O AK R IDGE N ATIONAL L ABORATORY U.S. D EPARTMENT OF E NERGY O AK R IDGE N ATIONAL L ABORATORY U.S. D EPARTMENT OF E NERGY O AK R IDGE N ATIONAL L ABORATORY.

1OAK RIDGE NATIONAL LABORATORYU.S. DEPARTMENT OF ENERGYOAK RIDGE NATIONAL LABORATORYU.S. DEPARTMENT OF ENERGYOAK RIDGE NATIONAL LABORATORYU.S. DEPARTMENT OF ENERGY BPWorkshop-2005 - LRBOAK RIDGE NATIONAL LABORATORYU.S. DEPARTMENT OF ENERGY

presented by L.R. Baylor

in collaboration with

P.B. Parks*, S.K. Combs, W.A. Houlberg, T.C. Jernigan,

S. Maruyama#, L.W. Owen, G.L. Schmidt, D.A. Rasmussen

Oak Ridge National Laboratory, *General Atomics, #ITER International Team

at the

Burning Plasma Workshop

5-July-2005

Tarragona, Spain

Pellet Fueling Technology Leading to Pellet Fueling Technology Leading to Efficient Fueling of ITER Burning Plasmas Efficient Fueling of ITER Burning Plasmas


• ITER requires significant fueling capability to operate at high density for long durations

• Gas fueling will not be able to sustain high density in ITER due to limited neutral penetration in the thick dense scrape off layer

• Pellet fueling from the inner wall looks promising for core fueling with high efficiency despite limited pellet speeds

• The ITER pellet injection system requires capabilities well beyond the current state-of-the-art

– Throughput enhancement of nearly an order of magnitude

– Reliability at high repetition rate is required for BP control

• The use of pellets for ELM triggering and amelioration remains a possibility for ITER

– Understanding pellet interaction with NTMs, ELMs, RWMs etc needed

Overview


• ITER plasma volume is 840 m3 and scrape-off layer is ~20 cm thick. This compares to 20 m3

and ~ 5 cm for DIII-D.

ITER Fueling Needs are Significant

ITER Cross Section

4 m

DIII-D Cross Section




• ITER is designed to operate at high density (> 1x 1020 m-3) in order to optimize Q.


ITER Cross Section

4 m





• Gas to be introduce from 4 ports on outside and 3 in the divertor region


ITER Cross Section

GasInjectors

4 m






• NBI fueling to be negligible (< 2 x 1020 atoms/s or < 3 torr-L/s )


ITER Cross Section

GasInjectors

4 m

Note that DIII-D at 10 MW is ~ 10 torr-L/s







• Inside wall pellet injection planned for deep fueling and high efficiency. Reliability must be very high.


ITER Cross Section

GasInjectors

PelletInjection

4 m







• Inside wall pellet injection planned for deep fueling and high efficiency. Reliability must be very high.

• Pellet injector must operate for up to 1 hour continuously and produce up to 4500 cm3 of DT ice per discharge.


ITER Cross Section

GasInjectors

PelletInjection

4 m


Gas Fueling in ITER is Much Less Efficient than in Current Machines

D+ 1

019 m

-3 s

-1

• This B2-Eirene slab calculation shows that gas puff core fueling in ITER will be much less effective than in current experiments such as DIII-D.

- Gas fueling rate of 100 torr-L/s for DIII-D

- Gas fueling rate of ~1000 torr-L/s for ITER case (L. Owen and A. Kukushkin) (see also Kukushkin & Pacher, Plasma Phys. Control. Fusion 44, 931, 2002 )

0.0 0.2 0.4 0.6 0.8 1.00.001

0.01

0.1

1

10

100

1000

DIII-D Gas

ITER Gas

normalized minor radius

Gas Fueling Efficiency < 1%

Gas Fueling Source Profile

Fueling efficiency is Nplasma / Nsource


Pellet Injection from Inner Wall Looks Very Promising for Tokamak Plasma Fueling

• Net deposition is much deeper for HFS pellet in spite of the lower pellet velocity used to survive curved guide tube

• Pellets injected into the same discharge and conditions (ELMing H-mode, 4.5 MW NBI, Te(0) = 3 keV)

