Analysis and fire resistance design of concrete filled steel tube reinforced concrete columns

Transcription

1 Analysis and ire resistance design o concrete illed steel tube reinorced concrete columns *Lei Xu 1) and Yu-Bin Liu ) 1), ) School o Civil & Architectural Engineering, DaLian Nationalities University, DaLian 1166, China 1) xulei@dlnu.edu.cn ABSTRACT The use o concrete illed steel tube reinorced concrete (CFSTRC) columns has been the interests o many structural engineers. This paper reports a inite element analysis o the compressive behaviour o CFSTRC columns subjected to ire. The inluences o important parameters that determine the ire resistance o the composite columns were investigated. On the basis o parametric analysis outcomes, simpliied ormulae were proposed, And good agreement were obtained either between the results predicted by omulae and numerical simulation results or between the results predicted by omulae and measured results. 1. INTRODUCTION Concrete illed steel tube reinorced concrete (The "Concrete illed steel tube reinorced concrete" will be abbreviated as CFSTRC or the citation purpose.) columns are a kind o composite structures, they taking concrete illed steel tube (CFST) as the core region, and placing reinorcing cage and casting concrete encircle the central region. Their cross sections are typically shown in Fig. 1. Considering the circular steel tube can provide more eective conines to inner concrete and make the structure acquire more avorable mechanic perormance, the paper mainly dealt with the columns with cross sections shown as Fig. 1(a) and 1(b). The structure developed on the basis o reinorced concrete (RC) and CFST structures. It remains advantages o two structures and overcomes their disadvantages. Its main superiorities are as ollowing: (1) compared with RC columns, a substantial increase in load bearing capacity and stiness o the columns and much better endurance characteristics can be obtained; () compared with CFST columns, CFSTRC is expected to have more avorable ire resistance, and the outward buckling and corrosion o the steel tube can also be avoided; (3) CFSTRC columns can be easily connected to RC beams, and higher construction speed the beam-column joint can be designed according to the knowledge o conventional reinorced concrete joint. 1) Proessor ) Proessor 89

2 The structure is innovated and developed irstly by Chinese scholars (Lin 8). Owing to the excellent mechanical perormance o CFSTRC columns, this kind o new column structures has be used more and more extensively in high-rise buildings, super high-rise buidings, and large span buildings. For the moment, more than 3 high-rising buildings have adopted CFSTRC columns (Lin 8). longitudinal bars concrete longitudinal bars concrete longitudinal bars concrete steel tube steel tube steel tube concrete stirrup concrete ( stirrup concrete stirrup a) (b) (c) Fig. 1 Typical cross-section o CFSTRC columns The CFSTRC column has also attracted more and more research interest since 1995 (Lin 8). Through 1 years o the theoretical studies and the constantly extensive use o CFSTRC columns, Technical speciication or steel tube-reinorced concrete column structure(cecs188:5) promulgated and enorced in 5. But above reseaches are all concerned with the mechanic behaviours o the columns under the condition o room-temperater, researches on the mechanical properties exposed (or ater exposure) to a ire o the structures are still in starting phase. It is also lack o ire design and calculation method in the relevant speciications and codes, including CECS188:5. For the moment, only Zhou (6) and Wu (7) have experimentally and theoretically investigated on the axial compressive perormance o the columns ater exposure to ire and the bonding slip behavior o the columns beore and ater exposure to a ire respectively, Liu (7) has numerically simulated the temperature distribution and ire resistance o axial compressive CFSTRC columns. From the literature review, it can be concluded that either the theoretical analysis methods on the mechanical characteristics or a simpliied model that can predict the ire resistance o the CFSTRC columns exposed to ire need to be developed. So, in the paper, a inite element model is proposed irstly. The theoretical method is validated by tests, and the calculated results agree well with those o tests. Secondly, the parametric analysis on ire resistance o CFSTRC columns subjected to ire are carried out by using the inite element model. Finally, or practical application considerations, a simpliied ire resistance ormula o CFSTRC columns is developed. The achievements in the paper can be reerred or the ire design codes making and the ire design o practical engineering.. FINITE ELEMENT MODELING 83

