Identification and classification protocol for complex systems
Résumé
This paper proposes a test protocol for drift identification and classification in a complex production system. The key objective here is to develop a classifier for failure causes where variables depend on a set of measured parameters. In the context of our work, we assume that the drift problem of a production system is generally observed in control products phase. The model proposed in this paper for failure causes classification is structured in the form of a causes-effects graph based on Hierarchical Naïve Bayes formalism (HNB). Our key contribution in this is the methodology that allows developing failure causes classification test model in the complex and uncertain manufacturing context.