1. ERMSAR 2012, Cologne March 21 – 23, 2012 CONDUCT AND ANALYTICAL SUPPORT TO AIR INGRESS EXPERIMENT QUENCH-16 J. BIRCHLEY 1, L. FERNANDEZ MOGUEL 1, C. BALS 2, E. BEUZET 3, Z. HOZER 4, J. STUCKERT 5 1) PSI, Villigen (CH) 2) GRS, Garching (DE) 3) EDF, Clamart (FR) 4) AEKI, Budapest (HU) 5) KIT, Karlsruhe (DE)

2. ERMSAR 2012, Cologne March 21 – 23, 2012 Background and objectives Planning analysis Experiment conduct and outcome Conclusions and outlook Outline

3. ERMSAR 2012, Cologne March 21 – 23, 2012 Air ingress issues have come into prominence in recent years – post RPV failure, spent fuel – several recent and ongoing programmes  separate effect and integral tests  model development – QUENCH-16 extends database of air ingress bundle data  performed in frame of EU-supposed LACOMECO  proposed and defined by AEKI, Hungary Objectives: examine reaction with air following mild pre-oxidation in steam and investigate reaction with both O 2 and N 2 – pre-oxidised layer 200 μm maximum – long period of steam starvation Stringent test objectives meant careful planning analyses needed Background, objectives

4. ERMSAR 2012, Cologne March 21 – 23, 2012 QUENCH containment and test sectionQUENCH bundle cross section QUENCH facility

5. ERMSAR 2012, Cologne March 21 – 23, 2012 Planning support performed by – GRS (ATHLET-CD) – EDF (MAAP4.07/EDF) – PSI (SCDAPSim/MOD3.5 and MELCOR 1.8.6) Strategy – define a pre-oxidation transient at T = ca. 1500 K to give pre-oxidised layer 150-200 μm – investigate different power levels and Ar, air flow rates to seek complete O 2 consumption long before nominal limit temperature of 1823 K Converged on a nominal test protocol supported by all simulations – power: 10 kW for 5000 s then 4 kW to end – flow rate (pre-ox):3 g/s steam + 3 g/s Ar – flow rate(air):0.2 g/s air + 1 g/s Ar – reflood:50 g/s water when T,max = 1823 K Planning analyses

6. ERMSAR 2012, Cologne March 21 – 23, 2012 All the codes used are lumped-parameter, system or sub-system level codes for transient analysis of nuclear plant accident sequences – two-phase transient thermal hydraulics – non-condensable species – metallic oxidation and core degradation All the codes have recently been (are being) improved – oxidation in steam and air using established correlations as baseline – modifications to represent breakaway oxidation – Zr + N 2 reaction (ATHLET-CD) Different levels of detail in treatment of thermal-hydraulics and other processes Different levels of detail in noding Key code features

7. ERMSAR 2012, Cologne March 21 – 23, 2012 GRS Fuel rod temperatures showing effect of onset of O 2 oxidation and local starvation Progression of local complete O 2 consumption; starvation period 920 s starvation starvation phase

8. ERMSAR 2012, Cologne March 21 – 23, 2012 EDF Fuel rod temperatures at 250, 650, 950 and 1250 mm for 3 g/s (solid) and 1 g/s (dashed) Ar flow Progression of local complete O 2 consumption (1 g/s Ar); starvation period 1150 s starvation phase

9. ERMSAR 2012, Cologne March 21 – 23, 2012 PSI SCDAPSim Effect of air and Ar flow on oxygen consumption and period of starvation Fuel rod temperatures showing effect of onset of O 2 oxidation and local starvation starvation period1540 s starvation phase

