Home Search Collections Journals About Contact us My IOPscience

Chapter 1: Overview and summary

This article has been downloaded from IOPscience. Please scroll down to see the full text article.

2007 Nucl. Fusion 47 S1

(http://iopscience.iop.org/0029-5515/47/6/S01)

View the table of contents for this issue, or go to the journal homepage for more

Download details:
IP Address: 93.180.53.211
The article was downloaded on 07/07/2013 at 09:47

Please note that terms and conditions apply.

IOP PUBLISHING and INTERNATIONAL ATOMIC ENERGY AGENCY NUCLEAR FUSION
Nucl. Fusion 47 (2007) S1–S17 doi:10.1088/0029-5515/47/6/S01

Chapter 1: Overview and summary

Editors of ‘Progress in the ITER Physics Basis’:
M. Shimada1,a , D.J. Campbell1,2 , V. Mukhovatov1 , M. Fujiwara3 ,
N. Kirneva4 , K. Lackner5 , M. Nagami6 , V.D. Pustovitov4 ,
N. Uckan7 and J. Wesley8
International Tokamak Physics Activity Topical Group Chairs,
Cochairs and Chapter Coordinators:
N. Asakura6 , A.E. Costley1 , A.J.H. Donné9 , E.J. Doyle10 ,
A. Fasoli11 , C. Gormezano12,13 , Y. Gribov1 , O. Gruber5 ,
T.C. Hender14 , W. Houlberg7 , S. Ide6 , Y. Kamada6 , A. Leonard8 ,
B. Lipschultz15 , A. Loarte2 , K. Miyamoto13,16 , V. Mukhovatov1 ,
T.H. Osborne8 , A. Polevoi1 and A.C.C. Sips5
1
ITER Organization
2
European Fusion Development Agreement Close Support Unit-Garching
3
National Institute for Fusion Studies, Japan
4
Kurchatov Institute, Russia
5
Institut für Plasmaphysik, Garching, Germany
6
Japan Atomic Energy Agency, Naka, Japan
7
Oak Ridge National Laboratories, USA
8
General Atomics, USA
9
FOM-Institute for Plasma Physics Rijnhuizen, Netherlands
10
University of California, Los Angeles, USA
11
Centre de Recherches en Physique des Plasmas, Ecole Polytechnique Fédérale, Switzerland
12
Ente per le Nuove tecnologie, l’Energia e l’Ambiente, Italy
13
Retired
14
United Kingdom Atomic Energy Authority, Abingdon, UK
15
Plasma Science and Fusion Center, Massachusetts Institute of Technology, USA
16
University of Tokyo, Japan
E-mail: [email protected]

Received 6 November 2006, accepted for publication 30 April 2007

Published 1 June 2007
Online at stacks.iop.org/NF/47/S1
Abstract
The ‘Progress in the ITER Physics Basis’ (PIPB) document is an update of the ‘ITER Physics Basis’ (IPB), which
was published in 1999 [1]. The IPB provided methodologies for projecting the performance of burning plasmas,
developed largely through coordinated experimental, modelling and theoretical activities carried out on today’s large
tokamaks (ITER Physics R&D). In the IPB, projections for ITER (1998 Design) were also presented. The IPB also
pointed out some outstanding issues. These issues have been addressed by the Participant Teams of ITER (the
European Union, Japan, Russia and the USA), for which International Tokamak Physics Activities (ITPA) provided
a forum of scientists, focusing on open issues pointed out in the IPB. The new methodologies of projection and
control are applied to ITER, which was redesigned under revised technical objectives. These analyses suggest that
the achievement of Q > 10 in the inductive operation is feasible. Further, improved confinement and beta observed
with low shear (= high βp = ‘hybrid’) operation scenarios, if achieved in ITER, could provide attractive scenarios
with high Q(> 10), long pulse (>1000 s) operation with beta <no-wall limit and benign ELMs.
PACS numbers: 28.52.−s, 52.55.Fa, 52.55.−s, 52.40.Hf
(Some figures in this article are in colour only in the electronic version)

a Author to whom any correspondence should be addressed.

