Train's Impulse Real-Time Detection And Analysis Based On Virtual Instrument

An analysis system about comfort by train was introduced by this paper based on virtual instrument. G language (LabVIEW) is used as software platform and monolithic processor and acceleration sensor (ADXL105) are used as hardware platform, composed of data acquisition, data preliminary handling, signal analysis(time domain analysis , frequency domain analysis)and impulse grade assessment, with the function of data acquisition, data storage and data analysis, and has friendly interface.This data acquisition block of this analysis system uses microcomputer(AT89C52), and other circuits as platform to data acquisition. The analogy signal is amplified and filtered by data preliminary procession block, such as amplify circuits and filter circuits.The signal time domain and frequency domain block are used to domain analysis and frequency analysis. The impulse grade assessment block of train is the core of this analysis system, using data procession functions in LabVIEW and LabWindows, and achieve the grade assessment.During the data acquisition, there are other elements except noise, which can make some difficulties in data analysis and data procession. So, it is not right to judge the impulse grade from the only accelerator. How to analysis the data and find the impulse exactly, we need some algorithm procession.An autocorrelation method was introduced by this article, and this method can detach impulse from vibration accelerator and assess the impulse grade quantitatively.The sampled accelerator are composed of accelerator signal, background noise,impulse accelerator. When the train run steadily, its R_x (0) is small. When there is impulse signal, there must be period signals in the sampled signal. According to the autocorrelation definition, the R_x (τ) will attenuate slowly in the 0 - ∞This system acquires the accelerator at different speed by train between XiangFan —WuHan. According to the impulse intensity and frequency, conducts corresponsive frequency domain analysis and autocorrelation analysis, and builds a new impulse grade assessment criterion based on the correlation peak and the width of the peak.