1. Evolution of Bootstrap-Sustained Discharge in JT-60U
EX1-4
Evolution of Bootstrap-Sustained Discharge in JT-60U
Y. Takase,a S. Ide,b Y. Kamada,b H. Kubo,b O. Mitarai,c H. Nuga,a
Y. Sakamoto,b T. Suzuki,b H. Takenaga,b and the JT-60 Team
aUniversity of Tokyo, Kashiwa Japan
bJapan Atomic Energy Agency, Naka Japan
cKyushu Tokai University, Kumamoto Japan
21st IAEA Fusion Energy Conference
Chengdu, China
October 2006
I am pleased to report the experimental results of bootstrap-dominated plasmas produced in JT-60U, on behalf of the JT-60 Team.

2. Outline Introduction Experimental Results Conclusions
Motivation and Objective
Experimental Results
Bootstrap-driven discharge (fBS ~ 1)
Evolution of self-sustained phase
Comparison with high fBS discharge (fBS ~ 0.9)
Bootstrap overdrive (fBS > 1)
Conclusions
The outline of my talk is as follows:
Motivation and objective of research are presented.
In the experimental results section, a fully bootstrap-driven plasma is described, and evolution of such a discharge is discussed. A comparison is made with a discharge which has a finite contribution of positive beam-driven current.
Evidence of bootstrap overdrive is presented.
And conclusions are given.

3. Motivations and Research Objectives
External power required to drive the necessary Ip has a large impact on the recirculating power fraction, and therefore on the cost of electricity.
Study characteristics and controllability of plasmas with fBS ~ 1
If BS overdrive (fBS > 1) could be achieved, this may be used for Ip ramp-up. In this case requirements for external current drive can be reduced substantially.
Demonstration of BS overdrive
External power required to drive the necessary plasma current has a large impact on the recirculating power fraction, and therefore on the cost of electricity.
Hence, it is important to study the characteristics and controllability of plasmas with nearly 100% bootstrap current fraction.
If bootstrap overdrive could be achieved, this may be used for plasma current rampup. In this case requirements for external current drive can be reduced substantially. This leads us to prove experimentally that bootstrap overdrive is indeed possible.
These results have a fundamental impact on the design of ST reactors and slim-CS tokamak reactors with limited central solenoid capability.
Fundamental impact on designs of ST reactors and “slim-CS”
tokamak reactors with limited central solenoid (CS) capability.

4. JT-60U Coil System and Operational Scenarios
JT-60U Coil Configuration
Surface loop voltage is given by
Vl = - Ml,OHİOH -  Ml,PFİPF - Lextİp
3 types of control scenarios:
constant Ip
Lext dIp/dt = 0
constant OH coil current
Ml,OH dIOH/dt = 0
constant surface flux
Vl = 0
co/ctr
The cross section of JT-60U is shown here.
The red coils comprise the Ohmic transformer, which is called the F-coil in JT-60U.
The yellow coils provide the main vertical field, whereas the blue coils control the triangularity.
Green lines are projections of NB trajectories.
The surface loop voltage is determined by induction from the OH coil, other PF coils, and the plasma current. 3 types of control scenarios are used.
In the constant Ip mode, the last term is zero.
In the constant OH coil current mode, the first term is zero.
In both cases, noninductive overdrive produces a negative surface voltage,
and in the constant OH coil current mode, Ip rampup should also be observed.
In the constant surface flux mode, the left hand side of the equation is zero, ensuring that no net flux crosses the plasma surface. In this case, noninductive overdrive results in Ip rampup.
Bt = T
Ip = MA
(q > 10)

5. Power Balance for Ramp-up Experiment
Poynting flux across the plasma surface is kept at zero.
VlIp = Pext – dWmext/dt
= dWmint/dt – Pel + V2/RSp
Ip = INI + V/RSp ; INI = ICD + IBS ;
Pext = VlextIp ; Pel = – VINI ;
Wmext = LextIp2/2 ; Wmint = LintIp2/2 ;
Wm = Wmext + Wmint
 
