摘要
剩余使用寿命(remaining useful life,RUL)预测是保障机械装备可靠性、可用性和安全性的重要技术.本文提出一种基于嵌套长短期记忆(nested long short-term memory,NLSTM)网络的机械装备RUL预测方法,它通过融合多传感器监测信号,实现对机械装备RUL的精确预测.区别于普通LSTM网络,NLSTM将存储单元进一步加深,将一个LSTM神经元结构嵌套在原有LSTM的存储空间中,实现对多传感器时间序列信号中长期依赖性的深度捕捉.本文使用涡扇发动机和加工刀具两个实验案例来验证NLSTM的预测性能;从涡扇发动机案例验证可知,相比于LSTM,NLSTM的预测性能在两个指标上分别整体提升了4.66%和15.18%,且NLSTM的预测结果也优于文献中的其他先进方法;从加工刀具案例验证可知,NLSTM的预测结果在六个刀具上的预测结果均优于LSTM.
Remaining useful life(RUL)prediction is an important technology to ensure the reliability,availability and safety of mechanical equipment.This paper proposes a RUL prediction method based on nested long short-term memory network(NLSTM)for mechanical equipment.This method realizes accurate prediction of RUL of mechanical equipment by fusing multi-sensor monitoring signals.Different from ordinary LSTM networks,NLSTM further deepens the storage unit and nests an LSTM neuron structure in the original LSTM storage space to achieve the deep capture of the long-term dependence in the multi-sensor time series signal.This paper uses two experimental cases of turbofan engine and machining tool to verify the prognostic performance of NLSTM.In the case of turbofan engine,it can be seen that compared with LSTM,the prediction performance of NLSTM has an improvement of 8.72%and 15.93%,respectively,on the two indicators,and the prediction results of NLSTM are also superior to other advanced methods in the literature.In the machining tool case,it can be seen that the prediction results of NLSTM are better than those of LSTM on all six machning tools.
作者
程一伟
朱海平
吴军
邵新宇
CHENG YiWei;ZHU HaiPing;WU Jun;SHAO XinYu(School of Naval Architecture and Ocean Engineering,Huazhong University of Science and Technology,Wuhan 430074,China;School of Mechanical Science and Engineering,Huazhong University of Science and Technology,Wuhan 430074,China)
出处
《中国科学：技术科学》
EI
CSCD
北大核心
2022年第1期76-87,共12页
Scientia Sinica(Technologica)
基金
国家自然科学基金(批准号:51875225,52075202)资助项目。
关键词
循环神经网络
嵌套长短期记忆网络
剩余使用寿命预测
多传感监测数据
机械装备
recurrent neural network
nested long-short term memory network
remaining useful life prediction
multi-sensor monitoring data
mechanical equipment
