Rule-based integrated building management systems

Abstract

The introduction of building management systems in large buildings have improved the control of building services and provided energy savings. However, current building management systems are limited by the physical level of integration of the building's services and the lack of intelligence provided in the control algorithms. This thesis proposes a new approach to the design and operation of building management systems using rule-based artificial intelligence techniques. The main aim of is to manage the services in the building in a more co-ordinated and intelligent manner than is possible by conventional techniques. This approach also aims to reduce the operational cost of the building by automatically tuning the energy consumption in accordance with occupancy profile of the building.

A rule-based design methodology is proposed for building management systems. The design adopts the integrated structure made possible by the introduction of a common communications network for building services. The 'intelligence' is coded in the form of rules in such a way that it is both independent of any specific building description and easy to facilitate subsequent modification and addition. This is achieved using an object-oriented approach and classifying the range of data available into defined classes. The rules are divided into two knowledge-bases which are concerned with the building's control and its facilities management respectively. A wide range of rule-based features are proposed to operate on this data structure and are classified in terms of the data classes on which they operate.

The concepts presented in this thesis were evaluated using software simulations, mathematical analysis and some hardware implementation. The conclusions of this work are that a rule-based building management system could provide significant enhancements over existing systems in terms of energy savings and improvements for both the building's management staff and its occupants.