Robust design of composites manufacturing processes with process simulation and optimisation methods

Due to the increasing variations in raw materials and manufacturing processes, composite manufacturing processes have more part-to-part variations compared with the metal manufacturing processes. To improve part quality consistency, tooling design optimisation is an imperative step for addressing the stochastic behaviour of composite manufacturing processes. This paper presents an optimisation approach for the typical composite manufacturing technique of resin transfer moulding (RTM), which minimises the sensitivity of the mould design to uncertain material properties by choosing appropriate locations of injection gates and vents. This paper proposes a stochastic simulation based approach for the RTM processes. Normal distribution and Weibull distribution were utilised as the statistical models for representing the permeability values for the main region and race-tracking, respectively. Based on the statistical properties of the permeability, a graph-based two-phase heuristic (GTPH) was adopted to minimise the flow dispersion value (a quantitative measure for part quality consistency) such that the process design is not sensitive to the material and process parameter variations.