Analysis Of Rail Crack Signal Based On Magnetic Flux Leakage Testing

Real-time detection of railway health status can detect potential hazards in time,which is of great significance to guarantee the safe operation of high-speed railway.The detection method based on the magnetic flux leakage,is a non-contact,fast and accurate detection means.In order to ensure the accuracy of detection,it is very important to analyze the magnetic flux leakage signal based on the reasonable design of magnetic flux leakage signal detection equipment.In this paper,the magnetic flux leakage signal is analyzed,and the magnetic flux leakage signal is identified based on SVM and the theory of sparse learning is firstly proposed to analyze the magnetic flux leakage signal of the rail crack,and good recognition result is obtained.Based on the principle of magnetic flux leakage detection,a multi-sensor magnetic flux leakage detection circuit is designed to detect the crack of different shapes and sizes.The characteristics of the signals collected by different sensors and the signal changes at different speeds are analyzed.Adaptive filtering and empirical mode decomposition are used to preprocess the collected signals.A magnetic flux leakage signal sample library with different cracks is constructed using the magnetic flux leakage signals collected by the experiment,which can lay a foundation for the follow-up signal analysis.Based on the characteristics extraction of magnetic flux leakage(MFL)signal,the multi-sensor information fusion classifier based on SVM algorithm is designed to classify the different cracks.By sparse representation of signal.The KSVD dictionary learning method is used to automatically extract the characteristics of magnetic flux leakage signals and classify the different cracks.In this thesis,the magnetic flux leakage signal detection equipment can collect the three-dimensional components of the rail crack magnetic flux signals,and can detect the cracks with minimum width of 0.2mm and depth of 0.2mm.The new features of the magnetic flux leakage signal of the rail crack are proposed.The multi-sensor fusion decision method based on SVM can classify the crack types accurately.The first time,DKSVD algorithm is applied to the magnetic flux leakage signal recognition,which can automatically extract the characteristic atoms of the signal and make good recognition of the magnetic flux leakage signal.