Simulation of Fuel Behaviours under LOCA and RIA Using ...€¦ · choose experts, find partners internal simulation of fuel behaviours under loca and ria using fraptran and uncertainty

CHOOSE EXPERTS, FIND PARTNERS

INTERNAL

SIMULATION OF FUEL BEHAVIOURS UNDER

LOCA AND RIA USING FRAPTRAN AND

UNCERTAINTY ANALYSIS WITH DAKOTA

IAEA Technical Meeting on Modelling of Water-Cooled

Fuel Including Design Basis and Severe Accidents,

28 October - 1 November 2013, Chengdu, China

Dr. Jinzhao Zhang Fuel Modelling & Safety Analysis [email protected]

INTERNAL

• Introduction/Objectives

• FRAPCON and FRAPTRAN Fuel Rod Codes

• Independent Validation of FRAPCON and FRAPTRAN

• Uncertainty/Sensitivity Analysis Method

• FRAPTRAN Simulation and Uncertainty Analysis of CIP3-1

• FRAPTRAN Simulation and Uncertainty Analysis of IFA-650.5

• Conclusions and Perspectives

TABLE OF CONTENT

29 Oct 2013 IAEA TM Fuel Modelling in Accidental Conditions, Chengdu, China 2

INTERNAL

• Modelling of fuel behaviours during Loss of Coolant Accident (LOCA) and Reactivity Initiated Accident (RIA) are important:

– High burnup LOCA/RIA tests: Halden, NSRR, CABRI…

• Better understanding of complex phenomena : fuel fragmentation, relocation, dispersal, cladding ballooning, burst, oxidation, and hydriding…

– Revision of LOCA/RIA acceptance criteria: USNRC, IRSN…

• Need improved fuel rod codes and uncertainty analysis methods

• As the Owner’s Engineer of all 7 Belgium plants, Tractebel needs to:

– Qualify the fuel rod codes FRAPCON/FRAPTRAN for simulation of high burnup fuel behaviours during LOCA/RIA conditions;

– Develop a safety evaluation method for margin assessment regarding to the new LOCA/RIA safety criteria;

– Develop a method for independent verification of the safety analyses for demonstrating the compliance with the new LOCA/RIA safety criteria.

INTRODUCTION


INTERNAL

29 Oct 2013 4 IAEA TM Fuel Modelling in Accidental Conditions, Chengdu, China

INTRODUCTION

• FRAPCON & FRAPTRAN fuel rod codes are being used for

– Independent verification of fuel rod design provided by fuel vendors;

– Independent verification of vendors LOCA/RIA safety analysis and reloads fuel safety evaluation;

– Generation of fuel rod input data for neutronics code;

– Feasibility studies for power uprate, burn-up extension and power modulation;

– Operational and licensing support.

• Qualification of the fuel rod codes by independent validation

– Assessing the applicability of both codes to specific applications

– Participation in international benchmarks

• Application of statistical uncertainty and sensitivity analysis method

INTERNAL

• Objectives of this presentation

– Demonstrate the capability of FRAPCON/FRAPTRAN to simulate the LOCA/RIA fuel behaviours of interest, based on OECD fuel rod benchmark cases

• CABRI RIA test CIP3-1, and

• Halden LOCA test IFA-650.5

– Identify relevant input parameters that influence the phenomena of interest

– Evaluate the impact of the fuel rod fabrication data, model and test uncertainties on the results of interest

To be used for further code qualification and method development

INTRODUCTION


INTERNAL


FRAPCON & FRAPTRAN FUEL ROD CODES

• Fuel rod performance and transient analysis codes developed by PNNL

– Used by USNRC and an international user group

– For both steady-state and transient conditions, including LOCA/RIA

• Major models and capability

– Fuel thermal models including thermal conductivity degradation;

– Mechanical models including FRACAS-I rigid pellet and 1D thin wall model, or the optional finite element analysis (FEA) model for the cladding stress-strain analyses;

– Fission gas release (Massih or FRAPFGR), rod internal pressure (RIP) and void volumes models;

– Cladding oxidation and hydrogen content models;

– Simplified thermal hydraulic (TH) model (« Coolant » and « Heat » Options)

