USING BEST ESTIMATE SYSTEM CODES IN NUCLEAR SAFETY APPLICATIONS - PowerPoint PPT Presentation

Loading...

PPT – USING BEST ESTIMATE SYSTEM CODES IN NUCLEAR SAFETY APPLICATIONS PowerPoint presentation | free to download - id: 1e0093-ZDc1Z



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

USING BEST ESTIMATE SYSTEM CODES IN NUCLEAR SAFETY APPLICATIONS

Description:

More detailed treatment of neutron transport. Computation Intensive. ... Model base on extended flow regime map and additional correlations for each regime. ... – PowerPoint PPT presentation

Number of Views:70
Avg rating:3.0/5.0
Slides: 43
Provided by: rafael85
Learn more at: http://lrs.web.psi.ch
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: USING BEST ESTIMATE SYSTEM CODES IN NUCLEAR SAFETY APPLICATIONS


1
USING BEST ESTIMATE SYSTEM CODES IN NUCLEAR
SAFETY APPLICATIONS
  • by Rafael Macián-Juan
  • Labor für Reaktorsicherheit und Systemsverhaltung
  • NES Colloquium 23rd February 2006

2
Outline
  • Nuclear Safety
  • Nuclear Systems
  • Accident Scenarios
  • Analysis of Nuclear Systems
  • Best Estimate Codes
  • Uncertainty Analysis
  • The System Code as a Tool
  • Outlook

3
Nuclear Safety
  • THE MAIN ISSUE TO BE ADDRESSED IN NUCLEAR SAFETY
    IS

TO ASSURE THAT THE RADIACTIVE PRODUCTS, AS FAR
AS REASONABLY ACHIEVABLE, REMAIN SAFELY CONFINED
AT ALL TIMES DURING Plant Operation. Accident
Scenarios. Refueling of the Reactor. Storage of
Spent Fuel. Transport of Spent Fuel. Reprocessing
of Spent Fuel.
4
Nuclear Safety
  • The containment of radioactive substances is
    achieved by
  • DEFENSE IN DEPTH
  • Multiple Barriers Isolate them
  • Fuel Structure.
  • Fuel Coating / Cladding.
  • Reactor Vessel.
  • Close Circuit Coolant System.
  • Containment.
  • Plant Siting.
  • Evacuation Plans.

Closed Space in Fuel Element
Fuel (Ceramic o Alloy)
Coolant (Closed Circuit)
1
2
3
4
5
Nuclear Fuel
Exterior
Containment Space
Clad and Fuel Coating (Metal o Ceramic)
Vessel (Steel or Concrete)
Containment (Steel and Concrete)
Fission Gases
Fission Products
5
Nuclear Systems
  • A Nuclear Power Plant consists of a number of
    systems that interact with one another to achieve
    the goal of safe electrical power production.
  • Reactor Core Systems (Power Production).
  • Reactor Coolant Systems (Energy Transport).
  • Instrumentation and Control.
  • Engineering Safeguards (Safety).
  • Additional Systems (Support).
  • Knowledge of their interactions is essential for
    the design and safe operation of the plant.

6
Modeling of Systems Interactions
  • Systems Behavior is determined by the
    interactions amongst the systems components
    (sub-systems).
  • The interactions can be described formally by
  • Mathematical modeling of relationships.
  • Mathematical modeling performance variables.
  • Analytical modeling permits the representation of
    complex systems by computer models
  • Objective modeling based on best-physics and
    measurements.
  • Subjective model inputs must be adequately
    quantified ? challenge.
  • Computerized Systems Analysis is a powerful tool
    to analyze how components and sub-system failures
    can propagate and affect system performance.

7
Analysis of Nuclear Systems
  • Based on Systems Analysis Computer Programs.
  • They are used to analyze
  • Incidents in Plants during Operation.
  • Components and Subsystems Response.
  • Progression of Accidents.
  • Consequences of Accidents.
  • Input from Experimental Facilities.
  • Used to gain valuable
  • Knowledge of the physical behaviour of nuclear
    systems (usually scaled).
  • Data for program assessment and validation
    models and integrated program use.
  • Input from incidents in Nuclear Systems (NPPs).
  • Used to gain valuable
  • Data for program assessment and validation
    integrated program use.
  • Knowledge of the physical behaviour of nuclear
    systems (full scale).

