Abstract:
This paper applies a successive linear response surface method (SRSM) to solve two complex shape optimization problems related to vehicle crashworthiness: knee impact on an instrument panel and head impact on an A-pillar. Geometric pre-processing and adaptive mesh topology are integrated to enable robust and accurate parametric modeling of non-linear impact events. Design variables include both geometric parameters and a discrete variable, enabling effective optimization in regions of high sensitivity and non-linearity. The study demonstrates that simulation-based shape optimization, when paired with adaptive meshing and discrete parameter handling, can lead to significant improvements in safety performance indicators such as knee force and head injury criteria. The approach ensures mesh consistency, improves convergence, and leverages the computational benefits of parallel processing.
