Automated reasoning techniques for intelligent control of building systems
Abstract
A self-organizing controller based on fuzzy set theory is applied to the problem of building temperature and lighting control. A three-level hierarchical control structure composed of supervisor, coordinator, and local layers is employed. Fuzzy sets representing zone occupancy, utility cost, and building dynamics are used along with weather forecast data as inputs to the supervisor, which generates local controller setpoint schedules based on accumulated knowledge of past building usage and behavior. The supervisor is autonomously optimized with respect to a performance index penalizing energy consumption, occupant discomfort, and fuzzy model error. The coordinator layer of the intelligent building system is responsible for short term error correction, and modifies the supervisor schedule when unexpected occupancies occur. Both lighting and temperature setpoints may be changed to maintain occupant comfort using a fuzzy rule-base, which relates setpoint errors with the appropriate corrections. Implementation of the system in an office building results in a 20 to 40% savings in energy as compared to a constant setpoint scheme. Improvement over traditional night setback schedules is also expected, as the controller is designed to accommodate variable occupancy loading conditions. The ability of the HVAC system to meet heating or cooling demand may also be enhanced since the intelligent building control system may dynamically shed load from unoccupied zones.
Degree
Ph.D.
Advisors
Shoureshi, Purdue University.
Subject Area
Mechanical engineering|Systems design
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.