Estimation of the effect of changing resistance parameters on engine efficiency in electrical vehicles with convolutional neural network

Kitiş U., Polat Ö.

ENGINEERING COMPUTATIONS : INTERNATIONAL JOURNAL FOR COMPUTER-AIDED ENGINEERING AND SOFTWARE, cilt.2025, ss.1-17, 2025 (SCI-Expanded)

Özet

Purpose As the temperature of asynchronous motors changes, the resistance values inside them also change, which affects the efficiency of the asynchronous motor. The purpose of this study is to estimate the effect of the change of stator resistance on motor efficiency in electric vehicles with a convolutional neural network (CNN). Design/methodology/approach In this study, the changing resistance parameters of the asynchronous motor were modeled, and the efficiency was estimated depending on the change of stator resistance. The system was labeled as efficient or inefficient according to the efficiency value. Then, spectrogram images were taken from the single-phase current values of the system and these images were used as dataset for CNN. The asynchronous motor was modeled in Simulink and Python programming language was used for CNN via Spyder IDE. Sequintal model was used for CNN and Relu was chosen as the activation function. MaxPooling, Flatten and Dense layers were used in the intermediate layers. Findings As a result of the study, the test accuracy of the CNN was found to be 95%, which can be considered a very good success. Originality/value This is a study in which the system outputs of the electric vehicle are combined with a CNN and the effects of changing the stator resistance are determined.