Bearing Fault Diagnosis Based On Null Space Operator-Frequency Weighted Energy Operator Method

With the continuous development of science and technology,more and more advanced technology has been applied to different fields of machinery and equipment,and the safe operation of machinery and equipment is related to personal and property safety,its importance is self-evident.Fault monitoring and diagnosis of mechanical equipment is a research hotspot at home and abroad,and mechanical fault diagnosis based on vibration signal is the most widely used method at present.Firstly,the significance and research background of the subject are described.Then,the null space theory,including local narrow-band signal,differential operator and integral operator,is elaborated.The difference between the two operators is compared,and the local narrow-band signal of the two operators is obtained.Inspired by the core idea of EMD,the null space dictionary is constructed.Then,the decomposition method of null space singular operator is discussed.And its numerical solution is given.By comparing the simulation signals,the results show that the null space singular operator decomposition method is more robust than the EMD method,and has no endpoint problem.It also essentially solves the problem of EMD without mathematical support and analytical solution.The limitation of the integral operator is optimized,which solves the problem that the integral operator can not decompose part of the signal accurately,and improves the computational efficiency.In view of the theoretical shortcomings of Hilbert transform,this paper introduces Teager energy operator,which takes advantage of its advantages in detecting transient components.It can solve the instantaneous frequency more effectively and is not restricted by Bedrosian equation and Nuttall theorem.Its energy spectrum is more accurate for frequency detection of high frequency part.For the problem of noise sensitivity and negative energy of Teager energy operator,this paper improves it and proposes a frequency domain weighted energy operator.By comparing the simulation signals,it shows that the improved method effectively reduces the problem of noise sensitivity and solves the problem of negative instantaneous energy of Teager energy operator.Based on the null space optimized integral operator and differential operator decomposition method and frequency weighted energy operator demodulation method proposed in this paper,a fault diagnosis method combining them is proposed.The feasibility and accuracy of this method in practice are verified by practical application and comparison.