A Battery Modeling Technique Based on Fusion of Hybrid and Adaptive Algorithms for Real-Time Applications in Pure EVs

Abstract

Battery management systems in electric/hybrid vehicles are entirely based on accurate and reliable state estimation techniques. Consideration of combined effects of temperature, internal resistance rise, and capacity loss is essential to accurately estimate the current state of charge (SOC) of the battery. Most of the existing algorithms fail to capture the combined effects of the aforementioned variables. Moreover, an analysis of the inter-dependencies between the states and parameters related to the driving conditions is required. The main contribution of this paper is to introduce a hybrid and adaptive method of SOC estimation which captures the effects of initial SOC deviation and changes in SOC due to loss in capacity and rise in internal resistance. Modeling parameters related to the change in SOC, temperature, and capacity loss are continuously updated for the implemented battery model, whereby the equivalent circuit can be tuned to recent state of the battery. Simulation results are verified with experimental data from literature to prove the efficiency of the approach in terms of real-time applicability and accuracy.