An intelligent fault diagnosis method for rotating machinery based on genetic algorithm and classifier ensemble

摘要

On the basis of extracting six time-domain features and five frequency-domain ones from vibration signals processed by empirical mode decomposition, an intelligent method to construct a rule-based classifier ensemble for fault diagnosis of rolling bearings was presented. The candidate reducts which could be used to build the base classifiers were found by applying a genetic algorithm for feature reduction on the decision table, and then the other genetic algorithm for diversity evaluation was used to search an optimal ensemble of base classifiers. Based on the results above and the improved weighted voting strategy, the final classifier ensemble for fault recognition could be set up. It was proved by the diagnosis experiment including normal condition, inner race fault, outer race fault and rolling element fault of SKF 6203 bearings that the method proposed in this paper was valid due to the satisfactory accuracy of classification.