Experimental Investigations on Reliability Assessment based on Hierarchical Maximum Entropy Bayesian Method

Owing to the influence of various factors, the vibration performance of rolling bearings is non-stationary degradation, and the reduction process possesses non-linear dynamic characteristics. Information, such as performance recession trajectory, probability distribution and vibration performance reliability function, varies with the movement process, which is difficult in analyzing using traditional reliability theory. So the hierarchical maximum entropy Bayesian method is proposed to establish a mathematical model, which is used to evaluate the vibration performance continuity reliability of rolling bearings under the condition of unknown priori information. A deep groove ball bearing (type SKF6205) is used as an example to perform an experiment investigation, which shows that the vibration performance continuity relative reliability of the rolling bearing keeps a "reclining chair" relationship with the wear diameter. The performance continuity reliability evaluation model has no requirements for the priori information of data samples, types and components of bearings.