8
Accident Scenarios
  • DESIGN BASIS
  • Included in the design process of nuclear
    systems.
  • Mechanical design of components.
  • Design of Accident Mitigation Systems
  • Redundancy.
  • Capacity.
  • The plant is designed to withstand these
    accidents without significant release of
    radioactivity to the environment.
  • They refer to
  • Nuclear Accident Scenarios
  • Reactivity Induced Accidents (RIAs).
  • Thermal-hydraulic Accident Scenarios
  • Loss of Coolant Accidents (LOCAs).
  • Coupled Accident Scenarios
  • Reactivity Insertion due to moderator density
    changes or low neutron poison concentrations.
  • Stability in BWRs.
  • Non-DESIGN BASIS
  • Accident sequences
  • Not fully considered in the design process.
  • Possible, but highly unlikely.
  • Included in the Regulatory process
  • Strives to be as thorough as possible.
  • Beyond design-basis" accident sequences are
    analyzed to fully understand the capability of a
    design.
  • The plant is NOT designed to withstand these
    accidents without significant release of
    radioactivity to the environment.
  • BUT
  • safety factors included in the design make
    systems resilient to catastrophic damage.

9
Transient Categories for LWR Transients
10
Systems Analysis and Computer Programs
  • CHARACTERISTICS of BE Codes
  • Physical Models
  • Empirical Correlations.
  • Mechanistic Models.
  • First Principles Models.
  • Numerical Models
  • PDE Conservation Equations.
  • Neutronic Description (Diffusion or Transport).
  • Control System Theory.
  • Numerical Methods
  • Finite Differences or volumes
  • Nodal Methods (Neutronics).
  • Implicit, Semi-implicit or Explicit time
    discretization.
  • Iterative solution methods convergence.
  • Component Based Codes.
  • CONSERVATIVE CODES
  • Based on models and methods with a high degree of
    conservatism.
  • Safety margins are usually very large.
  • Eg. LOCA App. K based codes.
  • BEST ESTIMATE (BE) CODES
  • Models and Methods based on the Best Available
    science
  • Physically realistic solutions.
  • Better performance to model complex systems
    interactions.
  • They produce a Best Estimate result.
  • Provide solutions that can be used to optimize
    safety and operation
  • The plants gain margin for operation without
    violating safety limits.

11
Systems Analysis and Computer Programs
vv
vl
Vessel (PWR)
Primary and Secondary System (PWR)
12
Best Estimate Thermal-Hydraulics
  • Modern Best Estimate System Codes Describe the
    Flow Field by
  • Set of Coupled PDEs which represent conservation
    laws for
  • Mass.
  • Energy.
  • Momentum.
  • The System is closed for solution by Closure
    Laws
  • Physical Models for Heat Transfer.
  • Interfacial (vapor-to-liquid).
  • Structures to fluids.
  • Physical Models for Momentum Transfer.
  • Interfacial drag.
  • Wall to fluid drag.
  • Pumps and turbines.
  • Especial Models for important physical phenomena,
    eg.
  • Critical Heat Flux (CHF).
  • Critical Flow.
  • Tracking of interfaces.
  • Thermodynamic Properties of the fluid(s).

Mass
Convective Transport of Mass
Energy
Momentum
13
Best Estimate Neutronics
  • Core Neutronic Behaviour
  • Solution of Neutron balance equation
  • Static Calculation (criticality).
  • Dynamic Behaviour (transient).
  • Neutron balance defined by
  • Leakage.
  • Fissions (POWER).
  • Absorptions.
  • Solution Methods
  • Nodal Methods for Diffusion Approximation.
  • Fast Solution Procedures.
  • Accurate power distributions.
  • Most Used in System Codes.
  • Advanced Transport Theory Methods.
  • More detailed treatment of neutron transport.
  • Computation Intensive.
  • More accurate ?

