Study on Fault Diagnosis for Gearbox Based on D-S Evidence Theory

摘要

This paper presents a new method based on the evidence theory of Dempster-Shafer in order to solve the problem of concurrent faults. The method takes the advantages of the techniques of information fusion to process the vibration signal of the gearbox, and then gains data of characteristic layer. The data are supplied to BP neural network and fuzzy classification, which is classed different of two teams. As different of evidence combination, the two teams data are fused in order to acquire the final result of diagnosis.