Short-Time Wind Speed Forecast Using Artificial Learning-Based Algorithms.

Area of Science:

• Renewable Energy Systems
• Artificial Intelligence in Energy
• Time Series Forecasting

Background:

• The increasing demand for industrial power necessitates reliable renewable energy sources.
• Renewable energy generation, particularly wind power, is subject to unpredictable environmental factors like wind speed, pressure, and humidity.
• Accurate forecasting of these variables is essential for efficient grid management and energy production.

Purpose of the Study:

• To investigate and compare the performance of various forecasting algorithms for wind speed data.
• To identify a forecasting model that balances prediction accuracy with computational efficiency for wind harvesting farms.

Main Methods:

• Evaluated five models: Artificial Neural Network (ANN), Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Convolutional LSTM (ConvLSTM), and Support Vector Machine (SVM).
• Utilized statistical and time indicators for comprehensive model evaluation.
• Focused on a case study involving wind speed data from a wind harvesting farm.

Main Results:

• Support Vector Machine (SVM) demonstrated the highest prediction accuracy.
• Convolutional LSTM (ConvLSTM) achieved high prediction accuracy while requiring less computational effort compared to other models.
• The study confirmed the effectiveness of deep learning models in handling complex, unpredictable data like wind speed.

Conclusions:

• While SVM offered superior accuracy, ConvLSTM presents a more practical solution for wind speed forecasting due to its efficiency.
• The ConvLSTM model is recommended for wind harvesting farms needing reliable and computationally feasible wind speed predictions.
• Advanced forecasting models are vital for optimizing renewable energy integration into the power grid.