Design optimization of Permanent Magnet Brushless Direct Current Motor using Radial Basis Function Neural Network

This paper is about a methodology for the optimization of a Permanent Magnet Brushless Direct Current (PM-BLDC) motor. The most advantage of this proposed method is its mathematical modeling effectiveness. In specific, it is focused on multi-objective optimization by using a Radial Basis Function (RBF) Neural Network simulated in the Matlab environment. The aim of this optimization process was to maximize the efficiency and to minimize the permanent magnet mass, active mass, and volume of the motor. In order to verify results, two-dimensional models were developed and thoroughly analyzed using Finite Element Analysis (FEA) in Ansys-Maxwell. Moreover, the comparison of the RBFNN and Genetic Algorithm (GA) results were also figured out and the comparison showed that the RBFNN has better ability in finding the optimal solutions and also has less computational time than GA.