【摘要】 掘进机截割减速器的齿轮容易过早损坏。提出一种基于齿轮油磨粒信息建立的门控双注意单元（GDAU）模型，可用于解决齿轮的剩余寿命预测问题。利用掘进机截割减速器试验台对该模型进行试验验证。与长短时记忆（LSTM）网络和门控递归单元（GRU）神经网络进行对比，结果表明，GDAU模型对齿轮剩余寿命的预测效果更好，在较短时间内能够迅速收敛，其损失量仅有0.0016，与LSTM和GRU模型相比有更快的收敛速度及更高的预测精度。更多还原

【Abstract】 The gears of the roadheader cutting reducer are prone to premature damage. A gated double attention unit(GDAU) model based on gear oil abrasive particle information was proposed, which can be used to solve the remaining life prediction problem of gear. The model was tested and verified by using the roadheader cutting reducer test bench. Compared with the long short-term memory(LSTM)network and gated recurrent unit(GRU) neural network, the results show that the GDAU model has a better prediction effect on the remaining life of the gear, can converge quickly in a short time, and its loss is only 0.0016, which has a faster convergence speed and higher prediction accuracy than the LSTM and GRU models.更多还原