Publications

Abstract : Battery technologies and advanced battery management systems are amongst the most trending research in automotive sectors as a result of the unprecedented push for electric vehicles. This paper, among existing methods for S tate of Charge (SoC) estimation of Lithium-Ion batteries used in Electric Vehicles (EV), explores various artificial intelligence-based and direct measurement techniques. A performance comparison of SoC estimation using the coulomb counting approach, Support Vector Machine (SVM) methods, and an optimal feed-forward artificial neural network (ANN) for different storage temperature, initial conditions, and stress tests have been presented for a Lithium-Ion battery for a variety of standard data sets. The stated models are trained using to predict SoC when voltage and current are given as inputs. Both the models are tuned and trained in a cloud-based open-source jupyter environment, collaboration. The results obtained post-performance analysis depicts the potential of ANN for accurate SoC estimation of battery used in EV. ANN has achieved a Mean Absolute Error (MAE) in a range of 0.5-1.4% over one complete cycle. This work can be further extended to validate the real-time performance of ANN with data collected from a hardware setup.