کنفرانس بین المللی ماشین و محرکه های الکتریکی

صفحه اصلی / The 5th International Conference on Electrical Machines and Drives

Data-Driven Prediction of Average Torque, Phase Resistance, and Coil Turns in 6/4 Switched Reluctance Motors Using ANN and EMDLAB

نویسندگان :

Nasrin Majlesi¹ Ali Jamali-Frad² Tohid Sharifi³

1- دانشگاه علم و صنعت ایران 2- دانشگاه صنعتی امیرکبیر 3- دانشگاه صنعتی امیرکبیر

کلمات کلیدی :

Switched reluctance motor (SRM)،Artificial Neural Network (ANN)،EMDLAB،Torque Prediction،data-driven approach

چکیده :

In this paper, a data-driven approach based on an artificial neural network (ANN) with high accuracy was developed to predict the average torque, phase resistance, and number of coil turns of a switched reluctance motor (SRM). An initial 6/4 SRM model was designed in ANSYS Maxwell, and to reduce the computational cost of finite-element analysis, the open-source EMDLAB software was used to generate the data required for the ANN. A total of 1,000 SRM samples with variations in stator pole arc angle, rotor pole arc angle, and stator inner diameter were designed and simulated in EMDLAB. These three geometric parameters were used as the input data to the ANN so that it could simultaneously predict the three important outputs, including average torque, phase resistance, and number of coil turns. Evaluation of the model on the test data demonstrates the model’s ability to generalize to new data. The trained network achieves a coefficient of determination (R²) close to one for all outputs, and the residual error is very small. This surrogate model significantly reduces the time required for multi-objective optimization. The results show that the proposed model is a reliable substitute for time-consuming simulations and can be effectively used in the design and optimization processes of SRM machines.