Study On Local Fault Diagnosis Of Rolling Bearings Under Non-stationary Conditions

The rolling bearing is one of the most important components of rotating machinery which is widely used in industrial production,and the stability of its operation is closely linked to the safety of the entire mech anical system.However,rolling bearings in its operation process will often produce fatigue peeling and other failures.The early fault signal is relatively weak and usually submerged in the background noise.And frequency estimation is inaccurate because the speed fluctuation caused by the non-stationary condition of the equipment such as speed up and down.These problems seriously restrict the accurate diagnosis of the fault bearing.The research of separation metho d about vibration signal of the rolling bearing and instantaneous frequency estimation method of non-stationary signal,contributed to accurately extract the fault characteristics and realize the accurate diagnosis of bearing fault,which is of great signi ficance to maintain good working condition of rotating machinery.The main contents of this thesis are as follows:(1)The separation method of rolling bearing fault sign al characteristics is studied.The background noises is weaken by EEMD.The MED is used,in order to eliminate the interference of the bearing assembly transmission path,and the Teager-Kaiser energy operator demodulation analysis method is used to extract fault feature components from independent carrier component,according to the signal frequency spectrum to determine fault types of rolling bearing.The experimental results show that the proposed method can effectively reduce the interference of irrelevant components and separate the fault features.(2)An improved method of order tracking for rolling bearings based on NCT is proposed.The non-stationary signal processing method of the rotating machine is analyzed,and the key and main points of the nonstationary signal.The NCT method is adopted to depict the non-stationary signal in the time-frequency domain;peak search method is used to estimate the instantaneous frequency of the rolling bearing and the least square fitting method is used to fit the frequency.The appropriate threshold is set and the instantaneous frequency is got properly through repeated cycles.The phase time discriminator is calculated by the instantaneous frequency and the signal resampling is realized by taking this time scale for sampling time.The simulation experiment proves that the metho d can obtain phase discriminator without tachometer and the result is of high precision.(3)The fault diagnosis method of rolling bearing under the condition of strong noise and variable speed is investigated.A non-stationary signal processing technique based on EEMD and improved NCT is proposed.By the combination of EEMD and MED,the fault signal is effectively separated;instantaneous frequency estimation based on improved NCT of the separated signal is carried out to get the phase time discriminator and demodulation analysis of the signal is achieved by using energy operator;By the calculated instantaneous frequency,the corresponding phase time discriminator is calculated.The phase time discriminator is used to analyze the demodulated signal for the order tracking,and the fault type is determined according to the amplitude spectrum.The experimental results show that the method can effectively diagnose the faults of rolling bearings.