BEHAVIOUR OF CONCRETE FILLED STEEL TUBULAR COLUMNS WITH DIFFERENT CROSS-SECTIONS

Abstract

In the modern world, there are severe constraints on building space. The skyrocketing prices of land and construction equipments have fuelled the movement of the construction industry towards high-rise buildings. There is also a need for more space for thoroughfare, which has led to the need for smaller cross-sectional sizes in columns. These new demands have researchers to innovate and develop new construction methods to satisfy these requirements. Concrete Filled Steel Tubular Columns (CFT) have slowly begun to emerge as an alternative to the traditional construction practices. Over the last two decades, CFT columns have been used in many tall structures and bridges all over the world. These columns are constructed by filling the hollow volume of steel tubes with concrete. A composite section such as CFT provides the advantages of both its constituents. Concrete imparts Compressive strength and stiffness while steel tube improves the tensile strength and ductility. The composite action between steel tube and concrete affords many other benefits like improved seismic resistance and reduction in member sizes. The present study investigates the behaviour of different types of Concrete Filled Steel Tubular Columns. Three different core configurations, namely concrete filled single steel tube (CFST), Concrete filled Double steel tube (CFDST) and Reinforced Concrete filled steel tube (RCCCFST) of such columns were investigated. The Double Skin consisted of a hollow steel tube (of smaller diameter) inside an outer steel tube. The annulus between the two tubes was filled with concrete. The RCC-CFST specimens were designed by redistribution of total area of steel. This was achieved by replacing a steel tube of higher wall thickness with a steel tube of lower wall thickness. The difference in area of steel between the two cross-sections was supplied by traditional longitudinal reinforcement. This allowed the author to investigate the difference in behavior of tubular columns having same area of steel but different core configurations, i.e. CFST and RCC-CFST. A total of 81 experimental specimens were tested in this work. Three different cross-sectional shapes which include circular, square and rectangular were used for every core configuration. Lengths, thicknesses and cross-sectional dimensions were also varied for individual specimens of each shape. Structural response of different cross-sectional shapes was studied and the results of all tested specimens were compared. The variations in load carrying capacity, mode of deformation and ductility with respect to core configuration, cross-sectional shape, thickness of steel tubes and length of the specimens were investigated and reported. ii The strength of the specimens obtained from experimental investigations was compared with the nominal load capacity obtained by direct summation of plastic strengths of the constituent materials. Increase in strength (over the nominal capacity) was reported for all tested specimens. The peak enhancement in strength of circular specimens was observed in RCC-CFST specimens, whereas the CFST single skin specimens showed the maximum enhancement in square and rectangular columns. The primary mode of deformation of the composite column was observed as the local buckling of steel tube accompanied by the crushing of concrete core. Breakage of bond was observed between steel and concrete at location of local buckling, leading to loss of confinement. Local buckling of circular specimens was initiated by yielding of steel tube followed by crushing of adjoining concrete core near the mid height in general and extended to the ends of the specimens From the experimental studies on ductility, it was concluded that the circular specimens showed predominantly strain hardening behaviour, while the square and rectangular specimens showed strain softening characteristics. The maximum ductility in circular specimens was observed in CFST specimens. On the other hand, RCC-CFST columns showed higher ductility in square and rectangular columns. The experimental investigations on effect of increase in length of specimens showed that strength and ductility decrease with increase in length of specimens. The decrease in ductility was more in samples with higher thickness of steel tube. The experimental studies on the redistribution of area of steel using longitudnal reinforcing bars as compensation for steel tube showed encouraging results for circular specimens. It was noted that for the circular specimens, it was feasible to replace a tube of higher thickness with a tube of lower thickness (and rebars) while maintaining the strength and ductility. However, loss of ductile behaviour and strength was reported in RCC-CFST square and rectangular specimens. Eight International Codes were used to evaluate the theoretical axial load capacities of tested specimens. The results showed that Canadian Code (S016) and Chinese Code (CECS) give the best estimate for strength of circular specimens. On the other hand, Chinese Codes GJB and CECS gave the closest approximations for square and rectangular columns respectively. iii Numerical simulations were performed for all the tested specimens. Three-dimensional nonlinear finite element models were prepared for all the specimens in the commercial FE software ABAQUS 6.8. The details of the modeling procedure and simulation of contact behaviour between steel and concrete are explained in detail. The numerical models were validated with fifty- four single skin specimens (circular, square and rectangular) selected from the literature. The proposed numerical model was then successfully extended to simulation of Double Skin and Reinforced CFST columns of different cross-sectional shapes and lengths. A new concept of study of confinement pressure was introduced for the circular CFST specimens. Using the results from numerical simulation, an attempt was made to calculate the lateral confining pressure for circular specimens using ABAQUS. The numerically obtained values pressure were then compared with value of lateral confining calculated using theoretical material model from literature. It was seen that the confining pressure for short columns from numerical simulations matched well with the corresponding values from theoretical model. However, the theoretical model was found to overestimate the confinement pressure for composite columns of higher lengths. At the end, an attempt was also made to calculate the strains induced during the actual experimental process, using a new technique known as Digital Image Correlation (DIC). A typical circular specimen was tested and analyzed using this process. The experimental strains were compared with strain values for the same from ABAQUS at specific locations. It was noticed that the values from numerical simulations agreed acceptably with the experimental (DIC) values.