Sökning: onr:"swepub:oai:research.chalmers.se:5fad8ed0-7039-40b9-8604-0da8f7ac200e" >
A deep learning app...
A deep learning approach to anomaly detection in nuclear reactors
-
- Calivá, Francesco (författare)
- University of Lincoln
-
- De Sousa Ribeiro, Fabio (författare)
- University of Lincoln
-
- Mylonakis, Antonios, 1987 (författare)
- Chalmers tekniska högskola,Chalmers University of Technology
-
visa fler...
-
- Demaziere, Christophe, 1973 (författare)
- Chalmers tekniska högskola,Chalmers University of Technology
-
- Vinai, Paolo, 1975 (författare)
- Chalmers tekniska högskola,Chalmers University of Technology
-
- Leontidis, Georgios (författare)
- University of Lincoln
-
- Kollias, Stefanos (författare)
- University of Lincoln
-
visa färre...
-
(creator_code:org_t)
- 2018
- 2018
- Engelska.
-
Ingår i: Proceedings of the International Joint Conference on Neural Networks. ; 2018-July
- Relaterad länk:
-
https://doi.org/10.1...
-
visa fler...
-
https://research.cha...
-
https://research.cha...
-
visa färre...
Abstract
Ämnesord
Stäng
- 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.
Ämnesord
- NATURVETENSKAP -- Data- och informationsvetenskap (hsv//swe)
- NATURAL SCIENCES -- Computer and Information Sciences (hsv//eng)
- NATURVETENSKAP -- Fysik -- Annan fysik (hsv//swe)
- NATURAL SCIENCES -- Physical Sciences -- Other Physics Topics (hsv//eng)
Nyckelord
- anomaly detection
- denoising autoencoders
- deep learning
- signal processing
- convolutional neural networks
- clustering trained representations
- nuclear reactors
- unfolding
Publikations- och innehållstyp
- kon (ämneskategori)
- ref (ämneskategori)