Energy management of battery electric vehicles hybridized with supercapacitor using stochastic dynamic programming

In this study, a stochastic Energy Management System (EMS) for Battery Electric Vehicles (BEVs) hybridized with the supercapacitor is proposed. At each moment, the EMS should determine an optimal power distribution between the supercapacitor and the battery. Because of the uncertain nature of the power demand, an effective EMS should be able to handle uncertainties. As a result, a Stochastic Dynamic Programming (SDP) approach has been proposed and demonstrated to be successful. In this investigation, the power demand has been predicted based on a Markov chain assumption using some real drive cycles data points. The used drive cycles are categorized in two groups, which are training drive cycles and test ones. The Transition Probability Matrix (TPM) is built by the training cycles; meanwhile simulation results are based on the test drive cycles. In comparison to the results of other methods, the SDP results show more improvements. In addition, in terms of computational costs, it has a significant advantage over the other rival approaches.