Docstoc

4-4 Severe Accident Codes for VTL2 PSA - DESCRIPTION

Document Sample
4-4 Severe Accident Codes for VTL2 PSA - DESCRIPTION Powered By Docstoc
					                    Severe Accident Codes for
                            L2 PSA
                     (ASTEC Requirements)
                           Gabor Lajtha1, Emmanuel Raimond2, Pal Kostka1,
                                                    1) VEIKI Budapest, Hungary
                                                   2) IRSN Fontenay-aux-Roses


                                            ERMSAR meeting, ,Bulgaria
                                                  September 1, 2008



                                                                                 1
ERMSAR Nessebar, Bulgaria, September 23-24, 2008
                                                   Background



  • The goal of the L2 PSA group in the SARNET project was to
    describe, compare the L2 PSA methods used by the partners,
    and then to identify some possibilities for harmonization



  • One of the most important parts of the L2 PSA methodology is
    the assessment of the accident progression and phenomena, by
    computer codes.




                                                                   2
ERMSAR Nessebar, Bulgaria, September 23-24, 2008
                                                   Background 2

           „to harmonise the methodologies
            applied for assessing risk and improve
            Level 2 PSA tools”

           „ASTEC is aimed to become the
            ultimate code for integration of the
            current research in severe accidents,
            and as the tool of choice for safety and
            risk assessment. It is strongly
            recommended to define a standard of
            acceptability for ASTEC models.
            Statements such as code results are
            “acceptable”,    “robust”,    “physically
            reasonable”, “satisfactory”, etc. are too
            subjective to be meaningful without a
            clear understanding of what degree of
            agreement is considered to represent
            the subjective attribute.” 050614
            SARNET Review Report (Consolidated
            - Final)



                                                                  3
ERMSAR Nessebar, Bulgaria, September 23-24, 2008
                            Tools Used by the SARNET Partners

                     Probabilistic software
                         Risk Spectrum (RELCON AB), IRRAS (Idaho National
                          Engineering &Environmental Lab.
                     CET software
                         EVNTRE (Sandia N.L), SOLOMON (AEA technology),
                          KANT (IRSN), SUPER (Westinghouse)
                     Deterministic tools
                         Integrated tools
                             ASTEC, MAAP, MELCOR
                         Detailed mechanistic codes
                             COCOSYS, CONTAIN, RELAP, ATHLET, CATHARE,
                               VICTORIA, WECHSL, TRAP-MELT…
                             3D codes: CFX, GASFLOW, TONUS, MC3D




                                                                             4
ERMSAR Nessebar, Bulgaria, September 23-24, 2008
                              Questionnaire on integrated codes


     The aim of this questionnaire is to collect the user’s
         need against integrated severe accident codes
         based on their present experience of L2 PSA
         calculations.

     The different needs are then identified and prioritized
         to help the ASTEC developers in their effort to
         improve the code for fitting the different purposes
         of L2 PSA.



                                                                  5
ERMSAR Nessebar, Bulgaria, September 23-24, 2008
                                   Status of ASTEC code in 2004
  •      „ASTEC is a computer code for the calculation of the entire
         sequence of a severe accident in a light water reactor from the
         initiating event up to the release of fission products into the
         environment, covering all important in-vessel and ex-vessel
         phenomena. The ASTEC code is composed with a number of major
         modules. ASTEC code is co-developed by IRSN and GRS in
         France and Germany.”

  •      „For IRSN, the main reason for the choice of ASTEC is a
         strategic reason.”

  •      „For GRS, ASTEC has been considered not yet developed
         enough to be a standard tool in a PSA.”

                                                   Serge Vidal: Tools used for PSA 2
                                                           SARNET 2004


                                                                                       6
ERMSAR Nessebar, Bulgaria, September 23-24, 2008
                                                   Questionnaire


                                    Organization          Type             Code
                                   AREVA            Konvoi         MELCOR (ongoing)
                                   GRS              Konvoi         MELCOR 1.8.4
   PWR                             IRSN             900            ASTEC V0
                                   IRSN             1300           ASTEC V1
                                   SwedPower        Westinghouse   MAAP
                                   AREVA            German         MELCOR (ongoing)
   BWR                             GRS              BL69           MELCOR 1.8.5
                                   SwedPower        ASEA-ATOM      MAAP
                                   UJV,             440/213        MELCOR
   VVER                            VEIKI            440/213        MAAP4/VVER
                                   UJV              1000           MELCOR 1.8.5
   Advanced reactor
                                   AREVA            EPR            MAAP4
   (EPR,ESBWR…)


