Machine Learning Modeling and Simulation of Asynchronous Electric Drive

This study focuses on simulating and modeling the behavior of a frequency-controlled asynchronous electric drive using MATLAB/Simulink. The control system is designed using the root locus method along with the Vyshnegradsky diagram. The research examines the system's dynamic response during startup under normal load. The results indicate that the electromagnetic torque can be optimized, with minimal overshot, by selecting specific Vyshnegradsky parameters. Experimental validation shows that the control system achieves an acceptable level of accuracy. Additionally, machine learning techniques such as support vector machine (SVM), random forest (RF), and k-nearest neighbor (k-NN) are used to predict the behavior of key system variables. Evaluation of the models was conducted using two frequently used criteria. i.e., root mean square error and correlation coefficient. RF provided lower RMSE and higher CC in comparison with SVM and k-NN. On the other hand, k-NN provided higher CC, but it generated higher RMSE in comparison with RF. Accordingly, in the training phase, RF was selected for the optimum modeling.