Application of neural network and computer simulation to improve surface profile of injection molding optic lens

A search method using neural network algorithm and computer simulation was proposed to solve multi-variables problems such as injection molding process. In the injection molding of optical lenses, injection conditions have critical effects on the optical quality of molded lens. High-precision injection molding techniques are required for the fabrication of plastic optical lens. Therefore, the suggested technique in this study was constructed to search optimum conditions for improving surface profile of optical lenses through restraining porosity creation and minimizing thickness reduction. Simulation for injection molding was conducted focusing on the influence of various process parameters on defects such as porosity creation and thickness reduction. After the simulation, optimum injection molding conditions were predicted using a neural network program based on leaning data extracted from simulation results. To demonstrate the effectiveness of this technique, a series of injection molding experiments were carried out, and experimental results under selected injection conditions were compared with the results output predicted by the neural network.