0029-5515/07/060001+17$30.00 © 2007 IAEA, Vienna Printed in the UK S1

M. Shimada et al

Contents
1. Introduction
2. ITER objectives and capabilities
3. Progress in key physics issues and its impact on the choice of the main design parameters of ITER
3.1. Core confinement and transport
3.2. Core confinement, transport and steady-state operation with weak or negative magnetic shear
3.3. Particle and impurity transport
3.4. Edge pedestal and ELMs
3.5. Stabilization of neoclassical tearing modes (NTMs)
3.6. Feasibility of sustained operation above the no-wall ideal MHD beta limit
3.7. Feasibility of disruption mitigation using massive gas injection
3.8. Digital plasma control systems
3.9. Particle control and power dispersal
3.10. Energetic particle physics
3.11. Diagnostics
4. Summary
Appendix A.
A.1. ITER
A.2. Operation scenarios and phases

1. Introduction The present ITER design stems from more than 15 years of
joint magnetic fusion reactor design activities and supporting
The objective of ITER is to demonstrate the scientific physics and technology research by the four original ITER
and technological feasibility of fusion energy for peaceful parties—the European Union (EU), Japan (JA), the Russian
purposes. Its inductive operation is expected to produce Federation (RF) and the United States of America (US)—who
significant fusion power (∼500 MW) through the D–T reaction are now joined by China, India and Korea in the collaborative
with high fusion gain Q ∼ 10 (the ratio of fusion power to international effort to construct the ITER tokamak device. For
the external heating power) for 300–500 s. In other words, the 3 years Canada was also a partner. Present plans call for first
majority of the heating power will be provided by alpha particle operation of ITER to commence 8.5 years after the start of
heating and the plasma largely determines its own profiles. construction.
ITER will be the first device in which this autonomous plasma Physics understanding and methodologies of projection
state is achieved. ITER will also aim at steady-state, high and control of such burning plasmas must be based on
gain operation lasting for ∼3000 s. In steady-state plasmas, a experimental, modelling and theoretical research. The
large fraction (>50%) of the plasma current will be driven by Progress in the ITER Physics Basis (PIPB) that follows in
spontaneous bootstrap current originating from the pressure chapter 2-9 of this issue consists of updates to the physics basis
gradient, which enhances the degree of plasma autonomy. for a burning plasma tokamak that was presented in the ITER
Impurity levels are also determined self-consistently by plasma Physics Basis (IPB) [1]. As the title of that document implies,
processes, such as sputtering, screening, transport and radiative the content of the IPB comprises an extensive compilation of
cooling: for example, excessive impurity levels would limit the the physics basis for the design and operation of a burning-
fusion reaction through dilution and radiative cooling, which plasma-capable tokamak, specifically the 1998 embodiment
would then reduce the impurity concentration to an equilibrium [2,3] of the International Thermonuclear Experimental Reactor
level. Since helium particles (alpha particles) are created by (ITER). Physics research after the IPB has been carried out
the fusion reactions, the determination of impurity levels will by four participating parties (EU, JA, RF and US). The
be more complex in fusion plasmas. Engineering tests of International Tokamak Physics Activity (ITPA) provided an
reactor-relevant components, such as breeding blankets, are excellent forum focusing on open issues pointed out in the
also an important mission of ITER, and require a reliable IPB, which has made significant progress. The PIPB provides
operation scheme with significant fusion power and long pulses an update—focusing on progress obtained since 1998—to the
(>1000 s). Projection studies show that such scenarios are physics basis considerations identified in the IPB and presents
possible with a modest requirement on confinement and beta application of those considerations to the current embodiment
(hybrid scenarios, which combine inductive and non-inductive of the ITER design [4]. It also discusses the role of ITER in
current drive at a plasma current lower than the inductive the strategy of fusion reactor development (chapter 9 of this
scenarios). While the physics basis for ITER’s nominal issue [5]).
inductive operation is relatively well established, the projection This chapter will provide an introduction and outline
of plasma performance is associated with some uncertainty, of the complete PIPB document. The mission, objectives
due to extrapolation of parameters to regimes unattainable and technical attributes of the design are described in detail
in present machines, and experiments in ITER are needed to in [4] and will be shown briefly in section 2 and described
lay the foundation for the operation of demonstration power in greater detail in appendix A. In section 3, the progress
reactors (Demo) that will follow ITER. in the understanding of physical processes and methodologies

