Detecting Snap-Through Boundaries Using an Adaptive Parameter Space Approach

Abstract

Parametric space investigation for nonlinear response of post-buckled structures is essential in understanding the dynamics and avoiding potential catastrophic failures. Typically, continuation type approaches are utilized for such studies. However, in case of randomly forced structures, brute force simulations are utilized to understand the parameter space of interest. In this paper an adaptive meshing approach is used in conjunction with parameter space sampling so as to save computational resources and yet provide enhanced insight in to the nonlinear phenomenon of interest. It is shown that the detection of snap-through response boundaries for a flat, buckled beam under random excitation can be efficiently and accurately generated using this approach. Some observations about the accuracy and performance of the approach are presented as well as comparison to a common brute force sampling algorithms.