Research On Ultrasonic Rail Detection Technique Based On LFM Signal

With the development of economy, the transportation capacity is also increasing. As oneof the most important means of transportation, railway transportation plays an important role.However, a variety of injuries appear in rain’s long-running process, which threatens runningsafety seriously. As a most widely used application, ultrasonic nondestructive testing is alsoused in the rail flaw detection. Ultrasonic nondestructive testing uses the theory that whenultrasonic wave transmits into the rail and encounters defects in medium interface, it willreflect. By detecting the echo time delay and amplitude, the information such as size, shape,properties and the location of the flaw can be determined. With the impact of large noise andmultipath, however, traditional detection methods can not accurately get the information offlaw echo signal’s time delay and amplitude. In this paper we designs two objective defectecho time delay and amplitude estimation system which are aimed at coping with multipathand large noise. The main contents of the paper read as follows：Firstly, this article adopts the active detection method and transmits linear frequencymodulation signal by ultrasonic probe, which effectively solves the irreconcilablecontradiction of the transmission distance and time resolution of the traditional single carrierfrequency signal;Secondly, in terms of time delay estimation, this paper combines with linear frequencymodulation signal in the ultrasonic detection environment and carries on the theoreticalderivation and proof of the three echo time delay method includes the matched filtering, thecross-correlation and the fractional Fourier transform. With comparative simulation of theerror precision and anti-noise reasons, finally we choose the matched filtering algorithm fortime delay estimation;In terms of amplitude estimation, this paper combines with the time delay estimation ofthe cross correlation, the matched filtering, the FRFT and proposes the ratio amplitudeestimation method. And in view of the low SNR environment, this paper puts forward theimproved algorithm of FRFT ratio method. This method firstly filters out noise and multipathsignals with the adaptive algorithm, then uses the method of the fractional Fourier transform ratio to determine the objective echo signal amplitude. This method not only solves theproblem that we can’t estimate amplitude with the distortion signals which is got by usingadaptive filter only, but also solves the problem that the useful signals is submerged in noisewhen it’s transformed by FRFTAt last, this paper combines ultrasonic detection environment and designs two differentkinds of delay-amplitude estimation system. By contrasting with and analyzing both systems,we can find the system based on cross-correlation/matched filtering is better. By applyingboth systems to the actual rail flaw detection experiment platform, we can get the estimationwhich is close to the theoretical value.