Multi‐Objective Optimization of Injection Compression Molding for Precision Plano‐Convex Lenses Using a Hybrid Approach Integrating Simple‐Attention‐Enhanced CNN and NSGA ‐ III Algorithm

ABSTRACT Injection Compression Molding (ICM) benefits from low injection pressure and uniform compaction, making it effective for producing high‐precision polymer lenses. However, the complex thermo‐mechanical coupling and numerous process parameters pose significant challenges to achieving consistent lens quality. To address this issue, this study optimizes volume shrinkage, warpage, residual stress, and molded mass of a plano‐convex lens in ICM. Key process parameters, mold temperature, melt temperature, injection time, packing time, packing pressure, compression distance, and velocity, are set as variables. A Taguchi‐based orthogonal simulation was conducted, followed by signal‐to‐noise ratio and random forest analysis, which revealed clear trade‐offs among the four quality objectives. Furthermore, a lightweight convolutional neural network incorporating a simple attention mechanism (SimAM‐CNN) was proposed to establish a nonlinear mapping model between the parameters and the multiple quality objectives. Compared with traditional convolutional and backpropagation neural networks, SimAM‐CNN demonstrated higher prediction accuracy and generalization capability for all four quality objectives. The fast elitist non‐dominated sorting genetic algorithm (NSGA‐III) was then employed for multi‐objective optimization, and the equal‐weight Technique for Order Preference by Similarity to Ideal Solution method (TOPSIS) was used to select the optimal process parameter combination from the Pareto optimal solution set. Compared with initial process parameters, the optimized parameters reduced volume shrinkage, warpage, and residual stress by 6.3%, 2.9%, and 8.0% respectively, while improving molded mass accuracy by 2.5%. Finally, high quality plano‐convex lenses were successfully manufactured via ICM. The results demonstrate that the proposed method can effectively enhance the molding quality of polymer lenses.