Application of Fuzzy Pattern Recognition in Ultrasonic Transverse Wave Detection of Wheel Flaws

Flaws in train wheel could cause wheel broken in highspeed running process, which seriously affect the traffic safety. Detection of wheel flaws is important and urgent. A fuzzy pattern recognition method is applied to flaw classification in ultrasonic transverse wave detection of wheel defect. The received flaw echoes are pre-processed by wavelet filtering and Fourier transform. The feature values of flaws are extracted in frequency domain, on the base of which the fuzzy sets of typical flaw are constructed. The concept of similarity degree and the choosing near principle have been used in recognition. Experimental results show that the method is efficiency and can achieve very promising performance. And comparing with the other means, it performs more efficient.