Compound fault diagnosis based on probability box theory

In view of compound fault of rolling bearing in mechanical transmission system, the fault signal was decomposed by probability box modeling method to obtain the different probability box models. The original data with abundant statistical information was taken as the research object. The raw-data-direct probability box modeling method was used to analysis the original data with data envelopment method. The probability boxes with similar similarity were enveloped by demodulation method. The results show that the proposed method can identify the fault characteristics in the compound fault of rolling bearings.