Development of eDART based online diameter prediction systems in injection molding with multiple linear regression and fuzzy logic

Abstract Injection molding is a versatile technique for processing a wide range of thermoplastic and thermosetting polymers, as well as their composites. Dimensional defects are a critical issue in injection molding. This research focuses on developing online diameter prediction systems via multiple linear regression (MLR) and fuzzy logic. The systems are developed using Delrin 311 DP material, with the Taguchi methodology employed to define optimized process parameters that ensure adequate process capability. Processing data were collected from the sensors embedded in the surface of mold cavity by the eDART system. Regression analysis was employed to build and test the relationship between the real-time data from in-mold sensors and the diameter of molded part. The real time data from the sensor-based monitoring system, including end of cavity, hydraulic injection pressure, and efficient viscosity, were selected as the inputs for the predictive model. Both MLR and fuzzy logic models were established to predict the outcomes, based on the data retrieved from the sensors, achieving the prediction accuracies of 99.09% for MLR and 99.98% for fuzzy logic, respectively. Fuzzy logic demonstrated its reliability in predicting diameters and minimizing dimensional defects in injection molding.