Viability of Long Short Term Memory Networks in the Health Prediction of Lithium-ion Batteries

Lithium-ion batteries play a crucial role in diverse aspects of present day life owing to their distinctive attributes and adaptable uses, spanning from everyday electronic devices to Electric vehicles(EVs) and sustainable energy systems.The life and operational features of rechargeable batteries are significantly influenced by Battery Management System (BMS). Implementation of BMS is essentially aimed at overseeing and regulating various aspects of a battery's functioning to optimize overall efficiency, ensure safety, and prolong battery life. The principal concern associated with batteries is battery degradation, which is the slow deterioration of a battery's capacity over time, making it harder for it to hold a charge. It is a natural and inevitable process that occurs due to various factors.The application of artificial intelligence for State of Health (SoH) prediction in batteries, a critical aspect in assessing their remaining useful life. The process entails thorough data collection from the battery system, including charge-discharge cycles, temperature profiles, voltage, and current data. Subsequently, relevant features are selected, and the data undergoes preprocessing to address missing values and inconsistencies.Long Short Term Memory (LSTM) networks are trained on the dataset, allowing them to learn patterns or characteristics features in voltage(v), current(I) and temperature(T) to predict battery's health.Successful implementation depends on the quality of training data and appropriate feature selection. The simulations were performed using MATLAB software and the results are presented.