2.7 mm pellets - HFS 45° vs LFS

HFS 45°vp = 118 m/sDt = 5 ms

0

5

10

0.0 0.2 0.4 0.6 0.8 1.0 1.2

LFS vp = 586 m/st = 1 ms

ne (

101

9 m

-3) DIII-D 98796 - measured ne

Calculated Penetration

HFS - 95%LFS - 55%

FuelingEfficiency:


ExB Polarization Drift Modelof Pellet Mass Deposition (Rozhansky, Parks)

• J= - 2p/RB and this balances the polarization

return current Jp = (/B2) dE/dt. (p is cloud

pressure and is cloud density)

• Therefore the pellet cloud motion equation is dV/dt = 2p/R

R drift distance is stronger at higher plasma due to higher cloud pressure

• Detailed model by P.B. Parks, [Phys. Rev. Lett. 94, 125002 (2005)].

-+

+

-

B 1/R

R

E

ExB

PelletAblatant (Cloud)

HFS LFS

• Polarization of the ablatant occurs from B and curvature drift in the non-uniform tokamak field:

• The resulting E yields an ExB drift in the major radius direction, V = (ExB)/B2

Theoretical Model for Pellet Radial Mass Drift

BBeB

2WWv

3

||B


Pressure Relaxation Lagrangian (PRL) Code Solves Coupled Drift and Parallel Dynamics for a Series of Cloudlets

• The PRL code uses the pellet size and plasma parameters at each point along the ablation track determined by PELLET code [Houlberg, Nucl. Fusion 1988] to initialize the cloudlet parameters using model of Parks, et al. 10-20 cloudlets are assumed per pellet. [see Phys. Rev. Lett. 94, 125002 (2005)].

• The cloudlets form a tube of high density plasma along the field lines. The ends of the cloudlet are sheared off as it drifts inward (mass shedding).

• The experimental plasma profiles are used by PRL to calculate the cloudlet pressure relaxation, drift velocity, and shedding location.

• The deposition profiles from each cloudlet are summed, yielding a net n deposition profile.

PELLET Ablation(No drift)

Cloudlet

ne

Deposition profile

Cloudlet drift

Cloudlet

Lagrangian cells of constant mass

QHeat flux

Pellet + Cloud


Experiment and PRL Model Compare Well

0

5

10

0.0 0.2 0.4 0.6 0.8 1.0

ne

(1019

m-3) DIII-D 98796

2.7mm pellet, vp = 586 m/sPRL Model

0

5

10

0.0 0.2 0.4 0.6 0.8 1.0

n

e (1

019 m

-3)

DIII-D 994772.7mm pellet, vp = 153 m/s

Data

Outside midplane launch Inside launch (45 deg above mid-plane)

Data

• Vertical arrows indicate pellet burnout location

• Fueling efficiency for inside launch is much higher (even with slower pellets)outside launch theory = 66% , exp = 46% (discrepancy due to strong ELM)

inside launch theory = 100% , exp = 92% (discrepancy due to weak ELM)

PRL model is a major breakthrough in understanding the physics of pellet mass drift

PRL Model

NGS AblationModelx0.3NGS

AblationModelx0.5


D+ 1

019 m

-3 s

-1

DIII-D

• Gas puff core fueling in ITER will be much less effective than in DIII-D

- ITER pellet profiles are from PRL (P. Parks) ( 5-mm @ 16 Hz )

- gas fueling rate of ~1000 torr-L/s for ITER case B2-Eirene slab calculation (L. Owen and A. Kukushkin)

0.0 0.2 0.4 0.6 0.8 1.0

1

10

100

1000

HFS Pellet

LFS Pellet

Gas

0.0 0.2 0.4 0.6 0.8 1.00.001

0.01

0.1

1

10

100

1000

HFS PelletLFS Pellet

Gas

ITER

Gas Fueling Efficiency < 1%

HFS pellet Gas puff

Pellet Injection is Crucial for Effective Core Fueling in ITER as Shown in H-mode Fueling Source Profile Comparison


Density Change in ITERas a Function of Inner Wall Pellet Size

• Pellet fueling deposition calculations from PRL for ITER with different size pellets. Larger pellet size yields marginally deeper mass penetration. Mass drifts well beyond the pedestal for all pellet sizes. Outside midplane injection deposition profiles (dashed) with no drift are shown for comparison.