INTERNAL


FRAPCON & FRAPTRAN FUEL ROD CODES

• Developmental code assessment

– FRAPCON3.4 code assessment database: 133 fuel rods

– FRAPTRAN1.4 code assessment: 43 integral assessment cases

The parameters of interest

• Fuel temperatures, FGR, cladding corrosion, cladding deformation and burst time

Assessment of bias and sensitivity on major models

INTERNAL


INDEPENDENT VALIDATION

• Selected cases on PWR and BWR fuel rods at high burnups

– Super-ramp cases : PK1-6, PW3 & PW5

– Accidental conditions : NSRR BWR RIA fuel rods FK1-3

– Normal operation : AREVA idealised case

• Focusing on capability for verification of design/safety criteria

– Fission gas release and rod internal pressure

– Fuel temperatures

– Stress and strain states for pellet to clad mechanical interaction (PCMI/PCI)

Comparison with available measurement data, and

Comparison of available models and sensitivity studies

Contribution to IAEA FUMEX-III

INTERNAL



• FRAPTRAN simulation of selected 4 RIA tests with high burnup fuel rods at different coolant temperatures – CABRI sodium loop test CIP0-1

– CABRI water loop test CIP3-1 (blind calculation)

– NSRR capsule tests : VA-1 & VA-3

All measurement data are not yet available

Comparisons code to code and between different users

• Focusing on capability for verification of RIA acceptance criteria – Fuel average enthalpy


– Cladding temperature

Contribution to OECD RIA benchmark

INTERNAL



• FRAPTRAN simulation of 3 Halden LOCA tests (PWR rods)

– IFA-650.3

– IFA-650.4

– IFA-650.5

• Focused on the relevant thermal and mechanical responses of fuel and cladding during LOCA


– fuel fragmentation and relocation

– cladding ballooning and burst (rupture)

– oxidation

• Objectives

– Check the ability of the codes to predict or reproduce the measurements

– to identify the improvements to be made in the codes

Simulation of OECD LOCA benchmark cases

INTERNAL

UNCERTAINTY ANALYSIS METHOD


• 1970’s - 1988: Conservative evaluation model (EM)

• 1988 – 2005: Best estimate code calculations for LOCA accident analysis

– RG1.157 : Best-estimate Calculations Of Emergency Core Cooling System Performance (May 1989).

• Development of LOCA analysis methodologies with deterministic or statistical unvcertainty

analysis: BELOCA (Westinghouse), DRM (Framtome), ASTRUM (Westinghouse)…

• 2005 – present: best estimate code calculations plus uncertainty analysis (BEPU) for LOCA and non-LOCA accident analysis

– RG 1.203: Transient And Accident Analysis Methods (December 2005)

• Application of deterministic or statistical uncertainty analysis method in fuel rod design and RIA analysis

Regulatory requirements and industrial trends

INTERNAL



• The sampled N input uncertainties are propagated through code calculations Y = f(X)

Propagation of input parameter uncertainties

INTERNAL


• Determination of the Wilks’

estimator (top rank) with

minimum number of

calculations (N )

– one-sided tolerance limit

1 − γN = β

– double-sided tolerance limits

1 − γN –N (1-γ) γN-1 = β


Non parametric order statistics

INTERNAL


• DAKOTA = Design Analysis Kit for Optimization and Terascale Applications


Use of DAKOTA tool from Sandia National Lab

INTERNAL

OECD RIA BENCHMARK CASE CIP3-1

• The CIP3-1 blind test case

– RIA test to be performed in the CABRI reactor with pressurised water loop

• At a pressure of 155 bars, an inlet temperature of 280 °C, and an inlet velocity of 4 m/s.

– The CABRI core power during the CIP3-1 test is assumed to be a 10 ms pulse.

– The used rodlet has been refabricated from a high burnup PWR fuel rod

• UO2 rod cladded with ZIRLO at a maximum local burnup close to 75 GWd/t

• The rodlet has a length of about 702 mm, with a plenum of about 2 cm3, and a He filling pressure of 20.5 Bar.

• FRAPCON3.4/FRAPTRAN1.4 simulation of CIP3-1

– FRAPCON calculation of the base irradiation of the fuel rod and the rodlet based on the specifications, with nominal rod data and operating conditions.

