DETERMINATION OF MAXIMUM COIL HEIGHT IN VARIOUS SLOT GEOMETRIES OF ROTARY ELECTRIC MACHINES USING SUPPORT VECTOR REGRESSION (SVR)
8th International Bursa Scientific Research Congress December 16-18, 2025, Bursa, TURKEY, 16 - 18 December 2025, pp.842-851, (Full Text)
- Publication Type: Conference Paper / Full Text
- Page Numbers: pp.842-851
- Open Archive Collection: AVESIS Open Access Collection
- Bursa Uludag University Affiliated: Yes
Abstract
ABSTRACT This study presents a Support Vector Regression (SVR)–based framework for predicting the maximum coil height in various slot geometries of rotary electric machines. Five practical slot forms were modeled using key geometric parameters, and a dataset was generated to reflect realistic dimensional ranges encountered in industrial machine design. The proposed SVR model, trained using MATLAB with 20-fold cross-validation, achieved excellent predictive performance (R² = 0.9998, RMSE = 0.021), demonstrating near-perfect agreement between predicted and true values. Residual trends, parity plots, and Shapley-value–based feature importance analysis confirmed the accuracy and physical consistency of the model, with slot-depth parameters (h3, h4) identified as the most influential predictors. The results show that the proposed approach provides a fast, reliable, and computationally efficient alternative to repeated numerical simulations, enabling rapid evaluation and optimization of different slot forms. Keywords: Support Vector Machines, Slot Geometry, Coil Height Prediction, Rotary Electric Machines, Machine Learning