The Advanced Tokamak:

Goals, prospects and research opportunitiesAmanda Hubbard

MIT Plasma Science and Fusion Centerwith thanks to many contributors, including A. Garafolo, C. Greenfield,

C. Kessel, D. Meade, M. Murakami, F. Najmabadi, T. TaylorOpinions are my own…

GCEP Fusion Energy Workshop on Opportunities for Fundamental Research and

Breakthrough in Nuclear FusionPrinceton, NJMay 1-2 2006

The Advanced Tokamak

• Introduction: What is an ‘advanced tokamak’?

• The AT vision for fusion energy– Drawing heavily on ARIES studies.

• Current results and near-term prospects– Focusing here on US program.

• AT on ITER: What we will (and won’t) learn.

• Research Opportunities: ideas to advance and accelerate fusion energy prospects. To start the discussion….

An “advanced tokamak” device is, in terms of magnetic configuration, simply a TOKAMAK

Pure toroidal field does not confine charged particles

Adding poloidal field does confine charged particles. Produced by toroidal current.

Tokamak needs a toroidal current for stability.Current conventionally driven by tranformer;- Current is driven around central solenoid.Inherently NOT steady-state.

Tokamaks lead other configurations in fusion performance, are approaching ‘breakeven’

ITER

D. Meade, ARIES workshop 4/24/05

“Conventional” tokamak operation will be primary mode of operation on ITER

• Heating applied mainly on-axis, inductive current drive, profiles relax to ‘natural’ state.

• Much experience worldwide, good confidence in extrapolation to burning plasma conditions.– This will allow critical exploration of burning plasma physics.

• Could probably be used to make a fusion power plant. – Advantages of relative simplicity, staying away from

performance limits. – BUT projected power plant not seen as economically

attractive (at least in prior assessments with low cost oil!)

Tokamak current does not have to be driven by a transformer!

• Alternative means of current drive:– External current drive, by neutral beams, or microwaves (various

ranges from ion cyclotron (~100 MHz), Lower Hybrid (~ 5 GHz), electron cyclotron (~100 GHz))

– “Bootstrap current”: Self-generated current due to temperature, density, pressure gradients in the plasma.

• All of these are fairly well understood, and have been demonstrated to work on many experiments.– Gives potential for steady-state operation.

• “The crucial distinguishing feature of an Advanced Tokamak over a conventional tokamak is …the use of active control of thecurrent or shear profile, and of the pressure profile or transport characteristics” (AT Workshop, GA, 1999)– Same tokamaks can (and do) operate in both conventional

and advanced regimes.

Vacuum VesselDoor

Vacuum Vessel

PF Coils

Central SolenoidTF Coils

Cryostat

MaintenancePort

First Wall and Blanket

Hardback Structure

Diverter Region

High Temperature Shield

Low TemperatureShield

SuperconductingToroidal Magnets

Cryostat

MaintenancePort

First Wall and Blanket

Hardback Structure

Divertor Region

High Temperature Shield

Vacuum Vessel

Vacuum VesselDoor

PF Coils

Central Solenoid

OPTIMIZATION OF THE TOKAMAK CONCEPTLEADS TO AN ATTRACTIVE FUSION POWER PLANT

● The U.S. ARIES system study

● Optimization of the tokamakconcept is known as theAdvanced Tokamak program

● Attractive features

— Reduced sizeHigher pressure, reduced heat loss

— Improved economics— Improved power cycle

Low Activation

130–02/TST/wj

Power cycle Pulsed Steady-state

COE ¢/kWhr ~13 ~7

Major radius (m) 8 5

Conventional Optimized

NATIONAL FUSION FACILITYS A N D I E G O

DIII–D 130–02/TST/wj

THE GOAL OF THE ADVANCED TOKAMAK PROGRAMIS TO OPTIMIZE THE TOKAMAK CONCEPT FORATTRACTIVE FUSION ENERGY PRODUCTION

● Steady state— High self-generated bootstrap current

● Compact (smaller)— Improved confinement (reduced heat loss)

