De la caractérisation des matériaux et simulation du procédé à l'optimisation de la fabrication des composites par injection sur renfort

The Liquid Composite Molding techniques (LCM) have become widely used composite manufacturing methods for medium to small production volumes. The objectives of this thesis consist of developing characteristic equations that describe the thermo-mechanical behaviour and phase changes of thermosetting matrix composites, create and implement new numerical models for process simulation and finally, propose a new and comprehensive optimization methodology based on genetic algorithms for the curing and cooling of LCM parts. In these processes, the fibrous reinforcement is progressively saturated by the liquid resin. Once the mold cavity is filled, the thermosetting resin polymerizes and solidifies. From the point of view of the numerical simulation, the process can be divided into three principal stages: (1) filling, (2) curing and (3) cooling and demolding. The material characterisation, numerical simulation and optimization of these three stages are included in the scope of works of this thesis.

The first part of this study concerns the characterization of the cure kinetics and the viscoelastic behaviour of a glass/polyester composite. A semi-empirical model was developed to take into account the effects of inhibitor decomposition and glass transition on cure kinetics. Two models are proposed to describe the chemical and thermal effects on the mechanical properties of the composite. The first one is a thermo-chemical elastic non-linear model that considers total matrix relaxation in the rubbery state above the glass transition temperature. The second one is a viscoelastic model based on the stress relaxation and the time/temperature superposition principle.

A one-dimensional solution of the heat equation coupled with stress analysis has been developed to simulate thermal, chemical and mechanical effects during cure. The calculation of residual stresses and the elastic model have been experimentally verified by molding thin composite plates. A study has been carried out on the evolution of internal stresses during the curing and cooling of thick parts.

The optimization of the pre-heating, curing and cooling stages in the manufacturing of composite parts by LCM is finally treated. This comprehensive approach incorporates several objectives in the same optimization procedure: minimization of cycle time, improvement of mechanical properties and reduction of laminate residual stresses. An evolutionary algorithm based on genetic algorithms, the code called LeCoq (Logical Evolutionary Curing Optimization and Quenching), has been developed to optimize the mold temperature transient profile. Finally, the curing of a thick composite part has been optimized with this algorithm. (Abstract shortened by UMI.)