关键词: 剩余使用寿命, 深度学习, 视情维护, 特征提取, 故障预测和健康管理

Abstract: Serious industrial accidents can be avoided in time by valid Remaining Useful Life (RUL) prediction for equipment.Besides,RUL prediction can also provide technical support for condition-based maintenance,which plays an important role in modern industry.In recent years,deep learning has been widely used in the field of RUL prediction by virtue of its unique advantages and potential in big data processing and feature extraction.In view of this,the latest research on deep learning in equipment RUL prediction was reviewed.Several typical deep learning methods applied to RUL prediction were introduced,whose basic principles and modeling methods were also described.Then,the applications and development trends of these typical deep learning methods in the field of RUL prediction in recent years were summarized in detail.The current challenges of deep learning-based RUL prediction along with relevant prospects were put forward and discussed.

Key words: remaining useful life, deep learning, condition-based maintenance, feature extraction, prognostic and health management