A Novel Model for Evaluating the Operation Performance Status of Rolling Bearings Based on Hierarchical Maximum Entropy Bayesian Method

Abstract

Information such as probability distribution, performance degradation trajectory, and performance reliability function varies with the service status of rolling bearings, which is difficult to analyze and evaluate using traditional reliability theory. Adding equipment operation status to evaluate the bearing operation performance status has become the focus of current research to ensure the effective maintenance of the system, reduce faults, and improve quality under the condition of traditional probability statistics. So, a mathematical model is established by proposing the hierarchical maximum entropy Bayesian method (HMEBM), which is used to evaluate the operation performance status of rolling bearings. When calculating the posterior probability density function (PPDF), the similarities between time series regarded as a weighting coefficient are calculated using overlapping area method, membership degree method, Hamming approach degree method, Euclidean approach degree method, and cardinal approach degree method. The experiment investigation shows that the variation degree of the optimal vibration performance status can be calculated more accurately for each time series relative to the intrinsic series.