Study On In-situ Detection Method Of Electromagnetic Ultrasonic Guided Wave For High-speed Railway Turnout

When the train is running on the railway,the rail is not only subject to the vertical pressure of the train wheels,but also subject to the horizontal thrust of the wheels,under the action of various factors,the rail is very prone to fatigue damage,crack expansion and other problems.In the railway system to high speed,heavy load development,turnout basic rail tread crack damage is mainly caused by the rolling contact fatigue of the rail,when the turnout basic rail tread crack expansion to a certain depth（6～10 mm）,quickly turn into the rail head transverse crack,and finally lead to rail head transverse fracture,endangering the safety of train operation.In addition,for the turnout base rail waist,as one of the important parts to bear the weight of the train,because the rail steel in the smelting process will appear bias,inclusions and other phenomena,after the rail is put into use,in the external load,the initial small rail waist damage gradually to the rail head and rail foot direction cracking,and eventually lead to rail fracture.At present,rail flaw detection mainly relies on rail flaw detection car,it needs to take up the time of the day,and there are problems such as detection blind area and low detection efficiency.For the turnout base rail waist,it is in the middle part of the rail structure,there are bolts,support rods and other characteristics of the structure,rail flaw detection vehicle detection can’t achieve the turnout base rail waist for in-situ detection,turnout base rail waist in-situ detection is one of the problems in the rail damage detection.In order to achieve fast detection of turnout base rail tread,this thesis proposes a signal processing method combining Synchrosqueezed Wavelet Transform（SWT）and Pulse Compression Technique（PCT）to improve the detection speed.In addition,for the turnout base rail waist in-situ detection problem,this thesis uses only in-plane displacement of the horizontal shear（Shear Horizontal,SH）wave guide for detection,avoiding the rail waist at the bolt,support rod and other features of the structure on the detection of the echo,in order to achieve in-situ detection.In this thesis,the research on electromagnetic ultrasonic surface wave imaging detection methods for rail tread cracks is firstly carried out.Firstly,the mathematical models containing Gaussian white noise,sinusoidal signal plus Gaussian white noise interference are established respectively,and the suppression effects of a total of three signal processing methods,namely PCT,SWT and combined SWT-PCT methods,on the noise in the echo of electromagnetic ultrasonic detection are analyzed.Secondly,in order to verify the noise suppression capability of the above method,a 1 MHz surface wave electromagnetic acoustic transducer（EMAT）was designed to detect rail treads containing cracks.Finally,the Hilbert-Huang Transform（HHT）method was used to compare the noise reduction capability and ultrasonic imaging effect of the ultrasonic echo signal corresponding to the Tone-Burst signal and the combined SWT-PCT method,using the ultrasonic echoes of the detected cracks as the object of study.The results show that the combined SWT-PCT method can detect oblique cracks in the rail tread of 10 mm and 20 mm in length and 6mm in depth by processing the electromagnetic ultrasonic surface wave detection echo signal,compared to PCT-only surface wave detection with electromagnetic ultrasound the SNR of the echo signal is improved by at least 6.82 d B,which significantly improves detection speed and B-scan image resolution.Subsequently,a study on the in-situ detection method of transverse cracks in turnout base rail waist by electromagnetic ultrasonic SH wave guide was carried out.Firstly,the rail waist is equated to a steel plate of the same thickness as its cross-section,the dispersion curve of the steel plate is solved,the excitation frequency that can excite the SH₁ guided wave is selected according to the dispersion curve,a finite element simulation is established and the dispersion effect of the SH guided wave in the actual rail waist and steel plate model is investigated to determine the EMAT design parameters for excitation of the SH₁ guided wave.Secondly,a unilateral excitation,received EMAT and a bilateral antisymmetric excitation,received EMAT were produced to test the steel plate and verify the finite element simulation results,and a bilateral antisymmetric excitation,received EMAT was used to test the transverse cracks in the steel plate.Finally,the detection echo signals under the two excitation methods of Tone-Burst and Barker code signals are collected separately and the detection effects are compared.The results show that the two-sided anti-symmetric excitation and receiving EMAT can suppress the SH₀ and SH₂ guided wave modes,obtain the enhanced SH₁ guided wave,and realize the in-situ detection of the transverse crack of the turnout base rail waist.The PCT-SWT combination method can be used to process the non-synchronous average Barker code ultrasonic echo to obtain a higher signal-to-noise ratio than the 64 times synchronous average Tone-Burst signal,which can increase at least 4.31 d B,significantly improving the detection efficiency of the turnout base rail waist injury.Finally,an in-situ magnetostrictive SH surface wave detection method for transverse cracks in the turnout base rail waist is investigated.Firstly,the rail waist was equated to a steel plate according to the principle of section equivalence and the excitation frequency of the SH surface wave was selected according to the dispersion curve of the steel plate.A finite element simulation model of SH surface wave detection of transverse cracks in steel plates was developed,and the signal-to-noise ratio of the detected echoes corresponding to Tone-Burst and 3×13-bit combined Barker codes was analyzed.Secondly,the SH surface wave EMAT with magnetostrictive mechanism was fabricated for the detection of transverse cracks in steel plates.The ability of the SH surface wave EMAT to detect transverse cracks in steel plates was analyzed for different combinations of the number of pairs of permanent magnets and the number of turns of the bending coil,and the optimum combination of the number of pairs of permanent magnets and the number of turns of the bending coil was determined.finally,the detection echo signals under Tone Burst and 3×13-bit combined Barker code excitation were collected respectively for analysis of the detection results.The results show that the intensity of the ultrasonic echo signal is proportional to the number of pairs of permanent magnets and the number of turns of the zigzag coil.When the permanent magnets are 3 pairs and the number of turns of the zigzag coil is 4,transverse cracks of5 mm×1 mm×1 mm and 10 mm×1 mm×1 mm on the rail waist can be detected,crack echo signal-to-noise ratio of not less than 8.60 d B.The SH surface wave is more sensitive to the detection of transverse cracks in steel plates,and the defect echo signal-to-noise ratio is improved by at least 2.01 d B compared to low-frequency SH₁ guided wave detection.The use of combined Barker codes can significantly improve the detection efficiency of turnout base rail waist injury.