Occupant behavior impact in buildings and the artificial intelligence-based techniques and data-driven approach solutions

Abstract

Occupant behavior in buildings might result into gap between predicted and actual energy use and cause indoor thermal comfort fluctuations, due to its uncertainty and unpredictability. Previous studies mainly focus on the investigations of occupant behavior impact and modeling methods, whereas few concludes strategic solutions. It is hard to narrow the negative impact of occupant behavior without a clear Impact-Modeling-Solution clue. Thus, the major objective of this research is to review and analyze the impact of occupant behavior and then to summarize modeling methods and strategic solutions by connecting specific techniques and approaches. This paper first investigates the characteristics of occupant behavior and then discusses the impact. It indicates that thermal comfort (environmental condition) triggers occupant actions with corresponding building systems, resulting in energy load changes. The occupant behavior in turn causes thermal comfort fluctuations. The strategic approaches of traditional methods such as surveys, experiments/tests and simulations, and artificial intelligence (AI)-based techniques are analyzed respectively. It is found that the intelligent approach shows high robustness in addressing the uncertain and unpredictable characteristics of occupant behavior in buildings, compared with traditional methods. The AI-based methods and data-driven approaches can enhance the prediction of building energy consumption and the recognition of occupant's thermal comfort. Typical modeling methods and flowchart for occupant behavior are presented with comparisons. The novel physics-based and data-driven model which performs strong adaptability, high decision-making efficiency and fast response of the controller is especially introduced. The corresponding improvements and future directions are also proposed to reach optimal effects.