Dual-Stage Hybrid State-of-Charge Estimation of Lithium-Ion Battery Pack

Estimation of State-of-Charge (SoC) of a battery pack is an essential function of Battery Management System in Electric Vehicles. This letter proposes a hybrid dual-stage estimation approach that combines data analytics with the classical OCV-SoC approach. The first stage of the algorithm estimates the OCV during the runtime of an EV, and based on the OCV, the SoC is estimated in the second stage. The framework uses piecewise linear regression to compute the runtime OCV in the first stage and a polynomial regression approach in the second stage to find the SoC using the OCV-SoC relation. The data required for training the model is obtained from the Urban Dynamometer Driving Schedule, and the trained model is tested on Dynamic Stress Test and other driving profiles in a Python environment. Further, an experimental setup consisting of 4 Li-ion cells arranged in a 2S2P battery pack configuration is developed to validate the proposed approach for SoC estimation.