Methodology for Operational Diagnosis of Corrosion Damage to Cladding of Fuel Rod in VVER Reactors Based on Artificial Neural Network

G. I. Sharaievskyi

Institute for Safety Problems of Nuclear Power Plants, NAS of Ukraine, 12, Lysohirska st., Kyiv, 03028, Ukraine

DOI: doi.org/10.31717/2311-8253.25.2.3

Abstract

This work examines the practical implementation of the author’s theoretical approach to solving the problem of early detection of pre-accident thermal state of the fuel rod cladding in a light water reactor with critical thermodynamic parameters of the coolant. The goal is to automatically detect the initial stages of vapor phase generation on the surface of the heat exchanger. The physical principle of the proposed diagnostic model involves timely, non-contact recognition of the activation regime of the first centers of vapor formation based on computer identification of spectral signatures of diagnostic signals of this abnormal thermal state of the fuel rod surface. The computer implementation of this preventive diagnostic approach involves the use of computational algorithms to ensure automatic recognition of neutron noise signals in the reactor core of a nuclear power reactor of the VVER type in a multi-dimensional space of spectral features. As indicators of thermohydraulic regimes subject to automatic recognition, the work proposes using a set of spectral components of noise diagnostic signals recorded at the outputs of measurement sensors, primarily electron-emission neutron flow detectors. The work demonstrates the possibility of using neutron diagnostic signals and their spectral signatures in computational procedures that ensure automatic recognition of non-regulated thermal contact regimes with the fuel rod surface. Practical verification of the developed mathematical support and the created software complex for automatic identification of non-regulated thermohydraulic regimes in the area of critical thermodynamic parameters of light water coolant has been carried out. The possibility of preventive detection of an emergency heat transfer crisis by the parameters of neutron noise in the heat-releasing assembly of fuel rod has been demonstrated.

Keywords: preventive diagnostics, corrosion damage, start of boiling coolant, neutron noise spectra, spectral signature recognition, automatic recognition of diagnostic signals, artificial neural network, Kohonen topology

References

1. Shapovalova S. I., Sharayevskyi G. I. (2007). [Computer modeling of a self-organization map for solving the problem of signal recognition]. Bulletin of the National University “Lviv Polytechnic”. Automation, Measurement and Control, vol. 574, pp. 75−81. (in Ukr.)

2. Tolubinsky V. I. (1980). Teploobmen pri kipenii [Heat transfer during boiling]. Kyiv: Naukova Dumka, 315 p. (in Rus.)

3. Nosovskyi A. V., Sharaevsky I. G., Fialko N. M., Zimin L. B., Sharaevsky G. I. (2017). Teplofizika resursa jadernykh energoustanovok [Thermal physics of NPP resource]. ISP NPP, NAS of Ukraine, 624 p. (in Rus.)

4. Kliuchnikov А. А., Sharaevskyi I. G., Fialko N. M., Zimin L. B., Sharaevskyi G. I. (2010). Teplofizika bezopasnosti atomnykh elektrostantsii [Thermophysics of NPP safety]. Kyiv: ISP NPP, NAS of Ukraine, 484 p. (in Rus.)

5. Nesis E. I. (1973). Kipeniye zhidkostey [Boiling of Liquids]. Moscow: Nauka, 280 p. (in Rus.)

6. Andreeva A. B., Maershina G. I., Kobylinsky G. P. (1989). Ochagovaya korroziya tsirkoniyevykh splavov. Faktory i mekhanizmy. Obzor [Localized corrosion of zirconium alloys. Factors and mechanisms. Review]. Moscow: TsNII-Atominform, 42 p. (in Rus.)

7. Sharaievskyi G. I. (2011). [Neural network software for automatic diagnostics of nuclear power plant equipment elements] (PhD Thesis). Kyiv: G. E. Pukhov Institute for Modelling in Energy Engineering, NAS of Ukraine, 24 p. (in Ukr.)

8. Hardware and software complex for noise diagnostics of VVER reactors. Technical description. Moscow: National Research Centre “Kurchatov Institute”, 1992, 235 p. (in Rus.

Full Text (PDF)