EjNIdV EjOHdV
power balance
for poloidal field
magnetic energy
Overdrive power (INI > Ip)
Pel = dWm/dt – Pext +V2/RSp > 0
for ramp-up
For constant flux control (VlIp = 0)
dWmint/dt = Pel – V2/RSp  Pel > 0
Pext = dWmext/dt > 0 is supplied by
the external circuit.
PNB
Pel Ph
Multiplying the expression for the surface voltage yields the power balance equation for the poloidal magnetic field. I have time for only a brief summary in this talk, but I would be happy to discuss the details if you could visit my poster. Noninductive overdrive means that the noninductive current exceeds the total plasma current. The work that the noninductive current does against the reverse loop voltage, denoted by P_electric, generates the poloidal field. Noninductive current includes the bootstrap current in addition to the beam driven current. Another way to generate the poloidal field is by induction from external coils, denoted by P_external. A small fraction of poloidal field energy is dissipated resistively, and the difference goes into increasing the poloidal field energy.
Pext Wm Wp
V2/RSp
Ploss
dWm/dt
dWp/dt

6. Fully Bootstrap-Driven Discharge
Fully BS-driven plasma is realized
Duration of self-sustained phase is limited by slow confinement degradation (this case), or by b collapses
slow decrease of Wp & Ip
I would like to turn now to experimental results. As Takenaga-san mentioned in the Overview talk this morning, steady sustainment of Ip by bootstrap current alone is extremely challenging.
The discharge shown on this page is an example of a fully bootstrap-driven plasma realized using the constant OH coil current mode. This was realized for about 0.2s indicated by the arrow. After this time the energy confinement degraded slowly, and it became impossible to keep the stored energy constant. The vertical field is ramped down in response to the decreasing stored energy, and caused the plasma current to decay slowly over several seconds.
The current density profile measured by MSE, and the q profile are shown.
The evolution of the ion temperature profile is also shown.

7. Loop Voltage is Kept Nearly Zero for 0.2 s
MSE-based reconstruction: Poloidal flux is nearly constant for ~ 0.2 s.
dIp/dt  0
dIp/dt < 0
Vl ~ 0 V
This graph shows the evolution of poloidal flux at different radial locations inside the plasma, calculated based on magnetic reconstruction using the MSE measurement.
It can be seen that the time derivative is kept nearly zero for 0.2s. During this time the calculated beam driven current is -33~-35kA, and the inductive current is -5~-30kA.
Iind = -5kA
-30kA
Iind > 0
INB = -35kA
-33kA

8. Bootstrap-Sustained Discharge
Total current = 543 kA (MSE)
Inductive current = -5 kA (calculated from Vl profile)
Beam driven current = -35 kA (calculated by OFMC/ACCOME)
 Bootstrap current = 583 kA (possibly slight overdrive)
E046687
@ 4.6 s
Profiles of the total current density (measured by MSE), the calculated beam driven current density, and the inductively driven current density, calculated from the loop voltage profile shown in the previous page, are shown. The inferred bootstrap current of 583kA is slightly greater than the total current of 543kA, possibly indicating a slight overdrive. Although there is a rather large uncertainty of nearly 50kA in the determination of the bootstrap current, it is clear that the bootstrap current fraction of approximately 100% was achieved.

9. Examples of BS-Sustained Plasma
all with const. OH coil current
Several discharges with fBS ~ 100% were obtained
E046293
constant Ip maintained at 0.51MA for 1.3 s noninductively
(BS + negative NB)
Bv ramp-up
~ 100% BS
In order to show that there is more than one discharge in which f_BS ~ 100% was achieved, three discharges are shown. All these discharges used the constant OH coil current mode. In particular, at a lower current of 510kA, a constant Ip was maintained for 1.3s.

10. Dynamics of BS-Driven Plasma
Stored energy Wp increases and Ip increases to 610kA
ITB shrinks at b collapses, then recovers partially
Ip is self-sustained (not driven by NBCD or OH coil)
const. IF
b collapse
610kA 1 2 3 4 5 6 7
It was shown earlier that a slow degradation of confinement resulted in the reduction of Ip that could be self-sustained. When beta is higher, beta collapses can occur. In the discharge shown here, several beta collapses were observed. At a beta collapse, the ITB shrinks in radius, but then recovers spontaneously. In this kind of discharge, the plasma current is self-sustained by the bootstrap current, and is not determined by external sources such as NBCD or the OH coil.

