Lithium Iron Phosphate (LiFePo4) Batteries Health Prognosis via Plateau Identification During High-Rate Charging and Discharging

The stability and performance of lithium-ion (Li-ion) batteries are significantly impacted by high-rate loading effects. The plateau voltage and capacity are a critical parameter when evaluating the performance, stability, and overall health of a battery, particularly in rechargeable Li-ion batteries. This paper focuses on a data-driven battery management system (BMS) approach for load-sensitive applications, such as battery energy storage systems (BESS) for electric vehicles (EVs) to ensure safe and stable performance during high-rate loading. It investigates the deterioration of lithium iron phosphate (LiFePO4) batteries, which are well-known for their high energy density and optimal performance at high temperature during charge-discharge loading variation above standard current-rate (C-rate). The paper proposes a plateau voltage and capacity identification model at different loading rates to accurately determine the degree of battery ageing in load-sensitive applications such as EV. In addition, the statistical linear regression (SLR) analysis model and the deep neural network (DNN) model are the proposed methods for identifying plateau voltage and capacity. It is found that at distinct loading rates, the DNN model outperforms the SLR model, which also functions exceptionally well with a maximum accuracy of 99.9%.