A novel interpolation scheme for strongly coupled multiphysics topology optimization with explicit sensitivity control

Abstract

Abstract This work presents a novel interpolation scheme for sensitivity control in density-based multiphysics topology optimization, with a focus on acoustic structure interaction problems. We first demonstrate the necessity of the new interpolation in response to challenges identified from prior research, particularly the critical, yet largely unexplored, requirement for explicit control over sensitivities arising from non-physical fictitious fields, which can mislead optimization efforts when using conventional interpolations and, in effect, lead to suboptimal designs. Through several numerical examples concerning acoustic structure interaction topology optimization problems, we highlight the benefits of the introduced interpolations in contrast to conventional interpolations, showing how they lead to better-performing designs and smoother optimization processes. Moreover, the introduced interpolations extend to other multiphysics topology optimization scenarios requiring explicit sensitivity control.