• Pellets injected into the same discharge conditions from the inner wall guide tube port. (H-mode, Te(0) = 20 keV, Tped = 4 keV, ped=0.04)

0.0

1.0

2.0

3.0

4.0

5.0

0 0.2 0.4 0.6 0.8 1

5mm

n

e (1

019

m-3)

3mm

ITER Deposition Profiles

Outside Midplane Pellets x 0.3

Inner Wall Pellets


0

2

4

6

8

0 0.2 0.4 0.6 0.8 1

n

e (1

019 m

-3) k=2

k=5 k=20

Weaker Shear Leads to Deeper Mass Deposition for ITER Inner Wall Pellet Injection

0 0.2 0.4 0.6 0.81

2

3

4

Saf

ety

fact

or,

q

• Pellet fueling deposition calculations from PRL for ITER with different plasma q profiles. Stronger shear at the edge leads to more rapid mass shedding of the cloudlets and hence shallower mass penetration.

• Pellets (6mm from inner wall) injected into the same discharge conditions (H-mode, Te(0) = 20 keV, Tped = 4 keV, ped=0.04)

1

k=2

qq0 (qa q0)k

k=5k=20

q0 1, qa 3.7


• ITER will initially have 2 pellet injectors that each provide D2, DT, T2 pellets (5mm @ 16Hz, 3mm @ 32Hz).

• Inside wall pellet injection for deep fueling beyond the pedestal and high efficiency. Reliability must be very high.

• Guide tubes bring the pellets in from divertor ports and routes them to the inner wall.

• Pellet injector must operate for up to 1 hour continuously and produce ~ 1.5 cm3/s of ice.

ITER Pellet Fueling Requirements

Pellet Path in ITER


Plasma Density (nGW) 0.4 – 1 ( 0.5-1.2 x 1020 m-3 )

Fuel Isotope Pellet D2, DT , T2(80%T/20%D)

3-5 mm diam => 1.25 - 6 x1021 atoms n/n ~ 1.3%-6.6%

Gas Fueling Rate (Pa-m3/s) Up to 400 (~3000 torr-L/s)

Pellet Fueling Rate (Pa-m3/s) 100 for D2, DT (~800 torr-L/s)

50 for T2 (~400 torr-L/s)

Pulse length (s) Up to 3000

Gas injection system» Supplies H2, D2, T2, DT, Ar, Ne, and He via a gas manifold

» Primary use for initial gas fill, control of SOL, and flushing impurities to divertor» Makes use of conventional gas handling hardware and requires minimal R&D

Pellet injection system» Supplies H2, D2, and DT pellets: 3 to 5 mm diam. (32 to 16 Hz, respectively)

» Only at pre-conceptual design level and some R&D still needed

Requirements refined at ITER Pellet Injector Workshop in Garching, May 2004

ITER Fueling Systems Requirements & Present Design


• Initial tests with 5.3 mm pellets

• Pellet speeds limited to ≈300 m/s for intact pellets

• Guide tube mass loss ≈10% at speed limit

ITER Inner Wall Guide Tube Tests Indicate 300 m/s Speed Limit

S. Combs, et al. SOFT 2004

Pellet Path in ITER


ITER initially will have 2 pellet injectors for deep core fueling as the primary fuel delivery system.

» Up to 6 injectors planned for future

Requires continuous, highly reliable, high throughput, tritium rich pellets

» significant throughput extension of present-day designs

» Centrifuge accelerator witha continuous screw extruder

» Inner wall pellet injection with curved guide tubes

» Maximum T concentration is ~80% due to tritium processing plant limitation

PIS to be enclosed in cask that rolls up to a divertor port

6.5 m 2.5m

3.6

m

ITER Pellet Injection System Conceptual Design

Rotor Arm

Extruder

D2, T2, DTSupply

Vacuum Pump Tritium

Reprocessing Plant

GuideTube Centrifuge

Guard Vacuum

Pellet Injection Cask

Cryocooler

Cutter

Rotor Arm

Extruder

D2, T2, DTSupply

Vacuum Pump Tritium

Reprocessing Plant

GuideTube Centrifuge

Guard Vacuum

Pellet Injection Cask

Cryocooler

Cutter


• Tritium pellet formation and acceleration were found to be readily achievable with present technology.