11. Evolutions of Ti and q Profiles
ITB is eroded (becomes narrower) at b collapses, and both Wp and Ip decrease.
Spontaneous recovery of ITB, Wp, and Ip occurs.
However, complete recovery to the original level is not achieved.
2nd collapse q
The evolutions of the ion temperature profile and the q profile are shown.
It can be seen that at sequential beta collapses, erosion of the ITB occurs and the q_minimum radius shrinks inward. Although there is spontaneous recovery following each beta collapse, the recovery is not perfect and settles at a state with lower Wp and Ip.
1st collapse
qmin radius

12. Stable Sustainment Aided by Co-NBI
Steady sustainment achieved for over 2s with INB ~ 40 kA (fBS ~ 0.9)
Confinement is better with co-NBI (1.2 MJ / 5.9 MW vs. 1.0 MJ / 8.1 MW)
dIp/dt > 0
dIp/dt  0
const. IF control
In order to lengthen the duration of stationary state, a co-tangential NB was added to a discharge otherwise identical to the fully bootstrap driven discharge (46293).
The robustness of steady sustainment was improved greatly by the addition of only a small fraction of co NBCD. The net beam driven current is about +40kA, less than 10% of the total plasma current in this discharge. In contrast to the slowly degrading energy confinement in the fully BS-driven discharge, energy confinement was maintained at a high level, and a nearly constant plasma current exceeding 515kA was sustained for over 2s.

13. Evidence of Bootstrap Overdrive
IAEA 2004: CS recharging with Vl < 0 achieved (const. Ip control)
Slow ramp-up at 10 kA/s for 0.5 s with zero inductive flux input
const. flux control
ctr &  NBI only
In the 2004 IAEA Conference, CS recharging with negative loop voltage, achieved in the constant Ip mode, was reported as a possible evidence of BS overdrive. In order to demonstrate BS overdrive more clearly, the constant surface flux mode was implemented. In this mode, real-time equilibrium reconstruction is used to compute the poloidal flux at the plasma surface, and this flux is kept at a constant level by feedback control using the OH coil current. This method ensures that there is no net flux crossing the plasma boundary, and the plasma current rampup becomes a direct indication of noninductive overdrive. In this example, a slow rampup at a rate of 10kA/s was maintained for about 0.5s.
For comparison, in the LHCD rampup discharge reported in the 2002 IAEA Conference, which was substantially overdriven, the rampup rate was about 40kA/s. Considering that the present discharge is only slightly overdriven, the rampup rate of 10kA/s is not an unreasonably small value.

14. Conclusions
A fully bootstrap-driven discharge with fBS ~ 1 was realized
Ip = 510 kA was maintained for 1.3 s (with net INB = -35 kA).
Dynamics of BS-sustained plasma
In discharges without a b collapse, slow degradation of confinement resulted in slowly declining Wp and Ip.
ITB shrinks radially at b collapses, resulting in Wp and Ip decrease. Subsequently, partial recovery of ITB and Ip occurs spontaneously.
Addition of positive INB (< 10% of Ip) helps steady sustainment of Ip greatly.
Evidence of bootstrap overdrive was observed
Slow Ip rampup (10 kA/s for 0.5 s) was observed with no inductive flux input.
In conclusion, a fully bootstrap-driven discharge with a bootstrap fraction of nearly 100% was realized. A nearly constant Ip of 510kA was maintained for 1.3s.
The evolution of such a BS-sustained plasma was studied.
In discharges without a beta collapse, a slow degradation of confinement resulted in slowly declining Wp and Ip.
At higher beta, beta collapses are observed. ITB shrinks radially at beta collapses, resulting in Wp and Ip decrease. Subsequently, partial recovery of ITB and Ip occurs spontaneously.
The addition of a positively directed beam-driven current of < 10% of the total current, helps steady sustainment of Ip greatly.
Evidence of bootstrap overdrive was observed using the newly implemented constant surface flux feedback control, which ensures no net flux input to the plasma. A slow Ip rampup at a rate of 10kA/s was achieved for 0.5s.
Extension of these results to higher Ip and a more complete characterization of controllability of such a plasma remain topics of further research.
Extension to higher Ip and a more complete characterization of controllability of such a plasma remain topics of further research.