74 Matching Results

Search Results

Advanced search parameters have been applied.

Reactor noise analysis by statistical pattern recognition methods

Description: A multivariate statistical pattern recognition system for reactor noise analysis is presented. The basis of the system is a transformation for decoupling correlated variables and algorithms for inferring probability density functions. The system is adaptable to a variety of statistical properties of the data, and it has learning, tracking, updating, and data compacting capabilities. System design emphasizes control of the false-alarm rate. Its abilities to learn normal patterns, to recognize deviations from these patterns, and to reduce the dimensionality of data with minimum error were evaluated by experiments at the Oak Ridge National Laboratory (ORNL) High-Flux Isotope Reactor. Power perturbations of less than 0.1 percent of the mean value in selected frequency ranges were detected by the pattern recognition system.
Date: January 1, 1976
Creator: Howington, L.C. & Gonzalez, R.C.
Partner: UNT Libraries Government Documents Department

Experience with a digital noise analysis system in subcriticality measurements on a mockup of the FFTF

Description: From nuclear science symposium; San Francisco, California, USA (14 Nov 1973). A digital Fourier analyzer was programmed to perform reactor neutron noise analysis measurements and on-line processing of the data to obtain the steady-state reactivity. The system is suitable for recovering cross spectral density with low correlatedsignal component and for repetitive measurements with efficient use of reactor time. (auth)
Date: January 1, 1973
Creator: Pare, V.K.; Kryter, R.C. & Mihalczo, J.T.
Partner: UNT Libraries Government Documents Department

Multivariate statistical pattern recognition system for reactor noise analysis

Description: A multivariate statistical pattern recognition system for reactor noise analysis was developed. The basis of the system is a transformation for decoupling correlated variables and algorithms for inferring probability density functions. The system is adaptable to a variety of statistical properties of the data, and it has learning, tracking, and updating capabilities. System design emphasizes control of the false-alarm rate. The ability of the system to learn normal patterns of reactor behavior and to recognize deviations from these patterns was evaluated by experiments at the ORNL High-Flux Isotope Reactor (HFIR). Power perturbations of less than 0.1 percent of the mean value in selected frequency ranges were detected by the system. 19 references. (auth)
Date: January 1, 1975
Creator: Gonzalez, R.C.; Howington, L.C.; Sides, W.H. Jr. & Kryter, R.C.
Partner: UNT Libraries Government Documents Department

On-site noise diagnostics at Palisades Nuclear Power Station

Description: From an applications symposium on power plant dynamics, control and testing; Knoxville, Tennessee, USA (8 Oct 1973). A real-time noise analyzer was used on-site at the Palisades nuclear plant to diagnose the cause of neutron flux oscillations. The analysis was compared with an earlier off-site analysis of tape recorded signals. The measurements and conclusions are described, and recommendations for application of on-site noise analysis in nuclear plants are presented. (HDR)
Date: January 1, 1973
Creator: Fry, D.N.; Kryter, R.C. & Robinson, J.C.
Partner: UNT Libraries Government Documents Department

Results in the application of pattern recognition methods to nuclear reactor core component surveillance

Description: From nuclear science symposium; San Francisco, California, USA (14 Nov 1973). Pattern recognition methods were applied to analyze and interpret neutron noise data from the High Flux Isotope Reactor (HFIR) at ORNL. The results show that it is feasible to detect some core component failures by means of machine- discernible differences in the time-dependent noise power spectra. These neutron spectra (signatures) were analyzed by using a clusterseeking algorithm to derive a set of templates for automatic computer evaluation of the reactor's mechanical integrity and soundness. (auth)
Date: January 1, 1973
Creator: Gonzalez, R.C.; Fry, D.N. & Kryter, R.C.
Partner: UNT Libraries Government Documents Department

On-line reactor surveillance algorithm based on multivariate analysis of noise

Description: A mini-computer based surveillance algorithm for monitoring signals from nuclear power plants to provide advanced warning of anomalous conditions has been developed and implemented for on-line applications. The reactor state is characterized by the analysis of noise signals. During an observation period when normal conditions prevail, the surveillance algorithm utilizes these noise characterizations to form a statistical description of normal reactor behavior. At the conclusion of this observation period, the surveillance algorithm examines each new incoming characterization to determine if it differs signicantly from the normal description. (auth)
Date: January 1, 1975
Creator: Piety, K.R. & Robinson, J.C.
Partner: UNT Libraries Government Documents Department

Application of noise analysis to safety-related assessments and reactor diagnostics

Description: Noise analysis methods were used to assess anomalous in-core temperature fluctuations at the Fort St. Vrain gas-cooled reactor and postaccident reactor conditions at Three Mile Island, Unit 2. In addition to these applications of noise analysis, the underlying technology is developed concerning (1) analytical methods for predicting noise signatures under postulated anomalous conditions, (2) techniques for on-line monitoring of boiling water reactor stability, (3) new methods for locating and characterizing loose or drifting metallic objects in reactor coolant systems, and (4) acquisition of baseline noise signatures for commercial pressurized water reactors.
Date: January 1, 1980
Creator: Dryter, R.C. & Fry, D.N.
Partner: UNT Libraries Government Documents Department

