![]() |
|||
![]()
|
![]() |
![]() |
![]() |
2. HISTORY OF EXPERT SYSTEMS WITHIN THE POWER INDUSTRYOne of the first areas where expert systems were utilized in the power industry was in the field of complex machine diagnostics. An early example was developed by Sakaguchi and Matsumo in 1983 (Sakaguchi and Matsumo, 1983). One of the next substantial systems was VIAD (Scheibel et al., 1986) which was later released commercially as EXPLORE-EX (a vibration advisor expert system) in 1990 by Entek Corporation. Later in 1990, AMETHYST appeared on the market and was, at the time, one of the most widely used expert systems for diagnosing problems with rotating machinery (Milne, 1990). Also in 1990, VARMINT (a vibration advisor for complex machinery) came onto the market. This was one of the first condition monitoring expert system to utilize the Microsoft Windows 3.0 interface (VARMINT, 1991). Table 1 gives a list of the major systems described in this chapter, with the year and continent in which they were developed. The power industry has been, in general, slow to embrace Artificial Intelligence (AI). Expert systems have to some extent been investigated, and their use is growing; the U.K. Institute of Mechanical Engineers organized a seminar on this specific issue in November 1993; this seminar contained a number of presentations on the application of expert systems within the industry.
Most of these systems relate to the provision of diagnostic information to power plant control room operators, to assist them in the decision-making process. For example, Gemmell et al. (1993) discuss an expert system that provides diagnostic information on the cause of faults observed from plant data on turbine generators; Lausterer et al. (1993) discuss another expert system for provision of information to operators, and particularly the practical difficulties experienced in deploying the expert system in a power plant environment. A slightly different application, being used by Scottish Power, is the system used for cost and performance modeling in power generation, which uses a rule-based approach to plant simulation (Adams and MacDonald, 1993). The level of use of AI techniques in the U.K. contrasts sharply with the power industry in the U.S., which is much more diverse, with many small private generating companies. There, the Electrical Power Research Institute (EPRI) has developed (and is continuing to develop) numerous advisory systems using expert system methods, and is actively investigating emerging AI techniques. Clearly, the uptake of expert systems/knowledge based systems (KBS) in the power industry has increased considerably in recent years. Also, the use of other AI techniques, such as neural networks and hybrid systems, is becoming popular. 3. APPLICATIONSExpert systems have been applied to various aspects of the power generation process, from diagnosing problems on steam turbines to monitoring and diagnosing problems in the entire generation process. Many of the systems described in this chapter involve a form of machine condition monitoring. Condition monitoring is not a single technique, but in fact is an "umbrella" term used to describe a wide variety of different but complimentary techniques that can be used to determine machinery condition. Other terms such as "predictive maintenance," "condition-based maintenance," and "reliability-centered maintenance" have been used to describe the same principles. The purpose of implementing condition monitoring within an industrial plant is to move away from the maintenance strategies of breakdown maintenance (where machinery is repaired or replaced after breakdown) or planned maintenance (where maintenance work is carried out at regular intervals, whether or not it is actually required). Condition monitoring, in simple terms, involves the capture of data that describes various parameters of machinery, and the analysis of these data to determine the condition of a given machine at a given point in time. Data analysis can be a complex process, often involving qualitative judgments made by experienced engineers who cannot always express fully the reasons for their interpretations; this presents a difficulty in the application of condition monitoring and an opportunity for the application of expert system technology. Examples of systems are given in the sections that follow. The systems are categorized into their relevant areas of operation within the power generation process.
|
![]() |
|
Use of this site is subject certain Terms & Conditions. Copyright (c) 1996-1999 EarthWeb, Inc.. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of EarthWeb is prohibited. Please read our privacy policy for details. |