Optimisation of material composition in functionally graded plates for thermal stress relaxation using statistical design support system

Abstract

Abstract This study addresses the optimisation of material composition in a functionally graded plate for thermal stress relaxation, subjected to through-thickness thermal gradients, with the aim of minimising a stress utilisation ratio. We simplify the problem by approximating the functionally graded plate as a multi-layered plate. Material compositions in individual layers are optimised using a statistical design support system (SDSS), incorporating design of experiments and mathematical programming techniques. The volume fractions of ceramic constituent in the respective layers are considered as design variables, and an analytical solution for transient thermal stresses is utilised to evaluate the objective function. The optimisation results obtained using the SDSS are compared with those from a genetic algorithm (GA) to validate the applicability of the proposed method. Our findings indicate that the SDSS replicates the ceramic volume fraction distribution optimised by the GA, while significantly reducing optimisation time.