A novel data-driven SOH prediction model for lithium-ion batteries

Authors:

Abstract
With the increasing utilization of lithium-ion batteries, prognostics and health management (PHM) of batteries has become inevitable in energy system studies. Amongst, state-of-health (SOH) analysis introduces several challenges in online applications. In this respect, this paper presents a novel data-driven SOH prediction model based on machine learning methods, namely, feed-forward neural network (FFNN), support vector machine (SVM), and relevance vector machine (RVM). A novel feature extraction technique is employed, and a large dataset including 44 batteries with different cycle patterns is considered to test the accuracy of the proposed method. The results show that the maximum mean-absolute-percentage error (MAPE) of the proposed models for test batteries is 1.95%. The results demonstrate that the accuracy of the proposed method is suitable, and it can be easily implemented in large-scale applications.

This publication has 13 references indexed in Scilit: