To address the residual stress induced during the cure of fibre reinforced thermoset polymer composites, two different approaches were suggested for coupling a non-dominated sorting genetic algorithm (NSGA-II) with finite element (FE) simulations based on a viscoelastic constitutive law. These two approaches were proposed with consideration of different ways of integrating NSGA-II and the FE model. In Approach A, NSGA-II was performed based on results from a series of simulations under various combinations of cure variables. Alternatively, Approach B employed NSGA-II to iteratively update and optimise the cure profile for subsequent simulations. Results indicated that both approaches achieved simultaneous reductions in cure time and macroscale residual stress, with Approach B showing further improvements due to the direct coupling between the NSGA-II and simulations. Specifically, the maximum residual stress and cure time optimised by Approach A were reduced by 5%–9% and 22%–50% respectively, while those obtained by Approach B were reduced by 7%–10% and 32%–49% respectively, compared to those based on the manufacturer recommended cure profile. The evolution of stress in composites based on optimised cure profiles from these two approaches was also elucidated. Additionally, microscale modelling further revealed a 3%–5% reduction in the average residual stress within a representative volume element (RVE) model was also shown, depending upon the approach adopted. Ultimately, by combining a NSGA-II and FE simulations, the optimisation of cure time and residual stress at the macroscale and cure time together with a reduction of microscale stress could be realised.
Read full abstract