Artificial Intelligence Techniques for a Scalable Energy Transition by Unknown

Author:Unknown
Language: eng
Format: epub
ISBN: 9783030427269
Publisher: Springer International Publishing

8.4 HI Trend Modeling for RUL Estimation

As illustrated in Fig. 8.3, the modeling approaches of HI trends for the estimation of RUL can be decomposed into three main families: physical approaches, data-driven approaches, and expert methods. Another classification is proposed in [2], where the RUL estimation methods are classified into: reliability based, similarity based, model based, and data-driven based approaches. The most used physical model form is the differential one, whose order and parameters are identified according to the physical knowledge and data available on the degradation process. The updating of the parameters makes it possible to compensate the modeling uncertainties and the adaptation of the model to changes in the degradation rate. Data-driven approaches are the most used and can be decomposed into five families: statistical models, stochastic models, deterministic models, probabilistic models, and machine learning model. The third family of trend modeling comprises those approaches that formalize the knowledge of industry experts through the tools of fuzzy logic and Bayes probabilities. Only data-driven methods able to include expert knowledge into the prediction models are presented in this paper.

Fig. 8.3Classification of RUL estimation approaches

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.