Publications

Abstract : Electric vehicles are becoming the front runners for both urban and rural mobility challenges on account of less pollution, rising fuel cost, better performance and environmental impacts. However, the aging and replacement cost of battery packs, resulting from dynamic and nonlinear behavior of the battery degradation is still an unresolved problem in electric automobile applications. The goal of this work is to introduce a novel machine learning based battery degradation control strategy to avert the rapid capacity loss of battery packs bearing in mind of vehicle performance. Battery currents and depth of discharge are chosen as the battery aging control parameters and performance validation is achieved by doing simulation on degrading battery pack in the electric vehicle model using various charging and discharging profiles. The proposed closed-loop control strategy is developed by evaluating different regression models using generated dataset, based on work related to data-driven power management strategy and controlling battery current limit values. From the comparison, it is observed that Gaussian process regression shows better precision over other regression models. The simulation outputs prove the ability of the proposed strategy to extend the electric vehicle battery life by 2.03% over 200,000km. This work can be extended by using deep learning based models and more charge and discharge profiles.