Optimal objective function for simulating endurance time excitations

Abstract

Endurance Time (ET) method is a dynamic analysis procedure in which increasing excitations are imposed on structures; these excitations are known as Endurance Time excitation functions (ETEF). This study presents a method to find the optimal objective function for simulating ETEFs which unconstrained optimization problems are. In optimization problems, equations are defined in term of an objective function. In the problem of simulating ETEFs, the objective function can be defined in many different ways regarding considered intensity measures and respective weighting factors. In addition, the type of calculating residuals (absolute way or relative way) diversifies objective function definitions. The proposed method for determining optimal objective function includes quantifying the accuracy of ETEFs in a scalar quantity regardless of their objective functions and introducing an approach to overcome the dependence of results on initial points of optimizations. The proposed method is applied and results are then presented. It is observed that considering only acceleration spectra and calculating residuals in the relative way creates more accurate ETEFs.