Energy-related occupant behavior in buildings: Approaches for monitoring and modelling

Abstract

Buildings have become one of the most energy-intensive sectors globally, contributing to about one-third of the worldwide energy consumption. The response to growing global concerns regarding energy-use has prompted the building sector to develop strategies for energy efficiency and investigate potential areas for optimizing building energy performance. The research efforts have highlighted several underlying factors that are attributed to this performance. Occupant Behavior (OB), or the way people interact with building systems, was found to have significant influence on building’s performance, and further investigations revealed the lack of understanding the sector has regarding this phenomenon. Moreover, considerations of building occupants during the design stage rely on simplistic or deterministic approaches that do not provide an accurate representation of occupants. This challenge is also present in building performance simulation tools that are used to develop policies and provide recommendations for enhancing building energy efficiency. The lack of provisions for a dynamic human-building interaction is often attributed to the gap in expected and actual performance of buildings. The difficulties in addressing the topic arise from OB being a complex culmination of several phenomenon, including the occupant’s presence, movement, actions, clothing, habits, social parameters and other contextual factors. Consolidated research efforts in this area, such as the International Energy Agency’s Annex reports highlight the need for improving the accuracy of monitoring and modeling OB. Key recommendations also include the investigation and inclusion of social factors underlying OB, which are often neglected or underreported in energy research.

This thesis aims to provide a better understanding of OB by analyzing several aspects of it, in diverse contexts and settings. The work carried out can be categorized into three topics: OB Monitoring, OB Profiles, OB Modelling. The first one of these comprises of case studies using diverse methods to monitor and analyze occupants in different built environments. The second one focuses on the preparatory work needed for developing OB models, by creating occupant profiles, databases, and frameworks. The last one aims to provide insight and recommendations regarding the modelling and simulation part. The four research questions framed under these categories and their outputs can be summarized as follows:

How to monitor and analyze the impacts of occupant behavior in different spaces: Three distinct case studies were carried out to answer this question. The work consisted of using multiple sensing modalities to monitor occupant presence, energy-use, movement, and activities. Each study provided a unique perspective, with the first one using a combination of PIR and environmental sensors to collect and examine granular device-level, office-level, and occupant-level data in a shared office space. The second one made use of depth registration to monitor occupant influences in an operating room, and the last study focused on capturing an extensive activity-based dataset for smart homes. Since the lack of shared standards and benchmark datasets is another challenge in the development of guidelines and models, one of the goals of these studies was also to provide benchmark datasets that are available for a broader use.

How can the social aspects of OB be evaluated for modelling/simulation purposes: The research question was divided into three tasks, starting with the development of a hypothetical framework and extended Theory of Planned behavior model, which was useful in quantifying the social aspects. The second task put this model in implementation, by collecting data through surveys and analyzing the factors that influence people's behavior. The investigation process involved creating structural equation models, regressions, and path analysis. Regression was found to be the best fit, while path analysis was acceptable. Influence factors for occupant's energy-related behaviors were obtained for each building system. The addition of additional variables improved the predicting power of current models. The framework and evaluation of variables led to the creation of OB profiles, supported by analysis highlighting valid and significant variables. The k-modes clustering technique yielded suitable clusters for modelling purposes.

How can the social aspects of OB be modelled/simulated: An important part of this thesis was incorporating an interdisciplinary approach, connecting social behavioral theories to the engineering aspects of modeling and simulation. Addressing this question required assessing modeling techniques that would be best suited to include the diversity and flexibility required to include social aspects. Agent-based models were used to simulate this aspect, which proved capable of executing the already defined fixed and static models while also accommodating the diversity of OB.

What kinds of prerequisites are needed to map environmental layouts around the occupant: This question was part of a broader study aimed at developing a hypothetical OB model that attempts to address multiple modeling requirements for OB, a part of which was to simulate the environment around the occupant. The work carried out in this thesis aimed to map occupant environmental layouts in a room by identifying seven variables for a library and developing a Matlab application for spatial information collection. The library was tested using a sample dataset of 80 offices at NTNU. A database was created to connect the information generated from this study to larger models, which were expanded to include other datasets from previous studies.

By addressing these questions, this thesis was able to contribute to a broader understanding on the subject of occupant behavior, providing insights about the monitoring and modelling process, and highlighting additional challenges in the subject. The findings may contribute to a better design of building operation and management that places its occupants at the center.