DRIVE-XAI: Demand Response Integrated Visibility and Energy Efficiency with Exploring Explainable AI in Smart Buildings

Abstract

Demand response optimization and energy minimization in smart buildings have become paramount due to the global rise in energy consumption and response to external signals, such as electricity price changes, grid conditions, or incentives from utility providers. The Demand Response (DR) management system is crucial in effectively managing smart device energy supply and demand, especially during peak loads. This thesis introduces an innovative technique that leverages Explainable Artificial Intelligence (XAI) techniques, such as SHAP and LIME, to enhance demand response strategies and minimize energy consumption in smart buildings. Utilizing these XAI algorithms, our methodology provides insights into decision-making processes and enables building managers to understand and trust the AI prediction algorithms that influence energy consumption behaviors. The potential of XAI to transform demand response strategies is a significant contribution to the field. This thesis uses the proposed XAI-driven smart building architecture to identify key factors influencing energy usage by analyzing real-time data from Internet of Things (IoT) devices in smart buildings. It enhances demand response initiatives by suggesting optimal load management strategies. This approach has significant practical benefits, substantially improving energy efficiency and reducing operational costs while ensuring occupant comfort. Moreover, the explainability mechanisms facilitate informed decision-making, fostering a more responsive and responsible energy management system. The findings suggest that integrating XAI reduces energy consumption during peak periods and sets the stage for future advancements in intelligent energy systems in smart buildings. This thesis underscores the potential for XAI to transform demand response strategies, promoting sustainability and resilience in the face of increasing energy challenges.

Date of Award	2024
Original language	American English
Awarding Institution	HBKU College of Science and Engineering