Online quality optimization of the injection molding process via digital image processing and model-free optimization

Injection molding is a widely used polymer processing technology that transforms plastics into products of various shapes and types. Defects may occur on molded products during production under improper settings. Online process optimization is a kind of systematic method that handles such problems by investigating the correlations of process variables and the quality of final products. However, this method requires accurate measurement of the quality index, which is challenging for some defects, such as shrinkage and flash. This study proposes a systematic method that combines digital image processing (DIP) and model-free optimization (MFO) to solve such problems. DIP technology is used to monitor the magnitude of surface defects. MFO uses online measurement as feedback to determine the optimal settings. The proposed method is successfully implemented on an injection molding machine. Experimental results show that it can effectively determine optimal solutions to eliminate defects of molded parts.