• Pellets with high T2 concentration are envisioned for fueling ITER using the isotopic tailoring scheme

– T2 rich pellets combined with D2 gas puffing (Gouge, et al., Fusion Tech. 1995)

• Multiple pellet injectors with different T fractions can be used to control fusion power

Tritium Pellet Injector at TSTA

Tritium Extrusion(≈8 mm)

Tritium Pellet Formation has been Proven and Can Be Used to Control Plasma Burn


D2

He

LHe

Screw

HeatExchanger

Solid D2

VacuumPump

Batch Piston and Continuous Screw Extruders are Possible to Meet the Needs for the ITER Pellet Injection System

• Both batch piston and continuous screw extruders have been developed as possible ITER ice sources

• Multiple batch extruders have produced 1.3 cm3/s (S. Combs, et al,

RSI (1998) while a continuous screw extruder by PELIN has produced steady-state H2 ice up to 0.3 cm3/s. (~1/5 of rate needed for ITER) (I. Viniar, SOFT 2004).

• Throughput enhancement may be possible or multiple such extruders could be used on the ITER pellet injection system.

• Simpler operation makes the screw extruder preferable over a batch extruder.

LHe

OFHC Copper

Batch Piston Extruder Continuous Screw Extruder

8mm H2 ice extrusion


Pellet ELM Triggering May Provide Tool for ELM Amelioration

• Pellet injection has been found to trigger ELMs in ELMing H-mode plasmas ( AUG, DIII-D, JET).

• LFS pellets trigger larger ELMs than the same pellets from the inner wall, leading to a possible sensitive LFS pellet ELM trigger.

• AUG has succeeded in increasing the ELM frequency and lowering the ELM size using small pellet triggers. (P. Lang et al., Nuc. Fusion 2004)

• ITER 3mm size pellet is for ELM triggering using a LFS guide tube.

• Further research is needed to investigate the pellet induced ELM mechanism and its scaling to ITER.

• Interaction of pellets with NTMs, RWMs, ELMs, etc. needs better understanding.

DIII-D 120775H-mode 5 MW NBI

0

2

4

61.0

1.5

2.0

Div

erto

r D

(a.u

)n e

L(1

014 c

m-2

)

3.6 3.7 3.8 3.9 4.0 4.1

1.8mm V+3 pellets


0

2

4

61.0

1.5

2.0

Div

erto

r D

(a.u

)n e

L(1

014 c

m-2

)

3.6 3.7 3.8 3.9 4.0 4.1

1.8mm V+3 pellets

Time (s)

LFS Pellets forELM triggering


ITER will require significant fueling beyond that provided by gas

» Gas fueling and recycling expected to be very inefficient

Inner wall injection port will allow up to 300 m/s pellet injection

Modeling of the proposed ITER pellet injection scenario looks promising for core fueling well beyond the H-mode pedestal

» Further validation of the ExB polarization drift model is needed with diagnostics and scaling studies

The pellet fueling system for ITER presents challenges for the technology developers in throughput and reliability, concepts look promising

» Development is underway and expected to take ~ 5 yrs

» Centrifuge and extruder prototypes will be produced which can be available to test on existing tokamak devices

ELM triggering by small LFS pellets also a promising technique for ITER

» Further research to optimize and understand physics of pellet induced ELMs and ELM amelioration is required as well as other MHD interactions.

Summary


Pellet Droplet Concept for ELM Triggering

V+3

Pellet DropletDevice • A batch extruder with curved nozzle and

pellet cutter can supply sub 1mm D2 pellets at up to 100 Hz for triggering ELMs on DIII-D

• The extruder is to be cooled with a GM cryocooler for simplified installation and operation

• Gravity acceleration limits speed to < 50 m/s

• Low speed and small pellet size minimize fueling, but make strong enough density perturbations to trigger ELMs

0.7-1.3 mm @ 50-100 Hz

SolenoidCutterD2

Ice


ITER Pellet Injector Pumping System

• The mass loss rate is on the order of ~12 Pa-m3/s (90 torr-L/s) per pellet injector when operated at the full fueling rate. (~10% of fueling rate)

• Assuming an operating pressure of 0.1 Pa (1 mtorr) implies a required pumping speed of 120,000 L/s.

