Using Machine Learning Techniques to Estimate the Remaining Useful Life of a System with Different Types of Datasets

Document Type

Conference Proceeding

Publication Date

1-7-2021

Publication Title

Lecture Notes in Networks and Systems

Volume

182

First page number:

139

Last page number:

147

Abstract

© 2021, Springer Nature Switzerland AG. Intelligent Prognostics and Health Management (PHM) technology aims to estimate the Remaining Useful Life (RUL) of a subsystem or a component using data collected by sensors. The rise in the complexity of systems requires new models to capture the relationship between the sensors and the RUL. Novel deep Convolutional Neural Networks (CNN) have been proposed as an approach for estimating the RUL. However, large amount of data is needed to use Machine Learning (ML) techniques. We explore current ML methods being used with different types of datasets and provide a conclusion on deciding what learning method works best with unique datasets. We find that, for most systems, the ML method used highly depends on the dataset and can greatly decrease the cost and increase the reliability.

Keywords

Convolutional neural networks; Deep learning; Degradation data; Fault detection; Prognostics and health management

Disciplines

Electrical and Computer Engineering

Language

English

UNLV article access

Share

COinS