columns were conned using two opposing or cross spirals (spirals crossing at lateral faces (CSCL) (1-6)), two of the columns were conned using two opposing or cross spirals (spirals crossing at highly stressed tension and compression faces (CSCLR) (1-2)) with only 90 spiral rotation with respect to the main rst six columns. The other six were conned with the standard regular (single) spirals (RSC (1-6)). Table 1 summarizes the specimen names and properties. Each specimen was named using RSC for regular spiral, CSCL for cross spiral and CSCLR for cross spiral with 90 spiral rotation. The numbers following the letters dene different load eccentricities (e/D), e as the eccentricity and D, the diameter of the specimen, and different spiral steel grades. The spiral spacing is 75 mm centre to centre for regular spirals as therefore for cross spirals is 150 mm. For example, CSCL-6 refers to column conned with cross spirals, by 100% eccentricity and lateral spiral yield stress equal to 435 MPa. Figure 1 shows a comparison between the regular spirals and cross spirals. The specimens had an overall diameter of 250 mm, which is one-third of a typical column of 750 mm diameter. As shown in Figure 2, the distance (198 mm) centre to centre of the longitudinal
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 247256 (2011) DOI: 10.1002/tal

CIRCULAR REINFORCED CONCRETE COLUMNS

249

Table 1. Specimen details. Eccentricity Number

1 2 3 4 5 6 7 8 9 10 11 12 13 14

Exp. code
RSC-1 RSC-2 RSC-3 RSC-4 RSC-5 RSC-6 CSCL-1 CSCL-2 CSCL-3 CSCL-4 CSCL-5 CSCL-6 CSCLR-1 CSCLR-2

Type
Regular Regular Regular Regular Regular Regular Cross Cross Cross Cross Cross Cross Cross Cross

mm
125 187.5 250 187.5 250 125 187.5 125 250 125 187.5 250 250 250

%
50 75 100 75 100 50 75 50 100 50 75 100 100 100

Spiral steel grade

G1 G2 G2 G1 G1 G2 G1 G2 G1 G1 G2 G2 G1 G2

28 days ( fc ) (MPa)
30.5 30.7 30.6 30.5 30.3 30.4 30.5 30.6 30.8 30.7 30.5 30.4 30.6 30.5

Test day ( fc ) (MPa)

33.1 33.4 33.2 32.9 33.3 32.8 33.4 32.8 32.9 33.1 33.0 33.2 32.9 33.1

Figure 1. Spiral comparison. rebars was kept constant for columns with cross spirals and regular spiral in order to make a fair comparison in terms of exural capacity. This led to a minimum cover of 20 mm to the outer edge of the outer spiral for columns with cross spirals; however, columns with single (regular spiral) had a minimum cover of 20 mm. The specimen details for the cross spiral and regular spiral columns are shown in Figure 2. The used reinforcements were tested at the Strength of Materials Laboratory of the Department of Mechanical Engineering of the Amirkabir University of Technology. The stress strain curves for the test material are shown in Figure 3. G1 ( fy = 323 MPa) and G2 ( fy = 435 MPa) were used for spirals, and G3 ( fy = 335 MPa) was used for longitudinal reinforcement. Each column had six 12 mm deformed steel bars for longitudinal reinforcement. The longitudinal reinforcement ratio for all columns was 0.014. The spiral spacing used is 75 mm for regular and 150 mm for cross spiral columns. The spiral was made of 6.5-mm diameter smooth bar. The transverse (connement) volumetric reinforcement ratio for all columns was 0.0086. This ratio was calculated based on the columns concrete core measured to the outside edges of spirals. For columns conned
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 247256 (2011) DOI: 10.1002/tal

250

GH. R. HAVAEI AND A. KERAMATI

Figure 2. Specimen details.

Figure 3. Steel stressstrain curves. with cross spirals, the concrete core was assumed to be the area surrounded by the average of the centrelines of the two spirals, which equals to the area surrounded by the outer diameter of the inner spiral, as shown in Figure 2. A ready-mix concrete with pea gravel was used for the specimens. For each column concrete, six standard cylinder specimens taken, three tested at 28 days and the other three tested at the test day (around 90 days). The average concrete strength are shown in Table 1. Each column was outtted with 24 PL-60-11(120 0.3 ) TML Series precision strain gauges. Six of strain gauges were placed on the compression side, eight on the lateral and ten on the tension side of the specimen as shown in Figure 4.
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 247256 (2011) DOI: 10.1002/tal

CIRCULAR REINFORCED CONCRETE COLUMNS

251

a) compression side

b) tension side

c) lateral side

Figure 4. Strain gauges and LVDT arrangement.