– FRAPTRAN transient calculation for the rodlet based on

• specified bet estimate testing conditions

• default model options (in particular, the « Coolant » option for the T-H model)


INTERNAL


• Objective

– to assess the ability of Reactivity Initiated Accidents (RIA) fuel rod codes to reproduce the results from experiments performed in different conditions in NSRR and CABRI test reactors…with a certain degree of adequacy.

• Preliminary uncertainty/sensitivity analysis is performed

– To consider the impact of the uncertainties in fuel rod data, operating conditions and model options on the code simulation results

– To provide certain confidence on the code simulation results

In line with the BEMUSE and PREMIUM project sof OECD/NEA/CSNI /WGAMA and the UAM project of NSC/EGUAM.


INTERNAL


Input uncertainty parameter Mean Standard deviation Lower bound Upper bound Distribution

Thermal conductivity model 0 1 -2 2 Normal

Thermal expansion model 0 1 -2 2 Normal

Fission gas release model 0 1 -2 2 Normal

Fuel swelling model 0 1 -2 2 Normal

Cladding creep model 0 1 -2 2 Normal

Cladding corrosion model 0 1 -2 2 Normal

Cladding hydrogen uptake model 0 1 -2 2 Normal

Multiplicative factor on the temperature history during base irradiation 1 0,00355 0,9929 1,0071 Normal

Multiplicative factor on the power history during base irradiation 1 0,02 0,96 1,04 Normal

Multiplicative factor on the power pulse 0,92976 0,0186 0,89257 0,96695 Normal

Coolant inlet enthalpy (J/kg) during the transient 1232080 5080 1221920 1242240 Normal

Cladding outside diameter (m) 0,0095 0,000019 0,009462 0,009538 Normal

Cladding inside diameter (m) 0,008357 0,000019 0,008319 0,008395 Normal

Dish radius (m) 0,002475 0,0000625 0,00235 0,0026 Normal

Fuel density (%) 95,5 0,75 94 96,5 Normal

Pellet diameter (m) 0,008192 0,000006 0,00818 0,008204 Normal

Cladding roughness (µm) 0,6355 0,31725 0,001 1,27 Normal

Fuel roughness (µm) 1,6005 0,79975 0,001 3,2 Normal

Cold plenum length during base irradiation (m) 0,029531 0,000884 0,0278 0,0301 Normal


Identification and definition of input uncertain parameters

INTERNAL



The Upper/Lower Bound Values (Double-sided tolerance, N=93)

INTERNAL



Impact of the uncertainty distribution and sample numbers

INTERNAL



Sensitivity on the importance of uncertainty parameters

INTERNAL

OECD HALDEN LOCA TEST IFA-650.5

21

• Halden LOCA Experimental set-up

– Fuel rodlet (0.5 m) installed in test rig

– LOCA activated by blowdown valve

– Fuel power controlled by reactor power

– Surrounding rods simulated by electrical heater

– Measurements of interest:

• Cladding temperatures (TCC1 & 3)

Used as boundary condition

• Inner flow channel temperature (TCC3)

• Fuel rod pressure (PF1)

• Cladding elongation (EC2)

• Test of interest

– IFA-650.5: PWR rod at 83 GWd/t, ~72µm oxide, Fill pressure at 70bar, peak cladding temperature at 1100°C


INTERNAL


22

• Test reseults

– Cladding temperatures increase after the end of blowdown until scram (max 1040 °C)

– Rod internal pressure reaches maximum ~171 s after LOCA Balloning

– Burst was detected at 750 °C, ~178 s after the start of blowdown

– Rod internal pressure reduces very slowly after burst small crack?


