Functionally graded structure design for magnetic field applications

This study presents topology optimization methodology of anisotropic magnetic composites, which consist of two ferromagnetic materials with low and high reluctivity values, considering the nonlinear magnetic saturation effect. Instead of employing the asymptotic homogenization theory, the representative volume element method combined with the machine learning is used to build the continuous function model and it is applied to obtain the material property according to the design variable change. Finally, the micro-scale functionally graded structure composed of two ferromagnetic materials with the macro-scale topological morphology is simultaneously designed to improve the magnetic performance of actuators. Numerical examples for symmetric and asymmetric magnetic actuator models are provided to validate the effectiveness of proposed design process. In the numerical results, optimized configurations and objective values obtained with the nonlinear magnetic composite material, which depend on the intensity of the magnetic flux density, are compared with those of the linear magnetic composite material. •An optimization method for magnetic problems using ML and RVE is proposed.•Linear and nonlinear magnetic reluctivity properties were derived using the RVE.•ML was used to build the continuous function model for the magnetic reluctivity.•FGS design provides improved performance than isotropic multi-material design.•The validity of the proposed method is verified through numerical examples.