1. Study of transient heat flux deposition for various ELM types in NSTX J-W. Ahn, R. Maingi, and T.K. Gray (ORNL) In collaboration with K.F. Gan (ASIPP) Work done through NSTX / ASIPP collaboration NSTX Monday Physics Meeting August 27, 2012 NSTX Supported by Columbia U CompX General Atomics FIU INL Johns Hopkins U LANL LLNL Lodestar MIT Nova Photonics New York U ORNL PPPL Princeton U Purdue U SNL Think Tank, Inc. UC Davis UC Irvine UCLA UCSD U Colorado U Illinois U Maryland U Rochester U Washington U Wisconsin Culham Sci Ctr U St. Andrews York U Chubu U Fukui U Hiroshima U Hyogo U Kyoto U Kyushu U Kyushu Tokai U NIFS Niigata U U Tokyo JAEA Hebrew U Ioffe Inst RRC Kurchatov Inst TRINITI NFRI KAIST POSTECH ASIPP ENEA, Frascati CEA, Cadarache IPP, Jülich IPP, Garching ASCR, Czech Rep

2. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 2 Motivation Engineering limit of steady state peak heat flux is generally ~10 MW/m 2 Transient heat flux from large ELMs is a great threat to the protection of divertor plates for ITER and other future devices  It is possible that only small ELMs may be allowed NSTX-U will be in higher I p and heating power with longer pulse length, therefore ELM heat flux problem will be much more serious  Need to understand heat deposition for various ELM types in NSTX  Possibility to extrapolate to or predict ELM heat flux in NSTX-U and future ST device First systematic investigation of ELM heat flux from IR measurement in NSTX

3. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 3 IR camera diagnostics for heat flux measurement Divertor surface temperature is monitored by fast IR camera Single band (8-12 μm) fast IR camera in 2009 1 – Spatial resolution: 1.7 mm – Temporal resolution: 1.6 – 6.3 kHz Dual band (4-6μm and 7-10 μm) IR adapter in 2010 2 – For lithiated PFC surface, 1.6kHz frame speed Heat flux calculation from the measured surface temperature – 1-D radial heat flux profile from THEODOR 3 – 2-D (r, Φ) heat flux profile from TACO 4, improved to incorporate thin surface layer effect 5 1 J-W. Ahn, RSI 81 (2010), 023501, 2 A.G. McLean, RSI 83 (2012), 053706 3 Collaboration with IPP Garching, A. Hermann 4 G. Castle, COMPASS Note 97.16, UKAEA Fusion (1997) 5 K.F. Gan, submitted to RSI (2012)

4. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 4 Outline Study of transient ELM heat flux 1 – temporal and spatial evolution of peak heat flux and heat flux width - Type-V ELMs - Type-III ELMs with high and low  p - Type-I ELMs - Relationship between q peak and q 2-D heat flux data produced by TACO (through collaboration with MAST Team), now upgraded to take account of thin surface layer effect 1 K.F. Gan, submitted to NF

5. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 5 TACO calculates 2-D heat flux distribution at divertor surface Surface heat flux 2-D (r, Φ) heat flux data are obtained from TACO Analysis results from 1-D radial profile data are presented here but 2-D data are also being analyzed for the ELM and 3-D physics study Measured surface temperature Calculated 2-D heat flux Negative heat flux; Non-physical

6. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 6 The effect of thin surface layer on heat flux calculation Thin layer of hydrocarbon composite often forms on the surface of PFC material Low thermal conductivity with poor thermal contact with the PFC substrate underneath  IR emission is much brighter when a heat pulse hits the surface than it should be w/o thin layer  interpreted as much higher temperature rise than real  Excessively high heat flux computed to explain this  T surf Also, the surface layer makes the drop of surface IR emission excessively rapid after the heat pulse  interpreted as a rapid T surf drop  Solution of heat conduction equation yields negative heat flux to accommodate this temperature change Bulk PFC material (k, ρ, c p ) Thin surface layer Incident heat flux

7. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 7 Implementation of heat transmission coefficient in TACO alleviates negative heat flux problem  : Heat transmission coefficient 1 k layer : surface layer thermal conductivity, d: surface layer thickness Surface heat flux with  incorporated in heat conduction equation With  implemented in the heat conduction equation, negative heat flux problem is alleviated. This also lowers the computed peak heat flux 1 A. Herrmann, PPCF (1995) International collaboration on THEODOR with IPP-Garching Calculated 2-D heat flux with  = 60 (kWm -2 K -1 ) Calculated 2-D heat flux with no 

8. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 NSTX Choice of  value is made by comparing several values of deposited energy from different  The  value makes a large influence on the heat flux calculation, and different values lead to very different results In NSTX, we choose the  value that keeps the deposited energy constant after the discharge 8

9. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 Comparison of new TACO calculation and THEODOR with same  value Comparison of 1-D radial profile results from TACO to those from THEODOR  Good agreement for both the temporal evolution of peak heat flux and the spatial heat flux profile at a specific time slice 9 Peak heat flux evolution Heat flux profile

10. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 10 Outline Study of transient ELM heat flux 1 – temporal and spatial evolution of peak heat flux and heat flux width - Type-V ELMs - Type-III ELMs with high and low  p - Type-I ELMs - Relationship between q peak and q 2-D heat flux data produced by TACO (through collaboration with MAST Team), now upgraded to take account of thin surface layer effect 1 K.F. Gan, submitted to NF

11. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 Characteristics of several ELM types in NSTX Type-I ELMs:  plasma stored energy (W MHD ) drop by up to ~10 % Type-III ELMs:  1-3 % drop of W MHD Type-V ELMs:  Less than 1 % drop of W MHD, individual D  peaks are usually undiscernible 11 Maingi, NF 2005 Type-I Type-III Type-V Mixed Type-I / Type-V

12. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 Summary of q peak and int change during ELMs and ELM-free phase ELM typeq peak λ int Type-VIncrease Type-III (low  p ) Increase Type-III (high  p ) IncreaseDecrease Type-IIncreaseDecrease ELM-freeIncreaseDecrease int increases for all ELM types relative to the value for ELM-free phase 12

13. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 Heat flux profile during ELM has multiple local peaks  exponential fitting can’t characterize heat flux width well  Integral heat flux width int Evolution of heat flux profile shows that additional heat flux by ELM filaments mainly occurs in a zone displaced into SOL from original strike point  profile broadens Transient heat flux for type-V ELMs 13 Integral mid-plane heat flux width (1) (2) (3) (1) (2) (3) Separatrix

14. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 Transient heat flux for type-V ELMs 14 Heat flux profile during ELM has multiple local peaks  exponential fitting can’t characterize heat flux width well  Integral heat flux width int Evolution of heat flux profile shows that additional heat flux by ELM filaments mainly occurs in a zone displaced into SOL from original strike point  profile broadens Each type-V ELM increases q peak by 20 – 50% and int by ~20% (1) (2) (3) (1) (2) (3) Separatrix Type-V ELMs

15. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 Comparison of heat deposition pattern during type-V ELMy to ELM-free phases L-H transition ELM-free Type-V ELMs During the ELM-free phase, q peak increases and λ int rapidly decreases  total deposited power is somewhat constant During the type-V ELMy phase, q peak decreases and λ int increases  deposited power is again constant 15 132406

16. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 Transient heat flux for type-III ELMs 132401 I p =600kA, P NBI =4MW  High  p 132460 I p =700kA, P NBI =2MW  Low  p 16

17. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 Heat deposition for type-III ELMs with low  p ELM filamentary deposition of heat flux is distributed farther (4-5 cm) from the strike point. The profile then slowly reverts to the peak near SP during the ELM recovery phase During inter-ELM period, int grows and q peak decreases Each type-III ELM increases both q peak (by ~100%) and int (by ~20%) 17 (1) (2) 273.279ms 273.438ms 273.597ms 273.756ms 273.915ms 275.505ms (3) (1) (2) (3)

18. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 Heat deposition for type-III ELMs with high  p ELM filaments increase the heat flux 7-8 cm away from original strike point but the peak heat flux reverts back to the strike point afterwards During inter-ELM periods, int rebuilds slowly to the nominal value of ~3 cm Type-V ELMs are present between ELM peaks and they increase int Each type-III ELM increases q peak by a factor of 4 – 6 and decreases int by ~50% 18 (1) (2) (3) (1) (2) (3)

19. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 Heat deposition for type-I ELMs ELM filaments increase the peak heat flux a few cm away from the original strike point but the peak heat flux reverts back to the strike point afterwards  Similar behavior to type-III ELM with high  p Each type-I ELM increases q peak by up to 1000% and decreases int by 30-40% 19 (1) (2) (3) (1) (2) (3)

20. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 Summary of q peak and int change during ELMs and ELM-free phase ELM typeq peak λ int Type-V Increase ( x 0.2 – 0.5)Increase ( x ~0.2) Type-III (low  p ) Increase ( x ~1)Increase ( x ~0.2) Type-III (high  p ) Increase ( x 4-6)Decrease ( x ~0.5) Type-I Increase ( x 10)Decrease ( x 0.3-0.4) ELM-free Increase ( x 5)Decrease ( x 0.7) int increases for all ELM types relative to the value for ELM-free phase 20

21. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 Type-V Type-III Type-I Relationship between q peak and int for all ELM types Relationship between q peak and int at ELM peaks for multiple shots Overall inverse relationship observed between q peak and int  not good trend for future machine No clear correlation for Type-V alone Type-V and type-III appear to follow same trend together Some type-III ELM int values for large q peak (> 5-6 MW/m 2 ) are smaller than for type-I ELMs!! 21 A strong inverse relationship is observed in the lowest range of q peak for type-I  Acceptable regime might exist below q peak ~5 MW/m 2 ?? int saturates for largest values of q peak (> 10MW/m 2 ) for type-I

22. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 22 Summary and conclusions TACO has been implemented to generate 2-D heat flux data with surface layer effect taken into account – Choice of surface heat transmission coefficient (  ) greatly affects heat flux calculation. A value to keep the deposited energy constant after discharge is chosen in NSTX Temporal and spatial variation of q peak and q during three types of ELMs (type-V, type-III, and type-I) have been investigated − Type-V ELMs: Both q peak and q increases during each ELM. q peak decreases and q increases during type-V ELMy phase − Type-III ELMs: Both q peak and q increase during each ELM for low  p. q peak increases and q decreases for high  p. − Type-I ELMs: q peak increases and q decreases during each ELM Inverse relationship is generally observed between q and q peak. More work needed to characterize ELM heat flux behavior with underlying physics 2-D ELM heat profile analysis is underway using TACO data

23. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 MAST G De Temmerman, PPCF, 52 (2010) 095005 (14pp) Choice of  value is made by comparing several values of deposited energy from different  Several techniques to determine optimal  value  In MAST, comparison of deposited energy from IR measurement with the energy derived from the power balance is used 23

24. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 Transient heat flux for type-I ELMs Type-I ELMs:  Up to 10% loss of plasma stored energy and large transient heat load 24

25. NSTX J-W. Ahn: ELM heat flux study in NSTX (NSTX Monday Physics Meeting)27 August 2012 Transient heat flux for type-V ELMs Type-V ELMs Type-III ELMs Type-V ELMs:  Small ELM with < 1% drop in plasma stored energy 25