Time dependent Neutron distribution n(t,r)
Diffusion D(t,r)
Fission Sf (t,r)
Absorption Sa (t,r)
14
Coupled Solutions
  • Coupled Solutions integrate the main descriptions
    of physical processes that determine the behavior
    of a nuclear system.
  • Based on the transfer of information
  • Between main code physical solution procedures.
  • Internal or external coupling.
  • Explicit, semi-implicit or implicit.
  • Time synchronization (time step selection).
  • Convergence and robustness.

15
  • HOW GOOD ARE THE RESULTS OF A BEST ESTIMATE
    SYSTEM CODE ?

16
Assessment and Validation
  • ASSESSMENT
  • comparison of predictions with validated data.
  • VALIDATION
  • determines whether a physical model, numerical
    method or computer program can properly describe
    the phenomena they are designed to simulate.
  • VALIDATION IS BASED ON
  • Theoretical Analysis and Experimental or
    Numerical Assessment.
  • ASSESSMENT IS BASE ON
  • Experimental Data Separate Effect Tests or
    Integral Tests (CSNI, Developmental Matrices)
  • Exact Analytical Solutions.
  • Comparison To Validated Numerical Solutions

ASSESSMENT Compares computation predictions
with validated data. VALIDATION Determines
whether a physical model, numerical method or
computer program can properly describe the
phenomena they are designed to simulate.
17
Separate Effect Tests
  • Designed to assess the quality of the predictions
    of individual physical models.
  • Provide valuable information on
  • Adequacy of the model to describe the physics.
  • Accuracy of the model to provide results close to
    the experimental measurements.
  • Establish the range of relevant physical
    parameters for the application of the models.
  • Advantages
  • Focused on a few physical processes.
  • Measurements can cover large regions of a models
    state-space.
  • Disadvantages
  • Do not address system behaviour.
  • Scaling issues make it difficult to extrapolate
    to full plant.
  • Idealized conditions make it difficult to apply
    conclusions to dynamic plant behaviour.

18
CHF Model Assessment (TRACE)
TRACE Model
RIT Facility
AEE Facility
19
CHF Model Assess. (TRACE)
  • Objective
  • Assess TRACE Biasi and CISE GE CHF Models.
  • Dependence of CHF predictions on pressure, mass
    flux, axial power profile, and inlet sub-cooling.
  • Results
  • The accuracy in the range of pressure, mass flux
    and inlet sub-cooling analyzed is within 20
  • Best predictions for pressures between 50 to 80
    bar, Mass fluxes between 2000 to 3000 kg/m2s.
  • Only the at low pressure (10-50 bar) and low mass
    fluxes (up to 1000 kg/m2s) the agreement with the
    RIT and AEE experimental data is poor.

20
Void Fraction Model Validation (RETRAN-3D)
  • Objective
  • Validation of the boiling and the vapor transport
    models in heated channels.
  • Transient and Steady-state separate effect tests.
  • Results
  • Acceptably good predictions of transient void
    fraction with steady-state models.
  • But
  • Condensation Problems.
  • Underprediction of void fraction overprediction
    of slip ratio.

(Macian R., Coddington P., Stangroom P., Trends
in Numerical and Physical Modeling for Industrial
Multiphase Flows, Cargèse, France SEPTEMBER
27-29th 2000)
21
Pressure Wave Propagation (TRACE, RELAP5)
  • Objective
  • Assessment of the capability of systems codes
    (TRACE and RELAP5) to simulate pressure wave
    propagation.
  • Facility
  • UMSICHT in Germany (Fraunhofer Institut)
  • Piping system for Waterhammer studies.
  • Test
  • Benchmark 2 of NURETH-11.
  • Rapid Closing of Valve, followed by a waterhammer
    upstream and pressure wave propagation and
    reflection downstream,
  • Results
  • Good timing of first and second pressure peaks.
  • Low first pressure peak.
  • Not able to model pressure damping (lack of FSI)
  • Modifications of Condensation and Flashing heat
    transfer resulted in much improved solutions.

22
Pressure Wave Propagation (TRACE, RELAP5)
Change in Condensation
Base Case
(Barten, W., Jasiulevicius, A, Zerkak, O. and
Macian,R., 2006)
23
Model development (RETRAN-3D)
  • Objective
  • Development of a npn-equilibrium model to improve
    the prediction of inter-phase heat and mass
    transfer in RETRAN-3D.
  • Results
  • Model base on extended flow regime map and
    additional correlations for each regime.
  • Improved prediction of condensation and of fast
    depressurization transients.

