2nd European International Conference on Industrial Engineering and Operations Management

A deep long-short-term-memory neural network for lithium-ion battery prognostics

Ahmed Zakariae Hinchi
Publisher: IEOM Society International
0 Paper Citations
1 Views
1 Downloads
Track: Big Data and Analytics
Abstract

With the increasing challenges in energy storage, the importance of lithium-ion batteries reliability cannot be understated. The prediction of the battery exact time of failure can provide a cost-efficient maintenance plan.  In this paper, we propose a novel data-driven approach based on deep long-short-term-memory neural networks LSTM for battery's remaining useful life (RUL) estimation. The suggested method uses the past battery capacity, the time to discharge and the operating temperature to directly predict the RUL. To validate the proposed model, we conduct experiments using the NASA lithium-ion battery dataset. The results show that our method produces exceptional performances for RUL prediction under different loading and operating conditions.

Published in: 2nd European International Conference on Industrial Engineering and Operations Management, Paris, France

Publisher: IEOM Society International
Date of Conference: July 26-27, 2018

ISBN: 978-1-5323-5945-3
ISSN/E-ISSN: 2169-8767