Publication

Abstract : For an electric vehicle (EV) performance, range and safety of that EV depends on the state of the battery involved. This paper presents a model that employs temperature as an important feature for estimating the RUL of Li-ion batteries. In this work, simulated a 14-cell battery pack thermally using Ansys software to determine the power dissipation. Then trained an LSTM model using battery temperature data and as a result achieved low mean square error of 0.00135. RUL was predicted using real-time temperature data of a 3-cell battery pack recorded by the thermistors and processed by an Arduino Uno. The combined method proved to have a mean absolute error of 2.3997 hours; the lower values are a favorable sign for early RUL of EV batteries, which could support effective battery maintenance.