Multi-objective co-optimization of powertrain sizing, energy management, and eco-driving for the architectural design of electric vehicles
Ye Gao, Lei Feng, Ding Liu, Zhiwu Li
Abstract
The architectural design of the electric vehicle (EV) powertrains profoundly impacts the EV’s cost, energy efficiency, range, and other properties. Integrated and optimization-based systems engineering is critical for the early design phase of EVs; however, conventional early-phase design methods are primarily subjective and qualitative. The optimal designs on structures, component sizes, efficient speed trajectories, and energy management control are performed either heuristically or separately. This paper presents a model-based systems engineering (MBSE) methodology for evaluating and optimizing EV powertrain architectures. The methodology contains three key contributions: (1) a formal MBSE toolbox that supports the design, code-generation, and co-optimization of EV architectures, (2) a simultaneous co-optimization method that integrates sizing, control, and eco-driving, and (3) a computationally efficient multi-objective optimization solution applicable for large-scale problems. The effectiveness and efficiency of this approach are demonstrated through the optimization and comparison of three distinct powertrain architectures. Compared with a reference EV, our co-optimization method reduces the energy cost by around 8 % for highway driving conditions and around 13 % for urban driving conditions. The component cost of the EV may also be reduced by around 10 % . Compared with the efficient multi-objective NSGA-II algorithm, the proposed method obtains equivalent results with more than 90 % time reduction.