3 A inite element analysis (FEA) model by using ABAQUS sotware (SIMULIA 7) was proposed in this paper to analyze the ire behaviours o the CFSTRC columns..1material properties (1) Steel For steel tubes and reinorcing steel, an elastic-plastic stress-strain relation model proposed by Lie (1993) was used. The details or the stress-strain relationships are described as ollowing: ( T,.1) s s.1 ( T,.1) p.1 6 in which, 4 1 ; p y s p [ T,(.1)] ( T,.1) T.1 5.4T 1 exp 3.3T s P , s p (1) The initial modulus o elasticity or steel was determined by the slopes o the stressstrain curves at high temperatures. The Poisson s ratio was taken as.3 in the model. () Concrete The damage plasticity model or concrete deined in ABAQUS was used in the analysis (SIMULIA 7). The description o plastic behaviours and the yielding surace o concrete were predeined according to the equivalent uniaxial stress strain relationships in the program. Typical stress-strain relationships o core concrete in steel tubes at dierent temperatures was proposed by Song (1), which has been successully used to analyze the ire-resistance behaviour o CFST members. For concrete out o steel tubes, the stress-strain relationships developed by Lie (1993) were used which had been proved to be well used in RC structure analysis. These stressstrain relations are: For core concrete in steel tubes, y x x x 1 (,a).1 1 q ( x 1) 1.1 y x <1.1 x 1 x x>1 (,b) in which, x /, y /, c / T 1, c T 4. 1 T, / x, c 1. 1 is the coninement actor o CFST core o CFSTRC columns, which consist o steel tube and the core concrete in steel tube. It is deined as ollows: 831

4 A c T s y (3) A ck in which, A s is the cross-sectional area o steel tube, A c is the cross-sectional area o concrete, ck is the characteristic compression strength o concrete which equals to 67% o the compression strength o cubic blocks ( cu ), y (T) is the yield strength o steel tube at temperature T, which is deined as: y T<C T.91 y y T C T 1 (4) For concrete out o steel tubes, c c c T T max c 1 max c max 1 3 max c c> max max (5) max.5 6T.4T 1 6 in which, c C<T<45C T c T c C T 874C 1 T>874C c is the cylinder compression strength o concrete. The tensile property o all concrete was deined as a stress and racture energy relationship provided in the concrete damaged plasticity model. The racture energy o concrete at elevated temperature was calculated by the ollowing equation (Lu 11): G 4 6 T G T 11 T (6) in which, G is the racture energy o concrete at ambient temperature, which is deined as: G.7 c ( N / mm) (7) 1 The inluences o thermal expansion o concrete and steel are considered according to that recommended by Lie (1993). 83

5 . Procedure o analysis A sequentially-coupled thermal-stress analysis procedure has been adopted to analyze the mechanical behaviour o CFSTRC columns exposed to ire. This procedure is to simulate the heat transer process irstly, ater that, the acquired nodal temperatures within menmbers will be imported into the structural analysis model to calculate the structural response o members at dierent temperature. In order to ensure the correctly transer o nodal temperatures rom the heat transer model to the mechanical model, the mesh and node numbering o these two models must be identical. The heat transer analysis model o CFSTRC columns has been prosposed by the author o this paper and the details o the model has been described in Xu (1). On the basis o it, the structural analysis will be modelled. In the structural analysis model o CFSTRC columns, the 8-node linear reducedintegration 3D solid element (C3D8R) was used or concrete and steel tube, and the - node linear 3-D truss element (T3D) was used to model the steel reinorcing bars. Two rigid end plates were added to column s model in order to apply load and they also adopted the 8-node linear reduced-integration 3D solid element with (C3D8R) or structural analysis. The interactions o steel tube and concrete either in or out o steel tube were simulated by contact interaction in ABAQUS (SIMULIA 7). The suraces o the concrete and steel in the contact were deined as a contact pair, one as master surace and another as slave surace. The master and slave suraces may contact or separate rom each other. The mechanical properties o the contact pair are deined in normal and tangential direction respectively. The ``hard contact'' model was selected or the normal direction and the Coulomb riction model was selected or the tangential direction. A riction coeicient o.5 was ound suitable or predicting the ire response o CFSTRC columns at elevated temperatures. The normal ``Hard contact'' relation was selected to simulate the interace behaviour between the end plate and concrete. The end plate connected with the steel tube by `TIE' constraint. The boundary conditions will be applied on the end plates to simulate the actual boundary conditions. For pinned boundary condition, the translation UX and UY will be constraint along the loading line on the top end-plate, and the translation UX, UY and UZ on the bottom end-plate will also be constrained. The typical inite element meshing and predicted ailure modes o CFSTRC columns with circular and square cross section by using above model were shown in Fig. (a) and (b), respectively. The ailure modes o steel tube in the columns were also been shown in the Figure. It can be seen rom Fig. that the columns reach their ultimate strength due to overall buckling. Because the existance o the concrete both inside and outside o steel tube, there are neither inward nor outward local buckling occurred in steel tube..3 Veriications o the FEA model A set o tests on CFSTRC columns with both circular and square sections were carried out by the present author in the ormer research stage. The detailed inormation or these tests was given in Table 1 in this paper. The predicted ire resistance o these specimens were also given in Table 1 or comparison. It can be ound the comparison that, a good agreement is obtained between the predicted and test results. 833