(Macian R., Cebull P., Coddington P., Paulsen M.,
Nuclear Technology, 128(2), 1999)
24
Integral Tests
  • Designed to analyze the behavior of systems under
    different accident scenarios.
  • Provide very valuable data for the assessment of
    system codes.
  • Advantages
  • Address complex system behaviour.
  • Offer insights on integrated system response to
    accident scenarios.
  • Useful to investigate accident mitigation
    procedures.
  • Disadvantages
  • Scaling issues make it difficult to extrapolate
    results to full plant behaviour.
  • Difficult to isolate physical processes to assess
    the influence of different codes physical models.

PKL Facility
25
SBLOCA in ROSA-IV (RETRAN-3D)
26
PKL Shutdown Studies (TRACE)
  • Objectives
  • Assess TRACEs capability to address phenomena in
    shutdown and SBLOCA scenarios
  • Test Facility
  • PKL test facility represents the entire primary
    system and most of the secondary system of a
    4-loop 1300 MW PWR plant.
  • The PKL III experiments are a part of SETH
    project, supported by NEA/CSNI (14 countries).
  • Tests Analyzed
  • Test E3.1 Failure of RHRS with secondary
    pressure control, Accumulator injection, and
    restoration of RHRS.
  • Test F1.2 Reduction of primary coolant inventory
    until core uncovery, then increase primary
    coolant inventory.
  • Results
  • Test E3.1
  • Correct Prediction of overall heat transfer and
    main TH phenomena
  • But Some Problems Where Found
  • Possible overprediction of condensation heat
    transfer.
  • Problems during injection phase (Phase III).
  • Flow in U-tubes of Steam Generator (Drag).
  • Test F1.2
  • Good agreement with the data from the PKL test
    F1.2.
  • Prediction of the key values for the main
    phenomena observed during the various phases of
    the test.

Test E3.1
Test F1.2
(Jasiulevicius, A, Zerkak, O. and Macian, R.
2005)
27
LOFT 2-5 (TRACE)
  • LOFT L2-5 TRANSIENT
  • LOFT Facility was a 50 MWth PWR with
  • Vessel and Nuclear Core with 1300 fuel rods.
  • Intact Loop with SG, Pressurizer and two parallel
    pumps
  • Broken Loop with simulated pump and SG
  • ECCS Accumulator, LPIS and HPIS
  • Double ended break in the Cold Leg
  • All typical phases of a LOCA were modelled.
  • ECCS accumulator injection controlled the first
    clad temperature peak.
  • HPIS and LPIS controlled second peak and
    reflooding of the core quenching the fuel rods.
  • TRACE Reproduced all main phases well, but
  • PCT were lower than experimental values.
  • Quench times were delayed.

(Macian, R., 2005)
28
Assessment with Plant Data
  • Data from operation and incidents in NPPs are
    very valuable for Best Estimate System Code
    Assessment
  • They are not affected by scaling issues (real
    system).
  • They are a reflection of the response and
    interaction of systems under real conditions.
  • Collected by Plants and transmitted to Regulatory
    Bodies.
  • Sometimes released for international Benchmarks.

29
KKG Pump Trip without SCRAM (RETRAN-3D)
  • Objective
  • Assess RETRAN-3D against plant data from a PWR
    Plant
  • Transient
  • Pump Trip Event with the plant was in normal
    operation at full power.
  • No SCRAM, only partial rod insertion to reduce
    power.
  • Flow to the turbine is regulated to maintain a
    turbine power about 25 of nominal.
  • The bypass control system keeps the secondary
    side pressure constant.
  • Results
  • RETRAN-3D model is capable of reproducing the
    thermal-hydraulic behavior of the plant during
    the Pump Trip transient in an acceptable way.
  • The Point Kinetics model simulates core power
    behavior
  • as modified by the movement of control rod banks
    and
  • by changes in the thermal-hydraulic state of the
    primary and secondary sides.
  • Further modifications and refinements of the
    model appear to be necessary
  • Better models of the plant controllers
    (especially those controlling the steam flow and
    by-pass valves),
  • Better description of the steam lines and the
    pressure drops along them,
  • Development of a 3-D core model for more accurate
    modeling of the individual control rod banks
    movements

