Analysis Of High-speed Train Radiated Noise Based On Wavelet Threshold And EMD

As the backbone of China's transportation industry,the railway has always played a decisive role in economic development.The development of high-speed trains in these years has further enhanced our country's railway industry.During the operation of high-speed railways,noise is also generated along with the operation of trains,which may increase noise in the car and reduce passenger comfort,especially affecting the normal life of residents living near high-speed trains.As a result,how to reduce the radiation noise of high-speed rail has become a problem that must be solved in train development.The radiation noise level of high-speed trains is one of the important performance indicators of high-speed trains,and the control of radiation noise of high-speed trains is the key to reducing the radiation noise levels of high-speed trains.Therefore,it is necessary to analyze high-speed train radiation noise during the design stage of high-speed trains.The purpose of the analysis is to study the noise generation mechanism,accurately locate its main noise source,analyze the frequency cycle of the noise signal,and provide basis for designing high-speed trains with low noise and low vibration levels.The radiation noise of high-speed trains is a typical non-stationary signal.This paper mainly studies the frequency of wheel-rail noise,analyzes its frequency cycle,and the high-speed rail-rail radiation noise is in the low frequency range.Therefore,this paper combines the classical theory of signal processing,simulation data and the processing method of modern signal to analyze the characteristics of radiation noise of high-speed trains to achieve scientific evaluation and application.This article mainly did the following research work:1.This paper introduces the basic theory of wavelet transform and mainly analyzes the principle of wavelet denoising,including the steps of wavelet denoising and the selection rule of wavelet threshold.In addition,the Rigrsure threshold method,Sqtwolog threshold method,Heursure threshold method,and Minimaxi threshold method are applied to the simulation signal,and the denoised power spectrum is analyzed to verify the ability of the wavelet threshold to enhance signal-to-noise ratio and line spectrum.2.Secondly,a new method of signal analysis—Empirical Mode Decomposition method—is introduced.Firstly,the paper introduces the definitions of instantaneous frequency and natural mode components.Then,it explains the empirical mode decomposition method,the filter principle based on empirical mode decomposition method,the screening process of empirical mode decomposition as well as its completeness and orthogonality.Secondly,the integrated empirical mode decomposition method and adaptive noise complete set empirical mode decomposition method are introduced.Finally,the EMD,EEMD,and CEEMDAN are used to decompose the signals with modal aliasing and verify the improvement of the modal aliasing of EMD by EEMD and CEEMDAN through comparison.3.The signal is processed by the wavelet Rigrsure,Sqtwolog,Heursure,and Minimaxi threshold methods to analyze the time-frequency spectrum and marginal spectrum of the processed signal.4.The mathematical model of high-speed rail radiation noise is given by the frequency characteristics of the high-speed rail.Based on EMD,EEMD and CEEMDAN,the lineal spectrum analysis was performed on the high-speed rail radiated noise simulation signal.Through the autocorrelation coefficient of the imf component,the real imf mode was selected and the autocorrelation spectrum analysis was performed on the real mode to analyze the frequency characteristics.5.In the case of low SNR,EMD-wavelet threshold-LMS spectral analysis method,EEMDwavelet threshold-LMS spectrum analysis method,and CEEMDAN-wavelet threshold-LMS spectral analysis method are used to repeatedly verify different SNRs and different signals.Also,it is compared with the classical spectrum analysis method of the simulated signal and the spectrum analysis method of the wavelet hard threshold denoising to verify the ability of the theory proposed in this paper to improve the spectral analysis resolution of the high-speed rail radiation noise.