Physics-Informed Deep Neural Network for Bearing Prognosis with Multisensory Signals  

作　　者：Xuefeng Chen Meng Ma Zhibin Zhao Zhi Zhai Zhu Mao 

机构地区：[1]School of Mechanical Engineering,Xi’an Jiatong University,Xi’an Shaanxi 710049,China [2]Department of Mechanical and Materials Engineering,Worcester Polytechnic Institute,Worcester,MA,USA

出　　处：《Journal of Dynamics, Monitoring and Diagnostics》2022年第4期200-207,共8页动力学、监测与诊断学报（英文）

基　　金：support in part by China Postdoctoral Science Foundation (No.2021M702634);National Science Foundation of China (No.52175116).

摘　　要：Prognosis of bearing is critical to improve the safety,reliability,and availability of machinery systems,which provides the health condition assessment and determines how long the machine would work before failure occurs by predicting the remaining useful life(RUL).In order to overcome the drawback of pure data-driven methods and predict RUL accurately,a novel physics-informed deep neural network,named degradation consistency recurrent neural network,is proposed for RUL prediction by integrating the natural degradation knowledge of mechanical components.The degradation is monotonic over the whole life of bearings,which is characterized by temperature signals.To incorporate the knowledge of monotonic degradation,a positive increment recurrence relationship is introduced to keep the monotonicity.Thus,the proposed model is relatively well understood and capable to keep the learning process consistent with physical degradation.The effectiveness and merit of the RUL prediction using the proposed method are demonstrated through vibration signals collected from a set of run-to-failure tests.

关 键 词：deep learning physics-informed neural network(PiNN) Prognostics and Health Management(PHM) remaining useful life 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]