Cheng and Chung 2003
A nonlinear force-deformation model to simulate shear transfer behavior in the panel zone of CFT beam-column connections is proposed. In this model, influence of axial load on the shear transfer behavior is accounted for. To validate the proposed theory, five circular CFT beam-column connections were constructed and tested.
Experimental Study, Results, and Discussion
Force-deformation models are developed for the CFT column in the connection panel zone and for the CFT-beam connection. Five circular through-column connections with four types of joint details were constructed; these specimens were of type B, C, D, and E. CFT tubes were 15.7 in. in diameter and made of a 0.4 in. thick steel plate. The beam-column connection consisted of internal and external diaphragms. Four types of connection details were tested. To connect the beam flange to the external diaphragm, single-V groove penetration welds for types B and C specimens and fillet welds for types D and E specimens were used. Column tubes were spliced in the of panel zone for types B and E and on diaphragms at both beam flange levels for type C. The tube walls for all specimens except type D were stiffened with internal diaphragms. The internal diaphragm of types B and E was fillet welded before column splice. External diaphragms in all types were fillet welded to the tube wall except type C specimens.
Tests were performed under constant axial loads that varied by specimen type. Cyclic loads were applied by two actuators at beam tip with a triangular wave of displacement increments. Results showed that all specimens failed due to welding fracture during nonlinear behavior. The authors believe that the residual stress in the tube wall, built up as a result of the welding of the diaphragm, remarkably reduced the deformation capacity of the columns. This would mean that higher axial loads would result in better ductility and larger panel distortion.
A theoretical model is proposed which breaks the connection system into three parts: beams, column, and panel zone. The column section is modeled using nonlinear fiber element analysis. Comparison of test results and predictions from the proposed model for shear transfer in the panel zone were within a reasonable range for high axial loads but tended to be conservative for lower axial loads. The proposed models have good agreement with the tests in terms of initial stiffness and ultimate strength; however, they are unable to simulate post-peak behavior due to residual stress.
Reference
Cheng, C.-T. and Chung, L.-L. (2003). "Seismic performance of steel beams to concrete-filled steel tubular column connections," Journal of Constructional Steel Research, 59(3), p. 405-426. doi:10.1016/S0143-974X(02)00033-0