车下设备承载结构疲劳试验载荷谱编制方法

Compilation Method of Fatigue Test Load Spectrum for Underfloor Equipment Bearing Structure

提出多载荷时间历程的外推方法,给出车下设备承载结构疲劳试验载荷谱编制方法。基于载荷时间历程的雨流计数矩阵,利用二维核密度估计方法(Kernel density estimation,KDE),提出在多载荷时间历程输入条件下的外推方法;应用平均应力修正公式、累积损伤理论和等效损伤准则,根据外推载荷数据,给出疲劳试验程序谱的编制方法;结合动车组车下设备线路测试数据对外推方法进行验证,编制了车下设备承载结构疲劳试验程序谱,并分别使用疲劳试验程序谱和振动测试数据对承载结构进行疲劳强度计算。数据外推结果表明:多载荷时间历程外推方法计算的损伤与实际载荷损伤误差为3.6%,大的扩展倍数下KDE可以给出低周、大幅值载荷;疲劳强度计算结果表明:由于KDE外推产生了大幅值载荷,疲劳试验程序谱计算的损伤比振动测试数据计算的损伤增加了43.8%,所以,在进行疲劳试验载荷谱编制时,需要计算可能出现的大幅值载荷。此研究为其他结构程序载荷谱的编制提供借鉴案例。

Extrapolation method of multiple load histories is presented, and compilation of fatigue test load spectrum for underfloor equipment bearing structure is given. The extrapolation method is based on multiple load histories rainflow matrix and two-dimensional kernel density estimation（KDE） theory. According to the rainflow counting matrix extrapolated above, the compilation of fatigue test load spectrum is completed by applying mean stress correction formula, the cumulative damage theory and equivalent damage criterion. The extrapolation method is verified by EMU underfloor equipment line test vibration data, then fatigue test load spectrum of equipment bearing structure is compiled, and the fatigue strength of bearing structure is calculated respectively under fatigue test program load spectrum and line test vibration data. The extrapolation results of line test data show that error of damage between KDE extrapolated load and actual load is 3.6%, the low cycles and large amplitude load can be obtained by KDE extrapolation method when the extension scale is big enough. Fatigue strength calculation results show that damage from fatigue test program spectrum increased by 43.8% than that from line test data due to the large amplitude load produced by the KDE extrapolation method. So, in the compilation of fatigue test program load spectrum, it is essential to calculate the large amplitude load which may occur. This research provides load spectrum compilation reference for other bearing structure.