S2
 Chapter 1: Overview and summary

for projection and control since the IPB are summarized. In density, which improved the feasibility of achieving ITER’s
section 4, conclusions are presented. goals. But since ITER will operate at Q ∼ 10 and fusion
power of ∼500 MW, there will still be an extrapolation to
2. ITER objectives and capabilities power reactors (Q ∼ 50 and fusion power of ∼3 GW).
ITER is based on the tokamak concept [7]. A combination
The objective for ITER is to demonstrate the scientific and of an externally-generated toroidal magnetic field and the
technological feasibility of applying fusion energy for peaceful poloidal magnetic field, generated by toroidal current flowing
purposes. In more direct terms, ITER will, for the first time, be in the plasma and coils wound toroidally around the torus,
able to produce a ‘burning’ deuterium–tritium plasma—that is forms a configuration of nested magnetic flux surfaces that is
one where the majority of the heating needed to sustain the capable of stably supporting the 0.7 MPa (= 7 atm) thermal
fusion reaction is self-produced from fusion-generated alpha pressure required for a self-heated 500 MW DT plasma. The
particles. The production and control of such a self-heated ITER design incorporates superconducting magnet systems,
plasma has been the long-standing goal for more than 50 years water-cooled divertor and plasma-facing first wall and nuclear
of magnetic fusion research. shield systems that are magnetically and thermally capable of
In 1998, six years of joint work originally foreseen supporting long pulse burning plasma operation (burn pulse
under the ITER engineering design activities (EDA) agreement durations of 7 min or more). Figure 1 shows a cut-away
culminated in a design [2] fulfilling all objectives and the depiction of the ITER device. Table 2 summarizes the main
cost target adopted by the ITER parties (the European Union, design features and operational capabilities.
Japan, Russia and the US) in 1992 at the start of the EDA. The ITER facility will provide the capabilities for
However, for financial reasons, the ITER parties recognized achieving sustained fusion burn in DT plasma with fusion
the need of a new design to meet revised technical objectives energy gain Q in the range Q ∼ 5–20. This range of
and a cost reduction target of about 50% of the previously Q, which corresponds to an alpha-heating fraction fα =
accepted cost estimate. The joint central team and home Q/(Q+5) of 50–80%, spans the range of fα needed for physics
teams elaborated revised technical objectives. The revised studies of self-heated plasmas to the higher fα projected
performance specifications adopted by the ITER Council in to be needed for a future power-producing tokamak fusion
June 1998 [6] are set out in full in table 1; in summary they reactor.
require ITER: With ITER, it will be possible to investigate a wide
• to achieve extended burn in inductively driven deuterium– spectrum of new phenomena arising from the full nonlinear
tritium (DT) plasma operation with Q
10, not precluding interplay between α-particle heating, transport, stability,
ignition (i.e. the plasma is sustained by fusion reactions pressure and current profile control, and their compatibility
only and without auxiliary power injected), with a burn with a divertor and plasma-facing materials in steady-state
duration of between 300 and 500 s; conditions. These new phenomena could become more
• to aim at demonstrating steady-state operation using non- complex due to plasma autonomy in plasma profiles, current
inductive current drive with Q
5. density and impurity levels as discussed earlier. These
new phenomena will excite high academic interest and
In terms of engineering performance and testing, the design could pose challenges. However, through investigation of
should: new phenomena, physics understanding will progress and
• demonstrate availability and integration of essential fusion operational regimes attractive for a reactor could be developed.
technologies, The operational capabilities of ITER will ultimately
• test components for a future reactor and depend on the plasma performance that can be obtained and
• test tritium breeding module concepts; with a 14 MeV on the degree that the plasma current can be sustained by
neutron power load on the first wall
0.5 MW m−2 and non-inductive means, i.e. other than continuing to increase the
fluence
0.3 MWa m−2 . magnetic flux supplied by the central solenoid (CS). Present
In addition, the device should: plans envision that ITER will be capable of operating with three
possible modes of plasma operation (commonly described
• use as far as possible technical solutions and concepts as operation ‘scenarios’) that will encompass an increasing
developed and qualified during the previous period of the degree of non-inductive current drive capability and hence
EDA and achievable burn pulse duration. Table 3 details some of
• cost about 50% of the direct capital cost of the 1998 ITER the physics and operational attributes of these three types of
design. operation scenario.
The new ITER design, whilst having reduced technical The physics considerations that underlie the design of
objectives from its predecessor, will nonetheless meet these three categories of scenarios are fully explained in
the programmatic objective of providing an integrated various ITER design documents and publications and are
demonstration of the scientific and technological feasibility further addressed in chapters 2–8 of this issue. But briefly
of fusion energy. Further, the standard operation regime of put, as table 3 demonstrates, the progress from limited-duration
ITER of the 1998 design assumed electron densities higher inductively driven burn in the 15 MA ‘ELMy H-mode’ scenario
than the Greenwald density. This raised a concern, since many to the indefinitely sustainable burn possible in the steady-
tokamaks exhibited deterioration of energy confinement as state scenario is marked by a progressive decrease in total
the density approached the Greenwald density. The standard plasma current, a corresponding progression in non-inductive
operation regime of the new ITER is at or below the Greenwald current drive fraction, increasing normalized confinement

