Statistical and sensitivity analysis of failing rods in EPR LBLOCA

Asko Arkoma, Timo Ikonen

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

Abstract

A statistical fuel failure analysis procedure has been developed [1], and now applied in full scale to a large break loss-of-coolant accident (LBLOCA) scenario in EPR-type nuclear power plant. The goal is to statistically determine the number of failing fuel rods. It is also important to bring out the underlying causes for rod failures, and therefore a sensitivity analysis procedure for the preceding complex calculation chain is outlined and adopted. In order to produce a statistically reliable estimation on the number of failing rods, a large number of simulations are required. In the analysis, single rods are simulated with the coupled fuel performance - thermal hydraulics code FRAPTRAN-GENFLO. The statistically varied factors are divided into global and local by their range of influence. Transient thermal hydraulic and power history boundary conditions and model parameters are global and affect all the rods. Fuel rod manufacturing parameters and other rod position related factors are local and are related to a certain rod. The number of global scenarios is 59 as dictated by the Wilks' formula, and 1000 randomly sampled rods are simulated in each scenario. The thermal hydraulic and transient power boundary conditions for the fuel performance code are calculated with a statistical version of the system code APROS. The steady-state irradiation histories are simulated with FRAPCON, and the steady-state power histories used as boundary conditions for FRAPCON are obtained from SIMULATE calculations. Thus, a multistage calculation chain is required to consolidate the procedure. As an outcome, in the worst global scenario, 1.4% of the simulated rods failed. It can be concluded that the Finnish safety regulations, i.e. max. 10% of the rods allowed to fail, are met. In order to find out which input parameters have significance to the outcome of the analysis, a sensitivity analysis is done. At the moment, only the worst global scenario is considered but later on it is possible to include also the sensitivities of the global factors. Due to complexity of the existing data, first the relevant input parameters for the sensitivity analysis have to be specified. Data visualization with a cobweb graph is used for the screening. Then, selected sensitivity measures are calculated between the chosen input and output parameters. The sensitivity indices calculated are the Borgonovo's delta measure, the first order Sobol' sensitivity index, and squared Pearson correlation coefficients. The first mentioned is a novelty in this context. As an outcome, the most relevant parameters with respect to the cladding integrity were determined to be the decay heat power during the transient, the thermal hydraulic conditions in the rod's location in the reactor, and the steady-state irradiation history of the rod as represented in this analysis by the rod burnup.

Original languageEnglish
Title of host publicationTop Fuel 2016
Subtitle of host publicationLWR Fuels with Enhanced Safety and Performance
PublisherAmerican Nuclear Society ANS
Pages1115-1124
Number of pages10
ISBN (Electronic)9780894487309
Publication statusPublished - 1 Jan 2016
MoE publication typeA4 Article in a conference publication
EventTop Fuel 2016: LWR Fuels with Enhanced Safety and Performance - Boise, United States
Duration: 11 Sep 201615 Sep 2016

Conference

ConferenceTop Fuel 2016: LWR Fuels with Enhanced Safety and Performance
CountryUnited States
CityBoise
Period11/09/1615/09/16

Fingerprint

Loss of coolant accidents
Sensitivity analysis
Paramagnetic resonance
Statistical methods
Hydraulics
Boundary conditions
Irradiation
Data visualization
Nuclear power plants
Failure analysis
Screening
Hot Temperature

Keywords

  • EPR
  • FRAPTRAN-GENFLO
  • Loss-of-coolant accident
  • Sensitivity analysis
  • Statistical fuel failure analysis