2.2. Test setup and loading The columns were tested by use of a hydraulic testing actuator at the structural laboratory of the Department of Civil Engineering. The testing frame setup is shown in Figure 5. The top head of the specimen is vertically adjustable and it is attached to an actuator, while the bottom head is xed. The connections of two heads were prepared as a hinge connection with predened eccentricity. The lateral stability of every specimen in and out of plane was xed by appropriate steel elements to a rigid frame. Figure 4 shows the arrangement of (Linear Variable Displacement Transducer) LVDTs and strain gauges. A total number of six LVDTs and 24 strain gauges were used for every specimen. The specimens were tested using a 600-kN capacity compression actuator, and the data were monitored using an automatic data collecting system. The displacements and strains were monitored by a digital data logger system. The tests were continued up to failure under a monotonically increased load under a displacement control mode. The force, displacements, and strains data were obtained during the test and were led by computer software. 3. EXPERIMENTAL RESULTS AND DISCUSSIONS 3.1. Overall behaviour The behaviour of the columns under eccentric loading had an overall similarity. At the early stages of loading of the specimens, the noise related to the microcracking of concrete was obvious, indicating the start of stress transfer to the conned core. The maximum lateral deection was seen at the midheight of the specimen. The cracking, failure and fracture were gradual and began on tension face of the column. Concrete cracking progressed up and down the specimen. The fracture was initiated at mid-height of the specimens until reaching the ultimate column strength and the decrease of the jack
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 247256 (2011) DOI: 10.1002/tal

252

GH. R. HAVAEI AND A. KERAMATI

Figure 5. Test setup.

Buckling of longitudinal bar a) Regular Spiral b) Cross Spiral c ) Cross Spiral (Rotation)

Figure 6. Regular and spiral column failures. applied load. Inspection of the fractured specimens showed yielding of longitudinal steel bars in the tension face and buckling of longitudinal steel bars in the compression face of the columns. It was clearly observed the better performance of cross spiral columns to the regular spiral columns, mainly on the crack width and spacings. Figure 6 shows the failure and fracture region of conned by regular spiral and conned by cross spiral specimens. 3.2. Loaddisplacement behaviour The loaddisplacement curves of specimens with 75% and 100% eccentricities are shown in Figures 7 and 8. The displacements were measured from the load actuator. All specimens have an approximate linear behaviour before the yield point Y. For example, for the specimen CSCL-6, the tension side longitudinal steel bars are yielded at the force of 100 kN and a displacement of 6.4 mm. The secant stiffness is equal to 16.01 kN/mm. After yielding of the tension bars, the stiffness is decreased but the load capacity is increased to the point F where the compression reinforced buckles and forms a plastic hinge at the force of 120 kN and a displacement of 16.6 mm. The specimens totally have a similar nonlinear behaviour based on Figures 7 and 8. The rst part of all curves roughly is linear to yield point Y where the steel bars in tension edge are yielded. After the cover spalling,
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 247256 (2011) DOI: 10.1002/tal

CIRCULAR REINFORCED CONCRETE COLUMNS

253

a) Steel type G1

b) steel type G 2

Figure 7. Load-displacement behaviour of specimens with 100% eccentricity (e/D = 1.0).

a) Steel type G1

b) steel type G 2

Figure 8. Load-displacement behaviour of specimens with 75% eccentricity (e/D = 0.75).

the spirally concrete core is effectively activated, so second part of all curves have not enormous stiffness degradation. The maximum load carrying capacity of every specimen is achieving at spiral yielding point F. 3.3. Momentcurvature behaviour Figures 9 and 10 show the momentcurvature behaviour of specimens at the mid-height of the test length. The moments were calculated by multiplying the forces by the eccentricities. The curvatures were obtained with dividing the differential longitudinal strain of tension and compression edges per height of the mid-section (250 mm). It shows that the tested cross spiral columns have improved curvature capacities but lower bending stiffness and moment capacity with respect to regular spiral columns. Table 2 shows the rst stiffness, yield point, second stiffness and failure point for all specimens.