                                                                                      7
ERMSAR Nessebar, Bulgaria, September 23-24, 2008
                                      IAEA Statements from 1995
  Please comment the following from the IAEA Safety series
  No 50-p-8 1995 Appendix II:
  « the PSA codes, intended for routine application in PSA, are designed to run fast, so that they can
         calculate many sequences (and a number of times for a single sequence, if uncertainty
         analyses are required). In order to achieve these shorter run times, the modelling has to be
         simpler than in mechanistic codes. »
   « By definition, PSA codes are designed so that a Level 2 PSA analyst with a good degree of
         familiarity with general accident phenomena can run them reliably without the same detailed
         knowledge as a specialist using a mechanistic code dealing with a particular phenomenon or
         a phase of severe accident. »

  Generally the participants agree with these statements, but
  •    there is no integral code existing which was developed only to serve for PSA
       needs
  •    it should not be forgotten that for some specific phenomenon (DCH, steam
       explosion…) occurring during a severe accident, simplified modelling is very
       unsatisfactory or very complicated to be used
  •    a level 2 PSA can be performed with a limited number of detailed calculations
       or a large number of more simplified calculations depending on the L2 PSA
       method




                                                                                                         8
ERMSAR Nessebar, Bulgaria, September 23-24, 2008
                                          Modelling of Systems

        Valid for MAAP, MELCOR and ASTEC:
        • Models are available for different reactors
        • The characteristics of these models are diverse, depending
                on the severe accident code
        • Quality of the models depends on the nodalisation
                determined by the use

            The users agreed that the quality of the system models is
              largely independent of the complexity of the model.

          All ASTEC models were not checked for Level 2 PSA use.



                                                                        9
ERMSAR Nessebar, Bulgaria, September 23-24, 2008
                                                   Modelling of Phenomena
  • The severe accident phenomena may depend on the reactor design

  • It can be found that most of the phenomena were modelled by simple
    physical or parametric models in the examined severe accident codes
    (MELCOR, MAAP and ASTEC).

  • The answers pointed out that preparation of an input was straightforward
    for an experienced user.

  • The codes went through a formal validation process which determined
    suggested parameter ranges and models. The calculated results seemed to
    be physically acceptable. There were some exceptions :
     – hydrogen production, core melt and relocation
     – hydrogen distribution, burn
     – HPME, DCH
     – MCCI
     – in-vessel and ex-vessel steam explosion.

                                                                               10
ERMSAR Nessebar, Bulgaria, September 23-24, 2008
                                               Prority of the phenomena




                                                                          11
ERMSAR Nessebar, Bulgaria, September 23-24, 2008
                                           General code performance

    • Most of the L2 PSA developers intend to make a full, “non-stop sequence”
      calculations, for each severe accident sequence.
    • L2 PSA needs sensitivity and uncertainty analyses. For this reason, the
      separate physical processes should be calculated.




                                                                             12
ERMSAR Nessebar, Bulgaria, September 23-24, 2008
                                   General code performance




                                                              13
ERMSAR Nessebar, Bulgaria, September 23-24, 2008
                                               CONCLUSION
     Since 2004 substantial progress has been achieved to improve ASTEC both in
     modelling of important phenomena and in particular with view to calculation time
     for coupled calculations and code stability with the latter being one of the major
     drawbacks at that time.

     Overview of the answers of the level 2 PSA users, it is turned out that the ASTEC
     code characteristics are similar or better than the other integral codes (MAAP,
     MELCOR) used for level 2 PSA supporting analyses. There are some remaining
     need for model modifications:
             • modelling the physical process after vessel failure for EPR and
             • the possibility to calculate the severe accident in case of shutdown state
             • the possibility to calculate BWR sequences
     The code general behaviour, as the user can choose which modules to activate or
     not is perfectly suited for level 2 PSA and its uncertainty or sensitivity studies.
     The code performances are generally good, better or equivalent of the other codes.
     It was found that :
             • code input description
             • code physical description
                           should be ameliorated.

                                                                                            14
ERMSAR Nessebar, Bulgaria, September 23-24, 2008
                                            Dream of the User



                                                                ASTEC 2.0

                           L2 PSA


     L2 PSA
    RESULTS

                               Thank you, your attention


                                                                        15
ERMSAR Nessebar, Bulgaria, September 23-24, 2008