Cite this

Arkoma, A., & Ikonen, T. (2016). Statistical and sensitivity analysis of failing rods in EPR LBLOCA. In Top Fuel 2016: LWR Fuels with Enhanced Safety and Performance (pp. 1115-1124). American Nuclear Society ANS.
Arkoma, Asko ; Ikonen, Timo. / Statistical and sensitivity analysis of failing rods in EPR LBLOCA. Top Fuel 2016: LWR Fuels with Enhanced Safety and Performance. American Nuclear Society ANS, 2016. pp. 1115-1124
@inproceedings{8dc1458a758f4e00bfb3cd867a29bb9b,
title = "Statistical and sensitivity analysis of failing rods in EPR LBLOCA",
abstract = "A statistical fuel failure analysis procedure has been developed [1], and now applied in full scale to a large break loss-of-coolant accident (LBLOCA) scenario in EPR-type nuclear power plant. The goal is to statistically determine the number of failing fuel rods. It is also important to bring out the underlying causes for rod failures, and therefore a sensitivity analysis procedure for the preceding complex calculation chain is outlined and adopted. In order to produce a statistically reliable estimation on the number of failing rods, a large number of simulations are required. In the analysis, single rods are simulated with the coupled fuel performance - thermal hydraulics code FRAPTRAN-GENFLO. The statistically varied factors are divided into global and local by their range of influence. Transient thermal hydraulic and power history boundary conditions and model parameters are global and affect all the rods. Fuel rod manufacturing parameters and other rod position related factors are local and are related to a certain rod. The number of global scenarios is 59 as dictated by the Wilks' formula, and 1000 randomly sampled rods are simulated in each scenario. The thermal hydraulic and transient power boundary conditions for the fuel performance code are calculated with a statistical version of the system code APROS. The steady-state irradiation histories are simulated with FRAPCON, and the steady-state power histories used as boundary conditions for FRAPCON are obtained from SIMULATE calculations. Thus, a multistage calculation chain is required to consolidate the procedure. As an outcome, in the worst global scenario, 1.4{\%} of the simulated rods failed. It can be concluded that the Finnish safety regulations, i.e. max. 10{\%} of the rods allowed to fail, are met. In order to find out which input parameters have significance to the outcome of the analysis, a sensitivity analysis is done. At the moment, only the worst global scenario is considered but later on it is possible to include also the sensitivities of the global factors. Due to complexity of the existing data, first the relevant input parameters for the sensitivity analysis have to be specified. Data visualization with a cobweb graph is used for the screening. Then, selected sensitivity measures are calculated between the chosen input and output parameters. The sensitivity indices calculated are the Borgonovo's delta measure, the first order Sobol' sensitivity index, and squared Pearson correlation coefficients. The first mentioned is a novelty in this context. As an outcome, the most relevant parameters with respect to the cladding integrity were determined to be the decay heat power during the transient, the thermal hydraulic conditions in the rod's location in the reactor, and the steady-state irradiation history of the rod as represented in this analysis by the rod burnup.",
keywords = "EPR, FRAPTRAN-GENFLO, Loss-of-coolant accident, Sensitivity analysis, Statistical fuel failure analysis",
author = "Asko Arkoma and Timo Ikonen",
year = "2016",
month = "1",
day = "1",
language = "English",
pages = "1115--1124",
booktitle = "Top Fuel 2016",
publisher = "American Nuclear Society ANS",
address = "United States",

}

Arkoma, A & Ikonen, T 2016, Statistical and sensitivity analysis of failing rods in EPR LBLOCA. in Top Fuel 2016: LWR Fuels with Enhanced Safety and Performance. American Nuclear Society ANS, pp. 1115-1124, Top Fuel 2016: LWR Fuels with Enhanced Safety and Performance, Boise, United States, 11/09/16.

Statistical and sensitivity analysis of failing rods in EPR LBLOCA. / Arkoma, Asko; Ikonen, Timo.

Top Fuel 2016: LWR Fuels with Enhanced Safety and Performance. American Nuclear Society ANS, 2016. p. 1115-1124.

