S变换的火车车轮扁疤信号检测

为准确、快速检测车轮扁疤以保障列车行车安全,针对振动加速度传感器采集的车轮扁疤信号持续时间短、突变快的特点,采用短时能量判决和S变换时频分析方法相结合的算法对含有扁疤的振动信号进行检测定位。对典型扁疤信号进行频谱分析发现扁疤信号主要集中在2500Hz以下的频带范围内。通过对低通滤波后的振动信号进行短时能量判决筛选出扁疤可能存在的数据段,经S变换后,信号的突变特性表现明显,可以确定扁疤的起振时间。MATLAB仿真证明：该算法能够准确地对扁疤信号进行检测定位,并较其他方法步骤更简洁、定位更准确。

Detection of wheel-flat signal based on S-transform

In response to short duration and fast mutation characteristic of vibration accelerometercollected wheel-flat signal,this paper is an attempt to ensure the train＇s traffic safety,as a result of an accurate and rapid detection of wheel-flat.The paper introduces the detection and location of vibration signal with wheel-flat by an algorithm combining short-time energy estimation with time-frequency analysis method based on S-transform.Spectrum analysis of typical wheel-flat signal shows that its frequency range is focused below 2 500 Hz.The algorithm consists of selecting data segment where wheel-flat may possibly arise by the short-time energy estimation of vibration signal subjected to low-pass filtering,making signal mutation prominent by S-transform,and determining the starting vibration time of wheel-flat.Simulation results of MATLAB shows that the algorithm,capable of effective detection and location of wheel-flat signal gives a more concise step and an accurate location than other methods.