Time Series Electrical Motor Drives Forecasting Based on Simulation Modeling and Bidirectional Long-Short Term...

Area of Science:

• Electrical Engineering
• Artificial Intelligence

Background:

• Accurate electrical signal forecasting in three-phase Direct Torque Control (DTC) induction motors is vital for performance optimization and condition monitoring.
• Conventional prediction methods face challenges due to the complexity of DTC motors and operational variability.

Purpose of the Study:

• To develop an innovative forecasting approach for electrical signals in DTC induction motors.
• To enhance the precision and reliability of motor signal predictions.

Main Methods:

• Preprocessing simulation data using Fast Fourier Transform (FFT).
• Utilizing a Bidirectional Long Short-Term Memory (Bi-LSTM) network for signal forecasting.
• Comparative analysis against Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU) models.

Main Results:

• The proposed FFT-Bi-LSTM approach demonstrated superior performance in forecasting induction motor signals.
• Achieved low Mean Absolute Error (MAE) of 92.6864 for stator current and 93.8802 for rotor current.
• Exhibited reduced prediction loss with Root Mean Square Error (RMSE) averages of 105.0636 and 85.7820.

Conclusions:

• The FFT-Bi-LSTM method significantly improves the accuracy and reliability of electrical signal forecasts for DTC induction motors.
• This approach offers a robust solution for complex motor control systems.
• The findings highlight the potential for advanced AI techniques in motor diagnostics and prognostics.