A study on the practical application of the integrated ANN system for manufacturing the target quality of the injection molded product

The quality of the products manufactured by injection molding is greatly influenced by the process variables of the injection molding machine used during manufacturing. It is very difficult to determine the process variables considering the stochastic nature of the manufacturing process, because the process variable complexly affects the quality of the injection molded product. In the present study, we used an artificial neural network (ANN)-based method to determine injection molding process variables to manufacture products of desired quality, as ANNs are known to be highly accurate in analyzing non-linear problems. To train the ANN model, a systematic plan was developed using a combination of orthogonal and random sampling methods to represent various and robust patterns with a small number of experiments. According to the plan, injection molding experiments were performed to generate data, which were separated into training, validation, and test sets to optimize the ANN model parameters and test its predicting performance. Multiple-input single-output (MISO) and multiple-input multiple-output (MIMO) models were developed to predict 8 process variables to manufacture a product with specific dimensions and provide user reference information (mass and pressure at the end of fill). The predicted process variables were applied to an injection molding machine to verify the predicted accuracy of the ANN system. Finally, it was confirmed that the determination of process variables using the ANN method meets the tolerances required in general industry practice.