— Discovering the Ultimate Potential of the Tokamak —

Key Elements

Fusion Ignition Requirement }

H = τE/τconv Size∼

E

● High power density— Improved stability

PFus ∝ (n T) 2 Vol ∝ β2 B4 Vol

3×1021 m–3 keV s < n Ti τ ∝ (H a B κ) 2

β = 2 µo ⟨P⟩

B2

SIMULATIONS PREDICT SELF-CONSISTENT EQUILIBRIAWITH NEARLY 100% BOOTSTRAP

➞ Steady state with lowrecirculating power

● Off-axis current drive to supply missing current— Provided by high power

microwaves in DIII–D

● Other benefits of negative central shear profile

— Reduced transport, improved confinement

— Improved stability to centralunstable MHD modes ★ Ballooning★ Tearing modes★ Sawteeth

fBS = 0.92

q

4

3

2

10

Radius 1

⟨J •B⟩ tot

⟨J •B⟩BS

130-02/TST/wjNATIONAL FUSION FACILITYS A N D I E G O

DIII–D

NegativeMagnetic Shear

NATIONAL FUSION FACILITYS A N D I E G O

DIII–D 130–02/TST/wj

NEGATIVE CENTRAL SHEAR AND SHEARED E×B FLOWLEAD TO IMPROVED CORE CONFINEMENT

● Key physics

— Measured turbulence reduction is consistent with theoretical prediction

— Negative magnetic shear contributes to reduced γITG

● Similar reduction is often observed in other transport channels

★ E×B shearing rate exceeds maximum growth rate of ion temperature gradient mode

keV

0 0.2 0.4 0.6 0.8 1.0ρ 0 0.2 0.4 0.6 0.8 1.0ρ0

2

4

6

8

10

keV

0 0.2 0.4 0.6 0.8 1.0ρ0

2

4

6

8

10

105 s

-1

0

1

2

3

4

5

6

+ =

Ti|ωE×B|

γITG

Ti

0.9 s 0.9 s

1.2 s

130–02/TST/wjS A N D I E G O

DIII–DNATIONAL FUSION FACILITY

A COMPACT STEADY STATE TOKAMAKREQUIRES OPERATION AT HIGH βN

● High power density⇒ high βT

● Large bootstrap fraction⇒ high βp

● Steady state ⇒ high βN

1 + κ2

22

Pfus

PCD

cur

nqQ

ss= εeff

1 –

2

( ) B3aκγ

∝βN

ξ √A q βN

βT βp ∝ ( )βN

β

ε 1 + κ2 2

T

βN = 3.5Eq

uilib

rium

Lim

it

q* = 4

βN = 5

Curr

ent L

imit

Pow

er D

ensi

ty

Bootstrap Current

Conventional Tokamak

Advanced Tokamak

Pressure Limit βN ∝ power density ×bootstrap current

βN = βT /(I/aB) εβp

2

Advanced Tokamak concept of fusion power plant.

• Embodied in ARIES design studies, ARIES-RS and ARIES-AT. Japan has similar studies.

• Material courtesy of F. Najmabadi, UCSD

ARIES-AT is an attractive vision for fusion with a reasonable extrapolation in physics & technology

∗ Competitive cost of electricity (5c/kWh);

∗ Steady-state operation;∗ Low level waste;∗ Public & worker safety;∗ High availability.

Approaching COE insensitive of power density

Evolution of ARIES Designs

10Cost of Electricity (c/kWh)

0.46Thermal efficiency

0.29Recirculating Power Fraction

237Current-driver power (MW)

1.5Avg. Wall Load (MW/m2)

16Peak magnetic field (T)

2% (2.9)β (βΝ) Plasma pressure/magnetic p

8.0Major radius (m)

ARIES-IA

1st Stability, Nb3Sn Tech.

Reverse Shear Option

8.2

0.49

0.28

202

2.5

19

2% (3.0)

6.75

ARIES-I

High-FieldOption

7.5

0.46

0.17

81

4

16

5% (4.8)

5.5

ARIES-RS

5

0.59

0.14

36

3.3

11.5

9.2% (5.4)

5.2

ARIES-AT

Our Vision of Magnetic Fusion Power Systems Has Improved Dramatically in the Last Decade, and Is Directly Tied to Advances in Fusion Science & Technology

Estimated Cost of Electricity (c/kWh)

02468

101214

Mid 80'sPhysics

Early 90'sPhysics

Late 90's Physics

AdvanceTechnology

Major radius (m)

