A Novel ANN Model for Performance Parameter Determination of Brushless Direct-Current Motors for Light Electric Vehicles

Abstract

The increasing adoption of light electric vehicles (LEVs) has intensified demand for efficient, high-performance electric drive systems. Brushless direct-current motors (BLDCMs) in outer-rotor configurations are the predominant drive solution for L7e-class LEVs owing to their high torque density, compact structure, and low maintenance requirements. However, electromagnetic design and performance analysis of outer-rotor BLDCMs involve complex, multi-parameter calculations and time-consuming finite element simulations. To the best of the authors’ knowledge, no prior study has simultaneously addressed multi-output performance prediction of outer-rotor BLDCMs across multiple slot–pole combinations and supply voltage levels using an ANN-based surrogate model. This study proposes such a model, capable of predicting six rated performance parameters (no-load speed, nominal speed, torque, phase current, input current, and output power) directly from design inputs, without requiring repetitive finite element analyses. Three outer-rotor BLDCMs were analytically designed and simulated for L7e-class LEVs operating at different voltage levels, and the resulting FEA dataset was used to train a feedforward backpropagation ANN. The model takes key geometric and electrical design parameters as inputs and delivers multi-output performance predictions with high accuracy. Validation against a physically manufactured motor confirmed prediction accuracy exceeding 92%, demonstrating the model’s reliability on real-world unseen configurations. The proposed model reduces electromagnetic analysis time from several hours to mere seconds, offering a practical and scalable AI-assisted framework for BLDCM design in sustainable transportation applications.