Fault Diagnosis and Prevention by Fuzzy Sets

Abstract

Fault diagnosis and prevention in engineering systems are associated with imprecision which can often be overcome by fuzzy-set theoretic techniques. One such technique is based on the concept of conditioned fuzzy sets: a set of fuzzified influences acting on a system, conditions a fuzzy relation existing between those influences and the symptoms of the state of the system. The conditioning materializes as a fuzzy intersection (here extended to cover real phenomena), and a partial ordering of the influence parameters by their importance for each symptom of the state of the system. This paper explains the application of the conditioning fuzzy technique to a real engineering device. In particular, the paper focuses on how to use pertinent fuzzy-set theoretic concepts for creating a fault diagnosis and prevention model of an engineering system, and how to develop the application induced fuzzy intersection formula. A step-by-step summary of the computational procedure, including a sample calculation, concludes the paper. The fuzzy-set approach to fault diagnosis and prevention has several advantages: * The use of linguistic hedges makes it possible to develop the model of a system based on verbal formulation, and subsequent quantification, of the state of the system under study. That formulation follows from observations of the system either on the test bench, or in actual operation. * The knowledge base thus generated can be refined, as more experience of the behavior of the system is gained, to yield more dependable results.