Design and realization of inference engine of fault diagnosis expert system for electronic recycling equipment of missile based on support vector machine

Abstract

To address the problems of “bottleneck” of expert system (ES) in acquiring knowledge as well as “match conflict” and “combinational explosion” in expert system inference, base on fundamental principles of Maintainability appraisal of electronic recycling equipment in a type of missile, the diagnostic method is proposed for locating the fault of module unit or key component in various PCB Maintainability appraisal models by embedding SVM into ES inference engine, by which the design incorporates construction of ES inference engine diagnosing the faults with electronic recycling equipment in a type of missile based on SVM, together with the method for realization of SVM inference engine. The experiments prove that this method effectively improves ES fault diagnosis rate and identification ratio.