Industrial process defect classification by exploiting PCA and fuzzy logic

This This paper describes a proposed monitoring approach destined for industrial process using the principal components analysis (PCA) and fuzzy logic. The aim of our work is to detect and locate defaults using PCA and then classifying the existed problems in terms of gravity using fuzzy logic. We introduce a historic data in the form of a matrix of m variables and N observations: measures were taken of the various existing sensors in the process. In our case, we study the quality criteria for flour production process. The obtained results are effective by comparing them with expert data.