关键词: 寿命预测, 多变量灰色模型, 经验模式分解法, 丝杠, 退化, 机床

Abstract: To study performance degradation trend and predict remaining life of ball screw, an original on-line life prediction system based on multi-variable grey model was proposed. Three acceleration sensors were installed on different places of the screw pair so that the real-time monitoring on changing trends of screw performance under different machining conditions was realized. Most sensitive characteristic parameters were extracted by empirical mode decomposition, and the non-linear mapping relationships among screw life, the three cutting elements and signal features was established by multi-variable grey model, finally life prediction model of the screw based on multi-variable grey model was set up to realize evaluation of remaining life of the screw effectively.Experimental results showed that life prediction model of the screw could effectively predict the remaining life of the screw.

Key words: life prediction, multi-variable grey model, empirical mode decomposition, screw, degradation, machining tool