Research On Rail Defect Localization Based On Ultrasonic Signal Parameter Estimation

The railway is an important national infrastructure and is closely linked with people's travel and national economic development.With the continuous development of China's railways,the mileage of railways has been increasing year by year,the train is constantly speeding up and the quantity of rail testing has also been increasing day b y day.During the locomotive running,the alternating load between the wheel and the rail and harsh natural environment will make the rail have some defects.Rail health directly affects the normal running of locomotives and lives and property of passengers.In order to ensure the safety of railway operations,it is necessary to regularly check the lines.Over the years,our country's rail detection adopt the hand-push flaw detector,supplemented by large rail flaw detection car.Hand-push flaw detector detection speed is slow.China's railway line work and the situation of rail injuries are different from the abroad,so the foreign test vehicle can not meet the requirements of domestic testing and is too expensive.Therefore,it is of great significance to design a rail testing instrument and equipment suitable for our country.In this paper,a small-scale double-track rail inspection experiment device is designed to realize the high-speed acquisition and localization of the rail defect signal.The device is characterized by simple structure,small size,low cost,and fast detection.During the rapid detection of rails,the rapid positioning of the defect signal is key and difficult.Aiming at dealing with the large error of the traditional positioning algorithm,an algorithm for estimating the ultrasonic signal parameters based on matching tracking is proposed.The Gaussian signal model is used as the atom to construct the complete dictionary.Based on the idea of sparse decomposition,the ultrasonic signal is represented with a set of optimal atoms,eliminating noise and extracting defect echo signal and the Time of flight(TOF)is estimated by the cross-correlation method.Through simulation experiments,this algorithm is compared with the traditional parameter estimation algorithm under different signal-to-noise ratio(SNR).The simulation results show that the method is accurate,and it is suitable for the case of low SNR.In the detection of rails,the phenomenon of multi-echo overlap often occurs and an ultrasonic signal parameter estimation algorithm based on fractional Fourier transform(FRFT)is proposed.The overlapped multi-echo signals are handled by fractional Fourier transform to obtain the FRFT spectra of each order,and then the improved Shannon entropy of FRFT spectrum in each order is calculated.The optimal transformation order is determined by the inproved Shannon entropy,the window function separate the overlapped multi-echo signals in the optimal FRFT,and finally the TOF of each component is estimated using the cross-correlation method.Compared with the classical least squares method,the local optimal solution is not generated and the computation quantity is small.The validity of the proposed algorithm is verified by simulation experiments,and the TOF estimation results of each echo component are superior to the classical least squares method.Ultrasonic probes with different angles are used for the detection of different types of defects in the rail head,rail waist and rail bottom.The proposed methods in this paper are applied to the measured ultrasonic signal.The experimental results show that the estimated results of the proposed TOF estimation algorithm are close to the physical dimensions of the measured workpiece,and the estimation error is obviously less to traditional method.