基于EMD与坐标寻优的松脱件定位方法

针对目前松脱件定位方法普遍存在的定位误差大的问题，采用多传感器布局方案，提出一种基于经验模态分解（empirical mode decomposition，EMD）与坐标寻优的松脱件定位方法。EMD处理可分离信号不同振动波成分并提取信号主要本征模函数（intrinsic mode function，IMF）分量，使到达时间提取及波速计算更为准确；多传感器布局能有效降低定位偏差，通过制定传感器选取策略选择最优的传感器组合用于定位分析，实现定位的进一步优化。最后搭建平板试验台并与不同定位方法进行对比，结果显示本文方法定位精度最高，最小定位偏差在0.1 m左右。

A loose parts location method based on empirical mode decomposition and coordinate optimization

In view of the problem of large location errors in conventional loose parts location methods, a multi-sensor layout scheme has been adopted in order to develop a loose parts location method based on empirical mode decomposition (EMD) and coordinate optimization. EMD processing can separate the different vibration wave components of the signal and extract the main intrinsic mode function(IMF) component of the signal, making the arrival time extraction and wave velocity calculation more accurate. The multi-sensor layout can effectively reduce the location error, and the optimal sensor combination for location analysis can be determined by formulating a sensor selection strategy giving a further reduction in location error. Finally, a flat-plate test rig was built and used to compare different location methods. The results show that the method developed in this paper has high location accuracy, with a minimum location error of about 0.1 m, and it can be employed in various drop tests.