Prediction of Electric Power Production and Consumption for the CETATEA Building Using Neural Networks.

Area of Science:

• Electrical Engineering
• Renewable Energy Systems
• Data Science

Background:

• Economic and social development are significantly influenced by electric power production and consumption.
• Intelligent monitoring and control of energy systems are crucial due to energy supply pressures.
• Accurate power prediction is essential for effective energy management.

Purpose of the Study:

• To develop and evaluate a highly accurate power prediction system.
• To leverage a unique, large-scale dataset for advanced energy forecasting.
• To assess the performance of neural network algorithms in predicting unified electric power production and consumption.

Main Methods:

• Utilized a 4-year continuous real-time dataset from the CETATEA photovoltaic power plant.
• Employed neural network-based prediction algorithms.
• Analyzed over 4.2 million unified power values recorded every 30 seconds.
• Evaluated prediction accuracy using metrics like mean bias error, mean square error, and convergence time.

Main Results:

• The prediction system demonstrated high accuracy in forecasting electric power.
• Predicted unified electric power production and consumption closely matched measured values.
• The large dataset granularity contributed to the overall prediction accuracy.

Conclusions:

• Neural network algorithms are effective for accurate electric power prediction.
• The study highlights the importance of high-granularity, large-scale datasets in renewable energy forecasting.
• The findings support intelligent monitoring and control for energy systems.