6 member CFST core member CFST core (a) square cross-section (b) circular cross-section Fig. Typical inite element meshing and ailure modes o CFSTRC columns Table 1 Testing member inormation and ire resistance Specimen size L B(D) D s (mm) t s (mm) y (MPa) yb (MPa) Concrete outside steel tube cu (MPa) Concrete inside steel tube n Measured ire endurance (min) Predicted ire endurance (min) PARAMETRIC ANALYSIS Above proposed inite element model is used to analyse the eect o parameters on the ire resistance ( ) o CFSTRC columns. All parameters that will inluence the ire resistance o CFSTRC columns include: axial load level (n), dimension o cross-section (C), steel tube ratio (α s ), sectional core area ratio (α sc ), slenderness ratio (λ), eccentricity ratio (e/r o ), yield strength o steel tube ( y ), yield strength o reinorcing bars 834

7 ( yb ), concrete strength ( cu ), ratio o longitudinal reinorcement (ρ s ). In the ollowing, the inluencing law o these parameters on ire resistance will be discussed through a typical example. The basic calculating conditions o the example are as ollows: the columns subjected to ISO-834 (1999) standard ire, initial delection is L/1, C=mm, s =.6, sc =., ρ s =1.%, y =345Mpa, yb =335Mpa, the cubic strength o concrete out o steel tube is 6Mpa,the cubic strength o concrete in steel tube is 4Mpa, =4. The eect o axial load level (n) on ire resistance is shown in Fig. 3. Axial load level has a signiicant inluence on the ire resistance o CFSTRC columns. Fire resistance decreases dramatically as the axial load level increases. When axial load level reach to.8, the columns will ailure ater exposing ire about 15 minutes. The eect o outer perimeter (C) o CFSTRC columns on ire resistance is shown in Fig. 4. It seems that the increase in the outer perimeter leads to a obviously increase in the ire resistance. And the ire endurance time o CFSTRC columns with square crosssection is longer than CFSTRC columns with circular cross-section when their outer perimeters are equal. The parameter o sectional core area ratio (α sc ) o CFSTRC columns relects the thickness o outer concrete insulating layer, and the eect o the parameter on ire resistance is showned in Fig. 5. It can be ound that the ire resistance o the CFSTRC columns increase as the sectional core area ratio increases when the value o α sc is less than.35. Once α sc is greater than.35, the ire resistance will decrease with the increase o α sc. Fig. 6 shows the inluence o slenderness ratio (λ) on the ire resistance o CFSTRC columns. Fire resistance decreases obviously as the slenderness ration increases when λ is less than 4, but the decreasing trend tends to be gently when λ exceeds 4. The eect o other parameters including steel tube ratio (α s ), eccentricity ratio (e/r o ), yield strength o steel tube ( y ), yield strength o reinorcing bars ( yb ), the strength o concrete in steel tube and the strength o concrete out o steel tube ( cu ), ratio o longitudinal reinorcement (ρ s ) on ire resistance o CFSTRC are shown Fig respectively. In general, the eect o these parameters on ire resestance is slight n n Fig. 3 The eect o axial load level on ire resistance 835

8 C C Fig. 4 The eect o outer perimeter on ire resistance α α sc Fig. 5 The eect o sectional core area ratio on ire resistance λ λ Fig. 6 The eect o slenderness ratio on ire resistance 836

9 α s (%) 5 1 α (%) Fig. 7 The eect o steel tube ratio on ire resistance e/r e/r Fig. 8 The eect o eccentricity ratio on ire resistance Fig. 9 The eect o yield strength o steel tube on ire resistance 837

10 Fig. 1 The eect o yield strength o reinorcing bars on ire resistance Fig. 11 The eect o the strength o concrete in steel tube on ire resistance Fig. 1 The eect o the strength o concrete out o steel tube on ire resistance 838

