Control of quality defects of the glass fiber injection molded product based on LSO-RF and NSWOA

This paper proposes a multi-objective optimization method for injection molding, combining the random forest (RF) model optimized using the light spectrum optimizer (LSO) with the non-dominated sorting whale optimization algorithm (NSWOA), to tackle the quality defects of warpage deformation and volume shrinkage commonly observed for the glass fiber injection molded product. Based on the simulation results of central composite face centered design (CCF), the model for control of injection molding quality defects using LSO-RF is constructed. NSWOA is used for multi-objective optimization to obtain the Pareto solutions, and the critic method is employed to determine the optimal process parameter scheme. Then, the optimal process parameter scheme is inputted to Moldflow for simulation verification. The validation results demonstrate that the LSO-RF model achieves prediction errors within 3 % for warpage deformation and volume shrinkage, indicating good predictive accuracy. The optimized warpage deformation and volume shrinkage are 0.3519 mm and 7.0930 %, respectively. Compared to the pre-optimized state, the warpage deformation decreased by 32.57 % and the volume shrinkage decreased by 15.93 %, indicating a significant improvement in molding quality for the glass fiber injection molded product.