Paper Info
Title
Online Transient Behavior Prediction in Nuclear Power Plants
Abstract
This article describes an online simulator that works on a fuzzy network model to predict transient behavior of a nuclear power plant. The model can be a causal fuzzy, qualitative fuzzy, or quantitative fuzzy network depending on which simulation, fuzzy, qualitative fuzzy, or quantitative fuzzy simulation, is requested. The networks are derived from underlying mathematical models about the system. This deep design knowledge empowers the simulator to do transient behavior prediction. Specifically, each network contains properly grained fuzzy operators to relate system variables together. The operators are represented as mapping tables so that time-consuming fuzzy computations can be reduced to efficient table-lookup during online simulation. In addition, the simulator is equipped with a mechanism to dynamically decide a best time increment between two simulation steps, which makes the simulation even more efficient. Finally, the simulator includes analysis knowledge, operational knowledge, and input-output records to support diagnosis, prediction, and learning capabilities, which allow the simulator to make fast responses, answer anytime status queries, propose reasonable solutions, and perform adaptive modeling. It has been shown that it can successfully predict transient behavior of the temperature and pressure of a steam generator with ruptured U-tubes undergoing the operator recovery process in a nuclear power plant. This helps the plant operator immensely in correctly making and successfully carrying out emergency response plans to deal with critical accidents, whether or not they are unprecedented.
Year
DOI
Venue
2000
10.1080/08839510050179464
APPLIED ARTIFICIAL INTELLIGENCE
Keywords
Field
DocType
network model,steam generator,input output,mathematical model
Data mining,Neuro-fuzzy,Design knowledge,Computer science,Fuzzy logic,Artificial intelligence,Operator (computer programming),Fuzzy control system,Adaptive neuro fuzzy inference system,Mathematical model,Machine learning,Network model
Journal
Volume
Issue
ISSN
14
10
0883-9514
Citations 
PageRank 
References 
0
0.34
8
Authors
2
Name
Order
Citations
PageRank
Fu-hua Chou161.72
Cheng-Seen Ho212515.78