Integration of stochastic process simulation and real time process monitoring of LCM

Liquid Composite Moulding (LCM) and its corresponding sub-processes of filling and curing involve several sources of uncertainty. The present study uses process conditions and material behaviour uncertainty measurements with Monte Carlo (MC) simulation to quantify the effect of variability on the final process outcome. Surrogate models of mould filling and curing based on Kriging have been constructed substituting Finite Elements (FE) solutions to achieve execution of the MC simulation with very small computational effort. Combination of stochastic simulation with on line monitoring results can narrow down gradually the envelope of possibilities predicted as the process progresses. This was carried out in this study using temperature and dielectric sensor signals and an inverse solution scheme based on Markov Chain Monte Carlo (MCMC). The integrated inverse solution is capable of predicting the process outcome with increased levels of certainty as the sensors uncover information gradually during the process. The use of surrogate models allows this solution to be carried out in real time in the manufacturing line.