Bidirectional performance of concrete-filled steel tubular column connections.
Thesis DisciplineCivil Engineering
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
Square concrete-filled steel tubular (CFST) columns have been widely used in construction because they have high strength in two horizontal loading directions and they use thin economical steel tubes which provide concrete formwork. However, the performance of the beam-to-column connections in two-way moment frames with square CFST columns is not well understood. Methods commonly used for design are modifications of methods for hollow steel columns and simple, effective and economical design methods for connections in two-way moment frames, which are based on rigorous research and are applicable to a wide range of column sizes, are not available.
In this study, several connection types, which have been previously used and tested, are shortlisted and qualitatively assessed based on their force transfer mechanism, two-way frame suitability, and construction simplicity. Among the candidates, three connection types are selected for further evaluation: (1) external diaphragm connections (EDCs); (2) bolted endplate connections (BECs); and (3) direct-welded connections (DWCs).
A new design method for estimating the tension capacity of the diaphragm plates in external diaphragm connections (EDCs) is developed using simplified finite element models of one diaphragm and one square CFST column. The method considers the effect of bidirectional loading on the tension capacity of the diaphragms. Experimental testing conducted of a threedimensional subassembly with EDC showed no strength degradation. The beam plastic hinges rotations contributed to the structure inelastic deformation; although the diaphragm tension capacity was designed to be slightly lower than the tension force at the beam yield moment. This indicated that the proposed design method is conservative. The effect of bidirectional loading was not significant. The effect of beam shear is then incorporated into the design method. Finite element analyses are conducted on three-dimensional beam-column-joint subassembly models with parametric variations which verify the proposed design methods.
Unstiffened direct-welded connections (DWCs) to square CFST column were tested to small lateral drift ratios to obtain the elastic stiffness of the beam-to-column connections. The connections were later improved with (i) threaded rods passing through the column, and (ii) with washer plates inserted between the tube and the nuts. The improved connections achieved significant stiffness increases. Inelastic mechanisms for the DWCs with washer plates and nuts mainly occurred due to steel tube out-of-plane deformation and fracture near the beam flange tips. Finite element models provided the stress distribution factor in the tube along the beam flanges and methods to compute elastic rotational stiffness of unstiffened and improved directwelded connections were developed.
In this DWC testing, sliding hinge joints with symmetric friction connections were utilized as the beam splices. The initial sliding equivalent coefficient of friction (equal to sliding force divided by the proof load, number of bolts and number of shear interfaces) was found to be 0.34 and that at ultimate strength was greater than 0.45.
The performance of beam bolted end-plate connections (BECs) in a three-dimensional square CFST column subassemblies was investigated experimentally. Through-bolts were used with different bolt-to-flange distances in orthogonal directions. The connections were designed using the AISC and HERA design guidelines. It was shown that up to a drift of 4%, no significant strength degradation occurred. The governing inelastic mechanism mainly involved end-plate yielding near the beam flanges. This mechanism is only considered in the HERA design formulations. The AISC strength formulation was non-conservative for the test conducted. Bolt pretension force decreased up to 40% due to applied cyclic loading. Finite element analysis was with the experimental yield pattern. Also, analyses with bolt tension ranging from 0% to 90% of the nominal proof load indicated the same strength but increasing initial stiffness.
Creep tests of CFST columns subject to axial compression from transversal bonded through-bolts subject to 30% to 100% of the nominal proof load had a reduction of bolt tension force of maximum 5% after eight months for the specimens considered.
The force transfer mechanisms in the beam-column joint were considered and analyzed using strut-and-tie models. The strut-and-tie configurations were developed for joints: (1) without through-bolts and without external diaphragm plates, (2) with through-bolts and without external diaphragm plates, and (3) without through-bolts and with external diaphragm plates. Formulations for critical strut compression forces in the concrete core and tie tension forces in the tube were obtained from nodal equilibrium analyses. Nominal capacities of the strut and the tie are determined based on standard approaches.