Modeling and Simulation of Speed and Efficiency of BLDC Motor as A Starter Motor Based on Multilayer Perceptron (MLP) Neural Network

Abstract

This paper present the mathematical model of Permanent Magnet Brushless DC (PMBLDC) motor which can replace the current model existing in simulation tools. Currently, PMDC motor is being used as starter motor in automotive industry. PMDC motors having brushes suffers from poor efficiency and shorter life time than BLDC motors. Validation of designed BLDC motors is traditionally being performed using simulation tools like Ansys Maxwell software which is time-consuming and not efficient when parameters change. So there is a need for developing more efficient model for BLDC motors. Here we present a mathematical model obtained using Artificial Neural Network (ANN) modeling methods which resulted in considerably faster simulation time than Maxwell software. For this paper, some geometrical parameters were set as the input of the ANN. Also, efficiency and speed were set as the ANN output. The ANN model showed considerably faster evaluation time with acceptable accuracy.