Estimation of Li-ion Battery State of Health based on Multilayer Perceptron: as an EV Application. Issue 28 (2018)

• Record Type: Journal Article
• Authors: Kim, Jungsoo; Yu, Jungwook; Kim, Minho; Kim, Kwangrae; Han, Soohee
• Is Part Of: IFAC-PapersOnLine. Volume 51:Issue 28(2018)
• Page Start: 392
• Page End: 397
• Publication Date: 2018
• Subjects: Li-ion battery -- EV -- State of health -- SOH estimation -- Neural Network -- MLP -- Driving pattern -- Real-time -- BMS; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2018.11.734
• Languages: English
• ISSNs: 2405-8963

Abstract

Abstract: State of health (SOH) is a key issue for saving cost and guaranteeing safety while using a rechargeable battery. Therefore, numerous studies on SOH estimation have been conducted intensively. However, most of the studies need the experimental data for whole lifetime of a battery, and adopt standard charge/discharge pattern that does not reflect the real world driving pattern. For these reasons, it is not suitable to apply the results into battery management system (BMS) of an EV. In this paper, a practical SOH classification scheme based on multilayer perceptron (MLP) is proposed. Assuming that there is no data in the whole life span, classification based on neural network was performed using only data of some discrete life span. As a result of using MLP, the SOH is estimated with high accuracy in trained life span. Moreover, it still shows admittable estimation accuracy even in untrained life span.