An Integrated Model of Product, Production Process and Quality Information for Failure Prediction

Abstract

This study proposes a methodology for predicting product failures occurring due to known failures and causalities. The prediction is based on failure propagation in an integrated model of a product, production process, and quality information (PPQ). This methodology enables a designer to predict new failures. One of the concepts for failure prediction is the PPQ for compiling referring to quality-related issues. The other is an analysis methodology of the failure propagation for predicting new PPQ-based failures. The failure propagation indicates the possible failures from the production process and their causes. Three critical issues for system implementation are discussed in detail: 1) integrated PPQ model; 2) methodology for knowledge reuse by developing a PPQ-based library, and 3) an analysis methodology for failure prediction. The prototype system of the PPQ model is implemented for failure prediction in the example of a production system of Auto-Breaker.