The stability and ductility of four buckling-restrained braces (BRBs) with brace joints were studied. The load-carrying element of BRB was fabricated with steel (Chinese Q235), and a layer of colloidal silica sheet (0.5 mm in thickness) or four layers of plastic film (0.2 mm in thickness) were used as unbonding materials to provide space to prevent the buckling of inner core in higher modes and facilitate its lateral expansion in case of compression. Based on the equation of BRBs with brace joints of different restrained stiffnesses, the buckling load is calculated considering the initial geometric imperfections and residual stress, and the theoretical values agree well with the experiment results. It is concluded that the buckling load and ductility of BRBs are influenced greatly by the restrained stiffness of brace joints. If the restrained stiffness is deficient, the unstrained segment of BRBs with less stiffness will buckle firstly. As a result, the ultimate load of BRBs decreases, and the maximum compression load is reduced to about 65% of the maximum tension load; the stiffness also degenerates, and there is a long decreasing stage on the back-bone curve in compression phase; the ductility decreases, i.e., the ultimate tension ductility and ultimate compression ductility are approximately 15 and 1.3 respectively, and the cumulative plastic ductility is only approximately 200. If the restrained stiffness of joint is large enough, the stability will be improved as follows: the yielding strength and ultimate strength of BRBs are nearly the same, and there is an obvious strain intensification in both tension and compression phases; the ductility of brace also increases obviously, i.e., the ultimate tension ductility and ultimate compression ductility are both approximately 14, and the cumulative plastic ductility reaches 782.
A 9-story concrete-filled steel tubular frame model is used to analyze the response of joints due to sudden column loss. Three different models are developed and compared to study the efficiency and feasibility of simulation, which include substructure model, beam element model and solid element model. The comparison results show that the substructure model has a satisfying capability, calculation efficiency and accuracy to predict the concerned joints as well as the overall framework. Based on the substructure model and a kind of semi-rigid connection for concretefilled square hollow section steel column proposed in this paper, the nonlinear dynamic analyses are conducted by the alternate path method. It is found that the removal of the ground inner column brings high-level joint moments and comparatively low-level axial tension forces. The initial stiffness and transmitted ultimate moment of the semi-rigid connection are the main factors that influence the frame behavior, and their lower limit should be guaranteed to resist collapse. Reduced ultimate moment results in drastic displacement and axial force development, which may bring progressive collapse. The higher initial stiffness ensures that the structure has a stronger capacity to resist progressive collapse.