0

1

2

3

4

5

6

7

8

9

10

Mid 80's Pulsar

Early 90'sARIES-I

Late 90'sARIES-RS

2000 ARIES-AT

Present ARIES-AT parameters:Major radius: 5.2 m Fusion Power 1,720 MWToroidal β: 9.2% Net Electric 1,000 MWWall Loading: 4.75 MW/m2 COE 5 c/kWh

Estimated range of COE (c/kWh) for 2020*

01234567

Natural Gas Coal Nuclear Wind(Intermittent)

Fusion (ARIES-AT)

AT 1000 (1 GWe)AT 1500 (1.5 GWe)

ARIES-AT is Competitive with Other Future Energy Sources

EPRI Electric Supply Roadmap (1/99):Business as usualImpact of $100/ton Carbon Tax.

* Data from Snowmass Energy Working Group Summary.

Estimates from Energy Information AgencyAnnual Energy Outlook 1999(No Carbon tax).Annual Energy Outlook 2005(2025 COE, 2003$)

Advanced Tokamak Research on current experiments

• What is needed? – key issues

• What results have already been obtained?

• Near-term plans and prospects.

Physics Requirements for Advanced Tokamak

• For STEADY STATE, want– 100% non-induction current drive (external + self-generated

“Bootstrap”

• For low recirculating power, good economics, want – High “bootstrap fraction” 80-90% self-generated.

• To get this, need high normalized pressure, βN. This requires low transport, to get high gradients, which in turn are enabled by optimized current profile. High pressure itself improves economics.

Highly coupled control of current, transport profiles needed –for times long compared to plasma time scales, eg. energy confinement time τE, current relaxation time τCR.

Many of these requirements have been demonstrated in present expts

As examples, show recent results from DIII-D tokamak, San Diego at 2005 APS-DPP meeting (A. Garafolo, Univ. Columbia, M. Murakami, ORNL) and from C-Mod, MITDIII-D results rely heavily on MHD stabilization techniques to reach high β. This important aspect of AT research will be covered this afternoon by G. Navratil.

Other world tokamaks, in particular JT-60U (Japan) and ASDEX-Upgrade (Germany) also have strong AT programs, range of control tools. Will not attempt a comprehensive review here!

Also important work on advanced scenarios in spherical (low aspect ratio) tokamaks NSTX (PPPL) and MAST (UK), which will (I presume) be covered in talk by Martin Peng.

• Negative Central Shear– Stability to high-n ballooning modes and neoclassical tearing modes– Suppression of transport– Good alignment of bootstrap current with total current– Hollow current profile, wall-stabilization of low-n kink modes

Advanced Tokamak Goal is Steady-state OperationCombined with High Fusion Performance

• Steady state operation– 100% non-inductive

current– High βP, high fraction

of bootstrap current

• High fusion gain– High β, high τE

– High normalized fusionperformance:

G = βNH89/q952

Bootstrap current

Fusi

on

po

we

r

Recent DIII-D Experiments Achieved High FusionPerformance at High Bootstrap Current Fraction

• Combination of high confinement,high beta, and high bootstrapfraction sustained for ~2 s

• Multiple control tools needed,including

– Simultaneous ramping ofplasma current and toroidalfield

– Simultaneous FeedbackControl of Error Fields andResistive Wall Mode

• Transport analysis confirmspresence of internal transportbarriers (ITBs) in high β discharges

• Stability analysis indicates potential for higher beta operation• High noninductive current fraction (~100%) has been achieved• Steady-state sustainment will be pursued with new DIII-D tools

High Normalized Beta (βN ~4) Sustained for ~2 s at HighSafety Factor and High Confinement

122004

• βN > 6li for ~2 s

– Relies on wallstabilization of then=1 external kinkmode (conventionalstability limit ~4li)

• High performancephase generallyterminated by currentprofile evolution

– (m,n) = (2,1) tearingmode

High β Discharge Profiles Show NCS and ITBs

• Strong gradientstypical of ITBs areobserved in Ti, ne

and rotationprofiles, but notin Te

• Pressure peakingfactor, P(0)/<P>varies in range2.6-3.2 duringhigh beta phase

• P(0)/<P>=2.9

• fGW~ 0.4-0.6

High βN Discharge with Constant Plasma Current DrivenNoninductively for ~0.5 s

• Surface voltage< 0 after Ip ramp ends

• Internal loop voltageprofile showsnoninductive currentfraction ≥100 %, althoughnot fully relaxed

