Evaluation of Anomalous Fast-Ion Losses in Alcator C-Mod

Transcription

1 Evaluation of Anomalous Fast-Ion Losses in Alcator C-Mod S. D. Scott Princeton Plasma Physics Laboratory In collaboration with R. Granetz, D. Beals, M. Greenwald MIT PLASMA Science and Fusion Center W. Rowan Fusion Research Center, University of Texas at Austin APS/DPP October 27-31, 2003

2 Abstract In recent Alcator C-Mod campaigns, the efficiency of H-minority RF heating has dropped to ~50%. To determine whether the poor heating efficiency is caused by anomalous losses of the energetic hydrogen tail, we compare the measured d-d beam-target neutron emission during injection of a perpendicular, 50~keV deuterium diagnostic beam against classical predictions by the TRANSP code. We compare the beam-target emission to classical slowing down calculations in both normal and locked-mode discharges in a variety of plasma conditions (nebar = m -3, I p = MA, P RF = 0 2 MW, B φ = Tesla) in the Ohmic and H-mode regimes both with and without locked modes. No difference in neutron emission is observed between plasmas with and without locked modes, suggesting that locked modes do not cause additional fast-ion losses in these plasmas.

3 Motivation: is RF Heating Compromised by Fast Ion losses in C-MOD? 96/03/01 L-mode 0.8 MA possible locked modes 1.0 MA n e (10 20 m -3 ) Trends: Density dependence Lower performance in plasmas with locked modes. dw/dt at RF turn-on is less than P RF.

4 There is also a Weak, Long-Term Trend Toward Decreased Performance 0.8 MA open-1996 solid= MA open-1996 solid-2001 n e (10 20 m -3 )

5 Technique: compare beam-target dd neutron emission during D-DNB injection to classical calculations. C-Mod Diagnostic Beam 50 kev, ~4 amps Injected radially on midplane Diameter ~8 cm Classical fast-ion confinement is excellent. Collisionless orbits of 50 kev deuterons injected at 90o, Z = 4 cm, R=75, 83 cm.

6 Deuterium DNB experiments Deuterium DNB Ip (MA) e+20 nl_ e+20 1e Te0 from ECE V_HVM (kv) Neutron rate (/s) e+12 2e Net RF power (MW)

7 KeV kev 10^20 MW Tesla MA no Hmode no Hmode no no no no yes yes yes yes yes yes 25 Mode? Regime Z eff T i (0) T e (0) N e (0) P RF B t I p Locked Plasma SHOT Plasma Conditions

8 no Hmode no Hmode no no no no yes yes yes yes yes yes 25 Ratio BT Total B-T B-B TX TOTAL Mode? Regime Measured TRANSP CX Orbit Shine Pinj Locked Plasma SHOT Neutrons (10 12 /sec) DNB Power (kw) Transp Beam and Neutron Calculations

9 First approach: global calculation of Expected neutron rate Comparison of measured and calculated neutron rates may give information on prompt losses of fast beam deuterons (and presumably on RF-heated H-minority tail) We will use TRANSP for the full-blown calculation, but here we present results of simplified calculations: single τ s for all fast DNB deuterons; ignore half- and third-energy components (< 10% effect) use T e (0) use <n e >, Z eff, and assume Z imp = 9 to get n D

10 Convolution integral R DD n D n fast σ DD (E) v R DD (t) n D 0 I beam (t τ) σ DD (E(τ)) v(τ) dτ where v(t) =v 0 e t/τ S E(t) = 1 2 M Dv 2 (t) and σ DD (E) =Duane formula in NRL formulary Fit the convolution integral (using IDL curvefit) to neutron signal to determine τ S and normalization for each shot

11 Convolution CURVEFIT examples Fit to rise, fall, and flattop

12 Determining deuteron slowing down time from neutron rate

13 Locked modes do not affect slowing down time Not locked Locked

14 Theoretical estimate of τ S (classical) v beam v D : τ ii s (s) 0.044/n D (10 20 m 3 ) v beam v e : τ ie s (s) T e (kev) 3/2 /n e (10 20 m 3 ) (I have used ln Λ = 20 in the preceding expressions.) Note that τ ii s τ ie s τ s = 1 τ ii s τ ie s

15 Calculated neutron rates from Global Model

16 TRANSP Analysis Differences from Global Model: Models relevant beam physics: deposition, shine-thru, orbit losses, charge-exchange losses. Uses full profiles of electron temperature, density, etc.

17 A Wide Range of Plasma Densities Were Studied

18 Beam Deposition Profiles Range from Centrally-peaked to Edge-peaked

19 Temperature Profiles Ion temperature profile is calculated from χ i = n χ i, neo with scale factor n adusted to match neutron emission during the Ohmic phase of the plasma. Calculated neutron emission varies ~ linearly with Te and is insensitive to Ti.

