Research On Noise Source Identification Method Of Permanent Magnet Synchronous Motor Driven By Pure Electric Vehicle

With the global ecological environment deteriorating and the exhausting resource,the requirement for electric vehicles is increasing.Electric vehicles are being vigorously promoted in China.Permanent magnet synchronous motors are widely used in pure electric vehicles due to their simple structure,high power density and peak efficiency.The gap between pure electric vehicles and traditional cars is that the driving device changes from an internal combustion engine to an electric motor.The lower background noise of electric vehicles makes the contribution of motor noise to the vehicle more prominent.Consumers are paying more and more attention to ride comfort when buying a car.The noise generated by the driving motor will make people feel uncomfortable,so the market and enterprises have paid attention to the NVH performance of electric vehicle driving motors.Therefore,accurate identification of noise sources of permanent magnet synchronous motors for pure electric vehicles is a prerequisite for low noise design.In this paper,a pure electric vehicle driven permanent magnet synchronous motor is used as the research object.This paper compares the performance of fixed-point independent component analysis(Fast ICA)and robust independent component analysis(Robust ICA)noise source separation methods of multi-channel noise source identification algorithms and compares their influence factors.In order to overcome the complexity and economics of multi-channel noise measurement,empirical mode decomposition(EMD)and ensemble empirical mode decomposition(EEMD)are proposed to separate the noise sources of electric motors.Specific research work includes:First,the mechanism of PMSM noise generation,excitation source and characteristic frequency of each noise source are analyzed.The types and principles of noise source identification are described.Analyze the advantages and disadvantages of various methods in the process of noise signal identification.Lay a theoretical foundation for accurate identification of noise sources.Second,the noise source simulation signal of PMSM is constructed.Randomly select a mixing matrix to mix signals to form a mixed signal.Taking the absolute value of the correlation coefficient as an evaluation index to evaluate the effect of noise source identification,the identification effect of the multi-channel noise signal identification algorithms Fast ICA and Robust ICA at different signal-to-noise ratios and sampling frequencies is revealed.It illustrates the advantages of Robust ICA in noise source identification.Thirdly,EMD and EEMD with the ability to transform into signals of different scales are proposed to be used to identify the noise source of the drive motor to overcome the complexity and economics of multi-channel noise measurement points.The effects of different stopping thresholds,iterations,maximum modulus,and interpolation methods on the recognition effect were studied.Finally,the test data is used to verify the noise source identification algorithm,and the independent component analysis theory is used to solve and analyze the noise source contribution in the entire working condition area.It is obtained that the contribution degree of the noise source of the driving motor is switching frequency noise,radial electromagnetic force noise and mechanical noise.The contribution of the three noise sources will show different changes with different working conditions,which is of great significance for the optimal control of the drive motor noise in the entire working condition area.