Comprehensive global optimization of an implicit constrained multi-physics system for electric vehicles with in-wheel motors

This study establishes an implicit constrained global optimization approach for electric vehicles with in-wheel motors. Limited wheel space and specified vehicle performance make the traditional global optimization algorithms ineffective. An innovative optimization approach is proposed to provide a full understanding of the design space and to find the optimal implicit constrained global solution. First, an in-wheel motor is modeled as a multi-physics system considering structural mechanics, electromagnetism and thermal physics. Then, an optimization problem is built by applying structural geometry and vehicle performance to the multi-physics system as constraints. The structural geometric relationships between the wheel and motor are applied as linear constraints. Vehicle performance derived from typical load cases is treated as an implicit constraint that cannot be satisfied directly in typical optimization algorithms. An innovative procedure is then suggested to explore the design space governed by implicit constraints. Finally, feasible subspaces are distinguished and illustrated in detail. Optimizations based on feasible subspaces are performed. The accuracy and computational costs of the proposed approach are also discussed. The results showed that the proposed approach provides a better solution for the implicit constrained global optimization of multi-physics systems, such as electric vehicle driving systems and energy management, at the same computational cost as that of existing approaches.