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

TY - GEN

T1 - Statistical and sensitivity analysis of failing rods in EPR LBLOCA

AU - Arkoma, Asko

AU - Ikonen, Timo

PY - 2016/1/1

Y1 - 2016/1/1

N2 - A statistical fuel failure analysis procedure has been developed [1], and now applied in full scale to a large break loss-of-coolant accident (LBLOCA) scenario in EPR-type nuclear power plant. The goal is to statistically determine the number of failing fuel rods. It is also important to bring out the underlying causes for rod failures, and therefore a sensitivity analysis procedure for the preceding complex calculation chain is outlined and adopted. In order to produce a statistically reliable estimation on the number of failing rods, a large number of simulations are required. In the analysis, single rods are simulated with the coupled fuel performance - thermal hydraulics code FRAPTRAN-GENFLO. The statistically varied factors are divided into global and local by their range of influence. Transient thermal hydraulic and power history boundary conditions and model parameters are global and affect all the rods. Fuel rod manufacturing parameters and other rod position related factors are local and are related to a certain rod. The number of global scenarios is 59 as dictated by the Wilks' formula, and 1000 randomly sampled rods are simulated in each scenario. The thermal hydraulic and transient power boundary conditions for the fuel performance code are calculated with a statistical version of the system code APROS. The steady-state irradiation histories are simulated with FRAPCON, and the steady-state power histories used as boundary conditions for FRAPCON are obtained from SIMULATE calculations. Thus, a multistage calculation chain is required to consolidate the procedure. As an outcome, in the worst global scenario, 1.4% of the simulated rods failed. It can be concluded that the Finnish safety regulations, i.e. max. 10% of the rods allowed to fail, are met. In order to find out which input parameters have significance to the outcome of the analysis, a sensitivity analysis is done. At the moment, only the worst global scenario is considered but later on it is possible to include also the sensitivities of the global factors. Due to complexity of the existing data, first the relevant input parameters for the sensitivity analysis have to be specified. Data visualization with a cobweb graph is used for the screening. Then, selected sensitivity measures are calculated between the chosen input and output parameters. The sensitivity indices calculated are the Borgonovo's delta measure, the first order Sobol' sensitivity index, and squared Pearson correlation coefficients. The first mentioned is a novelty in this context. As an outcome, the most relevant parameters with respect to the cladding integrity were determined to be the decay heat power during the transient, the thermal hydraulic conditions in the rod's location in the reactor, and the steady-state irradiation history of the rod as represented in this analysis by the rod burnup.

AB - A statistical fuel failure analysis procedure has been developed [1], and now applied in full scale to a large break loss-of-coolant accident (LBLOCA) scenario in EPR-type nuclear power plant. The goal is to statistically determine the number of failing fuel rods. It is also important to bring out the underlying causes for rod failures, and therefore a sensitivity analysis procedure for the preceding complex calculation chain is outlined and adopted. In order to produce a statistically reliable estimation on the number of failing rods, a large number of simulations are required. In the analysis, single rods are simulated with the coupled fuel performance - thermal hydraulics code FRAPTRAN-GENFLO. The statistically varied factors are divided into global and local by their range of influence. Transient thermal hydraulic and power history boundary conditions and model parameters are global and affect all the rods. Fuel rod manufacturing parameters and other rod position related factors are local and are related to a certain rod. The number of global scenarios is 59 as dictated by the Wilks' formula, and 1000 randomly sampled rods are simulated in each scenario. The thermal hydraulic and transient power boundary conditions for the fuel performance code are calculated with a statistical version of the system code APROS. The steady-state irradiation histories are simulated with FRAPCON, and the steady-state power histories used as boundary conditions for FRAPCON are obtained from SIMULATE calculations. Thus, a multistage calculation chain is required to consolidate the procedure. As an outcome, in the worst global scenario, 1.4% of the simulated rods failed. It can be concluded that the Finnish safety regulations, i.e. max. 10% of the rods allowed to fail, are met. In order to find out which input parameters have significance to the outcome of the analysis, a sensitivity analysis is done. At the moment, only the worst global scenario is considered but later on it is possible to include also the sensitivities of the global factors. Due to complexity of the existing data, first the relevant input parameters for the sensitivity analysis have to be specified. Data visualization with a cobweb graph is used for the screening. Then, selected sensitivity measures are calculated between the chosen input and output parameters. The sensitivity indices calculated are the Borgonovo's delta measure, the first order Sobol' sensitivity index, and squared Pearson correlation coefficients. The first mentioned is a novelty in this context. As an outcome, the most relevant parameters with respect to the cladding integrity were determined to be the decay heat power during the transient, the thermal hydraulic conditions in the rod's location in the reactor, and the steady-state irradiation history of the rod as represented in this analysis by the rod burnup.

KW - EPR

KW - FRAPTRAN-GENFLO

KW - Loss-of-coolant accident

KW - Sensitivity analysis

KW - Statistical fuel failure analysis

UR - http://www.scopus.com/inward/record.url?scp=85019018043&partnerID=8YFLogxK

M3 - Conference article in proceedings

SP - 1115

EP - 1124

BT - Top Fuel 2016

PB - American Nuclear Society ANS

ER -

Arkoma A, Ikonen T. Statistical and sensitivity analysis of failing rods in EPR LBLOCA. In Top Fuel 2016: LWR Fuels with Enhanced Safety and Performance. American Nuclear Society ANS. 2016. p. 1115-1124