Ben Westeneng (supervised by C-L Lee, G A MacRae), University of Canterbury (EQC funded project 15/U705)
(A thesis submitted in partial fulfilment of the requirements for the degree of Master of Engineering endorsed in Earthquake Engineering accepted in lieu of final report).
This thesis describes a technical study of gusset plate connections in buckling restrained brace (BRB) frames. BRBs are an innovative seismic resisting system with similar behaviour in tension and compression. Some experimental testing has shown that the gusset plate connections at each end of the BRB can buckle earlier than expected. Limitations in the design methods used to predict the gusset plate buckling capacity have been identified. A detailed computer model of the gusset plate was developed and shown to be in good agreement with experimental testing. A number of other gusset plate connections were analysed where it was found that current design methods under predict capacity compared to experimental tests and computer models. A major limitation of some design methods is the assumption the gusset plate to be fixed from rotation and displacement at both ends. Recent research has suggested that the gusset plate should be considered with one end free to rotate and displace. It was determined that this suggestion may also under predict capacity. A new stiffness based method was developed to consider the capacity of the BRB system as a whole, including the gusset plate. A sensitivity study found that buckling capacity significantly decreases if one element of the system has relatively lower stiffness.
This thesis describes a finite element study of BRB gusset plates under in-plane loading to determine how their strengths compare with those of experiments, and with those of methods in current design standards. Limitations of current standards, such as the calibration of the equivalent column method, with its effective length, Whitmore width and Thornton length, and the lack of explicit considerations for frame action effects are described. In addition, a stiffness based BRB system stability method is developed. It is shown that the finite element modelling replicated behaviour and strengths of gusset plates in BRB system tests. Also, a number of other gusset plate connections were analysed where it was found that current design methods predicted greater strengths than that obtained from the analysis, or from experimental tests, due to limitations in the development of these design methods. The major limitation of the design method is the calibration for the equivalent column strut using the Whitmore length and Thornton length together with a particular buckling curve, and an effective length factor. This effective length factor, K, has recommended values ranging from the fully fixed brace case (i.e. K = 0.5) to the fixed-free sway boundary condition case (i.e. K = 2.0), in the literature. In the NZ standard, it is recommended that K = 0.70. Current design methods for the fixed-free sway boundary condition case (i.e. K = 2.0) rather than 0.70 estimated greater capacities than those from previous experiments and analysis. Also, variations of Whitmore width and Thornton length did not always produce the expected changes in performance indicating that they may not be reliable for determining gusset plate buckling capacity. Finite element models of a BRB frame considering frame action forces acting, and not acting, showed that frame action can decrease gusset plate buckling capacity. A simple stiffness based BRB element stability method was developed to determine how different BRB system elements affected buckling capacity. Flexibilities of the frame beam-column joint and all other elements of the BRB system are considered. Effects of residual stresses and yielding from the column curve are considered. The buckling capacity of the system was determined based on the minimum eigenvalue of the system’s stiffness matrix. A sensitivity study found that buckling capacity significantly decreases if one element of the system has relatively lower stiffness.