100 mstriangularsmoothing

Current drive analysis

211-05/mm/jy

With Improved Confinement, fni=100% Achieved

with Good CD Alignment

• Equilibrium measurement: Jind = neoE|| neo pol/ t fNI = 1 – find

• find = 0.5%, fNI = 99.5%

• Inductive current is locally & globally close to zero

NI current aligned well to “desired” Jtot

“good CD alignment”

• T= 3.5%, N = 3.6, q95 = 5.0 G = N H89/q952 = 0.3

ITER steady state scenario requirements satisfied

0.0 0.2 0.4 0.6 0.8 1.0–50

0

50

100

150

RADIUS, ρ

Flu

x S

urf

ac

e A

ve

rag

ed

To

roid

al

Cu

rre

nt

De

nsi

ty (A

/cm

)2 ⟨J(ρ)⟩

(Eq. Measurement)

Jtot

Jind

Loc

al to

roid

al c

urr

en

t d

en

sity

(A

/cm

)2 200

150

100

50

01.6 1.8 2.0 2.2

Midplane major radius, R (m)

2.4

Jφ(R)MSE Array

Tangential Radial Edge

Measurement

211-05/mm/jy

Transport Code Carries Out Data Analysis Based on

Equilibrium Reconstruction with Kinetic Profile Information

• Measurements: find = 0.5%, fNI = 99.5%

• Analysis shows: fBS=59% fNB=31% fEC= 8% fNI= 98%

• Equilibrium reconstruction (EFIT) lacks spatial resolution

Makes the current balance calculations problematic

0.0 0.2 0.4 0.6 0.8 1.0–50

0

50

100

150

Radius, ρ

Flu

x S

urf

ac

e A

ve

rag

ed

To

roid

al

Cu

rre

nt

De

nsi

ty (A

/cm

)2 120096F05⟨J(ρ)⟩

Jtot

JEC

JbsJNB

Loc

al to

roid

al c

urr

en

t d

en

sity

(A

/cm

)2

200

150

100

50

01.6 1.8 2.0 2.2

Midplane major radius, R (m)

2.4

⟨J J ⟩ (calc.) (calc.)ECEC

Jφ(r)

MSE Array Tangential Radial Edge Analysis (EFIT)

Analysis

1040309029

0.70 0.75 0.80 0.85 0.90Midplane Major Radius [m]

0

1

2

3

4

5

TS datat=0.761s H-modet=1.094s ITBt=1.276s ITB with on-axis heating

TS data

Ele

ctr

on

De

nsit

y (

10

20

m-3

)

Internal transport barriers, and core transport control, have been produced in C-Mod with normal shear, by

varying heating profile• OFF-axis heating alone causes

density peaking, ~ const T.• ON-axis heating clamps n, but

increases T, neutron rate.

Elec

tron

Pres

sure

(MPa

scal

s)

0.15

0.10

0.05

0

0.20

0.0 0.2 0.80.60.4 1.0r/a

RF

Pow

er D

ensi

ty (M

W/m

)

3

10

0

20

30

1.5 MW central ICRFadded into fully formed ITB

ITB, 2.35 MW Off-axis ICRF

H-mode, No ITB

t=1.294 s

t=1.127s

t=0.894 s

0.0 0.2 0.80.60.4 1.0r/a

On-axis + off-axis, 4 MWtotal rf power at t=1.3 sOff-axis alone, 2.3 MWtotal rf power at t=1.1 s

Levels of heat and particle diffusivity can be reduced to neoclassical, or increased to stabilize density and impurities.

ne

Te

1040309029

0.70 0.75 0.80 0.85 0.90Midplane Major Radius [m]

0.0

0.5

1.0

1.5

Ele

ctr

on

Tem

pera

ture

(keV

)

TS dataEdge ThomsonGPC2 dataFRC data

t=0.794s H-mode

t=1.094s ITB

t=1.261s ITB with added central rf power

2.0

Issues and Near-term plans for Advanced Tokamak Research

• While much has been achieved, much more remains to be done to realize potential of advanced regimes on burning plasmas, and fusion reactors.

• Most scenarios are still non-stationary (t < τCR) , and/or rely on current profile control techniques (eg, tailored heating during current rampup, central NBI) which don’t extrapolate to steady state.

