Broken Rail Detection System For Continuously Welded Rails Based On Ultrasonic Guided Waves

Broken rail threatens the safety of railway operations. With the development of high-speed rail, CWR is widely used in rail transport industry. However, it will generate large thermal stresses in the rail when temperature changes, which may increase the risk of broken rail. Domestic detection methods, such as rail flaw detecting car and track circuit, have some shortcomings. Therefore, it’s necessary to design a real-time broken rail detection system for the safety of railway transportation.Thesis presents a broken rail detection system for CWR based on ultrasonic guided waves. The hardware platform is based on ARM and FPGA. The ultrasonic guided wave transducers are used to send and receive ultrasonic guided waves. And long distance detection is realized.Thesis firstly researches the theory of broken rail detection. The frequency of35kHz is chosen as the inspection frequency. The transducer is glued to the rail waist towards longitudinal direction. The attenuation curve of guided wave is measured in the testing field. Detection system based on ARM and PFGA is designed. A DC/DC boost circuit and high voltage oscillator circuit are designed to achieve a high voltage to excite transducers. A concatenation of four op amp circuits deals with the amplification and filtering of the receive signals. Then the extraction and recognition of receive signals are carried out in FPGA with the signal processing including FIR digital filter and FFT spectrum analysis. Timing optimization of the FPGA meets the requirements of real-time processing. FPGA communicates with ARM via bus structure and ARM judges whether the rail is broken or not. Then ARM uploads messages to PC. The software on PC completes the work including data acquisition, data storage and display of real-time status of rail.Thesis completes broken rail detection system for continuously welded rails based on ultrasonic guided waves. And the system is tested in different railway lines. In the experiment site of Beijing Central Iron, the false alarm rate is0.17%and the false negative rate is0. In the experiment site of Datong-Xi’an high-speed Railway achieve2km rail detection and the system is stable and accurate so far.