Dimensional accuracy of thermoset composites: Shape optimization

We present a design sensitivity analysis, which supports the shape optimization of the tool used in thermoset composite manufacturing. The method provides a systematic way to predict the optimal tool geometry that leads to the minimum difference between desired and final shapes of the manufactured part. The design process is formalized by integrating process modeling, design sensitivity analysis, and numerical optimization into a single framework. The process modeling is based on a transient coupled thermo-chemo-viscoelastic finite element scheme, which accurately captures the evolution of residual stresses throughout the manufacturing process, and their effect on the final shape of the composite part. Design sensitivity information is extracted efficiently from the primal analysis using an analytical, direct differentiation method. The sensitivities are then provided to a numerical optimization program to improve the tool design. Optimization results are presented for two specific applications involving mold design for cross-ply laminates and L-shaped composite parts.