You need JavaScript to view this

On-line plant-wide monitoring using neural networks. Paper submitted to 8th Power Plant Dynamics, Control and Testing Symposium, May 27-29, 1992, Knoxville, Tennessee, USA

Abstract

The on-line signal analysis system designed for a multi-level mode operation using neural networks is described. The system is capable of monitoring the plant states by tracking different number of signals up to 32 simultaneously. The data used for this study were acquired from the Borssele Nuclear Power Plant (PWR type), and using the on-line monitoring system. An on-line plant-wide monitoring study using a multilayer neural network model is discussed in this paper. The back-propagation neural network algorithm is used for training the network. The technique assumes that each physical state of the power plant can be represented by a unique pattern of instrument readings which can be related to the condition of the plant. When disturbance occurs, the sensor readings undergo a transient, and form a different set of patterns which represent the new operational status. Diagnosing these patterns can be helpful in identifying this new state of the power plant. To this end, plant-wide monitoring with neutral networks is one of the new techniques in real-time applications. (author). 9 refs.; 5 figs.
Authors:
Turkcan, E; [1]  Ciftcioglu, O; [2]  Eryurek, E; Upadhyaya, B R [3] 
  1. Netherlands Energy Research Foundation (ECN), Petten (Netherlands)
  2. Istanbul Technical University (Turkey). Electrical Engineering Faculty
  3. Tennessee Univ., Knoxville, TN (United States). Dept. of Nuclear Engineering
Publication Date:
Jun 01, 1992
Product Type:
Technical Report
Report Number:
ECN-RX-92-033; CONF-920538-
Reference Number:
SCA: 220400; 210200; PA: AIX-24:027053; SN: 93000951195
Resource Relation:
Conference: 8. power plant dynamics, control and testing symposium,Knoxville, TN (United States),27-29 May 1992; Other Information: PBD: Jun 1992
Subject:
22 GENERAL STUDIES OF NUCLEAR REACTORS; 21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS; REACTOR CONTROL SYSTEMS; NEURAL NETWORKS; ON-LINE CONTROL SYSTEMS; BORSSELE REACTOR; NUMERICAL DATA; REACTOR SAFETY; REAL TIME SYSTEMS; SIGNALS; 220400; 210200; CONTROL SYSTEMS; POWER REACTORS, NONBREEDING, LIGHT-WATER MODERATED, NONBOILING WATER COOLED
OSTI ID:
10129479
Research Organizations:
Netherlands Energy Research Foundation (ECN), Petten (Netherlands)
Country of Origin:
Netherlands
Language:
English
Other Identifying Numbers:
Other: ON: DE93618419; TRN: NL92C0492027053
Availability:
OSTI; NTIS; INIS
Submitting Site:
NLN
Size:
[13] p.
Announcement Date:
Jul 04, 2005

Citation Formats

Turkcan, E, Ciftcioglu, O, Eryurek, E, and Upadhyaya, B R. On-line plant-wide monitoring using neural networks. Paper submitted to 8th Power Plant Dynamics, Control and Testing Symposium, May 27-29, 1992, Knoxville, Tennessee, USA. Netherlands: N. p., 1992. Web.
Turkcan, E, Ciftcioglu, O, Eryurek, E, & Upadhyaya, B R. On-line plant-wide monitoring using neural networks. Paper submitted to 8th Power Plant Dynamics, Control and Testing Symposium, May 27-29, 1992, Knoxville, Tennessee, USA. Netherlands.
Turkcan, E, Ciftcioglu, O, Eryurek, E, and Upadhyaya, B R. 1992. "On-line plant-wide monitoring using neural networks. Paper submitted to 8th Power Plant Dynamics, Control and Testing Symposium, May 27-29, 1992, Knoxville, Tennessee, USA." Netherlands.
@misc{etde_10129479,
title = {On-line plant-wide monitoring using neural networks. Paper submitted to 8th Power Plant Dynamics, Control and Testing Symposium, May 27-29, 1992, Knoxville, Tennessee, USA}
author = {Turkcan, E, Ciftcioglu, O, Eryurek, E, and Upadhyaya, B R}
abstractNote = {The on-line signal analysis system designed for a multi-level mode operation using neural networks is described. The system is capable of monitoring the plant states by tracking different number of signals up to 32 simultaneously. The data used for this study were acquired from the Borssele Nuclear Power Plant (PWR type), and using the on-line monitoring system. An on-line plant-wide monitoring study using a multilayer neural network model is discussed in this paper. The back-propagation neural network algorithm is used for training the network. The technique assumes that each physical state of the power plant can be represented by a unique pattern of instrument readings which can be related to the condition of the plant. When disturbance occurs, the sensor readings undergo a transient, and form a different set of patterns which represent the new operational status. Diagnosing these patterns can be helpful in identifying this new state of the power plant. To this end, plant-wide monitoring with neutral networks is one of the new techniques in real-time applications. (author). 9 refs.; 5 figs.}
place = {Netherlands}
year = {1992}
month = {Jun}
}