A Data-Driven State-of-Health Estimation Model for Lithium-Ion Batteries Using Referenced-Based Charging Time

Abstract

Accurate online state of health (SOH) estimation is crucial for the efficient and safe operation of lithium-ion battery packs in electric vehicles and grid-connected energy storage units. This paper proposes a novel data-driven SOH estimation model for lithium-ion batteries based on a new health indicator, namely referenced-based charging time. The proposed model utilizes the referenced-based charging time of partial charging cycles to predict the SOH using a machine learning approach. A deep feed-forward neural network, characterized via testing 90 different shallow and deep architectures, is implemented, trained, and tested on 17 batteries, which are cycled differently. The results show that the root mean square percentage error is 0.43% overall and less than 1% for each test cell.