Melcor Code Assessment

This co-operative research project between the US Nuclear Regulatory Commission (NRC), the Swiss Federal Nuclear Safety Inspectorate (HSK), and the Swiss Federal Institute of Technology Zurich – Laboratory for Safety Analysis (ETHZ-LSA) deals with the assessment and validation of the NRC severe accident code MELCOR. The goal is to assist the NRC (US Nuclear Regulatory Commission) in performing the assessments of its own code, the findings will be made available to the international community and be further used for the MELCOR code consolidation effort, which is the long-term goal of the NRC. The project will last for four years, from July-1-2001 to June-30-2005. An overview of the work accomplished is given below. It is noted here that the reports are preliminary and the conclusions do not necessary reflect NRC's view.

1. Review of core package In MELCOR 1.8.5, - code consolidation issue - Made a review of the modeling of severe core degradation phenomena in MELCOR1.8.5 against SR5 Mod 3.3 SCDAP/RELAP5 Mod.3.3.

2. MELCOR Probabilistic Risk Assessment (PRA) associated with improper dry cask loading conditions, such as higher decay heat, less He-filled, filled with wrong gas, and insufficient drying the cask, the effect of blocked vents and external fire issues. The time dependent temperature and pressure behaviors of the cask were evaluated. For the cases of loading the cask with higher decay spent nuclear fuel of 0.5 and 1.0 year cooling time, the steady state temperatures experienced by the cask is substantially higher. Hence, the risk of cask failure is considerable. The results will be used by NRC-PRA group to predict failure probability of dry storage casks under improper loading conditions and subsequent severe accident conditions [2.1].

3. Simulation of phase 2 of the Three Mile Island Unit 2 (TMI-2) accident scenario using MELCOR and SCDAP/RELAP5 - Investigated the in-vessel core damage progression of the accident, using the MELCOR TMI-2 input deck, which is limited only to phase 2 of the TMI-2 accident [3.1]. The report presents some simulation results of this phase 2 of the TMI-2 accident using the NRC-severe accident computer codes MELCOR 1.8.5-RE and SCDAP/RELAP5-Mod-3.3. These results will be compared with the available indirect TMI-2 data. Phase 2 of the TMI-2 accident extends from 100 minutes to 174 minutes after the loss of main feedwater.

4. Extensive simulation of modified phase 2 of the TMI-2 accident using SCDAP/RELAP5 and MELCOR for code consolidation assessment - Considered modified phase 2 of the TMI-2 base case accident, e.g., the makeup/HPI flow rate of the base case of TMI-2 accident during phase 2 has been changed to include two test cases: 1 kg/s, and 2.5 kg/s. The simulations using makeup flow rates of 1 kg/s and 2.5 kg/s are calculated with MELCOR 1.8.5-RD4, MELCOR 1.8.5-RE, SR5 mod-3.3kn, SR5 mod-3.3kz, and SR5 mod3.3bf. The base case simulation is also performed [4.1], [4.2]. The results of these simulations were compared.

5. Lower Head Creep Rupture Failure Analysis Associated with Alternative Accident Sequences of the Three Mile Island Unit 2 - The objective of this analysis is to assess MELCOR 1.8.5-RG against SCDAP/RELAP5 MOD 3.3kz (SR5m33kz), and SCDAP/RELAP5 MOD 3.3bf (SR5m33bf). This lower head creep rupture analysis considers: (1) the Three Mile Island Unit 2 (TMI-2) alternative accident sequence-1, and (2) the TMI-2 alternative accident sequence-2. SCDAP/RELAP5 (SR5) model of the TMI-2 alternative accident sequence-1 includes the continuation of the base case of the TMI-2 accident with the reactor coolant pumps (RCP) tripped, and the High Pressure Injection System (HPIS) throttled after approximately 6000 s accident time. SCDAP/RELAP5 model of the TMI-2 alternative accident sequence-2 is derived from the TMI-2 base case accident by tripping the RCP after 6000 s, and the HPIS is reactivated after 12,012 s. MELCOR model of the TMI-2 alternative accident sequence-1 is based on MELCOR TMI-2 phase-2 model by tripping the RCP and throttling back the makeup flows to zero from 6000 s. In MELCOR model of the TMI-2 alternative accident sequence-2, the RCP are tripped from 6000 s and the constant makeup flow rate of 3.75 kg/s is activated from 6000 s. This makeup flow rate includes pump seal flow rate, but excludes the HPIS flow rate. The simulation is run until the lower head wall ruptures. In addition, the lower head penetration failure is also calculated with MELCOR for both TMI-2 alternative accident sequences. Lower head temperature contours calculated with SR5 are visualized and animated with open source visualization freeware 'OpenDX'. Significant findings of the analysis include: (1) the TMI-2 lower head wall fails by creep rupture with either deactivations or activations of the HPIS; (2) for the TMI-2 alternative accident sequence-1, the time to creep rupture calculated with MELCOR 1.8.5-RG, SR5m33kz, and SR5m33bf agrees reasonably; (3) for the TMI-2 alternative accident sequence-1, the calculation with MELCOR predicts that the lower head wall failure occurred earlier than penetration failure, while MELCOR predicts the opposite for the TMI-2 alternative accident sequence-2; (4) for the TMI-2 alternative accident sequence-2, the calculation with MELCOR shows that when the lower head wall fails the temperature is 1810.9 K, which exceeds the melting temperature of 1789 K for carbon steel; (5) for both TMI-2 alternative accident sequences, calculations with both SR5m33kz and SR5m33bf indicate that different lower head wall locations fail rapidly one after another by a delay of a few seconds, while this is not the case for MELCOR. This analysis assumes that the debris is in perfect contact with the lower head wall, thus neglecting the gap cooling between the debris and lower head wall. This is done in SR5 by setting the heat transfer coefficient of the gap between the debris and the lower head vessel wall to a value of 10,000 W/(m2.K). MELCOR does not model this gap heat transfer. This study considers only in-vessel phenomena. The results have been published in ICAPP '04 [5.1], and ICONE12 [5.2]