S3
M. Shimada et al

Table 1. ITER detailed technical objectives and performance specifications (from the ITER Special Working Group Report to the ITER
Council on Task #1 results [6]).
Plasma performance
The device should:
• achieve extended burn in inductively driven plasmas with the ratio of fusion power to auxiliary heating power of at least 10 for
a range of operating scenarios and with a duration sufficient to achieve stationary conditions on the timescales characteristic of
plasma processes;
• aim at demonstrating steady-state operation using non-inductive current drive with the ratio of fusion power to input power for
current drive of at least 5.
In addition, the possibility of controlled ignition should not be precluded.
Engineering performance and testing
The device should:
• demonstrate the availability and integration of technologies essential for a fusion reactor (such as superconducting magnets and
remote maintenance);
• test components for a future reactor (such as systems to exhaust power and particles from the plasma);
• test tritium breeding module concepts that would lead in a future reactor to tritium self-sufficiency, the extraction of high grade
heat and electricity production.

Design requirements
• Engineering choices and design solutions should be adopted which implement the above performance requirements and make
maximum appropriate use of existing R&D database (technology and physics) developed for ITER;
• The choice of machine parameters should be consistent with margins that give confidence in achieving the required plasma and
engineering performance in accordance with physics design rules documented and agreed upon by the ITER physics expert groups
(predecessor of ITPA Topical Groups);
• The design should be capable of supporting advanced modes of plasma operation under investigation in existing experiments, and
should permit a wide operating parameter space to allow for optimizing plasma performance;
• The design should be confirmed by the scientific and technological database available at the end of the EDA;
• In order to satisfy the above plasma performance requirements an inductive flat-top capability during burn of 300–500 s, under
nominal operating conditions, should be provided;
• In order to limit the fatigue of components, operation should be limited to a few tens of thousands of pulses;
• In view of the goal of demonstrating steady-state operation using non-inductive current drive in reactor-relevant regimes, the
machine design should be able to support equilibria with high bootstrap current fraction and plasma heating dominated by
α-particles;
• To carry out nuclear and high heat flux component testing relevant to a future fusion reactor, the engineering requirements are
average neutron flux
0.5 MW m−2
average fluence
0.3MWam−2 ;
• The option for later installation of a tritium breeding blanket on the outboard of the device should not be precluded;
• The engineering design choices should be made with the objective of achieving the minimum cost device that meets all the stated
requirements.

Operation requirements
The operation should address the issues of burning plasma, steady-state operation and improved modes of confinement and testing of
blanket modules.
• Burning plasma experiments will address confinement, stability, exhaust of helium ash and impurity control in plasmas dominated
by α-particle heating;
• Steady-state experiments will address issues of non-inductive current drive and other means for profile and burn control and for
achieving improved modes of confinement and stability;
• Operating modes should be determined having sufficient reliability for nuclear testing. Provision should be made for low-fluence
functional tests of blanket modules to be conducted early in the experimental programme. Higher fluence nuclear tests will be
mainly dedicated to DEMO-relevant blanket modules in the above flux and fluence conditions;
• In order to execute this programme, the device is anticipated to operate over an approximately 20-year period. Planning for
operation must provide for an adequate tritium supply. It is assumed that there will be an adequate supply from external sources
throughout the operational life.

