Theoretical Comparison and Machine Learning Based Predictions on Li-Ion Battery’s Health Using NASA-Battery Dataset

This article highlights the potential of predicting cell parameters, such as potential and current, for the optimization of electrochemical processes. We used Python to make complex ML predictions. Lithium-ion (Li-ion) batteries are rechargeable batteries used for a variety of electronic devices, which range from electric vehicles, smartphones, and even satellites. A set of Li-ion batteries run through different operational profiles (charge, discharge and impedance) at room temperatures, and time. Here, input (features) parameters like, Impedance measurement, Voltage measured, Current measured, Temperature measured, Current load, Voltage load, Time, under charge and discharge cycles, for batteries evaluations and predictions. This NASA dataset is classified as a time-series dataset. Electrochemistry in battery making is one of the prevalent and proven sciences for Efficient Batteries for the Automobile Industry. With the advent of Material Science and Technology, it also embraced the electrochemistry utility for several electrochemical parameter customizations. Not only that, electrochemistry has been a proven asset for sensor technology and budding defense technology. Meanwhile, the development of nano-/ultrafine-grain metallic materials with improved strength has improved ion trapping and high electrode-potential composites and are prominent potential candidates for future fuel cell and Hydrogen trapping based cell’s technologies. Here, we have taken the NASA Battery Dataset study that employs machine learning models, such as linear regression, decision tree, random forest, and support vector regression, to predict battery capacity based on features such as voltage, current, temperature, and time. Descriptive statistics, capacity vs. cycle plots, and capacity vs. state of health (SoH) plots are used to analyze the dataset and visualize the battery performance over time. The correlation between different features is examined using a correlation heatmap. The mean squared error (MSE), root mean squared error (RMSE), and high R2 scores are also calculated.