Through extracting the information exergy characteristics

the support vector machine (SVM) diagnosis method of rotor vibration faults based on integration of information exergy was put forward. First of all

the rotor unbalance fault

shafts misalignment fault

rotor crack fault and rubbing fault were simulated respectively on the rotor test bench

and vibration acceleration signals of these four typical kinds of faults under multi-point was gathered; secondly

the information exergy characteristics of the rotor vibration fault process change law based on time-domain singular spectrum entropy and frequency-domain power spectrum entropy was extracted; finally

using the extracted information exergy characteristic was used as a fault vector

the SVM fault diagnosis model was established

and the rotor vibration fault was diagnosed. The diagnostic results of examples indicate that the use of information exergy characteristic combined with SVM in rotor vibration fault diagnosis make the diagnosis accuracy rate reach 97% improving the accuracy of diagnosis effectively.