Optimization of Plug-in Hybrid Electric Vehicles

Doctoral thesis, 2015

The rising concerns about the global warming and emissions on one hand, and the limited resources of fossil fuels on the other hand, have made electrification of vehicles a necessary topic among researchers and companies. Hybrid electric vehicles (HEV), which in addition to a primary power source, such as internal combustion engine or fuel cell, have an electric motor and an electric energy storage, such as a battery, have proved to decrease the fuel consumption. This is mainly due to regeneration of the braking energy, possibility to turn the engine off at low power demands, and higher efficiency gained from the extra freedom in choosing the engine operating points and downsized engine. Plug-in hybrid electric vehicles (PHEV) have the additional ability to run on electrical energy charged from the electrical grid due to their large capacity batteries. However, having extra electrical components in these vehicles, which results in higher cost, opens new questions concerning both the energy management and sizing of the components. This thesis further develops the application of convex optimization to simultaneously minimize operational and components costs. This means that besides the optimal component sizes, the optimization gives the optimal energy management strategy. Two different configurations, namely parallel and series PHEVs, are investigated. For a parallel PHEV, the effect of different performance requirement levels and battery prices on the optimal costs and sizes are investigated. For a series PHEV, the effect of driver’s driving and charging behaviors, performance requirements, and pricing scenarios on the optimal component sizes in different configurations are studied. To generate driving cycles that reflect driving patterns of different drivers, a systematic method based on Markov chain is used. Moreover, the impact of reduction in modeling detail is investigated on both computational time and accuracy of the results in the optimal sizing of a fuel cell PHEV. To cope with the integer variables in the problem, an iterative method using dynamic programming and convex optimization is introduced.