PREDICTION FOR LITHIUM-ION BATTERIES BY MEANS OF A RANDOM FOREST ALGORITHM

Abstract

Lithium-ion batteries have a minimal environmental effect, a long cycle life, and a high energy density, making
them ideal for storing power. However, it is subject to ageing, which means that after a specific amount of years, its
capability can decline and it might fail more frequently. The reliability and safety of the device relies heavily on how
close the estimate of remaining battery life really is. This research argues that machine learning may be used to
develop a solution that can estimate how much longer a lithium-ion ones (li-ion) battery will last in service. The
voltage generated by the output levels were monitored in the thingSpeak software, and the traditional substances
connected by the discharge voltage information were considered to predict the battery's longevity after the li-ion
battery was connected to a load. The dataset was evaluated and trained using random forest methods to predict the
battery's lifetime.After factoring in everything discussed, the estimated number of remaining battery cycles is arrived
at.In this study, ML and edge impulse are used to enhance the accuracy of Lithium ion battery life prediction
algorithms. To better estimate the RUL of batteries made from lithium ion, this research has recently included the
random forest approach.