Diagnosing Voltage and Current Imbalance of Three-Phase Induction Motor with Artificial Neural Network Method

Uden, Suparman
Yahya, Sofian
Tahtawi, Adnan Rafi Al

Jurnal Asiimetrik: Jurnal Ilmiah Rekayasa Dan Inovasi
Volume 6 Nomor 2 Tahun 2024

Industries today are increasingly using three-phase induction motors. It is an important tool for production continuity and progress. Power quality issues, such as voltage and current imbalance, are prevalent today and can lead to motor overheating and inefficiency. Even worse, interruptions can impede the production process, resulting in losses and higher repair costs. This study uses MATLAB and microcontroller-based Artificial Neural Network (ANN) methods to identify voltage and current imbalances in three-phase induction motors, thereby preventing significant damage and preserving the service life. ANN works by learning and classifying the collected data. The testing flow uses 30% of the data, while the training flow uses the remaining 70%. The classification results showed that 60.49% of the voltages were balanced, and 31.59% were unbalanced. We found an accuracy percentage of 99.51% for both balanced and unbalanced voltages, a mean squared error (MSE) of 0.0167, and a root mean squared error (RMSE) of 0.1294.