Injection molding is a typical and important batch process. Process monitoring is of great importance in injection molding. Traditional process monitoring methods based on multiway principal component analysis (MPCA) and multiway partial least squares (MPLS) suffer from batch length inequality and huge data volume. To overcome these problems, a new process monitoring method based on injection feature is proposed, along with a novel process capability assessment model.

The entire research will be presented in three parts: feature variable, process monitoring, and process capability assessment. In the feature variable part, the definition and properties of feature variable are discussed by introducing three injection feature variables- filling work, packing work and injection wo...[ Read more ]

Injection molding is a typical and important batch process. Process monitoring is of great importance in injection molding. Traditional process monitoring methods based on multiway principal component analysis (MPCA) and multiway partial least squares (MPLS) suffer from batch length inequality and huge data volume. To overcome these problems, a new process monitoring method based on injection feature is proposed, along with a novel process capability assessment model.

The entire research will be presented in three parts: feature variable, process monitoring, and process capability assessment. In the feature variable part, the definition and properties of feature variable are discussed by introducing three injection feature variables- filling work, packing work and injection work. In the process monitoring part, a new process monitoring method, injection feature based statistical process monitoring is developed and compared with traditional MPCA monitoring methods. In the process capability assessment part, a regression model between the standard deviation of injection work and that of product weight is built to predict process capability (C_p). An overall equipment capability fuzzy evaluation model is further built based on the predicted C_p, and a decision-supporting matrix is proposed to identify the factors that need improvement in the production.

A series of experiments are conducted and show that injection feature variables are powerful in process monitoring and process assessment. This research paves the way for a new process monitoring system with an innovative combination of process monitoring and quality management in injection molding.