Optimum design and sizing approach for a series-hybrid vehicle

The need to reduce pollutant emissions has pushed the automotive industry towards sustainable mobility promoting new technological solutions, among which the use of hybrid powertrains stands out. The development of a hybrid architecture is very complex and demand proper components sizing and the determination of optimized power-split strategies among different power sources, for example: ICE, electric generator/motor and batteries. Moreover, the experimental analysis regarding performance and emissions requires that the whole configuration of the propulsive system must be installed on the test bench. Hence, it negatively affects the cost of the entire design phase. In this scenario, an optimum design and sizing approach for a series-hybrid vehicle is proposed. The presented procedure relies on numerical modelling of the hybrid powertrain and the optimization of vehicle autonomy, as key performance index. The series-hybrid architecture model is referred to a front-wheel-drive vehicle. It is developed by modelling the longitudinal dynamics, the driveline and the power generation system. A Power Management System (PMS) is adopted to manage the power fluxes among the on-board power devices, according to the imposed mission requirement and the dynamics behaviour of machinery. The model also includes a braking energy recovery mode allowing the battery to be recharged during the whole mission. The application of the Genetic Algorithm for optimization leads to an improvement in the overall efficiency of the powertrain, while also contributing to a reduction in CO₂-equivalent emissions. To illustrate the effectiveness of the proposed optimization approach, the BMW i3 Range Extender is analysed as a representative case study, testing its performance through a normed driving cycle. From this stage of design, the evaluation of performance and emissions for any hybrid configuration can be conducted through experimental measurements on engine test benches, providing a more cost-effective and considerably simpler assessment method.