11 ρ (%) ρ (%) Fig. 13 The eect o longitudinal reinorcement ratio on ire resistance 4. PRACTICAL DESIGN FORMULA FOR FIRE RESISTANCE Due to the protection o the concrete out o steel tube and the interaction o steel tube and core concrete, the CFSTRC columns has a avorable ire resistance. By using the analysis model proposed in the paper, the ire resistance o the columns can be calculated. But the calculation procedure tends to be complicated, so the method is not suitable or practice application. In order to provide reerence or the ire design o practical engineering, the simpliied calculating ormulae are proposed. From the analysis in section 3, the results can be conclude that axial load level (n), dimension o cross-section (C), sectional core area ratio (α sc ) and slenderness ratio (λ) are the mainly inluence parameters on the ire resistance o CFSTRC columns. Based on these parameters, a great quantity o calculation and simulation were carried out. On the oundation o these analyses, the practical calculating ormulae are put orward. In the range o normal parameter in the practical engineering, that is, s =.4~., y =35 ~4MPa, cu =3~1MPa,C=1~6mm,λ=1~, the ire resistance can be calculated by the ollowing ormulae: For square cross-section: n.745n n 1 in which,.763 n 1, 6. n sc , C t 3 sc C /, / 4, n n /.6, sc sc R 1.9C e (8,a) /. 839

12 For circular cross-section: n in which,.188n n sc , C t R 1.59C e 1 (8,b) n n sc C /, / 4, n n /.6, sc sc /. The ire resistances o CFSTRC columns predicted using Eq. (8) are compared with numerical model predicted results in Fig. 14(a) and (b) or CFSTRC columns with circular and square sections respectively, there is a good agreement. Fig. 15 shows the predicted ire resistances by Eq. (8) agree well with the experimental results. ire resistance by Eq. (8) (min) ire resistance by numerical model (min) ire resistance by Eq. (8) (min) ire resistance by numerical model (min) Fig. 14 Comparison o ire resistance between Eq. (8) and numerical analysis results ire resistance by Eq. (8) (min) ire resistance by testing(min) Fig. 15 Comparison o ire resistance between Eq. (8) and test results. 84

13 5. CONCLUSIONS Based on the results o this study, several conclusions can be drawn within the limitations o the research work in this paper: 1. A inite element analysis (FEA) model was established to predict the ire resistance o CFSTRC columns under axial compression. The calculated results using this model shows good agreement with test results.. Axial load level (n), dimension o cross-section (C), sectional core area ratio (α sc ) and slenderness ratio (λ) are the mainly inluence parameters on the ire resistance o CFSTRC columns: (1) The ire resistance decreases dramatically as the axial load level increases, () it will increase obviously with the increase o the outer perimeter, (3) when λ is less than 4, the ire resistance decreases obviously as the slenderness ration increases, but the decreasing trend tends to be gently when λ exceeds 4, (4) when the value o α sc is less than.35, the ire resistance increases as the sectional core area ratio increases, whereas the ire resistance will decrease with the increase o α sc when the value o α sc is greater than A simpliied model is developed to calculate the ire resistance o CFSTRC columns. The results predicted by simpliied model are compared with the numerical simulation results and tested results respectively. It was ound that the predicted results by simpliied model are in good agreement with both testing results and numerical simulation results. REFERENCES CECS188:5 (5), Technical speciication or steel tube-reinorced concrete column structure, Beijing, China. ISO (1999), Fire-resistance tests-elements o building construction-part 1: general requirements, International Standard ISO 834, Geneva. Lie, T.T. and Denham, E.M. (1993), Factors aecting the ire resistance o circular hollow steel columns illed with bar-reinorced concrete, NRC-CNRC Internal Report, No.651, Ottawa, Canada. Lin, L.Y. and Li, Q.G. (8), Design concept and analysis o technical economy or steel tube reinorced concrete column, Building Str, 38(1), Liu, D.P. (7), Nonlinear analysis o axially compressed concrete steel tube composite columns illed (CFST) under ire, Xian, China. Lu, H., Zhao, X.L. and Han, L.H. (11), FE modelling and ire resistance design o concrete illed double skin tubular columns, J Constr Steel Res, 67(11), SIMULIA (7). ABAQUS/standard user's manual, Version 6.7. SIMULIA, Providence (RI). Song, T.Y., Han, L.H. and Yu, H.X. (1), Concrete illed steel tube stub columns under combined temperature and loading, J Constr Steel Res, 66(3), Wu, B. (7), Study on bond prosperity at the interace o composite column with core o concrete illed steel tube round about ire, Shanghai, China. 841

14 Xu, L. and Sun, J.G. (1), Temperature Field Calculation and Analysis within Steel Tube Reinorced Columns, OCEJ, 1(6), 15-. Zhou, J. (6), Study on the column reinorced by inner circular steel tubes under axial orce in normal temperature and ater ire, Zhejiang, China. 84