• A continuously regenerating cryopump, i.e. a snail pump, is an ideal pumping scheme as the output can be fed back in the fuel input stream. The snail pump by CAF, Inc. (C. Foster) is the most promising technology for this task.

• Prototypes of this pump have been developed with recent tests at LANL achieving > 120,000 L/s pumping speed for D2 (S. Willms, C. Foster, et al.)

LHe

Inlet

Exhaust to blower

Snail in operation


ITER Inner Wall 6mm PelletDeposition as a Function of Pedestal Te

0.0

1.0

2.0

3.0

4.0

5.0

6.0

0 0.2 0.4 0.6 0.8 1

Tped = 1 keVTped = 2 keVTped = 4 keV

n

e (

101

9 m-3)

• Pellet fueling deposition calculations from PRL for ITER with different size pellets. Larger pellet size yields marginally deeper mass penetration.

• Pellets injected into the same discharge conditions (H-mode, Te(0) = 20 keV, Tped = 4 keV, ped=0.04)



0

2

4

61.0

1.5

2.0

Div

ert

or

D

(a

.u)

neL

(1

014

cm

-2)

3.6 3.7 3.8 3.9 4.0 4.1

1.8mm V+3 pellets

Vertically Injected Pellets on DIII-D from V+3 Port Trigger ELMs with Minimal Density Perturbation

• 1.8 mm pellets injected from the DIII-D V+3 port trigger ELMs with minimal fueling of the core plasma. Pellets injected at ~200 m/s through a curved guide tube.


Coldhead

Extruder

• Extruders for long-pulse pellet injectors are usually cooled with LHe (dewars or cryogenic plant)

• Pellet injectors with cryocooler ice formation offers some advantages – low operating cost after initial investment (up to several $k per day for laboratory operation

with LHe)

– independent from LHe dewars or facility cryogenic systems

• Both piston and screw extruders have been successfully mated to cryocoolers. Extrapolation to higher throughput extruders for ITER is under development. (Pfotenhauer, Combs, Viniar )

H2 Extrusions Achieved with Cryocooler Technology


Multiple Extruders Operated in Sequence Provided a Steady-State Supply of H2 Ice for Feeding a Pellet Injector

• ITER EDA prototype testing at ORNL showed that batch piston extruders could be employed for the needed pellet source.

• With the three-extruder prototype, a rate of 1.3 cm3/s was produced and ≈0.33 cm3/s was maintained for a duration of 1 hour (~1/4 of rate needed for 5mm@16Hz).

• Operation of batch piston extruders at > 1 cm3/s is possible with 4-6 such extruders operating in tandem. 8mm H2 ice extrusions

(S. Combs, et al, RSI (1998).

LHe

OFHC Copper


• Pellet injection through the central solenoid as shown in this ITER cross-section has been suggested. (F. Perkins, PPPL, see Poster EP1.114 Tuesday morning)

• The impetus for this idea is to maximize penetration with faster pellets.

• Penetration distance of pellet will scale linearly in velocity in ITER pedestal instead of the customary v0.3.

Inner Bore Pellet Injector for ITER?


ITER Inner Wall 6mm Pellet ComparisonInner Bore vs Curved Guide Tube

• Higher speed inner wall pellet injection made possible by placing the injector inside the central solenoid (F. Perkins, Poster EP1.114 ).

• PRL calculations show a deeper fuel penetration for the same size pellets assuming > 3x the speed.

0.0

1.0

2.0

3.0

4.0

5.0

0 0.2 0.4 0.6 0.8 1

Vp = 0.3 km/sVp = 1.0 km/s midplane

n

e (

101

9 m-3)

O AK R IDGE N ATIONAL L ABORATORY U.S. D EPARTMENT OF E NERGY O AK R IDGE N ATIONAL L ABORATORY U.S. D EPARTMENT OF E NERGY O AK R IDGE N ATIONAL L ABORATORY.

Documents

lrb o ak r idge n ational

diiid cross section

iter plasma volume

significant iter cross

iter international team

iter pellet injection

high density

layer pellet