At a time when the automobile must comply with increasingly stringent environmental issues, the powertrain sizing is a major concern in the design of a hybrid electric vehicle which is at the heart of the energy savings challenge. In this context, the present work introduces a methodological approach to pre-size the drivetrain elements (engine, electric machine and battery) given the mission profiles related to the vehicle use (urban, extra urban) and considering no prior choice in selecting either the powertrain structure (serial or parallel) or the components rated power. The originality of the work is focused on two prime advances. On one hand, a model of the vehicle use is implemented to characterize a mission defined by the triplet of variables {speed; acceleration; road slope}. This model, based on the Markov matrix formulation, preserves the correlation between these variables and their statistical characteristics. Further to this modeling, a large family of vehicle missions can be randomly generated. On the other hand, the hybrid powertrain is modeled along a power flow approach for both series and parallel structures. Generic per-unit models of the components are used in order to avoid an a priori choice of their rated power and an online energy management maximizing the drivetrain efficiency is proposed. Finally, a sizing algorithm is implemented to minimize the vehicle fuel consumption on a set of simulated vehicle missions. The resulting sizing is thus optimized with respect to the intended vehicle use.
