Research And Development Of The Monitoring System For Locomotive Wheel Tread Scrape

The detection of the locomotive wheel tread scrape is one of the key research points in the railway field home and abroad. But because the bad surroundings and the disturbance sources are too complicated, nowadays there is no such equipment that can measure the scrape of the wheel tread accurately. As far as the detecting methods are concerned, there are the methods of the rail electricity disrupting, vibration detecting, CCD image, supersonic wave detection, strain gage and others. In this article, the testing system of the wheel tread scrape is developed based on the vibration detecting method.According to the calculating model of the wheel-rail system, we analysis the relationship between the wheel tread scrape and rail impact so as to monitor the scrape degree by the vibration parameters. But in consideration of the limitation of the conventional signal processing methods, this article expounds the fault information extraction method of the tread scrape based on the discrete cosine transform and Hilbert-Huang transform.In order to simulate the faults accurately of the tread, bearing and other parts of the wheel in laboratory, we develop the wheel fault diagnosis experiment device which contains of mechanical supporting, electrical transmission and electrical control. The electrical transmission includes: friction wheel, motor and compressor while the electrical control involves frequency controller, computer and the control interface circuit.Then the monitoring system which composed with three parts is designed according to the characteristics of the tread scrape signal. The three parts are sensors, data acquisition modulate and signal processing circuits. The signal is collected by vibration sensors that fixed on the bearing base. Then through the sensor interface and channel selecting, and must be dealt by current/voltage converting circuit, coupling type chosen, derivation and some other signal processing circuits. Finally the digital signal is analyzed by IC TMS320VC33.Finally, in the testing chapter, the wheel fault diagnosis experiment device is tested firstly and proved that it can satisfy the design requirement. Then the three referred methods are used to analysis the time domain signal of the tread scrape. And the results present that the discrete cosine envelops and Hilbert-Huang transform can recognize the tread scrape accurately while the FFT can't. Meanwhile, these two methods can satisfy the real-time and on-line detection by virtue of its simple algorithm.