Study Of Real-time Broken Rail Detection Denoising Algorithm Based On Ultrasonic Guided Waves

With the rapid development of Chinese railway in the direction of high speed and more goods,puts forward higher requirements on the integrity of the track structure,the importance of railway transport safety work is higher and higher.And real-time broken rail detection is cater to the requirements of the country,the railway safety checks,for there is a problem of railway do early found early repair,to avoid the occurrence of major disasters.However,such a complex natural environment on the railway,the noise signal is an inevitable problem,due to the interference of noise signal,can cause the influence of various we cannot imagine.Especially in broken rail detection,because of the interference of noise that we detect the broken rail position have the corresponding error,its impact is not just a few Numbers so simple,it may threaten our property and life safety.This thesis mainly involved in the real-time broken rail detection of ultrasonic guided wave denoising algorithm is studied,in order to more accurately judge the broken rail and more accurately position the broken rail.Due to the system of ultrasonic signal is a kind of weak nonlinear nonstationary signals,are highly susceptible to noise interference,makes to the broken rail judgment is not accurate positioning is not accurate.In order to avoid the occurrence of the above situation,this article from the traditional soft and hard threshold function,the advantages of extracting the two new threshold function,the use of this new threshold function creates a new wavelet threshold denoising algorithm function.Found through MATLAB simulation,relatively soft threshold function method,after the new denoising method of threshold function of signal to noise ratio increased by 11%to 22%,the mean square error(MSE)reduced by 21%-29%;Relatively hard threshold function method,the new denoising method of threshold function of signal to noise ratio increased by 11%-29%and the mean square error is reduced by 21%-30%,and the new threshold function method after denoising signal amplitude is more close to the original signal amplitude.Because the wavelet denoising has its corresponding faults,in view of its shortcomings and puts forward the wavelet packet threshold denoising.Verified and found a new threshold function not only in the wavelet threshold denoising have obvious advantages,and the function not only in the wavelet threshold denoising have obvious advantages,and the wavelet packet threshold denoising also has one of the more significant of the parameters in the optimization.Track damage location,the most important parameter is the echo time,in this paper,using the Hilbert transform and the signal envelope,find the signal peaks,then got the echo time.Involved in the paper uses MATLAB to the system modeling and simulation of algorithm,using DSP TMS320F2812 hardware validation.Can be found that the ultrasonic guided wave denoising algorithm of real-time broken rail detection makes positioning damage location more precise orbit,determine whether damage is more accurate,has great application value and significance.