INTERNAL


23

• Simulation with FRAPCON/FRAPTRAN

– Pre-irradiation with FRAPCON

– Transient simulation by FRAPTRAN with imposed cladding outside temperatures

– Focus on the fuel rod responses of interest: rod internal pressure, ballooning, burst, ECR

• Modelling assumptions

– FRAPCON simulation of the refabricated rodlet at normal operation conditions

– Modification of the FRAPCON restart file used for initialization of FRAPTRAN model

• Refabricated rodlet pressure and gas content

– Use of FRAPTRAN “Heat” option for thermal mechanical calculations only

• Cladding temperature history imposed as the coolant temperature on the base of TCC1 measurements

• High heat transfer coefficients (HTC) imposed identical cladding and coolant temperatures


INTERNAL


24

• Chosen models

– Fuel clad deformation: FRACAS–I Rigid pellet model (default)

– Clad ballooning/burst: BALON2 failure model with empirical stress & strain limits (default)

– Fission gas release: Massih model (default)

– High temperature oxidation: Cathcart-Pawel model (C-P)

• Plenum gas temperature model modification

– The original rod gas plenum temperature model gave unsatisfactory results: too high temperature and rod internal pressure

– Modifications made to allow specification of an external plenum volume held at a defined constant gas temperature

– A arbitrary gas temperature of 127 °C assumed for the whole transient

• Major source of uncertainties as the plenum gas temperature varies with time!

– Possible further improvement to the code by

• Imposing evolution of plenum gas temperature during transient, or

• Improving the gas plenum temperature model to calculate the plenum gas temperature


INTERNAL


25


Imposed cladding temperatures in two zones

INTERNAL


26


Evolution of best-estimate rod internal pressure

Burst Ballooning

INTERNAL


27

• Objectives

– Identify the most important input parameters influencing the result of interest

– Evaluate the impact of the fuel rod data, model and test uncertainties on the uncertainties of the calculation results

• Identification of uncertainty parameters in three categories – Fuel rod fabrication data

– Models

– Operation or test boundary conditions

• Selection of important uncertainty parameters – Some parameters added for confirmation of importance

– Distributions and ranges taken as usually presented in literature

– Material properties not included


Uncertainty analysis

INTERNAL


28


Uncertainty analysis: parameter range and distribution

INTERNAL


29

• Method and assumptions

– Monte-Carlo simple random sampling of all parameters with 93 FRAPCON/FRAPTRAN runs

– Use of first order: Min/Max are the lower and upper bounds (5/95 and 95/95, double-sided)

– Use of Pearson’s and Spearman’s correlation coefficients for sensitivity analysis

• Identification of the most influential parameters on the results of interest


Uncertainty analysis

INTERNAL


DAKOTA

Clad inner diameter

Pellet outer diameter

Resintering

Cladding roughness

Fuel thermal conductivity

FGR model

Fuel thermal expansion

Corrosion model

Steady-State Power

Fabrication

Clad inner diameter

Pellet outer diameter

Resintering

Cladding roughness

Models

Fuel thermal conductivity

FGR model

Fuel thermal expansion

Corrosion model

Boundary conditions

Plenum temperature

Cladding temperature

Steady state power

Transient power

Restart file

FRAPCON input

FRAPTRAN input

DAKOTA

Responses: - ECR, - strain, - pressure,…

UA/SA Results: - Lower/upper bounds - Correlations

Cladding roughness

Transient power

Plenum temperature

Cladding temperature


Uncertainty analysis: DAKOTA UA/SA process

INTERNAL


31


Uncertainty analysis: evolution of rod internal pressure

INTERNAL


32


Uncertainty analysis: evolution of average fuel temperature

INTERNAL


33


Uncertainty analysis: evolution of cladding radial strain

INTERNAL


34


Uncertainty analysis: evolution of Cathcart-Pawel ECR

INTERNAL


35

• Pearson’s linear correlation coefficients

– Designate the linear correlation between one input and one output.

• Absolute values less than 0.25 indicate week correlation.

• Absolute values between 0.25 and 0.75 indicate moderate correlation.

• Absolute values above 0.75 indicate strong correlation.

• Example: Fuel temperatures before burst at node 6 (close to burst position)


Sensitivity analysis

Instant = 180 s, node 6 Av. Fuel T. Center T.

