Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Analysis of neutron noise spectra using neural networks

Conference · · Transactions of the American Nuclear Society; (United States)
OSTI ID:5778048
 [1];  [2]
  1. Oak Ridge National Lab., TN (United States)
  2. Univ. of Tennessee, Knoxville (United States)
Neural network architectures based on backpropagation paradigm have been developed to recognize the features, and detect resonance shifts in, power spectral density (PSD) data. The goal is to advance the state of the art in the application of noise analysis techniques to monitor nuclear reactor internals. The initial objectives have been to use PSD data, acquired over a period of {approximately} 2 years by the power spectral density recognition system (PSDREC), to develop neural networks that are able to differentiate between neutron PSD data and anomalous spectral data (such as malfunctioning instrumentation) and detect significant shifts in the positions of spectral resonances while reducing the effect of small shifts (in neutron noise analysis, shifts in the resonance(s) present in a neutron PSD spectrum are the primary means for diagnosing degradation of reactor internals). Neural network systems referred to in this paper as spectral feature detectors (SFDs) and integral network filters (INFs) have been developed to meet these objectives. The performance of the SFDs is the subject of this paper.
OSTI ID:
5778048
Report Number(s):
CONF-910603--
Conference Information:
Journal Name: Transactions of the American Nuclear Society; (United States) Journal Volume: 63
Country of Publication:
United States
Language:
English