Research On Rolling Element Bearing Fault Diagnosis Based On Time-frequency Transform And Order Analysis Under Time-varying Rotational Speed

Rolling bearings are the most common and indispensable parts of rotating machinery in many engineering fields.They are often used under varying speed and variable load conditions,and the environment is very harsh,resulting in becoming the most vulnerable parts.The quality of its operation directly affects the operation of the system.Therefore,the fault feature extraction of rolling bearings in non-steady state is the most critical point for early monitoring and post fault diagnosis.Resonance demodulation technology is the most common method for bearing fault diagnosis,it requires constant speed as a precondition.Actually,the variable speed operation and the interference of gear meshing will eventually lead to constant speed as represented by resonance demodulation technology and others traditional fault diagnosis methods of the preconditions for constant speed is no longer applicable.In this paper,variable-speed rolling bearings with gear mesh interference are the research object and main tasks as follows:First of all,in the case of gearless mesh interference,in order to solve the inapplicability of the traditional diagnostic method in this environment,the paper first considers the problem from the perspective of time-frequency analysis.The paper proposes a method for obtaining the instantaneous fault characteristic frequency change line from the envelope time-frequency diagram that solve the problem of signal characteristics are extracted difficultly.Using the improved STFT peak search algorithm(PSA)to extract the instantaneous fault characteristic frequency change line;comparing with the rotating frequency curve to achieve fault diagnosis.The order analysis method is the most common and effective in solving the variable speed problem.This paper uses the calculation order tracking(COT)method,using instantaneous fault characteristic frequency change line to estimate the rotating frequency for bearing,in order to realize the equal-angle resampling of the vibration signal,achieving fault diagnosis by analyzing the characteristic order spectrum.Simultaneously,the paper proposes a fault angle domain resampling algorithm based on the instantaneous fault characteristic frequency change line of the sampling frequency re-modulation to realize the resampling in the angular range of the vibration signal,achieving fault diagnosis by analyzing the fault angle order spectrum.Secondly,in the case of gear mesh interference,the gear meshing frequency of the signal has an amplitude advantage.The improved STFT-PSA can extract the instantaneous dominant meshing multiply line from the envelope time-frequency diagram,the equal-angle resampling of the signal is realized by the COT method,achieving fault diagnosis by analyzing the characteristic order spectrum.Using the ensemble empirical mode decomposition(EEMD)method to eliminate the gear noises,achieving fault diagnosis by analyzing the characteristic order spectrum.The results show that the proposed method is effective and feasible.It has important application value for fault diagnosis of rolling bearings under variable speed conditions under the interference of gear noise.