Research on ultrasonic sparse DC-TFM imaging method of rail defects

Abstract

Rail damage is usually the main factor that affects the safe operation of trains, which may even lead to catastrophic accidents, such as train derailment and capsizing. It’s important to accurately detect the internal damage of the rail in real time. Ultrasonic testing is a main technology for the detection of rail damage in China, but there are some problems of misdetection and omission due to irregular rail shape, poor coupling, difficulty in automatic alignment, and inconvenient adjustment of the incident angle of the ultrasonic wafer, which results in a damage review rate of rail-defect detection lower than 78% in China. In this study, the phased array total focus method (TFM) imaging was applied to rail-damage detection, and the directivity correction-based sparse TFM imaging method (DC-TFM) was proposed to solve the problems of the existing phased array TFM imaging methods, such as inaccuracy and long computing time. The directivity function of array elements was used to correct the sound field intensity of different pixels to improve the accuracy of the TFM imaging. The experimental results show that the accuracy of the TFM imaging of defects is improved by 28% with correction of the sound field directivity. A sparse array designed based on the particle swarm optimization (PSO) algorithm was used to replace the full matrix data for the DC-TFM imaging, ensuring the imaging quality and reducing the imaging computing time. When the sparse rate is 75%, the array performance indicator (API) and the signal-to-noise ratio (SNR) are reduced only slightly, but the imaging computing time is shortened substantially by more than 40%, which indicates that the proposed sparse DC-TFM method can effectively improve the speed of rail-damage detection.