Performance Enhancement of Li Ion Battery through SOC Prediction at Various Temperature: A Machine Learning-based Approach

"Battery SOC", refers to Battery State of Charge is a measurement used to indicate energy left in a battery relative to its capacity. One of the most crucial parameters in an electric vehicle’s battery management system is SOC and represents available energy hence range of the vehicle or time before the battery needs recharging. Predicting Battery SOC by traditional method using a voltage and current produce highly inaccurate results due to the battery’s nonlinear characteristics, coupled with driving behavior and ambient temperature. The research focuses on predicting SOC at different ambient temperature of 0°C,10°C, 25°C,40°C, -10°C and -20°C, using a machine learning approach. Applying Random Forest Regression, Gradient Boosting Regression, and Polynomial Regression to a data of "3Ah LG HG2 cell" for battery SOC prediction, it has been found that RFR achieves the best possible error estimates at -20°C ranging from 0.003-0.02 for different performance measures like R-squared (R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>), Mean Squared Error, Mean Absolute Error, and Root Mean Squared Error with consistent trends across other temperatures. Feature importance analysis reveals voltage’s crucial role in SOC prediction at higher temperatures, while current and temperature gain significance at lower temperatures. RFR emerges as the best model, with a high R-squared value (0.9991).