Development of a Dynamic Model of Lithium Ion Battery Pack for Battery System Monitoring Algorithms in Electric Vehicles

The main objective of this paper is to develop an accurate and a self-corrective model for lithium ion battery pack, based on the analysis of properties and performance of equivalent circuit models of pack's cells and the use of artificial neural networks. This model is expected to meet the requirements for improving the accuracy of battery state estimation required for the monitoring algorithms of the battery system in electric vehicles.The use of the neural network will make it possible to adjust and correct modelling deviations and errors in equivalent circuit model parameters by comparing the modelled cells voltage and the measured voltage and generating modified model parameters based on this comparison.The validation experimental results show that the parameters of the modified model contributed to reducing the error rate in the cell voltage by approximately 70-80% of the error rate in the equivalent circuit model voltage.