|The built-in Energy Management System (EMS) of Plug-in Hybrid Electric Vehicles (PHEVs) plays an important role in the fuel efficiency of these vehicles. Recently, it has been revealed that prior knowledge of the upcoming trip can assist EMS to enhance the distribution of power between the energy sources, i.e. the engine and the motor-generators used in PHEVs, resulting in lower fuel consumptions. This dissertation intends to further investigate on a Trip Planning-assisted EMS (TP-assisted EMS), by studying its feasibility for online implementation, and evaluating its performance and robustness with respect to the trip data uncertainties in various practical scenarios, to ultimately answer this question: Does the TP-assisted EMS function as a reliable system for
PHEVs which can outperform conventional methods?
This research starts with improving upon an existing Trip Planning module with an emphasis on its online integration with the EMS module. In particular, the power-balance model of PHEVs is introduced, which is computationally inexpensive and yet adequately accurate to be used for the optimizations involved in the Trip Planning module. To speed up the optimizations, the use of Particle Swarm Optimization (PSO) algorithm is suggested. These modifications result in the reduction of computational time, making TP-assisted EMS module suitable for online implementations.
Once the TP-assisted EMS module has been integrated with a high-fidelity model of the baseline PHEV, namely, 2013 Toyota Prius PHEV, its performance and sensitivity/robustness have been extensively studied through Monte Carlo simulations, where numerous samples of standard as well as real-world drive cycles have been tested. However, in order to use these data for Model-in-the-Loop (MIL) and Hardware-in-the-Loop (HIL) tests, a Micro-trip Generator block has been developed. This block automatically segments the drive cycles, similar to the way that trip information is obtained in practice, making the simulation samples compatible with the Trip Planning module.
Statistical analyses of the simulation results show that the TP-assisted EMS is a superior controller compared to the conventional EMS strategies. Moreover, these simulations present one of the first sensitivity analyses that have been performed in the context of TP-assisted EMS for PHEVs, showing that this system is robust despite the existence of random disturbances and meanwhile has low sensitivity against variations of the design parameters.