• Most present experiments have plasma conditions quite different from burning plasmas. Eg.– Uncoupled (τe-i >> τΕ ) vs coupled (τe-i << τΕ ) electrons and ions

(lower vs higher density)– Core particle and momentum sources (vs RF, alpha htg.)– Both factors strongly affect transport barrier formation.

• Handling of high heat loads in divertor – common to all attractive configurations and will be covered in talk by Mike Ulrickson.

Experiments at Higher qmin and Higher βN

Will Address Steady-state Demonstration

• New divertor and improved density control will slow down qminevolution– By allowing higher temperature at lower density– By allowing higher ECCD at lower density

• Additional ECCD power will improve current profile control

• Higher βN at higher qmin will give higher bootstrap current– Will reduce Ohmic current at large radii– Will overdrive at small radii

• Compensate overdrive using ECCD, FW, Counter NBI

Advanced Tokamak Research on C-Mod

• AT research is an increasingly important focus on C-Mod, which is a compact (R=.68 m), high B (5-8 T), high ne (1020-1021 m-3) tokamak.

• Unique among world divertor tokamaks, can test AT physics and scenarios– At ITER field and density (key wave physics parameters).– Without core particle or momentum sources (all RF heating)– Strongly coupled ions and electrons (τe-i << τΕ )– Pulse lengths >> current relaxation times, routinely.

(ie., steady-state, relaxed j(r)).– ITER-level divertor fluxes.

Important challenge and test: Will AT regimes scenarios work as well in these conditions, typical of ITER and reactors??

• Program focuses on control of current and magnetic shear as well as transport and kinetic profiles with various shear profiles.– RF systems (ICRF +LHCD) provide key control tools.

LHCD is highest efficiency technique for current drive far off axis.– Also adding new cryopump, important for density control.– In near term, rely on shape and profile control to maximize no-wall

limit βN~3. Longer term, would like to add active stabilization.

New LHCD system on Alcator C-Mod

12 klystrons × 0.25 MW = 3 MW

@ 4.6 GHz

Transmission waveguides

Coupler grill – MIT/PPPL collab’n.

96 waveguide outlets, allowing flexible phasing to launch spectrum

Initial experiments in progress and first, significant, LHCD recently seen!

Example of non-inductive AT target scenario on C-Mod

• One of many optimized scenarios modelled with ACCOME.– ILH=240 kA– IBS=600 kA (70%)

J (M

A /

m2 )

r / a

Ip = 0.86 MA Ilh = 0.24 MA fbs = 0.7

r / a

Saf

ety

Fac

tor

- q(

r)

q(0) = 5.08

qmin = 3.30

q(95) = 5.98

• Double transport barrier • BT=4 T• ICRH: 5 MW• LHCD: 3 MW, N//0=3• ne(0)= 1.8e20 m-3

• Te(0)=6.5 keV (H=2.5)• βN=2.9

Scenarios without barrier, or only an ITB, have similar performance.

P. Bonoli, Nucl. Fus. 20(6) 2000.

Advanced Scenarios on ITER

As a burning plasma experiment, ITER will explore a range of physics parameters and scenarios. To guide planning, currently focusingon three main target scenarios, all at BT=5.3 T. Still some flexibility/uncertainty in sources, parameters.

TRANSP/TSC simulation of ITER S-S scenario.Houlberg&ITPA, IAEA04.

j(x)1. Conventional H-mode: “Baseline” Scenario, Q=10.Positive shear, q95=3, βN=1.8, HH~1, n~1020 m-3. fNI~ 0.25.

MAIN ITER GOAL!

2. “Hybrid” Scenario: Q=10Weak core shear, q95=4, qmin~1, βN=2.8, HH~1.2, fNI~ 0.5.

3. Steady-state: Q=5, long pulseWeak or negative shear, q95~5, qmin~2, βN=3.0, HH~1.2, fNI~ 1.0.

SECONDARY GOAL

What we can (and can’t) expect from ITER

• Demonstration of advanced, high non-inductive scenarios on ITER would be an extremely important step towards an AT DEMO reactor!– Would resolve many uncertainties about applicability with BP-

relevant plasma parameters and control tools, as outlined above.– Would start addressing key control issues with self-heating.

