Gear Fault Diagnosis Correlation Analysis Based on Probability Box Theory

Measuring signal in gearbox fault diagnosis low signal-to-noise ratio, high frequency of characteristic signal and noise reduction was difficult. And general gear fault diagnosis method was hard to solve the space-time registration problem of multiple source information. So probability box theory and information fusion methods were applied in this paper, which introduced the basic development process of the probability boxes and the methods of information fusion. However, independent irrelevance of measured signal was assumed before information fusion, which was unreasonable. For this reason, classification strategy and pattern of correlation based on probability box theory was put forward. The experiment proved that correlation classification for multiple source information during gear fault diagnosis improved fusion efficiency and science, can better met the authenticity of the space-time registration problem of multiple source information. which developed a new approach to gear fault diagnosis or bear fault diagnosis research to solve the space-time registration problem of multiple source information.