Clad inner diameter 0,88 0,89

Pellet outer diameter -0,71 -0,73

Resintering 0,42 0,43

Cladding roughness 0,10 0,09

Fuel thermal conductivity -0,97 -0,98

Relative power during transient 0,94 0,96

Relative power during base irradiation 0,89 0,91

FGR model 0,54 0,57

Fuel thermal expansion 0,99 0,99

Steady state corrosion model 0,80 0,82

Plenum temperature 0,17 0,18

Cladding temperature 1,00 0,99

INTERNAL


36

• Example: Rod internal pressure/burst time before burst at node 6 (close to burst position)

– RIP/burst time impacted significantly by plenum gas temperature and cladding temperature

– RIP impacted also by fuel thermal expansion model

– Cladding elongation and radial strain impacted by more parameters and models



Instant = 180 s, node 6 Internal P. Elongation R. strain Burst time

Clad inner diameter -0,34 0,93 0,85 -0,32

Pellet outer diameter 0,22 -0,26 0,01 -0,11

Resintering -0,02 -0,01 -0,04 -0,05

Cladding roughness -0,24 -0,23 -0,20 0,16

Fuel thermal conductivity 0,72 0,49 0,59 0,00

Relative power during transient 0,31 0,98 0,75 -0,17

Relative power during base irradiation -0,35 0,89 0,81 0,06

FGR model -0,04 0,14 0,13 -0,07

Fuel thermal expansion -0,91 -0,80 -0,85 0,10

Steady state corrosion model -0,38 1,00 1,00 0,19

Plenum temperature 1,00 1,00 1,00 -0,99

Cladding temperature 1,00 1,00 1,00 -1,00

INTERNAL


37

• Example: Cladding oxidation at node 6 (close to burst position)

– Before transient (ECR650.5): impacted only by cladding diameter, initial power and steady-state corrosion model

– After transient (ECR650.5e): impacted also by cladding transient temperature



Node 6 ECR 650.5 ECR 650.5 e

Clad inner diameter -0,99 -0,81

Pellet outer diameter -0,57 0,00

Resintering 0,05 0,09

Cladding roughness -0,05 0,00

Fuel thermal conductivity 0,06 -0,11

Relative power during transient 0,02 0,08

Relative power during base irradiation 1,00 0,99

FGR model -0,10 -0,25

Fuel thermal expansion -0,05 -0,01

Steady state corrosion model 1,00 1,00

Plenum temperature -0,13 0,29

Cladding temperature 0,10 1,00

INTERNAL


CONCLUSIONS

• FRAPCON & FRAPTRAN predict quite well fuel thermal behaviour during RIA (by comparing with other codes)

Adequate for design/safety criteria verification

• FRAPCON & FRAPTRAN mechanical models need to be improved to predict cladding deformation and PCMI failures during RIA

Further benchmark needed (OECD RIA benchmark phase II)

• The failure of fuel rod is sensitive to the initial conditions and the following parameters/models: – Initial gap thickness;

– Initial cladding spallation;

– Void volume;

– Fission gas release…

Uncertainty and sensitivity analysis needed (OECD RIA benchmark phase II)

INTERNAL

CONCLUSIONS

39

• With the measured cladding temperatures and imposed plenum gas temperature as boundary conditions, FRAPTRAN is able to simulate the Halden LOCA test IFA-650.5, in particular:

– Fuel pellet temperature;

– Rod internal pressure;

– The ballooning and burst.

Further model improvement needed

• The important parameters influencing the calculation results of interests during LOCA are identified:

– Plenum gas temperature;

– Cladding temperature;

– Cladding inner diameter;

– Initial power and steady-state corrosion for oxidation…

Uncertainty analysis needed


INTERNAL

PERSPECTIVES FOR FUMAC

40

• The accurate simulation of physical phenomena during LOCA/RIA is essential for further uncertainty analysis

Improvement of models for FGR, plenum gas temperature, axial gas

transportation, cladding ballooning and burst, fuel relocation and dispersal… ?

• Detailed measurements and/or their uncertainties in LOCA/RIA tests are important for fuel modelling and uncertainty analysis applications

Focus on a few, but well instrumented tests (e.g., Halden LOCA tests)?

• Thermal Hydraulic models needs to be improved to better simulate the test and transient conditions during LOCA/RIA

Coupling with a qualified system or sub-channel T/H code?


Documents

Simulation of Fuel Behaviours under LOCA and RIA Using ...€¦ · choose experts, find partners internal simulation of fuel behaviours under loca and ria using fraptran and uncertainty