(H98(y,2) factor, the energy confinement time normalized by with improved confinement of energy and particles. It is
the H98(y,2) scaling), increasing normalized plasma beta (βN ) characterized by a steep gradient in plasma pressure at a radial
and burn time. Here, β is the plasma pressure normalized zone (called edge pedestal), typically several cm wide inside
by magnetic pressure. The normalized plasma beta βN is the separatrix surface. Its reduced transport is explained by
the beta normalized by the Troyon scaling (Ip /(aB) in %, Ip E × B shear. It is often associated with periodical bursts of
(plasma current) in MA, a (plasma horizontal minor radius) energy and particle due to instability (edge localized mode,
in m and B (toroidal magnetic field) in T). The ITER steady- ELM), localized around the edge pedestal. The density and
state scenario can be represented as being prototypical of temperature at the pedestal top provide boundary conditions
the steady-state scenarios being advocated for future fusion of the core plasma, characterizing the plasma performance in
power reactor designs. ‘ELMy H-mode’ is a discharge mode the case the profile is stiff.

S4
 Chapter 1: Overview and summary

Central Blanket Module

Solenoid 421 modules
Nb3 Sn, 6
modules
Vacuum Vessel
9 sectors
Outer Intercoil
Structure Cryostat
24 m high x 28 m dia.
Toroidal Field Coil
Nb3Sn, 18, wedged Port Plug (IC
Heating)
6 heating
Poloidal Field Coil 3 test blankets
Nb-Ti, 6 2 limiters/RH
rem.
diagnostics
Machine Gravity Support
(recently remodelled) Torus Cryopump
8, rearranged

ertor
54 cassettes

Figure 1. ITER tokamak and major components.

Table 2. ITER parameters and operational capabilities. Here the magnetic shear s is defined as s = (r/q)dq/dr.
Parameter Attributes In a discharge with a positive magnetic shear, the current
density peaks on the centre (magnetic axis), monotonously
Fusion power 500 MW (700 MW)a decreasing with radius. In a discharge with a reverse magnetic
Fusion power gain (Q)
10 (for 400 s inductively
driven burn);
shear, the current density has a maximum at an off-axis

5 (steady-state objective) position. In a discharge with a weak magnetic shear, the current
Plasma major radius (R) 6.2 m density is almost constant from the centre up to typically about
Plasma minor radius (a) 2.0 m the half minor radius.
Plasma vertical elongation 1.70/1.85 Configurations with moderate or weak reversed shear have
(95% flux surface/separatrix)
permitted the development of plasmas whose characteristics
Plasma triangularity (95% flux 0.33/0.48
surface/separatrix) are close to the one required for steady-state scenarios: full
Plasma current (Ip ) 15 MA (17 MA)a non-inductive current, high confinement and high bootstrap
Safety factor at 95% flux surface 3 (at Ip of 15 MA) fraction (section 3.7, chapter 2 [8], chapter 6 [9]). They are
Toroidal field at 6.2 m radius 5.3 T also characterized by the development of internal transport
Installed auxiliary heating/ 73 MW (110 MW)b barriers when proper conditions are met. Internal transport
current-drive power
Plasma volume 830 m3 barrier is a zone in the plasma core with a steep gradient in
Plasma surface area 680 m2 plasma pressure. More recently, the development of magnetic
Plasma cross section area 22 m2 configurations with a wide volume of low magnetic shear
a
and a central value of q close to 1 has resulted in quasi-
Increase possible with limitation on burn duration.
b
stationary discharges with improved confinement and high
A total plasma heating power of 110 MW may be installed in
subsequent operation phases.
values of normalized beta. They are also characterized by
a low level of MHD activity. These discharges extrapolate
to the performance needed for the ‘hybrid’ scenarios foreseen
Plasma current profiles allow us to classify scenarios for for ITER.
ITER, since the safety factor q profile seems to be the dominant As was described in the IPB, there are a number of physics
parameter, although several physics phenomena are involved, basis considerations that are applicable to all three of the
often interlinked, and have to be taken into account. The safety proposed ITER scenarios, and there are also other physics basis
factor is defined by q = d/d, where  is the toroidal flux considerations that are directly relevant to only one or two of
and  is the poloidal flux enclosed by the magnetic surface. the scenarios. These scenario relevancy aspects enter in the
In simpler terms, at a rational q surface, q is the ratio of the presentation of key issues and progress that follows below and
number of toroidal turns to the number of poloidal turns of a into the detailed discussion of PIPB that appears in chapters 2–
field line. In other words, q is inversely proportional to the 8 of this issue.
rotational transform, i.e. the pitch or twist of the field line; q
is inversely proportional to the average current density inside 3. Progress in key physics issues and its impact on
the volume enclosed in a flux surface. Figure 2 illustrates the choice of the main design parameters of ITER
the variation of safety factor profiles observed in tokamak
experiments. In the reference H-mode scenario for ITER, the What follows below are brief narrative accounts of the status
plasma current is fully diffused and the q profile is monotonic and progress in key elements that collectively comprise the
with a large positive magnetic shear. body of the physics basis for a burning plasma tokamak and

