Intelligent Optimization of the Replication of Injection Molding Light Guide Plates Using Rapid Mold Surface Inducting Heating

This study applies a soft computing based optimization methodology, Evolutionary Regional Neural network with Genetic Algorithm, (ERNGA), to the parameter optimization of the injection molding of light guide plates. High frequency induction heating is applied to control the surface temperature of mold insert to improve the replication of micro features. The proposed scheme first establishes a neural network (NN) model from a small experimental design to simulate the system response, and searches for the model optimum using genetic algorithm (GA). To avoid imperfection of model due to small learning samples, this work sets up an evolutionary network model to confine the search of quasi-optimum using GA. The verification of the searched optimum serves as additional training samples to retrain the evolving model, and the process iterates until the reach of convergence. The design of a 2-inch light guide with micro V-grooves is used to illustrate the application of the proposed methodology. Converged optimum is reached at 12 iterations with an outstanding feature replication ratio of 92%, which outperforms the result from Taguchi¡¦s method and demonstrates the design efficiency.