10. ERMSAR 2012, Cologne March 21 – 23, 2012 Comparison - 1 Partner Code PSI SCDAPSIM PSI MELCOR1.8.6 GRS ATHLET-CD EDF MAAP 4.07 Experiment Heat-up Pre-oxidation Power Ar + steam Tmax (5000 s) 0-5000 s 10 kW 3 g/s + 3 g/s 1440 K 0-5000 s 10 kW 3 g/s + 3 g/s 1422 K 0-5000 s 10 kW 3 g/s + 3 g/s 1440 K 0-5000 s 10 kW 3 g/s + 3 g/s 1480 K 0 – 6300 s 10 – 11.5 kW 3 + 3.3 g/s 1489 K Cooldown Power Ar + steam Tmax (6000 s) 5000-6000 s 4.0 kW 3 g/s + 3 g/s 1061 K 5000-6000 s 4.0 kW 3 g/s + 3 g/s 1098 K 5000-6000 s 4.0 kW 3 g/s + 3 g/s 1090 K 5000-6000 s 4.0 kW 3 g/s + 3 g/s 1100 K 6300-7300 s 4.0 kW 3 g/s + 3.3 g/s 1067 K Air phase Power Ar + air 6000 - 9260 s 4.0 kW 1 g/s + 0.2 g/s 6000 - 8350 s 4.0 kW 1 g/s + 0.2 g/s 6000 - 9420 s 4.0 kW 1 g/s + 0.2 g/s 6000 - 8750 s 4.0 kW 1 g/s + 0.2 g/s 7300 - 11135 s 4.0 kW 1 g/s + 0.2 g/s

11. ERMSAR 2012, Cologne March 21 – 23, 2012 Summary of results Partner Code PSI SCDAPSIM PSI MELCOR1.8.6 GRS ATHLET-CD EDF MAAP 4.07 Experiment Quench(temp) Fast refill + 50 g/s water Power 9260 s (1823 K) 4 kW 8350 s (1823 K) 4 kW 9420 s (1823 K) 0 kW 8750 s (1823 K) 0 kW 11335 s (1883 K) 4 kW H2 mass, Max. oxide after preox 13 g 186 µm 15 g 190 µm 11 g 190 µm 19 g 242 µm 14 g 133 µm Duration air phase starvation 3260 s 1540 s 2350 s 1660 s 3420 s 920 s 2750 s 1150 s 4035 s 835 s H2 mass (reflood) 2g16 g1 g 128 g Remarksno influence of 0/4 kW during quench ZrN model would increase starvation time Comparison - 2

12. ERMSAR 2012, Cologne March 21 – 23, 2012 Test conduct showing electric power input and selected temperatures Off-gas mass composition showing O 2, N 2 consumption and H 2, N 2 release QUENCH-16 conduct O2 starvation N2 consumption release of H 2 and N 2 during reflood

13. ERMSAR 2012, Cologne March 21 – 23, 2012 Post-test videoscope inspection (front view) at elevation 550 mm, showing spalling of oxide scale Post-test videoscope inspection (side view) at elevation 790 mm, showing nitride formation and partial spalling Bundle examination - 1 rod #5 shroud

14. ERMSAR 2012, Cologne March 21 – 23, 2012 Bundle cross section at 430 mm: frozen melt relocated from upper elevations Bundle cross section at 830 mm: minor melting of some cladding segments Bundle examination - 2 oxide metallic

15. ERMSAR 2012, Cologne March 21 – 23, 2012 Bundle elevation 350 mm, cladding of rod #5: nitrides between two oxide layers Bundle elevation 550 mm, cladding of rod #9: nitrides between inner dense and outer porous oxide layers Bundle examination - 3 porous ZrO 2, probably containing reoxidised ZrN dense ZrO 2

16. ERMSAR 2012, Cologne March 21 – 23, 2012 QUENCH-16 adds significantly to knowledge on air ingress transient behaviour – complements previous experiments – minor pre-oxidation and long O 2 starvation period maximised influence of N 2 – significant ZrN formation and re-oxidation Coordinated pre-test planning analysis facilitated successful experiment – pre-reflood quite well predicted but models did not capture the strong reflood excursion which included significant oxidation of both solid and molten material – starvation, ZrN formation, or the two together might have been influential as a trigger – code models are not yet able to represent these effects reliably Post-test analyses are underway at several institutes – benchmark on QUENCH-air is being performed within WP5.1/JPA3 – answer “do we need to model effects of starvation and ZrN on oxidation during reflood?” – and we hope will point the way to how to do it Conclusions, outlook

17. ERMSAR 2012, Cologne March 21 – 23, 2012 The LACOMECO programme is performed by KIT with financial support from the HGF Programme NUKLEAR and the European Commission. Technical support is provided by institutes with the European Economic Area The development and validation of the code ATHLET-CD are sponsored by the German Federal Ministry of Economics and Technology (BMWi). PSI acknowledges financial support of ENSI, the Swiss nuclear regulatory organisation Thank you for your attention Acknowledgements