Quality analysis and prediction for multi-phase multi-mode injection molding processes

In batch processes, multiple process modes usually exist because operation conditions frequently change to satisfy different market requirements. Besides multiple modes, the multiplicity of operation phases within each batch is an inherent nature of many batch processes. Different between-batch modes and different within-batch phases may exhibit significantly different underlying behaviors and have different effects on the final product quality. In this work, based on multi-phase quality analysis and between-mode quality analysis at the same time, a new quality prediction strategy is developed for injection molding processes. Firstly, a multi-phase residual recursive model is established using each quality residual of the phase regression models, so the contributions of main phases would be combined together. Also, between-mode quality analysis is conducted to judge the relationship between the new mode and the historical modes. Online quality prediction is completed using the selected model built based on the new mode or the historical modes. The simulation results illustrate the feasibility of this strategy.