Temperature Prediction of Automotive Battery Systems under Realistic Driving Conditions using Artificial Neural Networks

Abstract

The accurate prediction of the battery temperature in an electric vehicle is crucial for an effective thermal management of the battery system. Here, a nonlinear autoregressive exogenous network is used to model the complex thermal behavior of a battery cell. It is trained with conventional driving data and uses input parameters that are easy to obtain. Its accuracy is proven for a wide range of temperatures, showing the simple, general and robust applicability of the approach.