Optimal Design of Surface-mounted Permanent Magnet Synchronous Motor for Robot Joint using Subdomain-Subdivide Assisted Method

This paper deals with the optimal design of SPMSM for robot joints using subdomain-subdivide assisted method (SSAM). SSAM is a multi-objective optimization algorithm that can simultaneously consider various objective functions. A meta model is used to solve the problem of calculation time, which is a problem of existing algorithms. In addition, the periphery of the predicted solution is defined as subdomain and subdomain is subdivided to increase the accuracy of the solution. This method has the effect of shortening the interpretation time as well as the accuracy of the solution because it does not subdivide the entire area. In the optimal design of SPMSM, design variables sensitive to objective functions are extracted using the Taguchi method and applied to the proposed algorithm. As a result, it was confirmed that the electromagnetic characteristics of the optimal model were superior to the initial model.