Computational vademecums for a fast and reliable simulation of RTM processes

Abstract

Proper Generalized Decomposition methods allow to obtain an efficient solution for multi-parametric problems without the need for simulating numerous problems to obtain a response surface. Instead, PGD obtains a priori a reduced solution in the form of a finite sum of separable functions, easy to store in memory so as to be evaluated under real-time constraints. The present work proposes to use this tool to optimize the main RTM process parameters, the injection flow rate and the injection/mould temperature, in order to ensure the complete filling of the mould and reasonable fabrication costs (fabrication time, mould heating). To do so, the two process parameters should be introduced in the model as new coordinates, and the Proper Generalized Decomposition method used to solve the multiparametric model then obtained. By using this procedure, we could build computational vademecums, having the two parameters of interest as coordinates, allowing the fabricant to define the best compromise between injection time and process cost (mould heating) while ensuring the complete filling of the mould. In this work, after revisiting some applications of PGD in RTM processes, the separability of parametric RTM solutions will be evaluated.