S5
M. Shimada et al

Table 3. ITER operation scenarios.

Scenario Plasma current (MA) Non-inductive fraction H98(y,2) li βN Burn duration (s)
Inductive (Scenario 2) 15 0.15 1.0 0.8 1.8 ∼400
Hybrid (Scenario 3) ∼12 ∼0.50 1–1.2 0.9 2–2.5
1000
Steady-state (Scenario 4) ∼9 1.00
1.3 0.6
2.6 3000a
a
3000 s limit is imposed by the cooling system.

Figure 3. H-factor versus normalized density for the IP scan and q95
Figure 2. q profiles. scan in JET [11] (chapter 2, section 4.1.3 [8]). Diamonds are
discharges with Ip = 2.5 MA and Bt = 2.25 T (q95 = 3). Solid
ITER. The projection to ITER is discussed where possible. circles are discharges with Ip = 3.0 MA and Bt = 2.7 T. Stars are
The summary accounts given here are necessarily brief. Full discharges with Ip = 3.5 MA and Bt = 3.2 T. Open triangles are
discharges with Ip = 2.5 MA and q95 = 3.6. Open squares are
scientific details and consideration of many additional aspects discharges with Ip = 2.5 MA and q95 = 4.6.
of physics basis progress will be found in chapters 2–9 and
references therein of this issue.
Confinement degradation relative to scaling prediction
was observed at densities close to the Greenwald density
3.1. Core confinement and transport
nG = I /πa 2 (where the units are 1020 m−3 , MA, m), which
At the time of IPB, the main approach used in predicting was a concern pointed out in IPB (chapter 2, section 5.3.2 [8]).
the performance of ITER in its main regime of operation, However, recent experiments on many tokamaks demonstrate
the steady H-mode, was the global energy confinement time that a good H-mode confinement can be obtained at densities
scaling approach. Here, global energy confinement time is close to or exceeding the Greenwald density by increasing the
defined as W/(P −dW/dt), where W is thermal energy stored triangularity of the plasma cross section and by using pellet
in the plasma and P is the total heating power (i.e. the sum injection or impurity gas puffing (figure 3) [11] (chapter 2,
of alpha-heating power and auxiliary heating power). Five section 4.1.3 [8]).
empirical log-linear (power law) scaling expressions for the The improved confinement mode, H-mode, is character-
energy confinement time were presented in the IPB [10]. ized by an abrupt reduction in heat and particle transport orig-
The Confinement Database and Modelling Expert Group inating at the plasma edge (pedestal) and propagating into the
recommended the IPB98(y,2) scaling as reference scaling for core. Experiments show that this transition requires some level
ITER design. Thermal energy confinement time is described of heating power (threshold power). The main approach to
by the IPB98(y,2) scaling as the projection of the H-mode threshold power Pthr in future
= 0.05621Ip0.93 BT0.15 P −0.69 n0.41
IPB 98(y,2) 0.19 1.97 large devices is at present a derivation of empirical scalings for
τE,th e M R
Pthr expressed in global plasma and device parameters, since
ε 0.58 κx0.78 , (3.1-1) a tested quantitative theory is not yet available (chapter 2, sec-
where Ip is the plasma current, BT is the toroidal field (TF), tion 4.3 [8]). Using an improved and expanded database, the
ne is the volume-averaged density, M is the averaged mass latest projection for ITER is a threshold power in the range
number, R is the major radius and ε is the aspect ratio (a/R, of ∼50 MW, within the capability of the ITER heating sys-
a is the horizontal minor radius). The units are (s, MA, T, tem (73 MW). However, there is uncertainty in the projection
MW, 1019 m−3 , AMU, m) and the elongation κx is defined and experiments suggest that heating powers ∼50% above the
as κx = So /(πa 2 ) with So the plasma cross-sectional area threshold power appear to be required to reach good H-mode
(chapter 2, section 5.3 [8]). The energy confinement time confinement. This indicates the need of further investigation
predicted for ITER is 3.7 s. The standard deviation of the in this area.
residuals for the standard data set with respect to IPB98(y,2) The transport of plasma heat, momentum and particles is
is +14%/ − 13%. enhanced over the prediction from collisional transport theory

