Abstract
Plug load energy consumption has been steadily rising over the recent years in the workplace and currently accounts for up to 33% of the energy consump tion in commercial buildings. To address this rising consumption, plug load management systems (PLMS) have been touted as a promising solution in the research community. However, past studies reported low adoption rates of ex isting systems as they heavily rely on users to actively manage their plug loads despite their lack of motivation. This highlights the need for more intelligent PLMS capable of reducing user burden by automating their plug loads to re duce energy wastage. Fortunately, advancements in Internet of Things (IoT) technology has led to the development of occupancy-driven smart energy man agement systems (OSEMS) for plug loads which can automate the users’ plugs load based on their presence information. However, the application of OSEMS for plug load management is still in its infant stages due several unique tech nical challenges that has yet to be addressed. This includes issues related with detection errors and the system’s inability to account for the plug load type and user ownership information due to its low resolution. Given these limitations highlighted, the objective of this thesis is to reduce plug load energy consumption in the workplace by developing an advanced IoT-based occupancy-driven plug load management system, known as Plug Mate, capable of automating users’ plug loads based on their high-resolution occupancy information while accounting for their diverse control preferences for different plug load types. The initial development process of Plug-Mate begins with a comprehensive user study to understand users’ perceptions of adopting PLMS in the work place, with the goal of developing a system that reduces user burden while increasing energy savings. Through a series of focus group discussions, online surveys, and laboratory studies, we identified several design implications to guide the design of Plug-Mate, as well as future PLMS. Based on the insights obtained from this user study, we developed Plug-Mate. iii Plug-Mate follows a novel 4-layer IoT-based framework, which is generalis able to different building systems for occupancy-driven smart energy manage ment. By applying this framework to plug loads, Plug-Mate is made up of sev eral key components, such as the Energy Monitoring & Control Network, the Bluetooth Low Energy (BLE) Occupancy Sensor Network, the Plug Load Iden tification, Occupancy Detection, and Control Modules, as well as the Database and web-based User Interface. Through a real-time automatic plug load identi fication model, Plug-Mate is capable of automatically identifying the different types of plug loads connected to the system based on their energy signatures captured at low sampling frequencies. This feature enables the system to auto matically assign different control rules based on each plug load type and elim inates the need for manual assignment. On top of that, the proposed system comes with a scalable and high resolution occupancy detection model which leverages existing BLE technologies found in users’ smartphone devices to per form zone-level indoor localisation without requiring them to install a mobile application or put on wearable sensors. By integrating a high-resolution and ac curate occupancy detection system to support the automation of plug loads, we overcame the detection and resolution issues encountered in existing systems. Finally, the system consists of an intuitive web-based user interface where users can utilise different control features to manage their plug loads and adjust the system’s automated controls based on their unique preferences. To demonstrate the proposed system’s feasibility, we conducted 3 field stud ies to evaluate several key components within Plug-Mate such as the plug load identification and occupancy detection models, as well as an overall system as sessment based on different control strategies. The final field assessment of the system showed that Plug-Mate was able to achieve energy savings up to 51.7% among different plug load types and the highest user satisfaction score of 4.77 out of 5. Based on the cost analysis and deployment considerations reported, it was also found that the system is financially viable for real-world implemen tation with a simple payback period of 3.21 years, based on a system lifespan of 10 years, and a savings-to-investment ratio of 3.1. Given the satisfactory per formance of the proposed system, Plug-Mate serves as a first-of-its-kind high resolution occupancy-driven system for plug load management and is expected to guide the development of future occupancy-driven plug load management systems in the workplace