Reserach On The Virtual Alarm System Of Approaching Train Based On Detection Of The Wheel/rail Noise

Safety is the lifeline of railway traffic. With the sixth railway big acceleration and the Electricity Multiple Unit (EMU) operation, the train operation maximum speed has 250km/h. On one hand it greatly enhance the convenience, on the other hand increase the dangerous of railway constructor .Therefore the protection of construction puts forward a higher request. For the sake of reducing the loss of state and people's property and safeguarding the people's life safety, a lot of units and scientific research institution of this issue with developed a variety of trains approaching alarm device, but they all existed some disadvantages.Alarm problem of approaching train was studied by adopting the method of detecting and analyzing the vibration information of the rail, the main work and conclusions are as follows.(1) Understood the home and abroad research situation, the existing Alarm methods of approaching train were introduced and analyzed, the ideas to alarm the approaching train was be put forward by detecting and analyzing the solid sound in the rail.(2) The mathematical model of the vibration by wheel-rail driving was established, through the Euler beam approximation and the reduced order processing, the mathematical model can be expressed a second-order differential equations with constant coefficients.(3) Solving mathematical model, the natural frequencies of long rail were obtained; experimental results show that the rail vibration signal below 1000Hz, its characteristic frequencies are coincident with some order natural frequencies of the rail.(4) The train approach alarm system's overall scheme design was completed .selection of the main hardware, developed the Data Acquisition card(UA302H) driver suitable for the LabVIEW platform, developed the data acquisition and data analysis program by MatLab and LabVIEW, constructed the virtual alarm system of approaching train, after tested, the system can normal operation.(5) Worksite data was acquired and off-line analyzed by adopting piezoelectric acceleration sensor instead of the solid acoustic sensor, including power spectrum analysis and wavelet analysis and obtained some valuable conclusions.â‘ The rail vibration signal below 1000Hz distinct contain the natural frequencies of long rail, especially the 14th 18th 21st 24th 29th order natural frequencies.â‘¡From the characteristics frequency's amplitude variation, rule is not clear when the train far distance from the sensor, but when the train distance sensor is close, amplitude change a big trend, especially the 24th, 29th order natural frequencies'amplitude variation trend evident when the train distance to the sensor the 1000-100 meters.Due to restrictions, the experimental analysis not only in one place and trains under the condition of low speed, but also were used the ordinary piezoelectric acceleration sensor, although obtained some valuable conclusions, the conclusion was applicable to different site environment needed further validation; In addition, using solid voice signal detection sensor can realize more remote alarm, whether the characteristics more distinct needed further experimental demonstration.