4. NUMERICAL RESULTS AND DISCUSSION 4.1. Geometric modelling The ABAQUS (6.8.1) nite element (FE) software was used for 3D modelling conned by spiralling RC columns. The experimental specimens were modelled and subjected to eccentric compression loading.
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 247256 (2011) DOI: 10.1002/tal

254

GH. R. HAVAEI AND A. KERAMATI

a) Steel type G1

b) steel type G2

Figure 9. Moment-curvature behaviour of specimens with 100% eccentricity (e/D = 1.0).

a) Steel type G1

b) steel type G2

Figure 10. Moment-curvature behaviour of specimens with 75% eccentricity (e/D = 0.75).
Table 2. Experiment results. First Stiffness (KN/mm)
30.3 23.72 15.28 23.21 11.94 22.65 18.21 28.15 11.11 24.62 23.97 14.51 16.23 14.23

Yield Point (Y) Force (KN)

236.46 191.45 110.94 169.2 102.82 293.66 153.36 265.5 94.68 243.25 163.02 108.81 96.73 111.86

Specimen
RSC-1 RSC-2 RSC-3 RSC-4 RSC-5 RSC-6 CSCL-1 CSCL-2 CSCL-3 CSCL-4 CSCL-5 CSCL-6 CSCLR-1 CSCLR-2

Disp (mm)
7.8 8.07 7.26 7.29 8.61 12.96 8.43 9.43 8.52 9.88 6.8 7.5 5.96 7.86

Second stiffness (KN/mm)

26.42 18.02 9.191 16.84 6.94 21.00 13.01 26.06 6.80 21.59 15.66 7.30 9.34 8.63

Failure point (F) Force (KN)

286.66 209.98 131.62 184.86 111.49 311.12 174.34 266.88 117.56 263.22 175.27 120.2 121.96 125.24

Disp (mm)
10.85 11.65 14.32 10.98 16.06 14.81 13.4 10.24 17.28 12.19 11.19 16.46 13.05 14.51

A 3D FE mesh used for geometric modelling of concrete column, which consists of 13520 hexahedral solid elements type C3D8 and 1760 wedge solid elements type C3D6 is used for the ABAQUS/ standard analysis. No mesh convergence studies have been performed, but the reasonable agreement between the analysis results and the experimental data suggests that the mesh is adequate to predict
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 247256 (2011) DOI: 10.1002/tal

CIRCULAR REINFORCED CONCRETE COLUMNS

X Y Z

255

Figure 11. FE modelling of specimen.

overall response parameters with usable accuracy. The reinforcement is modelled using 1442 truss elements type T3D2 that are embedded in the host solid elements used for concrete column. Two different boundary conditions are applied on two sides of the column using a loading setup consists of a 12-mm thickness plate, which is modelled as a rigid body with a reference point. FE discretization is shown in Figure 11. 4.2. Material properties The mechanical behaviour of the concrete material is modelled using the concrete damaged plasticity constitutive model. Some of these data are assumed values because they are not available for the concrete used in the experiment. The assumed values are taken from typical concrete data. Of particular interest is the calibration of the concrete tensile behaviour. The tensile strength is estimated to be 10% of the ultimate compressive strength. The concrete has a Youngs modulus of 30 000 MPa, a Poissons ratio of 0.20, a density of 24 000 N/mm3, a cracking failure stress of 3.33 MPa, and a Mode I fracture energy of 0.124 N/mm. The fracture energy value denes the area under the post-cracking stressdisplacement curve. The effect of different post-cracking softening behaviour is the subject of studies carried out in this investigation. The mechanical properties of steel material used for reinforcement modelled using classical metal plasticity constitutive model .The steel has a Youngs modulus of 206 900 MPa, a Poissons ratio of 0.30 and a density of 8750 kg/m3. Two types of reinforcement were used in simulations. 4.3. Boundary conditions and solution Since the displacement control approach is used for the loading of columns, an amplitude curve that consists of values of displacement versus time is used for each of simulations. The reference point for rigid body at the bottom of the column is fully xed, except that the rotation about the Z-axis is not constrained. In the analysis, the reference point for the rigid body at the top of the column moves down in the Y-direction by an amplitude curve dened for each simulation, thus loading it in compression; Furthermore the rotation of the column about the Z-axis is also free for this point. Since considerable nonlinearity is expected in the response, including the possibility of unstable regimes as the concrete cracks, the displacement at the top reference point is applied slowly, and automatic incremental and iterative schemes are used in order to avoid convergence problems. A general static analysis is, therefore, performed to seek the overall response of the concrete column.

