Temporal reasoning in real-time expert systems
Loading...
Files
Date
1991
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This thesis presents research and development of a temporal logic for use within realtime expert systems. The temporal logic provides a framework to represent both past and present temporal information. A prototype expert system was built to demonstrate the feasibility of these ideas in the domain of sootblowing within a coal-fired power generating station. A real-time expert system prototype has been developed to advise operators of sootblowing strategies by monitoring the performance of the boiler and the build-up of slag. This prototype called the Sootblowing Advisory Expert (SAX) includes several components. The first component is the Temporal Preprocessor which generates temporal events and relationships from operator entered information and simulated process I/O. The second component, Temporal History, maintains the history of temporal events and relationships. The Query Processor provides the means to transparently query both the history and current temporal history databases. The fourth component of SAX implements the domain specific dynamic and static expert knowledge in the form of facts and rules. The last element of SAX is the User Interface which provides the operator with the means to enter interactively plant operating information unavailable through process I/0. As well, the user interface provides the sootblowing advice and status information. This system was developed using the ART-1M Automated Reasoning Tool developed by Inference Corporation and is functional on a PS2/Model 70.