The diagnostic reasoning based on fuzzy self-organizing neural network and its application

Abstract

By means of the maturational fuzzy theory and the self-organizing map (SOM) neural network which is prepotent in the way of information mapping and self-organizing characteristics, this paper combines the fuzzy theory with SOM neural network having the feature of pinpoint nonlinearity, and applies it to the field of fault diagnosis for large-scale electro-mechanical equipments. The network structure and theory of learning of SOM are exhausted in detail, and the diagnostic principle and the implementation strategy of fault reasoning based on fuzzy self-organizing map (FSOM) neural network are researched. Finally, we verify the correctness and the practicability of this method by the instance during the procedure of diagnostic reasoning.