A two-stage optimization methodology for gate and vent locations and distribution media layout for liquid composite molding process

Official URL: http://dx.doi.org/10.1016/j.compositesa.2021.106522

The resin flow control strategies of the Liquid Composite Molding (LCM) process are affected by the mold geometry, resin, and preform properties as well as the process parameters such as the gate and vent locations and the placement of a highly permeable layer on the top of the preform. To achieve a successful mold filling process with a minimum level of macro void content and fill time with viscosity variation and the significant flow disturbances due to the race-tracking, a robust and computationally efficient optimization methodology is necessary. In this study, a new two-stage optimization methodology is proposed for successful and accelerated resin impregnation in the LCM process. The two-stage optimization approach provides a robust control tool for successful resin impregnation by mitigating the process-induced risks. The first stage of the optimization is carried out by choosing optimum locations for multiple gates and vents to reduce the void content of the part and fill time of the impregnation as the objectives by considering viscosity variation as a function of time and race-tracking occurrence. Subsequently, the second stage of the optimization is implemented to determine optimum layout for highly permeable distribution media (DM) whereby the resin flow is enhanced thereby leading to the achievement of the target void content and fill time. Genetic algorithm (GA) and Tree Search Algorithm are combined in a sequentially to find the optimum arrangement of gates and vents locations as well as DM layout. The two-stage optimization approach reduces the complexity of a single cycle optimization process which is composed of highly coupled constraints (void content, fill time, gate/vent locations, and DM layout) and thereby results in a robust methodology that ensures targeted void content and minimum fill time with reduced computational cost.