The object of this research is to investigate the feasibility of applying artificial intelligence and advanced adaptive control techniques to building systems to develop an integrated expert control system for buildings. This approach will have a hierarchical structure with two distinct features, namely a foreground/background optimization scheme which handles dynamics of the building system with time constants of different orders of magnitude; and a hybrid algorithmic/rule-based decision-making scheme. The background supervisory controller utilizes the detailed building system models to make decisions regarding the slow dynamics and disturbances of the order of days, while the foreground control schemes tune the HVAC equipment and lighting system to compensate for fast dynamics, such as occupancy variations. The significant features of this research include: (1) development of a three-level hierarchical intelligent control scheme, which compensates for system and environmental uncertainties; (2) formulation of a performance index for global optimization of a building system with emphasis on energy management, occupant comfort and operating and maintenance costs; (3) formulation and implementation of HVAC equipment self-tuning local controllers with set points determined by the second level controller; (4) formulation and implementation of an on-line diagnostic scheme which utilizes sensory information to calculate instantaneous statistical characteristics of the system for real-time failure detection and failed component isolation; (5) formulation of an algorithmic decision-making scheme based on the minimum uncertainty (entropy) concept. The resulting techniques will be simulated and to some extent implemented on small scale subsystem of a building. Industrial cooperation will be sought to plan implementation of the results on an actual building after the two-year duration of this study.
