DEVELOPMENT AND EVALUATION OF NOVEL AIR HANDLING UNIT CONTROL STRATEGIES FOR IMPROVED BUILDING DEMAND FLEXIBILITY AND THERMAL COMFORT

Abstract

Buildings account for a significant share of electrical power demand, up to 80% in some regions in the United States. With many flexible loads, buildings have the potential to reduce power demand during peak times and provide demand flexibility. Building enables demand flexibility through measures such as shifting energy use to another time and shedding of load. Applying advanced control strategies to heating, ventilation, and air conditioning (HVAC) loads allows buildings to reduce the electricity consumed during peak hours. While various demand control (DC) strategies have been employed in commercial buildings, they often fail to accurately shift and shed loads while maintaining occupant thermal comfort. Additionally, existing DC strategies face challenges in quantifying load reduction, achieving precise DC and ensuring even distribution of the reduced cooling load across thermal zones. These limitations highlight a critical research gap in effectively integrating demand flexibility with occupant comfort in building energy management systems. This dissertation develops and evaluates advanced control strategies for building demand control and thermal comfort using both simulation and experimental test beds. An innovative Energy Feedback (EF) control strategy and a novel Duct Static Pressure (DSP) cascade control strategy were first proposed, and their individual control performances were assessed using an experimental test bed. Subsequently, the integrated dynamic performance of the proposed EF and DSP cascade control strategies was evaluated against a baseline control strategy using a calibrated Modelica-based virtual test bed. Performance metrics such as settling time, relative error (RE), and the range of average zone temperature (RZT) are used as measures of quick response, control accuracy, and even cooling load sharing, respectively. The most effective strategy for minimizing the negative impact on occupant comfort during demand control is identified. The proposed control strategy demonstrated the fastest response and achieved precise load shedding, with a settling time of 0.03 hours, an RE of 0.01%, and an RZT of 0.00448°C. In contrast, the baseline strategy had a long settling time of 5.05 hours, an RE of 12.52%, and an RZT of 0.05°C. An optimization-based control strategy was incorporated into the EF Controller to provide effective setpoint control for achieving load shifting through precooling and load shedding during peak demand periods while minimizing electricity cost. The optimization-based control strategy enabled the modulation of the cooling load setpoint based on weather conditions and the building's daily load profile. This involved developing a reduced-order thermal model, estimating its parameters for control purposes, and developing and implementing the Optimization-Based EF Controller on the calibrated virtual test bed. The performance of the proposed Optimization-Based EF–DSP Cascade Control strategy was evaluated and compared with the baseline demand control strategy and a non-optimized EF–DSP Control strategy that utilized constant cooling load setpoints. The Optimization-Based EF Controller yielded a total electricity cost saving of 12.8% compared with the non-optimized controller, including an 8.7% reduction during the on-peak period and a 15.4% reduction during the off-peak period. The proposed Optimization-Based EF Control provides an effective approach for achieving demand-flexible HVAC operation by simultaneously reducing cooling energy consumption, minimizing electricity cost, maintaining occupant thermal comfort, and mitigating rebound effects. Finally, the performance of the Optimization-Based EF Controller was experimentally validated, and two practical implementation approaches for applying the optimization-derived EF setpoints on an HVAC system, namely the Schedule-Based EF Control and the Adaptive EF Control, were evaluated. Among the key contributions of this dissertation are the following: (i) the development and experimental evaluation of novel EF and DSP cascade control strategies to achieve accurate demand control and evenly distribute the reduced cooling load among thermal zones; (ii) the development and calibration of a Modelica-based virtual test bed that comprises of an AHU system and building envelope, and its application to the integrated performance assessment of the proposed control strategies; and (iii) the development and implementation of an optimization-based control framework that generates and communicates optimized control signals to the EF controller.