EXPERT SYSTEMS AND NEURAL NETWORKS IN. PLANT OPERATIONS [4]

Introduction

14.31. Expert systems and neural networks were introduced in Chapter 8 as information tools. Expert, or so-called “knowledge-based” systems play an important role in nuclear operator support systems and provide diagnostic aids in plant maintenance activities. Their role is to take ad­vantage of various knowledge sources and to provide guidance in solving problems. Neural networks, on the other hand, utilize highly intercon­nected elements that process information in a manner similar to that used in the human brain.

14.32. A feature of neural networks is their ability to recognize patterns, even when the incoming data are incomplete or distorted, which is not possible with normal computer programs. Therefore, the information­gathering power of expert systems, with the disadvantage of some input uncertainty or “fuzziness,” complements the processing power of neural networks, which can quickly yield results despite incomplete input. Al­though application of expert systems to nuclear power plants has become fairly common, neural network use has been mostly developmental in nature. Our primary purpose is to draw attention to the features of such applications.