5. COMPARISON WITH EXPERIMENTAL RESULTS 5.1. Comparison with experimental results The specimens of RSC-1, RSC-6, CSCL-1, CSCL-5 and CSCLR-2 were modelled and analysed using the numerical method. The load capacity of every specimen was calculated. It was seen that the
Copyright 2009 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 20, 247256 (2011) DOI: 10.1002/tal

256

GH. R. HAVAEI AND A. KERAMATI

numerical results are very close to the experimental results. This means that the numerical modelling is validated by the experimental results. Hence, the numerical method can be used to conduct a parametric analysis.

6. CONCLUSION From the testing of 14 RC columns with regular and cross spiral with the same volumetric spiral ratio subjected to different eccentric compression force, as recorded in the paper, the following conclusions were drawn: (1) The general performance of columns with regular spirals and cross spirals were similar. (2) For e/D between 0.5 and 1.0, the ultimate strength of regular spiral columns with steel grade G2 were about 10 to 15% higher than the similar ones with steel grade G1. (3) For e/D between 0.5 and 1.0, the ultimate strength of cross spiral columns were almost the same, an increase of only 2 to 5% using G2 versus G1. (4) Observing the experimental results of columns CSCLR (spirals crossing at highly stressed tension and compression faces) versus CSCL (spirals crossing at lateral faces) shows an increase of ultimate strength about of 4% for CSCLR. (5) Tested cross spiral columns have improved curvature capacities (higher rotational ductility) but lower bending stiffness and lower moment capacity with respect to regular ones.

ACKNOWLEDGEMENTS

The authors acknowledge the nancial assistances of the Amirkabir University of Technology (Tehran Polytechnic), which enabled conducting this research. The assistance of Prof. A. R. Rahai, the university chancellor, Dr. T. Taghikhani and Dr. S. Erfani, Mr. M.S. Khaleqi and Mr. M.S. A. Nikoukar of the structure laboratory is appreciated.

REFERENCES
ACI Committee 318. 2008. Building Code Requirements for Structural Concrete (ACI 318-08) and Commentary (318R-08). American Concrete Institute: Farmington Hills, MI. Budek AM, Priestley MJN, Lee CO. 2002. Seismic design of columns with high-strength wire and strand as spiral reinforcement. ACI Structural Journal 99(5): 660670. Hindi R, Al-Qattawi M, Elsharief A. 2005. Inuence of different connement patterns on the axial behavior of R/C columns. In Proceedings of the ASCE-SEI 2005 Structures Congress, New York, 2024 April, 2005. Hindi R, Turechek W. 2006. Experimental behavior of circular concrete columns under reversed cyclic loading. Construction and Building Materials 22(4): 684693. Kim J, Park C. 1998. The behaviour of concrete columns with interlocking spirals. Engineering Structures 21: 945953. Kunnath SK, El-Bahy A, Taylor A, Stone W. 1997. Cumulative seismic damage of reinforced concrete bridge piers. Technical Report NCEER-97-0006. University of Central Florida, National Institute of Standards and Technology, Orlando, FL. Lambert-Aikhionbare N, Tabsh SW. 2001. Connement of high strength concrete columns with welded wire fabric. ACI Structural Journal 98(5): 677685. Saatcioglu M, Grira M. 1999. Connement of reinforced concrete columns with welded reinforcement grids. ACI Structural Journal 96(1): 2939. Turechek W. 2006. Cyclic behavior of R/C circular columns conned with opposing spirals. MS Thesis, Department of Civil Engineering and Construction, Bradley University, Peoria, IL.

Copyright 2009 John Wiley & Sons, Ltd.

Struct. Design Tall Spec. Build. 20, 247256 (2011) DOI: 10.1002/tal