S6
 Chapter 1: Overview and summary

Figure 5. Q versus Tped (ion temperature at the pedestal top)

predicted for ITER by the MM, IFS/PPPL and GLF23 transport
models at the same input parameters [13]. Also shown are
Figure 4. A comparison of a GAM E × B oscillation observed in a predictions of the Weiland model at similar input parameters.
turbulence simulation of experimental discharge using the BOUT Horizontal bars indicate the ranges of pedestal temperatures
code and a comparison with measured turbulence poloidal velocity predicted by different pedestal scalings (chapter 2, section 4.2.1 [8]).
spectrum [12].
JET DT discharge, also support the possibility of achieving
(neoclassical transport theory) by a large factor, typically one Q > 10 in ITER (chapter 2, section 5.4 [8]).
or two orders of magnitude. The exceptions are improved Toroidal momentum transport in tokamaks is generally
confinement modes, where the ion heat transport is often found to be anomalous and the viscosity is reduced in the
reduced to a neoclassical level. This transport enhancement internal transport barrier, as are other transport coefficients of
was attributed to turbulent processes, but prior to the writing heat and particles (chapter 2, section 3.5 [8]). Spontaneous
of the IPB, little comparison was made between experimental rotation is observed without any momentum input, which
measurements and turbulence theory. More recently a wide can be qualitatively understood as a result of off-diagonal
range of experiment to theory/simulation comparisons have elements of the transport matrix. The understanding of toroidal
been conducted in core plasmas (chapter 2, section 2.3 [8]) momentum transport is still at a rudimentary stage; quantitative
using a variety of fluctuation diagnostics. Our confidence in prediction to ITER requires further study.
the transport model driven by drift wave turbulence is improved
by the correlation of core turbulence reduction and confinement 3.2. Core confinement, transport and steady-state operation
improvement with relative changes in the growth and damping with weak or negative magnetic shear
rates of the instabilities (ion temperature gradient (ITG)
mode, trapped electron mode (TEM), electron temperature At the time of writing the IPB, improved confinement regimes
gradient (ETG) mode), the identification of zonal flow activity with a weak or negative magnetic shear were being developed.
(e.g. geodesic acoustic mode (GAM)), the evidence for E × The enhanced performance of these plasmas was transient at
B velocity shear suppression of turbulence and turbulent that time. In the last few years, significant progress has been
transport and its effect upon the fluctuation parameters. achieved in developing the weak magnetic shear regimes with
Figure 4 shows a reasonable agreement of the frequency high βN , large fraction of the bootstrap current and improved
predicted for GAM oscillation and measurement [12]. energy confinement over the H-mode scaling sustained for a
Theory-based core modelling of ion temperature profile quasi-steady-state period. These regimes are promising for
shows good agreement with measurement, suggesting that the hybrid and steady-state operation in ITER.
ion temperature profile is stiff (i.e. the profile shape is quasi- The ordinary hybrid operating mode planned for ITER is
invariant) as predicted by ITG mode theory. This suggests based on a combination of inductive and non-inductive current
that temperatures at the top of the pedestal play a determining drive, leading to a long pulse operation (>1000 s) with a
role on the plasma performance. Figure 5 shows the fusion significant fusion power (>300 MW, Q = 5) at a medium
gain projected from theory-based core transport models, as safety factor (q95 = 3.3) and conservative confinement
a function of ion temperature at the pedestal top [13]. This assumption (HH98(y,2) = 1) [4].
analysis shows that Q > 10 can be obtained in ITER inductive Recently, high beta and high confinement have been
operation for pedestal temperatures >4 keV (see section 3.4). obtained in many tokamaks with weak magnetic shear with
Attempts have been made to develop integrated models which HH98(y,2) = 1.2–1.6, q0 = 1–1.5 and q95 = 3.5–4.5, in the
incorporate the core, pedestal and SOL (scrape-off layer, the absence of sawteeth (chapter 6 of this issue [9]). Figure 6
plasma connected to the divertor or limiter along the field shows that high βN of 2.5 is maintained for 16.5 s [14].
line)/divertor regions, and these predict Q ∼ 10 in the A higher beta limit is a key feature of these scenarios. The
reference inductive operation (chapter 2, section 5.5.5 [8]). combination of a lower current and a lower loop voltage
Results of dimensionless analysis, based on an appropriate would allow operation with a high fusion gain for very long

