Smart Building Energy Management Using IoT Sensor Networks and Artificial Intelligence

Thesis Abstract

The escalating demand for sustainable energy solutions in urban environments necessitates the adoption of intelligent building management systems that optimize energy consumption while maintaining occupant comfort. Traditional energy management approaches are often reactive, inefficient, and lack the precision needed to respond dynamically to real-time occupancy patterns and environmental conditions. This study aims to develop and evaluate a comprehensive IoT-based sensor network integrated with artificial intelligence algorithms to enable proactive, data-driven energy management in commercial buildings. The primary objectives are to design a scalable IoT sensor infrastructure for real-time environmental and occupancy monitoring, develop machine learning models for predicting energy consumption patterns, and implement an adaptive control system that optimizes energy use without compromising occupant comfort. The research adopts a mixed-methods design comprising quantitative data collection for system development and predictive modeling, alongside qualitative assessment of user acceptance and system usability. The study population includes active tenants and building management personnel within a 10-storey commercial office building located in a metropolitan city, with a total population of approximately 300 occupants. A stratified random sampling technique was employed to select 150 participants to ensure proportional representation across different user groups. Data collection involved deploying a network of IoT sensors—comprising temperature, humidity, light intensity, occupancy detectors, and energy meters—over a 12-month period, complemented by structured interviews and questionnaires to gather insights into user interactions and satisfaction. Data analysis utilizes descriptive statistics to profile environmental conditions and occupancy trends, followed by regression analysis and time-series forecasting to develop predictive energy consumption models. The core analytical framework is grounded in the Theory of Planned Behavior to assess user acceptance of the developed system, complemented by machine learning techniques such as Random Forests and Support Vector Machines for predictive accuracy and multi-variable analysis. Model validation is conducted through cross-validation and performance metrics including RMSE, MAE, and R-squared, to ensure robustness and reliability of the energy optimization algorithms. Expected findings suggest that IoT sensors provide high-resolution, real-time data enabling more accurate energy consumption predictions. Machine learning models are anticipated to improve energy efficiency by identifying optimal control strategies tailored to occupancy and environmental dynamics, resulting in an estimated 20-30% reduction in energy use compared to baseline levels. The integration of adaptive control algorithms is also expected to enhance occupant comfort levels, as evidenced by increased satisfaction survey scores. The study aims to demonstrate that a data-driven, AI-enabled energy management system is feasible and beneficial in the context of modern smart buildings, offering scalable solutions adaptable to diverse operational conditions. This research contributes to the existing body of knowledge by empirically validating the effectiveness of IoT and AI integration in building energy management, providing a scalable framework and specific algorithmic models for practitioners and researchers. It extends theoretical understanding of behavioral adoption in smart building environments through the application of the Theory of Planned Behavior, offering insights into factors influencing user acceptance. The study's practical implications include guiding policy formulation on sustainable building practices, informing the design of intelligent control systems for energy savings, and enhancing occupant engagement via user-centric interfaces. The conclusions emphasize the potential for IoT and AI to revolutionize building energy management, advocating for increased investment in sensor infrastructure and predictive analytics. Recommendations include implementing standardized protocols for sensor deployment, investing in training for building management teams, and fostering occupant engagement initiatives to maximize system benefits. Further research is suggested to explore long-term system resilience, integration with renewable energy sources, and cross-building interoperability to establish comprehensive urban energy management frameworks capable of supporting sustainable development goals.