20 Beam Shine-Thru is Small Except at the Lowest Density Beam Shine-Through Power Beam Shine-Through Fraction P DNB ~ 135 kw Power loss (kw)

21 Classical Beam Orbit Losses are Less Than 15% Except at Highest Densities P DNB ~ 135 kw Power loss (kw)

22 Transp Neutron Simulations (ordered by density) L-Mode locked Shot 25 Neo 0.5 Ip 0.98 L-Mode locked TOTAL BEAM-TARGET Shot 23 Neo 1.0 Ip 0.97 MEASURED BEAM-BEAM THERMONUCLEAR L-Mode locked Shot 18 Neo 1.0 Ip 0.97 L-Mode locked Shot 24 Neo 1.3 Ip 0.97

23 Transp Neutron Simulations (ordered by density) L-Mode Shot 17 Neo 1.3 Ip 0.98 L-Mode TOTAL Shot 31 Neo 1.7 Ip 0.59 BEAM-TARGET MEASURED BEAM-BEAM THERMONUCLEAR L-Mode locked Shot 11 Neo 1.7 Ip 0.97 L-Mode locked Shot 09 Neo 1.8 Ip 0.97

24 Transp Neutron Simulations (ordered by density) L-Mode Shot 32 Neo 1.8 Ip 0.38 L-Mode Shot 29 Neo 1.9 Ip 0.96 H-Mode Shot 30 Neo 3.0 Ip 0.96 H-Mode P rf = 2.3 Shot 15 Neo 4.0 Ip 0.98

25 Measured Beam-Target D-D Neutron Emission is a Factor ~2 less than Classical TRANSP Predictions for N eo > m -3

26 TRANSP Error Analysis in Medium-Density, Locked L-Mode Shot 09 L-mode Locked-mode Ip 0.97 Neo 1.8 Zeff 1.6 Nominal Analysis (black) Error Analysis Te +/- 10% Ne +/- 10% Zeff +/- 15% <Z> = 6, 20 edge attenuation Attenuation of beam by edge neutral gas might explain the low neutron emission.

27 TRANSP Error Analysis in Medium-Density L-Mode shot 17 L-mode Ip 0.98 Neo 1.3 Zeff 2.8 MEASURED Nominal Analysis (black) Error Analysis Te +/- 10% Ne +/- 10% Zeff +/- 15% <Z> = 6, 20 edge attenuation Measured neutron rate is just within measurement uncertainty. Dominant uncertainty is composition of impurity mix. Possible beam attenuation at plasma edge would more than reconcile measured neutron rate with classical predictions.

28 TRANSP Error Analysis at Low Density shot 23 L-mode Locked-mode Ip 0.97 Neo 1.0 Zeff 3.6 Nominal Analysis (black) Error Analysis Te +/- 10% Ne +/- 10% Zeff +/- 15% <Z> = 6, 20 edge attenuation Dominant uncertainty is composition of impurity mixture Zeff is high, so dilution is significant. Measured neutron emission is less than predicted by TRANSP, but within diagnostic uncertainty. Attenuation of beam by edge gas pressure is a negligible effect.

29 TRANSP Error Analysis in High Density H-mode Shot 30 H-mode Ip 0.96 Neo 3.0 Zeff 2.1 Nominal Analysis (black) Error Analysis Te +/- 10% Ne +/- 10% Zeff +/- 15% <Z> = 6, 20 edge attenuation The customary diagnostic uncertainties do not reconcile the measured neutron rate with TRANSP. Edge neutral pressure is low, so attenuation of beam by edge gas pressure is a negligible effect.

30 TRANSP Error Analysis in low-ip, Medium Density L-mode Shot 32 L-mode Ip 0.38 Neo 1.8 Zeff 2.0 Nominal Analysis (black) Error Analysis Te +/- 10% Ne +/- 10% Zeff +/- 15% <Z> = 6, 20 edge attenuation Excited-State Depo Measured neutron rate is anomalously low, outside of measurement uncertainties considered.

31 High Edge Neutral Pressure May Reduce Expected Neutron Emission by 20-40% in Some Plasmas Range of neutral pressure in various plasmas Edge Pressures and expected attenuation are small at low- to moderate density

32 High Edge Neutral Pressure is Observed in Some High Density Plasmas

33 Estimated Beam Attenuation through Plasma Edge (full energy component)

34 Conclusions Both a simplified global analysis and standard TRANSP analysis show that the measured d-d beam-target neutron emission is a factor ~2 less than expected from classical fast ion confinement and thermalization. In some -- but not all plasmas, the discrepancy is within diagnostic uncertainty. Error analysis is ongoing. These results disagree with long-standing observations of classical fast ion confinement and thermalization reported on a number of tokamaks and thus would appear to suggest a machine-specific anomaly, e.g. ripple due to coil misalignment. These experiments do not distinguish between anomalous loss of the beam ions versus anomalous slowing down. Anomalous losses of energetic ions might explain the RF heating results, but anomalous slowing down would not.

35 File: scott APS 2003 as presented.ppt Date: November 3, 2003