Neutron chain length distributions in subcritical systems

Description: In this paper, the authors present the results of the chain-length distribution as a function of k in subcritical systems. These results were obtained from a point Monte Carlo code and a three-dimensional Monte Carlo code, MC++. Based on these results, they then attempt to explain why several of the common neutron noise techniques, such as the Rossi-{alpha} and Feynman's variance-to-mean techniques, are difficult to perform in highly subcritical systems using low-efficiency detectors.
Date: September 27, 1999
Creator: Nolen, S.D. & Spriggs, G.
Partner: UNT Libraries Government Documents Department

THE DETECTION OF BOILING IN A WATER-COOLED NUCLEAR REACTOR

Description: Measurements made at ORNL to study the feasibility of boiling detection in a water-cooled nuclear reactor are described. The methods selected for the detection of boiling include measurement of the acoustical noise produced by the generation of bubbles and measurement of changes in the reactor-power spectral density produced by bubbles. Preliminary results indicating that both methods could detect boiling are shown. (auth)
Date: August 17, 1962
Creator: Colomb, A.L. & Binford, F.T.
Partner: UNT Libraries Government Documents Department

Absolute subcriticality measurement without calibration and detection efficiency dependence by the /sup 252/Cf source-driven noise method

Description: The /sup 252/Cf-source-driven noise analysis method determines the subcriticality of a system containing fissionable material from the ratio of cross power spectral densities between the detectors that detect particles from the fission process and between these detectors and an ionization chamber containing a spontaneously fissioning neutron source which provides neutrons to induce fission in the system. This method has two advantages: (1) a calibration is not required and thus subcriticality can be determined from measurements only on the subcritical system of interest, and (2) the subcriticality is independent of the type of detector or its efficiency. These properties of this technique are illustrated by measurements.
Date: January 1, 1984
Creator: Mihalczo, J.T. & King, W.T.
Partner: UNT Libraries Government Documents Department

On-line analysis of reactor noise using time-series analysis

Description: A method to allow use of time series analysis for on-line noise analysis has been developed. On-line analysis of noise in nuclear power reactors has been limited primarily to spectral analysis and related frequency domain techniques. Time series analysis has many distinct advantages over spectral analysis in the automated processing of reactor noise. However, fitting an autoregressive-moving average (ARMA) model to time series data involves non-linear least squares estimation. Unless a high speed, general purpose computer is available, the calculations become too time consuming for on-line applications. To eliminate this problem, a special purpose algorithm was developed for fitting ARMA models. While it is based on a combination of steepest descent and Taylor series linearization, properties of the ARMA model are used so that the auto- and cross-correlation functions can be used to eliminate the need for estimating derivatives.
Date: October 1, 1981
Creator: McGevna, V.G.
Partner: UNT Libraries Government Documents Department

Multilevel crossing rates for automated signal classification

Description: An investigation was made of multilevel crossing rates as a means of time series analysis of random signals. Pattern recognition techniques based on the Mahalanobis distance were implemented as a means of evaluating the discriminating power of level crossings. Measurement of multilevel crossing rates was found to be an easily implementable means for detection of changes in general frequency content. Level crossing analysis was also shown to be applicable for the study of conductivity measurements of two-phase flow of air and water, where knowledge of the relationship between amplitude and frequency was beneficial in characterizing the process.
Date: January 1, 1978
Creator: Mitchell, R.J. & Gonzalez, R.C.
Partner: UNT Libraries Government Documents Department

An evaluation of neural networks for identification of system parameters in reactor noise signals

Description: Several backpropagation neural networks for identifying fundamental mode eigenvalues were evaluated. The networks were trained and tested on analytical data and on results from other numerical methods. They were then used to predict first mode break frequencies for noise data from several sources. These predictions were, in turn, compared with analytical values and with results from alternative methods. Comparisons of results for some data sets suggest that the accuracy of predictions from neural networks are essentially equivalent to results from conventional methods while other evaluations indicate that either method may be superior. Experience gained from these numerical experiments provide insight for improving the performance of neural networks relative to other methods for identifying parameters associated with experimental data. Neural networks may also be used in support of conventional algorithms by providing starting points for nonlinear minimization algorithms.
Date: December 31, 1991
Creator: Miller, L. F.
Partner: UNT Libraries Government Documents Department

System for unattended surveillance of nuclear reactor behavior

Description: A multivariate statistical pattern recognition system for reactor noise analysis is presented. The basis of the system is a transformation for decoupling correlated variables and algorithms for inferring probability density functions. The system is adaptable to a variety of statistical properties of the data, and it has learning, tracking, updating, and dimensionality reduction capabilities. System design emphasizes control of the false-alarm rate. Its abilities to learn normal patterns and to recognize deviations from these patterns were evaluated by experiments at the ORNL High-Flux Isotope Reactor. Power perturbations of less than 0.1% of the mean value in selected frequency ranges were readily detected by the pattern recognition system.
Date: January 1, 1977
Creator: Gonzalez, R. C. & Howington, L. C.
Partner: UNT Libraries Government Documents Department

Development of an automated diagnostic system for BWR stability measurements

Description: An algorithm capable of automatically evaluating BWR stability has been developed. Main advantages are: Conservative estimate (asymptotic), adjusts to solve difficult conditions, confidence level, and error estimate. The apparent decay ratio (DR) is not a conservative estimate of the reactor stability. The asymptotic decay ratio must be used. Long enough record lengths must be used to reduce the uncertainty of the estimated DR.
Date: October 1, 1984
Creator: March-Leuba, J. & Smith, C.M.
Partner: UNT Libraries Government Documents Department