Fuzzy Rule-Based Approach for Diagnostics in Nuclear Plant Diesel Generators

Abstract

This chapter presents a comprehensive method of fault diagnosis that will enhance the condition-based maintenance capability of engineering system and reduce the dependence on human experts for fault diagnosis using system modelling and fuzzy rule-based integrated approach. This method can be used in a complex system where failure data is either unavailable or scarcely available, which does not allow us to use extensive data-driven techniques. This methodology uses the integrated approach wherein reliability-based techniques like failure mode and effect analysis (FMEA) and the fault tree analysis (FTA) is utilised. The FMMEA is used for finding various failure modes and mechanism and identification of appropriate parameters which can be used to assess impending failure. The fault tree analysis (FTA) is utilised for the qualitative analysis of various failure modes and mechanism and finding various relations between them. This knowledge is then utilised for making fuzzy rules in fuzzy rule-based diagnostics system. Further, mathematical model has been developed using laws of physics, which relates various parameters of the system. This mathematical model gives the healthy baseline and failure definition of various failures, this information will be used in fuzzy rule-based system for making membership function. Finally, a case study is carried out on diesel generator used in nuclear plant.