• A key goal of the current US program is to conduct research that will support AT on ITER, and to push for needed hardware.

BUT…• Because steady-state mission on ITER is secondary, it is NOT an

optimized machine for AT. For example,– Shaping flexibility is limited.– Heating and current drive likely underpowered.

• Much depends on hardware decisions not yet made, eg.– Will ITER have LHCD, needed for off-axis CD? How much?– Will ITER have active control coils to reach highest β ?– US will be pushing for these, but we won’t call all the shots!

Coming year or two is critical.

Research Opportunities

• What are key questions/topics which would take advanced tokamak from interesting and attractive scientific research to afusion energy source?

• Which are NOT likely to be funded in near-term US-DOE program?

• As a general principle, I assume that issues of direct application to ITER will likely get priority, and (hopefully) adequate funding.

• More general – but still important – issues or those aiming at steps beyond ITER, and fusion energy application, are likely to get less resources.

“CONTROL” ISSUES TOP MY LIST

• A fusion reactor or DEMO would need to run for long periods, close to stability limits, and with all profiles well optimized and controlled.

• In high-bootstrap scenario, these are tightly coupled– Current profile derived mainly from n, T profiles, which in turn depend

on both sources and transport. MANY interactions!– Limited external control of j(r). How much is needed?– Pressure/current profiles need to be aligned for stability.

• Heat mainly coming from fusion burn – reduces controllability.

• I would like to see a more focussed effort on demonstrating active, integrated profile control – not just ‘tailoring’ of profiles applicable to a specific machine or experiment (eg by adjusting heating times in rampup). – This likely won’t ‘just happen’, even with ITER coming.– Would benefit from an interdisciplinary approach, from plasma physics

experimentalists and theorists, plus engineering, controls, power systems experts.

D. Moreau IEA W60 Burning Plasma Physics and Simulation, Tarragona, July 2005

Actuators and nonlinear couplings in abootstrap-dominated steady state burning plasmas

Figure from P. Politzer et al., ITPA meeting Lisbon 2004

Can we develop a transport control tool?

• Part of the difficulty in profile control is the indirect nature of controlling temperature and density profiles.– We control heat and particle sources and, if we are fortunate,

current profiles. – Plasma transport determines T, n profiles. While there is good

progress in understanding transport (to be covered in talks by Tynan, Dorland), it is highly complex, with gaps in our knowledge; not easily amenable to control algorithms! χ D have been shown to be affected by j(r), by heating profile, by shear flow…

– The goal is NOT minimum transport, but optimum transport –otherwise pressure limits exceeded, impurities and ash accumulate.

• A “holy grail” of transport and control research is an active control tool for transport, independent of heat sources. Best hopes are for RF tools, which could eg. drive flow shear, modifying χ at specific location.

• There are ideas (eg, Ion Bernstein Waves), but currently not a focused experimental and theoretical effort. – Could I think be done with modest funding, including expt-theory

collaboration and small-scale lab tests.– Would be high leverage for AT fusion development.

Other pressing issues…

Most of these will I expect come up in other talks, and are active areas of international fusion research

• Improved divertor solutions and materials for steady state, reactor-level heat fluxes.

• Compatibility of core and edge plasmas in advanced modes (closely related to divertor issue)

• MHD control tools for sustaining high beta, suppressing code instabilities.

• Disruption avoidance and mitigation. • Extension to longer pulse lengths – EAST and KSTAR will

play important roles, though experience suggests it will take several years to develop needed AT tools.

Summary: The Advanced Tokamak

• The “advanced tokamak” is a tokamak operational scenario characterized by high degree of control of current and pressure profiles.

• Leads to optimized fusion reactor designs (eg ARIES-AT, RS) which are steady-state, have low recirculating power and lower size and cost than conventional tokamaks.– Extrapolates to competitive cost of electricity.

• Current experiments have demonstrated many key needed features, such as high β, reduced transport and non-inductive current drive.– Near-term research aims at extending such results to steady-state

and demonstrating in more burning-plasma relevant conditions.• ITER will be an important test of advanced scenarios in a burning

plasma! But, this is a secondary mission and the device may not have optimal design and tools.

• Further research is needed to go from current experiments to confidence in an advanced tokamak DEMO reactor, in particular more tools, understanding and experience in active control.