Recent progress in neutral beam current drive …. Hopf IAEA, Vienna, 01—04 Sept. 2015 Max-Planck-Institut für Plasmaphysik Recent progress in neutral beam current drive experiments

C. Hopf IAEA, Vienna, 01—04 Sept. 2015

Max-Planck-Institut

für Plasmaphysik

Recent progress in

neutral beam current drive experiments on

ASDEX Upgrade

C. Hopf, D. Rittich, B. Geiger, A. Mlynek, M. Reich,

A. Bock, A. Burckhart, C. Rapson, F. Ryter,

the ASDEX Upgrade and EUROfusion MST1* Teams

* See http://www.euro-fusionscipub.org/mst1

1 of 15

http://www.euro-fusionscipub.org/mst1




NBI and current drive capabilities on ASDEX Upgrade

Versatile NBI system 8 beams, 2.5 MW each, on two injectors: 2 “radial” beams @ 60 kV 2 “intermediate” beams @ 60 kV 2 “intermediate” beams @ 93 kV* 2 “tangential, off-axis CD” beams @ 93 kV*

Diagnostics ( information related to)

• Loop voltage

total driven current

• Motional Stark effect (MSE) current profile

• Fast ion D-alpha (FIDA) radial fast ion profile

• Faraday rotation polarimetry current profile • Neutron rates number, energy, radial distribution of fast ions Practical approach: Comparison to synthetic diagnostics

2 of 15


Off-axis NBCD on AUG: historic inconsistencies

MSE does not fit: S. Günter et al., NF 47 (2007) 920

3 of 15

H-mode, 600 kA



MSE does not fit: S. Günter et al., NF 47 (2007) 920

3 of 15

H-mode, 600 kA



“non-neo-classical case”

B. Geiger, PhD thesis (2012)

S. Günter et al., NF 47 (2007) 920

4 of 15

H-mode, 800 kA



“non-neo-classical case”

B. Geiger, PhD thesis (2012)

S. Günter et al., NF 47 (2007) 920

Immediate question: Are there cases where the fast ion distribution is neo-classical while the driven current does not agree with the neo-classical prediction? Simultaneous MSE and FIDA measurements needed!

4 of 15


The “new” discharge

• Continuous MSE and FIDA (beam 3 running through) • Preemptive NTM and sawtooth avoidance by ECCD • Real-time Te feedback control by RT-ECE/ECRH in off-

axis phase

5 of 15

Ip = 800 kA, H-mode, #31453


The “new” discharge

• Continuous MSE and FIDA (beam 3 running through) • Preemptive NTM and sawtooth avoidance by ECCD • Real-time Te feedback control by RT-ECE/ECRH in

off-axis phase

5 of 15

Ip = 800 kA, H-mode, #31453


Constant Te and ne profiles

Spline-fitted profiles as input to TRANSP, Te and ne from integrated data analysis. Ti from core CXRS

6 of 15

Te vs rtor

3.5 s 5.6 s


Very little MHD activity: Some fishbones, nothing else

SXR

Magnetic Spectrum

15 kHz

on axis off axis on axis

7 of 15


TRANSP predicted currents and loop voltage

TRANSP predicts: • ≈ 95/160 kA NBCD in on/off-axis phase • no change in bootstrap current

Measured loop voltage: Agrees best with TRANSP prediction assuming no or moderate anomalous diffusion

8 of 15


TRANSP predicted currents and loop voltage

Loop voltage sensitive to Zeff Zeff profiles calculated from CXRS-derived impurity concentration profiles

Measured and calculated loop voltage agree in absolute values!

Measured loop voltage: Agrees best with TRANSP prediction assuming no or moderate anomalous diffusion

8 of 15


Fast Ion D spectroscopy: radial distribution of fast ions

on axis NBI @ 3.5 s

off axis NBI

9 of 15

on axis NBI @ 7.5 s

FIDA 4D fast ion distribution function from TRANSP, post-processed using FIDASIM synthetic radial intensity profiles • agrees best with 0.0–0.2 m2/s anomalous diffusion

in on-axis phases



on axis NBI @ 3.5 s

9 of 15

on axis NBI @ 7.5 s

FIDA 4D fast ion distribution function from TRANSP, post-processed using FIDASIM synthetic radial intensity profiles • agrees best with 0.0–0.2 m2/s anomalous diffusion in

on-axis phases • unresolved discrepancy around rpol = 0.5 with off

axis NBI

off axis NBI


Interlude: Physics-based estimates of diffusion coefficients

10 of 15

M.J. Pueschel et al., NF 52 (2012) 103018

Based on gyrokinetic (GENE) simulations Pueschel at al. give approximate formulae for D describing micro-turbulent transport as function of experimentally accessible parameters: eff effective heat diffusivity

fast ion pitch angle EFI fast ion energy Te electron temperature plasma beta crit critical beta w.r.t. KBMs

Plotted values: • Energy-averaged • electrostatic and electromagnetic contributions added • finite Larmor radius approximation taken

Can be used in TRANSP as tabulated D(r, t, , E)

only for EM part

DES(r, t) + DEM(r, t) [m2/s]

rto

r



FIDA 4D fast ion distribution function from TRANSP, post-processed using FIDASIM synthetic radial intensity profiles • agrees best with 0.1–0.3 m2/s anomalous diffusion

in on-axis phases • unresolved discrepancy around rpol = 0.5 with off

axis NBI • Very little difference between neo-classical (D = 0),

D = 0.1 m2/s and DPueschel(r, t, E)

on axis NBI @ 3.5 s

11 of 15

on axis NBI @ 7.5 s

off axis NBI


Motional Stark effect spectroscopy: Radial current profile

12 of 15

Comparison with MSE angles calculated from TRANSP equilibria:

