Optimization based load forecasting and demand management in smart building microgrids with Greylag Goose and Bi...

Area of Science:

• Smart Grid Technology
• Artificial Intelligence in Energy Systems
• Renewable Energy Integration

Background:

• Smart Building Microgrids (SBMGs) face challenges in energy management due to inaccurate load prediction, demand-supply mismatches, and battery degradation.
• Existing forecasting and optimization methods often fail to ensure system reliability and battery resource durability in dynamic SBMG environments.

Purpose of the Study:

• To propose a novel domain-adapted forecasting and optimization framework for SBMGs.
• To enhance the accuracy of load prediction and demand response management.
• To improve the reliability and battery durability in SBMGs.

Main Methods:

• A hybrid framework combining Relational Bi-Level Aggregation Graph Convoluted Network (RBAGCN) with Greylag Goose Optimization (GGO).
• RBAGCN was re-engineered to integrate physical and operational interrelations of energy variables.
• GGO was employed to stabilize network weight convergence for non-stationary loads.
• Data preprocessing included Fast Resampled Iterative Filtering (FRIF) and Prairie Dog Optimization (PDO) for feature selection.

Main Results:

• The proposed framework achieved an average forecasting accuracy of 98.3%, significantly outperforming benchmark models (e.g., RNN at 82.6%).
• Demonstrated lower error metrics: Mean Absolute Error (MAE) of 0.0164, Mean Absolute Percentage Error (MAPE) of 0.0128, and Mean Squared Error (MSE) of 0.0069.
• Reduced prediction variance by 26.5% and convergence iterations by 17.8%, indicating enhanced statistical stability and learning efficiency.

Conclusions:

• The integrated RBAGCN and GGO framework offers a robust solution for accurate load forecasting and demand management in SBMGs.
• The methodology improves system reliability and extends battery resource lifespan through optimized energy utilization.
• This approach represents a significant advancement in intelligent energy management for smart buildings.