VPRS Model Approach to Fault Feature Selection and Diagnostic Rules Extraction
Wang Zhi-quan · Acta Simulata Systematica Sinica · 2003
A method based on VPRS(Variable Precision Rough Set) model theory is proposed to fault features selection and diagnostic rules extraction in engineering systems. The approach is realized by selecting proper classificatory error parameter b and using knowledge reduction. It can not only extract effectively the optimal fault features, but also find the simplest diagnostic rules for the system. Simulation results for rolling bearings show that the method improves the rate of fault diagnosis, decreases the number of feature parameters and diagnostic rules, and reduces the cost of diagnosis, with more applicability than the classical RS-method in practical applications.