A New Multi-Expert System Applied in Fault Diagnosis
DU Yan-bo · Electronics Optics & Control · 2009
In order to make a fast fault diagnosis when a complex system fails,we studied the characteristics of Support Vector Machine(SVM),Fuzzy Integral(FI) and expert system,and put forward an integral structure and diagnosis model of multi-expert system based on Least Square SVM(LSSVM) and FI.LSSVM has the advantages of simple structure,good generalization and globally optimal solution etc., which can realize the functions of obtaining knowledge automatically and logic reasoning quickly when used with expert system.The using of FI can improve the cooperation diagnosis precision of multiple experts.The expert system implements management and maintenance of the repository,symbol reasoning diagnosis and explanation.This method overcame the shortcomings of traditional expert system,such as lack of the mechanism of effective self-study and self-adapt,and difficult to resolve the nonlinear system fault problems.Therefore,it is a convenient way to deal with the multi-sign and multi-reason fault problems of a complex system.