Load optimization for connected smart green buildings using machine learning
Date
2025
Authors
Moghimi, Seyed Morteza
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Energy efficiency plays a crucial role in mitigating Greenhouse Gas (GHG) emissions, particularly in the building sector, where residential buildings are among the largest energy consumers. Despite the potential of buildings to generate Renewable Energy (RE), increasing energy demands pose both environmental and economic challenges. This dissertation presents a Machine Learning (ML)-based framework for optimizing energy consumption in Connected Smart Green Townhouses (CSGTs), focusing on efficiency, cost-effectiveness, emission reduction, and occupant comfort. A comprehensive study of adaptive, occupant-aware, and ML-based energy optimization is presented for Smart Green Townhouses (SGTs) and CSGTs. The goals are prediction, optimization, and real-time management of energy consumption, with a focus on sustainability, occupant comfort, and system intelligence.
A model for CSGTs operating in grid-connected mode is presented. This model incorporates sustainable building materials, smart sensors, Photovoltaic (PV) systems, and energy-efficient components. A hybrid Long Short-Term Memory–Convolutional Neural Network (LSTM-CNN) model is considered with real utility datasets. The results show that this approach outperforms traditional ML models such as Linear Regression (LR), CNN, LSTM, Random Forest (RF), and Gradient Boosting (GB). The Mean Absolute Percentage Error (MAPE) is below 5%, and the coefficient of determination (R^2) is above 0.85, which validates the accuracy for different bedroom configurations.
A robust ML-based optimization framework is proposed for CSGT energy and load management in island mode. The integration of Electric Vehicles (EVs) with Vehicle-to-Grid (V2G) functionality is shown to improve system resilience. The LSTM-CNN model provides a MAPE of 4.43% and a Root Mean Square Error (RMSE) of 3.49 kWh for the four-bedroom unit. The results confirm that occupant-aware optimization significantly improves performance under isolated conditions.
An adaptive control framework to enable automatic transitions between grid-connected and island modes is developed. By incorporating occupancy, weather, and electricity price data,
the system dynamically optimizes load consumption using LSTM-CNN and Multi-Objective Particle Swarm Optimization (MOPSO). Efficiency gains of 3–5% in grid-connected mode and 10–12% in island mode are observed with a 4–6% reduction in carbon emissions, demonstrating the value of real-time adaptive management.
An occupant-centric load optimization system leveraging real-time Internet of Things (IoT) data is proposed. This human-centric approach significantly improves comfort and operational efficiency. Energy loads are reduced by 7–13%, peak loads by 11%, and carbon emissions by 15–24%. Cost savings of 13–21% are achieved, and occupant satisfaction increases with a 19% improvement in thermal comfort and 14% better lighting adequacy.
The results presented highlight the effectiveness of advanced, applicable, and scalable ML-driven energy optimization in SGTs. The proposed approaches offer scalable, adaptable,
and occupant-centric solutions for energy-efficient, cost-effective, and environmentally sustainable residential buildings. Future research directions include integrating advanced renewable energy storage management, real-time grid interaction, federated learning, and edge Artificial Intelligence (AI) deployment to improve the adaptability and efficiency of
smart energy and load management in CSGTs. Integrating advanced ML models, real-time sensor data, and adaptive control techniques will provide solutions to address the economic,
environmental, and social challenges in sustainable urban housing.
Description
Keywords
Machine learning (ML), Smart green buildings, Load optimization, Renewable energy systems (RES), Connected smart green townhouses (CSGTs), Occupant-centric energy management, Energy efficiency, Sustainability, Burnaby, British Columbia, Predictive analytics, Performance metrics, Model accuracy, Hybrid LSTM-CNN, Energy consumption forecasting, Grid and island modes