Numerical Modelling and Optimization of Non-isothermal, Rigid Tool Liquid Composite Moulding Processes

Abstract

The term Liquid Composite Moulding (LCM) is used to define a class of composites manufacturing methods in which a liquid resin is injected into a closed mould containing a dry fibrous reinforcement. Resin Transfer Moulding (RTM) and Compression RTM (CRTM) are rigid-tooled subsets of this class which allow the production of high quality composite parts that are finished to tight dimensional tolerances. However, the design of the manufacturing cycle for these rigid-tool processes involves a large number of design variables which must be carefully chosen in order to minimize cycle time, capital layout and running costs, while maximizing final part quality. These objectives are principally governed by two separate phases of the cycle, namely the resin filling and curing phases, which are strongly coupled in the case of entirely non-isothermal production cycles. The distribution of design variables over largely disparate phases has so far prevented researchers from developing an integrated approach towards complete process optimization. The work presented in this thesis is focussed towards filling this lacuna, and can be divided into two major sections. The first deals with the investigation of accurate Finite Element (FE) models for solving the coupled flow/energy/species equations, which govern the convection-dominated resin flow through the fibrous medium during the filling phase. It is well known that Partial Differential Equations (PDEs) of this kind are difficult to solve using the classical Galerkin FE technique as it leads to spurious oscillations in the field variable. Therefore, a set of advanced stabilized techniques, most of which allow discontinuities either in the FE weighting functions or in the field variable itself, are suitably modified and tested in order to find the one scheme that best suits non-isothermal LCM simulations. In the second part of the thesis a comprehensive framework for optimizing the complete composites manufacturing cycle is developed, one which encompasses both the mould filling and the resin cure phases. Independently optimizing either phase often leads to conditions that significantly restrict or adversely affect the progress of the other. In light of this fact, a novel approach is suggested for optimizing the dual-phase multicriteria problem by modelling it as a static Stackelberg game with two virtual decision makers (DMs) monitoring the filling and curing phases, respectively. The game is simulated through a Bilevel Multiobjective Genetic Algorithm (BMOGA), in conjunction with an Artificial Neural Network (ANN) based surrogate model as a substitute for the otherwise computationally expensive FE mould filling simulation.