• “Neoclassical” does not fit well in outer channels

• Better (best) fit: D ≈ 0.5 m2/s

• “No NBCD” does not fit well in inner channels

• DPueschel in between neo-classical and D = 0.5 m2/s

Note: Compared to previous studies • Sign of Er (TRANSP)

corrected • Corrected geometry of

beam 8

sh

ifte

d

to

mat

ch

he

re

rpol = 0.12 rpol = 0.23 rpol = 0.47 rpol = 0.58 rpol = 0.68



12 of 15

rpol = 0.12 rpol = 0.23 rpol = 0.47 rpol = 0.58 rpol = 0.68








beam 8



12 of 15

rpol = 0.12 rpol = 0.23 rpol = 0.47

rpol = 0.58 rpol = 0.68








beam 8









beam 8


12 of 15

B. Geiger et al., PPCF 57 (2015) 014018

rpol = 0.12 rpol = 0.23 rpol = 0.47

rpol = 0.58 rpol = 0.68



12 of 15

rpol = 0.12

rpol = 0.23 rpol = 0.47 rpol = 0.58 rpol = 0.68








beam 8



12 of 15

rpol = 0.12

rpol = 0.23 rpol = 0.47 rpol = 0.58 rpol = 0.68








beam 8

Caution! MSE not absolutely calibrated (yet). We compare only temporal evolutions from the on-/off-axis transition onwards.


Faraday rotation polarimetry

Faraday rotation:

neBu ds

u: unit vector along LOS

13 of 15

All simulations shifted by –1.6°!



Faraday rotation:

neBu ds


13 of 15

• Generally reasonable agreement with TRANSP-based simulation

• Cannot resolve between neo-classical simulation and D = 0.5 m2/s




Faraday rotation:

neBu ds


13 of 15

• Generally reasonable agreement with TRANSP-based simulation

• Cannot resolve between neo-classical simulation and D = 0.5 m2/s

• No agreement w/o NBCD



Neutron rates

Unfortunately absolute measured neutron rates cannot be trusted currently. However, the relative integrated signal of the neutron spectrometer is reliable. Overall reasonable agreement. Unfortunately not sensitive to small variations of radial transport. Simulations have considerable uncertainty due to error in Zeff and H dilution of the D plasma. No conclusion supported.

14 of 15

1 2 3 4 5 6 7 8 90.0

0.2

0.4

0.6

0.8

1.0

1.2

ne

utr

on r

ate

(a.u

.)

realtime

neutron spectrometer (integrated)

TRANSP neo-classical

TRANSP D = 0.5 m2/s

all rates scaled to coincide in on-axis phases


Summary, conclusions, and outlook

Summary • Loop voltage, MSE, and Polarimetry clearly show that NBI drives a current. • FIDA during on-axis NBI compatible with no to mild (0.0–0.2 m2/s) anomalous transport • D 0.2 m2/s agrees well with “Pueschel values” • Disagreement of simulation with off-axis FIDA at medium r not understood. • Temporal evolution of MSE angles after on-/off-axis transition fits best with D 0.5 m2/s.

Conclusion Initial question: Do FIDA (radial fast ion profile) and MSE (current profile) contradict each other?

My tentative answer: We have no reason to assume so! Outlook • Diagnostics enhancements will reduce dependency on beam 3 and add redundant

information. • Absolute calibration of MSE expected soon Will allow separation of on- and off-axis

transport effects • Next step: Parameter study to find the transition from neo-classical to diffusion-dominated

regime

15 of 15



Current density profiles (TRANSP neo-classical)


Current density profiles (TRANSP neo-classical)


q profiles (TRANSP neo-classical)

1.85 sec (before NBI) 3.50 sec (on-axis NBI) 5.30 sec (off-axis NBI)

rtor

q

B. Geiger September 04, 2015, Fachbeirat, Greifswald 32

Comparison with TRANSP modelling

• Synthetic MSE –angles

can be calculated from

TRANP equilibria

• But offset per channel

needed for comparisons!

• Agreement with the neo-

classical simulations

(except the innermost

channel)

• Assumption of localized

anomalous fast-ion

transport has no clear

effect

• Earlier analysis: wrong

sign of Er in TRANSP and

geometry of NBI Q8 5 cm

too low


Improved analysis of radial MSE profiles

• Correct geometry

description of MSE LOS

allows study of radial MSE

profiles

• Bz can be determined

based on the MSE data

• -> radial current profiles

can be calculated

O. Ford


Change of radial current density profiles

• Relative change of the

neo-classical simulation

within the statistical error

bars of the measurement

• Similar agreement when

considering localized

anomalous transport

Petty, C et al., Nucl. Fusion 42 (2002) 1124 – 1133

• Given j~rot(B), the radial derivative of Bz is

roughly proportional to the current density

Recent progress in neutral beam current drive …. Hopf IAEA, Vienna, 01—04 Sept. 2015 Max-Planck-Institut für Plasmaphysik Recent progress in neutral beam current drive experiments

Documents