The method for aircraft unit fault diagnosis, which is based on the self‐organising feature map (SOM) artificial neural net and expertise, has been presented in this paper. A fault probability of components, of which an aircraft unit is comprised, is generated by the SOM. The qualitative possibilities that the components of a specific unit may cause a failure were estimated by an expert in natural language. It was converted into numeric value by means of Zadeh's fuzzy logic. Both the fault probability and the expertise value of the specific components have been considered in this method based on SOM too. The technical feasibility of this method has been shown by the example. The method is useful when a model based on the First Principles or a causal model of a diagnostic object is not available. It is also applicable to the fault diagnosis of other complicated electromechanical equipments.
MCB UP Ltd
Copyright © 2003, MCB UP Limited