Uncertainty of the Level 2 PSA for NPP Paks Gabor LAJTHA , Attila - PowerPoint PPT Presentation

Uncertainty of the Level 2 PSA for NPP Paks Gabor LAJTHA , Attila BAREITH, El ő d HOLLÓ, Zoltán KARSA, Péter SIKLÓSSY, Zsolt TÉCHY VEIKI INSTITUTE FOR ELECTRIC POWER RESEARCH Institute for Electric Power Research Co Budapest lajtha@aed.veiki.hu

Outline • Introduction • Uncertainties propagated from level 1 to level 2 PSA • Uncertainties considered in CET • Melt progression arrested (ECC restoration) • Hydrogen burn, early containment failure • Late containment failure • Propagation of uncertainties to containment failure states • Summary Workshop on Evaluation of Uncertainties in Relation to Severe Accidents & Institute for Electric Power Research Co Level 2 Probabilistic Safety Analysis 2 Budapest Aiix-en-Provence, 7-9 November 2005 lajtha@aed.veiki.hu

Introduction • Level 1 PSA for internal initiators and internal hazards – Nominal power and shutdown state • Over 500 core damage sequences considered • Level 2 PSA is based on Level 1 – 17 PDS, 13 release categories • Uncertainty analysis (aleatoric and epistemic uncertainties) Workshop on Evaluation of Uncertainties in Relation to Severe Accidents & Institute for Electric Power Research Co Level 2 Probabilistic Safety Analysis 3 Budapest Aiix-en-Provence, 7-9 November 2005 lajtha@aed.veiki.hu

Level 1 – Level 2 Interface: Uncertainty Analysis • Aleatory uncertainties were propagated numerically from level 1 PSA results to PDS frequencies • Basic event level uncertainty parameters were taken from level 1 model • Additional estimations were made for component failures and human actions not included in level 1 analysis • Quantification was performed on PDS level minimal cut sets • Special purpose computer programme was developed due to complexity of model and limitations of PSA software applied • Monte Carlo simulation was applied for quantification Workshop on Evaluation of Uncertainties in Relation to Severe Accidents & Institute for Electric Power Research Co Level 2 Probabilistic Safety Analysis 4 Budapest Aiix-en-Provence, 7-9 November 2005 lajtha@aed.veiki.hu

Level 1 – Level 2 Interface: Uncertainty Analysis cont’d 1e-4 95%, 5% 75%, Median, 25% 1e-5 1e-6 1e-7 1e-8 1e-9 PDS_05C PDS_05J0 PDS_05JA PDS_00B PDS_00E PDS_02A PDS_12B PDS_05F PDS_03B PDS_02B PDS_08B PDS_13C PDS_05B PDS_11B PDS_09F PDS_13F PDS_17F Workshop on Evaluation of Uncertainties in Relation to Severe Accidents & Institute for Electric Power Research Co Level 2 Probabilistic Safety Analysis 5 Budapest Aiix-en-Provence, 7-9 November 2005 lajtha@aed.veiki.hu

Level 1 – Level 2 Interface: Sensitivity Analysis • Focus on operator action for primary depressurisation upon severe accident signal – New EOP action not considered previously in level 1 PSA – Interest in examining changes in profile of dominant plant damage states as a function of this action • Importance and sensitivity measures were calculated for the given human failure event – Re-generation of PDS level cut sets with modified assumptions on failure probability – Calculation of most common measures of change • Results show that no significant changes can be expected upon moderate changes in human error probability except for one high pressure PDS Workshop on Evaluation of Uncertainties in Relation to Severe Accidents & Institute for Electric Power Research Co Level 2 Probabilistic Safety Analysis 6 Budapest Aiix-en-Provence, 7-9 November 2005 lajtha@aed.veiki.hu

Uncertainties in the Containment Event Tree • Severe accident simulations (MAAP4/VVER code) for each PDS by sampling important process parameters as random variables – 40 MAAP parameters – 10 parameters for hydrogen ignition and containment fragility – Latin hypercube sampling – 200 simulations for each PDS/branch in CET • Generation of uncertainty distributions for CET headings – Use of results from multiple severe accident analyses – Considerations of human failure probabilities, structural and equipment failures • Propagation of uncertainties from plant damage states to containment states and release/consequence categories Workshop on Evaluation of Uncertainties in Relation to Severe Accidents & Institute for Electric Power Research Co Level 2 Probabilistic Safety Analysis 7 Budapest Aiix-en-Provence, 7-9 November 2005 lajtha@aed.veiki.hu

Containment Event Tree Workshop on Evaluation of Uncertainties in Relation to Severe Accidents & Institute for Electric Power Research Co Level 2 Probabilistic Safety Analysis 8 Budapest Aiix-en-Provence, 7-9 November 2005 lajtha@aed.veiki.hu

