计算机科学
推论
数据挖掘
软件
贝叶斯概率
贝叶斯推理
机器学习
人工智能
作者
Teng Wang,Zheng Liu,Min Liao,Nezih Mrad
出处
期刊:Proceedings of the Annual Conference of the Prognostics and Health Management Society
[PHM Society]
日期:2020-11-03
被引量:1
标识
DOI:10.36001/phmconf.2020.v12i1.1261
摘要
Nowadays, the concept of “digital twin” has received great attention from both academia and industry. However, few methodological solutions have been reported in existing studies. This paper presents a life prediction method for aircraft structure, and illustrates how this method can be embedded into a “digital twin” framework. This method can fuse heterogeneous information acquired from inspected physic entity, fifinite element software, historical database and predictive model, giving an accurate and real-time prediction of remaining useful life (RUL) for aircraft structure. In the operation of this method, the degradation behaviour of inspected structure is observed in an online manner. Historical record document is used for generating prior knowledge. The external load condition is fed into fifinite element software for calculating the stress intensity factor. The well-known Paris law is adopted as predictive model. Finally, the Bayesian inference is used to integrate the information and predict the future degradation of inspected structure. Theoretical deviation and experiment on a public database demonstrate the effectiveness of this method, facilitating the implementation of “digital twin” in real-world scenario.
科研通智能强力驱动
Strongly Powered by AbleSci AI