Exploiting driving history for optimising the Energy Management in plug-in Hybrid Electric Vehicles

Abstract

This paper proposes an Energy Management Strategy (EMS) for a plug-in parallel Hybrid Electric Vehicle (pHEV) with the goal of minimising the fuel consumption while fulfilling the constraint on the terminal battery State-of-Charge (SoC). The proposed strategy assumes that the route was previously covered several times by the vehicle, in order to extract information about the feasible operating conditions in the driving cycle. Note that this situation is usual in commuting and daily trips. In this sense, the history of vehicle speeds and positions are used to build space-dependent transition probability matrices that are latter used for driving cycle estimation by means of Markov-Chain approach. Once the driving cycle is estimated, the torque-split problem in parallel hybrid powertrain is addressed using the Equivalent Consumption Minimisation Strategy (ECMS), where the associated boundary value problem of finding the weighting factor between battery and fuel cost that drives the SoC to the desired level at the end of the estimated cycle is solved and applied to the system. Finally, in order to make up for cycle estimation error, the ECMS is solved recurrently. For the sake of clarity, the proposed strategy is initially developed and analysed in a modelling environment. Then tests in an engine-in-the-loop basis are done for validation. In order to show the potential of proposed strategy, results are presented using a trade-off between the fuel consumption and the terminal-SoC for four different methods: the optimal power-split that requires a priori knowledge of the driving cycle for benchmarking, online ECMS with a fixed cycle estimation (average speed profile obtained from previous trips on the route), the proposed method, i.e. online ECMS with a dynamic cycle estimation and finally a rule-based charge depleting and charge sustaining strategy. The results demonstrate that the online ECMS outperforms the rest of online applicable methods.