Multi-objective optimization of energy-dissipating steel plate fuse links using response surface method

Abstract

Steel plates with openings are used to provide energy dissipation in earthquake-resistant structures. Shear-acting steel plates with butterfly-shaped or straight links are particularly used as easily replaceable fuses in low-damage design of structures. Sensitivity analyses have shown that the cyclic response of steel plate fuses varies depending on the shape of the fuse link. However, simultaneous response optimization approaches have not been applied in the past research on steel plate fuses. A multiple-response optimization study is needed to find optimal link shapes with improved cyclic response characteristics while simultaneously considering different optimization objectives. In this study, experimentally validated finite element simulations and response surface methodology are used to predict the cyclic load-displacement response of steel plate fuses. The behavior of steel plate shear links is evaluated in terms of initial stiffness, yield strength, ultimate stiffness, effective damping, peak strength, and ductility. Predictive equations are derived and verified for each response characteristic in terms of three influential factors, including the length to end-width ratio (L/b), end-width to thickness ratio (b/t), and mid-width to end-width ratio (a/b) of shear links. The predictive equations are then employed to perform multi-objective optimization studies using a desirability approach. Four different optimization objectives are considered, such as the maximization of effective damping and ductility. Optimal ranges are determined for each factor. The results show that the energy dissipation, strength, and ductility of the fuse can be simultaneously maximized if using relatively thick and moderately long links (with b/t smaller than 6.6, L/b between 2.7 and 6, and a/b between 0.33 and 0.96). Optimal fuses links are those with an intermediate a/b between 0.37 and 0.89 and a small to moderate b/t between 3.0 and 6.6.