Comprehensive Forecasting of Electrical Quantities in an Educational Building via Artificial Intelligence-Driven...

Area of Science:

• Sustainable energy systems
• Artificial Intelligence in power grids
• Building energy management

Background:

• Growing environmental concerns necessitate sustainable solutions in all sectors, including power systems.
• Buildings are increasingly designed as net-zero emitters or retrofitted for energy efficiency.
• Effective energy management is crucial for optimizing power system performance and reducing environmental impact.

Purpose of the Study:

• To enhance the energy management capabilities of an educational building through a smart infrastructure.
• To implement and validate an AI-driven framework for forecasting electrical parameters.
• To explore the generalization of AI models across different measurement nodes within the infrastructure.

Main Methods:

• Installation of photovoltaic panels and a distributed measurement system to collect voltage, current, and power data.
• Development and application of an AI-driven framework for data management and forecasting.
• Experimental validation of the AI model and correlation analysis of collected data.

Main Results:

• The implemented system provides reliable forecasts of electrical parameters.
• The distributed measurement system and data analysis offer valuable insights for energy optimization.
• The AI model demonstrated generalization capabilities across measurement nodes, even with data gaps.

Conclusions:

• The smart infrastructure effectively supports informed actions for optimizing energy management and system performance.
• The study highlights the potential for accurate predictions and highlights ease of deployment for scalability.
• The findings support the adaptation of this system in similar infrastructures for enhanced energy efficiency.