(Macian, R., 2001)
30
KKL (TRACE)
  • PRESSURE AND WATER LEVEL IN THE FW TANK
  • After the FW pump trip, the control-system
    reduces turbine load demand to 69 of rated
    value.
  • Overall pressure decrease trend well predicted
    for case at 100 load conditions.
  • Discrepancy in the pressure slopes for case 89
    (between 60s and 140s)
  • Relatively poor prediction of the condensation
    when high pressure steam from the HP-turbine is
    injected in the FW tank.

O. Zerkak and P. Coddington (2004)
31
BWR TURBINE TRIP (RETRAN-3D)
  • Objective
  • Assess and use RETRAN-3D for applications with
    dynamic coupling between 3D-kinetics and
    plant-system thermal-hydraulic descrption.
  • Coupled analysis using a 1 to 1 core model.
  • Plant
  • Peach Bottom 2 BWR-4.
  • Rated Power 3300 MWth.
  • Transient
  • Turbine Trip with closure of MSV at 60 power.
  • Pressure wave propagating down the core collapsed
    vapor.
  • Void fraction reactivity dominant.
  • Results
  • Power peak well predicted timing and maximum
    (Phase 3).
  • Reactivity effects well modeled by 3D-Neutron
    kinetics (Phase 3).

(Barten W., Coddington, P. and Ferroukhi, H.,
Ann. Nuc. Eng., 33, 2006)
SCRAM
32
Application to Nuclear Plant Analysis
  • Best Estimate System Codes are currently applied
    to the analysis of postulated accident scenarios
    of Nuclear Systems.
  • The trust and interpretation of the results is
    supported by
  • The description of the code provided by the code
    developers
  • Methods.
  • Models.
  • Applications.
  • The systematic asesment and validation during the
    development and testing phased of the code
    (assessment matrices)
  • The experience of the users (user effect).
  • The experience gained in assessment and
    validation studies.
  • The application of an uncertainty and sensitivity
    analysis methodology.

33
Thermal Expansion during Shutdown (RETRAN-3D)
  • Objective
  • Analysis of thermal expansion of the primary
    inventory in a PWR during shutdown with RHRS
    failure.
  • Results
  • Expansion of primary side inventory increases
    pressure after przer fills up.
  • Operator has at least 2.5 h to initiate
    mitigating actions.

(Macian R., Nechvatal L., Proceedings of ICONE-9,
Nice, France , April 2001)
34
PWR MSLB (RETRAN-3D)
  • Objective
  • Assess and use RETRAN-3D for applications
    requiring a dynamic coupling between 3D-kinetics
    and plant thermal-hydraulics systems analyses.
  • Coupled analysis using a 1 to 1 core model.
  • Results
  • Verification of the ability of RETRAN-3D to
    perform a fully coupled PWR MSLB accident
    analysis.
  • Return-to-power is very sensitive to the mixing
    assumption (conservatism of the no mixing
    assumption).
  • Perspectives
  • The 3D power distribution should permit accurate
    local DNBR evaluations (Departure from Nucleate
    Boiling Ratio).
  • Improve the modeling of the mixing phenomena in
    the lower and upper plenums (division of the
    plenums in more radial nodes).

(Zerkak, O., 2004)
35
Uncertainty Analysis
  • Uncertainty Analysis is necessary if useful
    conclusions are to be obtained from BE
    Calculations.
  • Approximations are made at every step of a BE
    Analysis.
  • Quantification of the uncertainty in code results
    helps to provide confidence in the interpretation
    and use of the results.
  • For instance, USNRC rules establish a
  • Conservative Analysis (eg Appendix K), or
  • 95 probability statement that the licensing
    limits are not exceeded.

