Noise Techniques in Nuclear Systems

Pázsit, Imre; Demazière, Christophe

doi:10.1007/978-0-387-98149-9_14

Cited by 38 publications

(24 citation statements)

References 81 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this method, data is processed to obtain average and variance of counts for different time interval [12]. Let T ⋅ be the time average operator for a time interval T. The following estimator is constructed:…”

Section: Feynman-methodsmentioning

confidence: 99%

See 1 more Smart Citation

Pile noise experiment in MINERVE reactor to estimate kinetic parameters using various data processing methods

Geslot

Gruel

Pépino

et al. 2015

2015 4th International Conference on Advancements in Nuclear Instrumentation Measurement Methods and Their Applications (ANIMMA

View full text Add to dashboard Cite

MINERVE is a two-zone pool type zero power reactor operated by CEA (Cadarache, France). Kinetic parameters of the core (prompt neutron decay constant, delayed neutron fraction, generation time) have been recently measured using various pile noise experimental techniques, namely Feynman-, Rossi-and Cohn-. Results are discussed and compared to each other's.The measurement campaign has been conducted in the framework of a tri-partite collaboration between CEA, SCK•CEN and PSI. Results presented in this paper were obtained thanks to a time-stamping acquisition system developed by CEA. PSI performed simultaneous measurements which are presented in a companion paper.Signals come from two high efficiency fission chambers located in the graphite reflector next to the core driver zone. Experiments were conducted at critical state with a reactor power of 0.2 W. The core integral fission rate is obtained from a calibrated miniature fission chamber located at the center of the core. Other results obtained in two sub-critical configurations will be presented elsewhere.Best estimate delayed neutron fraction comes from the Cohnmethod: 747 ± 15 pcm (1 ). In this case, the prompt decay constant is 79 ± 0.5 s -1 and the generation time is 94.5 ± 0.7 μs. Other methods give consistent results within the confidence intervals.Experimental results are compared to calculated values obtained from a full 3D core modeling with the CEA-developed Monte Carlo code TRIPOLI4.9 associated with its continuous energy JEFF3.1.1-based library. A very good agreement is observed for the calculated delayed neutron fraction (748.7 ± 0.4 pcm at 1 ), that is a difference of -0.3% with the experiment. On the contrary, a 10% discrepancy is observed for the calculated generation time (104.4 ± 0.1 μs at 1 ).

show abstract

Section: Feynman-methodsmentioning

confidence: 99%

“…Pile noise techniques allow estimating the kinetic parameters of a neutron multiplying system in a stationary state at critical or excited by an external neutron source [12]. Delayed neutron fraction and generation time can be linked to second order estimator functions based on neutron measurements (variance, correlations, power spectrum).…”

Section: A Principlesmentioning

confidence: 99%

Pile noise experiment in MINERVE reactor to estimate kinetic parameters using various data processing methods

Geslot

Gruel

Pépino

et al. 2015

2015 4th International Conference on Advancements in Nuclear Instrumentation Measurement Methods and Their Applications (ANIMMA

View full text Add to dashboard Cite

show abstract

“…During the forward problem, the reactor transfer function thoroughly captures the response of the neutron flux, which is induced by the known distribution of perturbations. If the perturbations reduce to a Dirac function applied to the point (r p ) at a given angular frequency (ω), then the transfer function is the Green's function of the system ([1], [15]). Considering that the effect of the perturbation can be assessed from any spatial point r, the induced neutron noise can be measured as in (2), where V refers to the volume of the reactor core.…”

Section: The Examined Scenariomentioning

confidence: 99%

A Deep Learning Approach to Anomaly Detection in Nuclear Reactors

Calivá

Ribeiro

Mylonakis

et al. 2018

2018 International Joint Conference on Neural Networks (IJCNN)

View full text Add to dashboard Cite

In this work, a novel deep learning approach to unfold nuclear power reactor signals is proposed. It includes a combination of convolutional neural networks (CNN), denoising autoencoders (DAE) and k-means clustering of representations. Monitoring nuclear reactors while running at nominal conditions is critical. Based on analysis of the core reactor neutron flux, it is possible to derive useful information for building fault/anomaly detection systems. By leveraging signal and image pre-processing techniques, the high and low energy spectra of the signals were appropriated into a compatible format for CNN training. Firstly, a CNN was employed to unfold the signal into either twelve or forty-eight perturbation location sources, followed by a k-means clustering and k-Nearest Neighbour coarse-to-fine procedure, which significantly increases the unfolding resolution. Secondly, a DAE was utilised to denoise and reconstruct power reactor signals at varying levels of noise and/or corruption. The reconstructed signals were evaluated w.r.t. their original counter parts, by way of normalised cross correlation and unfolding metrics. The results illustrate that the origin of perturbations can be localised with high accuracy, despite limited training data and obscured/noisy signals, across various levels of granularity.

show abstract

“…Neutron noise theory has its origins in the experiments performed at the "Clinton Pile" of Oak Ridge in the 1940s (Pázsit and Demazière, 2010), whose goal was the evaluation of absorption cross sections of various elements. Although they did not concern diagnostics at all, these experiments contemplated, for the first time, neutron oscillations.…”

Section: Introductionmentioning

confidence: 99%

A MOC-based neutron kinetics model for noise analysis

Gammicchia

Santandrea

Zmijarevic

et al. 2020

Annals of Nuclear Energy

View full text Add to dashboard Cite

A 2-D noise model is implemented in the deterministic reactor code APOLLO3 R to simulate a periodic oscillation of a structural component. The Two/Three Dimensional Transport (TDT) solver, using the Method of Characteristics, is adopted for the calculation of the case studies, constituted by a moving detector and control-rod bundle. The periodic movement is built by properly linking the geometries corresponding to the temporal positions. The calculation is entirely performed in the real time domain, without resorting to the traditional frequency approach. A specifically defined dynamic eigenvalue is used to renormalize in average the reactivity over a period. The algorithm is accelerated by the DP N synthetic method. For each cell of the domain, the time values of fission rates are analysed to determine the noise extent. Moreover we propose a systematic approach to the definition of the macroscopic cross sections to be used in dynamical calculations starting from library data. As an aside of our work we have found that even in static calculation this approach can produce significant changes.

show abstract

Noise Techniques in Nuclear Systems

Cited by 38 publications

References 81 publications

Pile noise experiment in MINERVE reactor to estimate kinetic parameters using various data processing methods

Pile noise experiment in MINERVE reactor to estimate kinetic parameters using various data processing methods

A Deep Learning Approach to Anomaly Detection in Nuclear Reactors

A MOC-based neutron kinetics model for noise analysis

Contact Info

Product

Resources

About