Example 1 :Melt progression arrested and spray system recovery • Uncertainty analysis – Variability in available time (i.e. time window) for ECCS and spray system recovery actions was considered using the results of MAAP4/VVER calculations – Probability of recovery was calculated from the time window values of the sampled MAAP analyses. – Variability in the context of recovery actions (performance influencing factors other than time) was not assumed in quantitative uncertainty analysis. • Sensitivity analysis – Studying sensitivity of overall results to likelihood of recovery (by numerical analysis) – Basemat melthrough is largely affected but not shown in release categories Workshop on Evaluation of Uncertainties in Relation to Severe Accidents & Institute for Electric Power Research Co Level 2 Probabilistic Safety Analysis 9 Budapest Aiix-en-Provence, 7-9 November 2005 lajtha@aed.veiki.hu

Example 1: Melt progression arrested and spray system recovery cont’d PDS_05C Estimate Core melt starting time (200 calculated sequence) * type of failures (recoverable or not) 95 100 * time for recovery Number in groups 80 T T − 54 − 60 = + ⋅ + ⋅ 3 , 17 40 2 P A B e C e 22 13 − 9 20 5 non re cov ery 2 0 0 0 0 22063 23753 25443 27133 28823 30513 32203 33893 35583 37272 38962 s 1 2 3 4 5 6 7 8 9 10 time groups Uncertainty in ECCS Recovery •1 •Average •1.25E-02 •0.9 •Median •1.30E-02 PDS_05C Relocation time of core •5% •5.76E-03 •0.8 •95% •1.83E-02 (200 calculated sequence) •0.7 •0.6 91 100 Number in groups •0.5 80 •0.4 60 36 32 40 •0.3 12 11 20 8 5 2 1 2 •0.2 0 •0.1 22905 25696 28487 31279 34070 36861 39652 42443 45235 48026 50817 s 1 2 3 4 5 6 7 8 9 10 •0 time groups •0 •0.005 •0.01 •0.015 •0.02 •0.025 Workshop on Evaluation of Uncertainties in Relation to Severe Accidents & Institute for Electric Power Research Co Level 2 Probabilistic Safety Analysis 10 Budapest Aiix-en-Provence, 7-9 November 2005 lajtha@aed.veiki.hu

Example 2: Hydrogen burn, early containment failure • A complicated process appears in a simplified manner in the CET. For the quantification of burn probabilities and in order to evaluate consequences the DET concept is introduced. • Hydrogen mole fraction – different hydrogen quantities are produced in each sample (MAAP4/VVER code calculation) • Ignition – probability of ignition depends on the existence of igniting sources (spontaneous ignition, recombiner) and also on the hydrogen concentration • Combustion mechanism – three combustion mechanisms are distinguished (burn, accelerated flames and DDT) for the determination of containment pressure load the H2AICC is used with Modified Adiabatic Isochoric Complete Combustion (AICC) model • Containment failure - Joint treatment of containment loads and fragility curves Workshop on Evaluation of Uncertainties in Relation to Severe Accidents & Institute for Electric Power Research Co Level 2 Probabilistic Safety Analysis 11 Budapest Aiix-en-Provence, 7-9 November 2005 lajtha@aed.veiki.hu

Example 2: Hydrogen burn, early containment • Hydrogen Mass (kg) failure cont’d • PDS_05 C (no ECC recovery) • Uncertainty calculation • Hydrogen Production in Vessel Phase PDS_05C Generated H2 Mass at Vessel Failure (200 calculated sequence) •600 60 Number in groups •500 •H2 total in vessel production 48 50 •H2 production until lower grid failure 39 37 40 32 30 22 •400 20 11 5 10 2 3 1 0 •300 1 21 1 59 1 97 235 273 31 2 350 388 426 464 502 kg 1 2 3 4 5 6 7 8 9 10 H 2 mass groups •200 •100 HYDROGE N LOAD 40 •0 35 •0 •20 •40 •60 •80 •100 •120 •140 •160 •180 •200 • No of Calculation (ordered according to H2 in vessel ) 30 Number in groups 25 Hydrogen production - MAAP4/VVER 20 Hydrogen concentration - MAAP4/VVER 15 Containment load - H2AICC 10 5 Ignition probability - 4 variables (LHS) 0 1 2 3 4 5 6 7 8 9 10 Num ber in groups 4 13 24 24 34 32 35 21 8 5 P ressure (bar) 1.8 2.2 2.6 3 3.5 3.9 4.3 4.8 5.2 5.5 Workshop on Evaluation of Uncertainties in Relation to Severe Accidents & Institute for Electric Power Research Co Level 2 Probabilistic Safety Analysis 12 Budapest Aiix-en-Provence, 7-9 November 2005 lajtha@aed.veiki.hu