1. Learning-based demand-driven controls for energy-efficient buildings
- Author
-
Peng, Yuzhen
- Subjects
- Energy-efficient buildings, Occupant behavior, Thermal comfort, Building control, Data-driven modeling, Demand-driven control, Buildings, Electric engineering, Architecture
- Abstract
In seeking to mitigate the increasing pressures on buildings and facilities for requirements of energy and comfort, this research focuses on exploring means to improve the energy efficiency of heating, ventilation, air-conditioning (HVAC) and thermal comfort for occupants in buildings. The occupants are the direct end users of HVAC systems and the corresponding indoor climate. To better understand occupant behavior within buildings, this research conducts an in-depth analysis of their demand for energy and comfort services through building sensor data. Such data includes the information of occupant movements, human interaction with the conditioned indoor environment, indoor and outdoor climate. This thesis also presents three different HVAC control methodologies with learning capacities, which have been implemented in a commercial building under real-world conditions. The first two control strategies aim to make the HVAC systems adapt to room occupancy for saving energy without compromising room temperatures during occupied periods. Across the entire case study rooms, the experimental results report up to 21% energy reductions as compared to the conventionally-scheduled HVAC systems. To enhance occupants’ indoor thermal comfort, the third control strategy aims to make the HVAC systems automatically respond to occupants’ comfort-related behavior (i.e. indoor temperature preferences) under dynamic contexts. The experimental results report 4% to 25% energy savings as compared to static temperature setpoints at the low values of preferred temperature ranges for space sensible cooling. The field test also shows that the active learning based control reduces the need for occupant interventions in adjusting room temperatures to fit their preferences. This thesis not only introduces the proposed methodologies but also illustrates and discusses extensive test results. It provides view and information on energy conservation of building systems and comfort improvement for occupants. Moreover, the relevant design processes can be extended to other building systems to achieve demand-driven or occupant-centric design and operation.
- Published
- 2019