A Novel Compliance Constrained Mass Optimization Framework for Vehicle Suspension Subframe Structures

Traditionally, vehicle subframe mass optimization process is achieved by an iterative process, which is usually conducted with virtual test using an initial flexible suspension structure while satisfying compliance constraints via multibody dynamic simulation software. The optimization process is typically performed via a multibody dynamic simulation software, and ideally, with an adequate flexible subframe model, the design problem is formulated and solved via a traditional optimization procedure. In reality with a complex model, optimization is in general extremely cumbersome, time-consuming, and with no guarantee of an optimal solution. For this reason, this article presents a novel, rapid, and accurate design framework for vehicle subframe structural mass optimization consisting of five main components: (1) a meta-model is constructed to directly approximate the relationship between the subframe geometry and compliance characteristics of the corresponding suspension. (2) An optimization search is performed on the meta-model directly to identify the subframe geometry with minimal mass and satisfied compliance constraint. In the meta-model, the geometric dimensions of the subframe and suspension compliances are formulated respectively as design variables and constraints in the optimization process. (3) Global sensitivity analysis is then performed to reduce the insignificant design variables of the meta-model. (4) Using discriminant classification analysis, the compliance constraint is modeled as a function of subframe stiffness at the attachment point. (5) The attachment stiffness is a combination of the real bushing element and the equivalent stiffness, which is due to the flexibility of the subframe and can be determined from a finite element analysis based on the geometry. The results presented in this article indicate that a reduction in mass of the optimized geometry can always be achieved despite the initial design and constraint level.