Localizing Perturbations in Pressurized Water Reactors Using One-Dimensional Deep Convolutional Neural Networks

@article{Pantera2021LocalizingPI,
  title={Localizing Perturbations in Pressurized Water Reactors Using One-Dimensional Deep Convolutional Neural Networks},
  author={Laurent Pantera and Petr Stul{\'i}k and Antoni Vidal-Ferr{\`a}ndiz and Amanda Carre{\~n}o and Dami{\'a}n Ginestar and George Ioannou and Thanos Tasakos and Georgios Alexandridis and Andreas Stafylopatis},
  journal={Sensors (Basel, Switzerland)},
  year={2021},
  volume={22}
}
This work outlines an approach for localizing anomalies in nuclear reactor cores during their steady state operation, employing deep, one-dimensional, convolutional neural networks. Anomalies are characterized by the application of perturbation diagnostic techniques, based on the analysis of the so-called “neutron-noise” signals: that is, fluctuations of the neutron flux around the mean value observed in a steady-state power level. The proposed methodology is comprised of three steps: initially… 
View PDF
4 Citations
Methods Citations
1

4 Citations

Deep learning techniques for in-core perturbation identification and localization of time-series nuclear plant measurements

Machine learning for analysis of real nuclear plant data in the frequency domain Annals of Nuclear Energy

A domain adaptation methodology is subsequently developed to extend the sim- ulated setting to real plant measurements, which uses self-supervised, or unsupervised learning, to align the simulated data with the actual plant data and detect perturbations, whilst classifying their type and estimating their location.

Multivariate Time Series Prediction for Loss of Coolant Accidents With a Zigmoid-Based LSTM

To enable an ability promotion of previous LOCA prediction models, a new gate function called zigmoid is introduced and embedded to the traditional long short-term memory (LSTM) model, which amplifies the gradient at the far end of the time series.

Machine learning for analysis of real nuclear plant data in the frequency domain

References

SHOWING 1-10 OF 39 REFERENCES

Deep Learning-based Anomaly Detection in Nuclear Reactor Cores

In this work, a methodology is proposed for the classification of different perturbation types and their position in a nuclear reactor core. More specifically, it is based on a Convolutional Neural…

Towards a Deep Unified Framework for Nuclear Reactor Perturbation Analysis

The first steps towards a novel unified framework for the analysis of perturbations in both the Time and Frequency domains are taken, showing that the perturbation type can be recognised with high accuracy, and frequency domain scenario sources can be localised with high precision.

NEUTRON NOISE-BASED ANOMALY CLASSIFICATION AND LOCALIZATION USING MACHINE LEARNING

The proposed methodology allowing the classification of anomalies and subsequently their possible localization in nuclear reactor cores during operation was demonstrated to be insensitive to parasitic noise and will be tested on actual plant data in the near future.

3D convolutional and recurrent neural networks for reactor perturbation unfolding and anomaly detection

An extended Deep Learning framework as part of the CORTEX Horizon 2020 EU project for the unfolding of reactor transfer functions from induced neutron noise sources is proposed and achieves state-of-the-art results.

Putting Together Wavelet-based Scaleograms and Convolutional Neural Networks for Anomaly Detection in Nuclear Reactors

This work presents a novel technique for anomaly detection on nuclear reactor signals through the combined use of wavelet-based analysis and convolutional neural networks, and indicates that the trained network achieves high levels of accuracy in failure detection, while at the same time being robust to noise.

INTELLIGENT TECHNIQUES FOR ANOMALY DETECTION IN NUCLEAR REACTORS

The obtained results of an initial stage of analysis on neutron flux signals captured at pressurized water reactors are encouraging, underlying the robustness and the potential of the proposed approach.

FEATURE EXTRACTION AND IDENTIFICATION TECHNIQUES FOR THE ALIGNMENT OF PERTURBATION SIMULATIONS WITH POWER PLANT MEASUREMENTS

In this work, a methodology is proposed for the comparison of the measured and simulated neutron noise signals in nuclear power plants, with the simulation sets having been generated by the CORE SIM+…

On the Neutron Noise Diagnostics of Pressurized Water Reactor Control Rod Vibrations —IV: Application of Neural Networks

A neural network-based identification method was investigated by constructing a network that was trained to determine the rod position from the detector spectra and found that all shortcomings of the traditional localization method can be eliminated.

Noise-based core monitoring and diagnostics: overview of the cortex project

An overview of the CORTEX project is given, which aims at developing an innovative core monitoring technique that allows detecting anomalies in nuclear reactors, such as excessive vibrations of core internals, flow blockage, coolant inlet perturbations, etc.

CORE SIM+: A flexible diffusion-based solver for neutron noise simulations