36
Uncertainty in Models
  • Objective methodology
  • Quantification of uncertainty in best estimate
    code physical models.
  • information provided by code model assessment
    with separate-effect test data.
  • Uncertainty Quantification
  • based on a new non-parametric estimator
  • quantifies the uncertainty by means of an
    estimated pdf of the statistical distribution of
    the models accuracy.
  • Uncertainty in State Space
  • Base on a multidimensional clustering technique
  • It assigns uncertainty to the model while
    accounting for the particular region of the
    assessment state-space in which the model is
    being applied.
  • Result
  • More detailed and accurate representation of
    model uncertainty. significantly reducing the
    corresponding uncertainties in code simulation
    results

Application to RETRAN-3D Drift Flux Model
(Vinai, P., Macian R. and Chawla, R. Proc.
ICAPP06, 2006)
37
Uncertainty Propagation
  • BASIS (McKay et al. 1988 GRS 1994)
  • Uncertainties in code inputs are treated as
    Stochastic Variables
  • Deterministic Code transforms Stochastic INPUT in
    Stochastic OUTPUT
  • Uncertainty in INPUT is PROPAGATED to OUTPUT
  • Statistical Methods extract uncertainty
    information from OUTPUT

38
Code Uncertainty Measures
  • BASIS
  • Uncertainty in Output Variables based on
  • Non-Parametric Statistical Methods
  • Produce TOLERANCE INTERVALS with
  • PROBABILITY CONTENT b
  • LEVEL OF CONFIDENCE g
  • Determination of SAMPLE SIZE N based on b and g

Two Side Tolerance Intervals r m 2 b 95 ,
g 95 N 93
39
Application Results (BEMUSE TRACE)
  • BEMUSE Objectives
  • Application of Uncertainty Propagation to
    Integral Test (LOFT L2-5).
  • Comparison of Methodologies.
  • Draw Conclusions and Propose Recommendations.
  • Results
  • Uncertainty propagation methodology developed for
    TRACE.
  • Application showed good comparison with other
    participants.
  • Analysis of problems in the code modeling of
    physical processes should not be replaced by
    uncertainty.

40
Putting it all Together
  • Best Estimate Systems Analysis is based on the
    use of computational tools with the adequate
  • Physical modeling and
  • Numerical solution methods.
  • to simulate the physical processes expected
    during the analyzed scenario.
  • A Best Estimate Methodology is a combination of
  • Computer codes with more realistic physics.
  • Validation and assessment procedures.
  • Evaluation of uncertainties.
  • Estimation of sensitivities.
  • Identification of assumptions and modeling
    shortcomings.
  • Conservative initial and boundary conditions are
    commonly used in combination with BE computer
    codes in most applications.

41
The System Code as a Tool
  • Deterministic Part in Dynamic PSA
  • Provides realistic simulation of System Behaviour
    acting on externally and internally generated
    events.
  • Dynamic branching of events based on Control
    Modules interacting with Schedulers.
  • Provider of Boundary Conditions to CFD
    calculations
  • Presents CHF solvers with system boundary
    conditions.
  • Coupling to CFD solvers gives local physical
    detail to System Codes.
  • Provider of Pressure, velocity and Temperature
    Fields for Fluid Structure Interaction analysis
  • Real time coupling to mechanical codes can
    improve the solutions of both kinds of analyses
    carried out separately.
  • Nuclear Plant Simulation Engine for Nuclear
    Simulators
  • Established application with simplified
    versions.
  • More powerful computers and faster, more robust
    numerical methods may permit use of full code
    capabilities.

42
Future Outlook
  • Best Estimate System Codes will continue to be
    used for Nuclear Systems analysis in future
  • Integrated solutions with fast running and robust
    solution methods.
  • Relatively easy to incorporate the latest
    advances in physical modeling.
  • Modular philosophy and GUI as running-environment
    facilitate their use in production environments.
  • Easy coupling to a variety of codes.
  • Development of Uncertainty methodologies to
    support Best Estimate approach.
  • Possibility to adapt them to study new reactor
    designs.
  • CFD will complement the use of System Codes in
    Future
  • Provide detailed local descriptions of physics.
  • Provide reference solutions and act as a research
    tool for first-principles model development.
  • Flexibility of application will necessitate
    advances in computer power and easier user
    interfaces for input and output processing.
  • Need to develop robust and accurate solutions for
    two-phase flow.
  • Need to develop and apply uncertainty
    methodologies and ambitious assessment programmes.
About PowerShow.com