Coupled Analysis and Metamodel-Based Optimal Design of Interior Permanent-Magnet Synchronous Motor Considering Multiphysical Characteristics

A metamodel-based optimal design was proposed and coupled analysis was performed to evaluate the multiphysical characteristics of an interior permanent-magnet synchronous motor, including its electromagnetic, thermal, and mechanical stress fields. To enhance the predictive performance of the metamodel in the complex design problem based on coupled analysis, seven metamodeling techniques were considered instead of a single approach. Subsequently, the metamodel was combined and designed with an evolutionary algorithm to optimize and improve the specific power of the motor; the objective here was to maximize the specific power by selecting the average torque, torque ripple, efficiency, winding temperature, and rotor safety factor as the constraints. The most suitable metamodeling technique was then selected for each output variable by comparing the coefficient of prognosis as the evaluation metric of the predictive performances of the seven metamodeling techniques. The optimal design resulted in 9.1% improvement of the specific power of the motor compared to the original design. The error rates were compared between the predicted values and finite-element analysis results at the derived optimal point, which confirmed that consideration of the most suitable metamodeling technique for each output variable can be useful in generating a metamodel with superior predictive performance.