Enhanced intelligent algorithm and data-driven approach for driving cycle efficiency optimization of interior permanent magnet machines

To address the high computational cost of optimizing interior permanent magnet (IPM) machines for electric vehicles (EVs), this study proposes an enhanced intelligent optimization algorithm and a data-driven artificial intelligence approach, aiming to improve driving cycle efficiency. The machine topology and vehicle model are defined, and efficiency maps and driving cycle efficiency for IPM machines are established. Two distinct acceleration strategies are introduced to overcome the computational burden of finite element analysis (FEA)-based optimization. The first strategy employs the enhanced intelligent optimization algorithm to simultaneous optimize four IPM topologies through a unified ranking mechanism, significantly reducing FEA evaluations compared to conventional optimization. The second strategy adopts a pretrained artificial neural network to replace FEA in the optimization loop, enabling highly efficient individual topology optimization. The optimized designs are validated by fabricating prototypes and conducting performance tests. Comparative analysis between FEA and experimental results confirms that the proposed methods are effective and significantly reduce computational time.