6. Ongoing and future activities - Currently, ongoing assessments are performed for a number of alternative events to the TMI-2 accident scenarios as described above to study the sensitivity of the time step with regard to the time to creep rupture. Besides, SCDAP/RELAP5 (SR5) stratified molten pool natural circulation models are used to assess SCDAP/RELAP5-MOD3.3 for the ongoing MELCOR code consolidation effort, which is intended to bring MELCOR into the state of parity with SCDAP/RELAP5 MOD3.3. Two issues of concern are the model of the natural convection driven heat transfer through the relocated stratified molten pool into the lower head and the model of gap heat transfer at the interface of debris and lower head that have been implemented in SCDAP/RELAP5 MOD3.3. These assessments include the use of the default steady state well-mixed (oxidic) molten pool model. In addition, the transient oxidic and stratified molten pool models are used for the analyses. Next, the Lower head creep rupture analysis is done for the TMI-2 base case accident. It is noted that all calculations performed so far assume that the lower head wall is in perfect contact with the relocated debris. Thus, these calculations did not include the SR5 gap heat transfer model [6.1], which evaluates the cooling of the lower head associated with a narrow gap [6.2, 6.3] that may form between the debris and the vessel wall. In the near future, this model will used to analyze the lower head creep rupture failure of the TMI-2 alternative accidents and TMI-2 base case accident. Furthermore, the station blackout assessment for the US nuclear power plant 'SURRY' will be studied, which is a pressurized light water reactor. The scenario is concerned with the creep rupture of the surge line, hot leg, and steam generator tube. In addition, the lower head creep rupture will be tackled. Sandia National Laboratories (SNL) is developing a new input decks for the TMI-2 base case accident that uses 2-dimensional lower head model similar to SR5. SNL may make this input deck available to the USNRC, probably next year. Again, the USNRC is interested to assess this new MELCOR TMI-2 deck against SCDAP/RELAP5. It is noted that MELCOR does not have the gap heat transfer model like SCDAP/RELAP5. It is suggested to SNL to include this model in MELCOR.

References:

• [2.1] 'Thermal Loads for Evaluating the Failure Probability of the HI-STORM 100 Cask under Severe Accident Conditions', July 31, 2001. Jason H. Schaperow (NRC/SMSAB/DSARE)
• [3.1] 'TMI-2 Analysis Exercise', Final Report, TMI-2 Joint Task Group, Principle Working Group No. 4, OECD-NEA
• [4.1] RELAP5/MOD3.3Beta, NUREG/CR-5535/Rev 1-Vol IV, 'RELAP5/MOD3.3Beta CODE MANUAL, VOLUME IV, MODELS AND CORRELATIONS', prepared for the Office of Nuclear Regulatory Research, US NRC, Washington DC
• [4.2] R. O. Gaunt et al,'Melcor 1.8.5 Simulation of TMI-2 Phase 2 with an enhanced 2-Dimensional In-Vessel Natural Circulation Model', ICONE10-22321, Proc. of ICONE 10; April 14-18, 2002, Arlington, Virginia, USA
• [5.1] S. L. Chan, 'Assessment of MELCOR 1.8.5 Versus different versions of SCDAP/RELAP5 MOD 3.3 with Lower Head Creep Rupture Analysis of Alternative Accident Sequences of the Three Mile Island Unit 2', ICONE12-49032, Proc. Of ICONE12; April 25-29, 2004, Arlington, Virginia, USA
• [5.2] S. L. Chan, 'Lower Head Creep Rupture Failure Analysis Associated with Alternative Accident Sequences of the Three Mile Island Unit 2', ICAPP04-4004, Proc. Of ICAPP '04 , June 13-17, 2004, Pittsburgh, PA USA
• [6.1] L. J. Siefken, E. W. Coryell, E. A. Harvego, and J. K. Hohorst, 'Modeling of Reactor Core and Vessel Behavior During Severe Accidents', SCDAP/RELAP5/MOD3.3 CODE MANUAL VOLUME 2, NUREG/CR-6150, INEL-96/0422, Vol. 2, Rev. 2, INEL-96/0422.
• [6.2] K. L. Thomsen, 'Percolation cooling of the Three Mile Island Unit 2 Lower Head by Way of Thermal Cracking and Gap Formation', Nuclear Technology Vol. 137, Jan 2002, pp 28-46.
• [6.3] J. Rempe et al., 'Margin-to-Failure Calculations for the TMI-2 Vessel', Nuclear Safety, 1994, Vol. 35, No. 2, pp 313-327.