A New Statistical Model for Rolling Element Bearing Fault Signals Based on Alpha-Stable Distribution
Changning Li, Gang Yu · 2010
A new statistical model for rolling element bearing fault signals is proposed based on alpha-stable distribution. Such a non-Gaussian model can accurately describe statistical characteristic of bearing fault signals with impulsive behavior. The characteristic exponent alpha of bearing fault signals with different fault degree is estimated by a stable distribution parameter estimation method. Estimation result explains the bearing fault signals belongs alpha-stable process. At the same time, alpha-stable density of every bearing fault signal fit well the empirical probability density in log-log plots, and their tail possess the same heavy tail behavior. Then the statistical model for different fault degree bearing signals all are valid.