S7
M. Shimada et al

Figure 6. Sustainment of high βN in JT-60U; the sustained βN is

plotted against the sustaining duration. The closed circles indicate
the results obtained before the 2002 IAEA Conference, while the
open circles represent the results after the 2002 IAEA FEC [14].
The improvement in heating and control has resulted in the longer
sustainment of high beta state. The gradual reduction in beta is due
to deterioration of confinement as a result of wall saturation.

tBURN(s)

Figure 8. Plasma performance figure of merit versus duration

normalized to current diffusion time (DIII-D). The filled squares are
ITER baseline scenario discharges, the filled diamonds are hybrid
scenario discharges and the open circles are other types of
discharge [18].

scaling. This figure of merit ranges typically from ∼0.25

Figure 7. Fusion gain and burn time versus NBI power with a
plasma current of 12 MA at HH = 1.2 and density of 85% of (steady state) to 0.4 (conventional H-mode) and shall be
Greenwald density [16].
0.4 for the ‘hybrid’ scenarios. The development of hybrid
scenarios on several tokamaks, allowing steady operation at
pulse duration. Strong impurity accumulation has not been higher beta limits than those for the reference ELMy H-mode,
observed in this mode. The quasi-stationary improved hybrid has been quite remarkable in the recent years. Steady values
regimes with ∼50% non-inductive current fraction at βN ∼ 3 of G
0.4, corresponding to Q ∼ 10 in ITER, have been
that is close to the no-wall beta limit have been obtained. achieved on several experiments for many current relaxation
HH98(y,2) = 1.2 is observed at n/nG = 0.85 [15]. If ITER times (figure 8) [18].
could achieve similar normalized parameters, fusion powers Thanks to a better understanding and modelling of heating
of ∼350 MW, Q > 10 would be expected at βN  2.2 and current drive actuators and to real-time data processing
(figure 7) [16]. The required βN is well below the no-wall ideal on key parameters such as the current profile, sophisticated
MHD limit. A burn time longer than 1000 s with Q ∼ 20 could algorithms have been developed for the active control of
be reached (figure 7). This operation scenario is a potential steady-state and hybrid scenarios (chapter 8 of this issue [19]).
candidate for an operation mode with high Q, long pulse, One of the main remaining issues is the development of
benign ELMs and no strong impurity accumulation. scenarios, or of algorithms, resulting in the lowest possible
For the steady-state (SS) operation, the total plasma demand on control in terms of additional power.
current at the current flat-top phase should be generated non- Many areas require further work, in particular, the
inductively by the bootstrap effect, neutral beam injection extrapolation of improved hybrid regimes to lower ρ ∗ ; the
(NBI) and RF waves. SS scenarios rely upon discharges with size scaling of ITB formation and sustainment of ITBs at
a relatively low plasma current, high safety factor q95
4,
high plasma density, Ti ∼ Te and slow toroidal rotation with
improved confinement (HH98(y,2)
1.3) and high beta (βN

prevention of impurity accumulation. An important effort
2.5) (chapter 6 of this issue [9]). The improved confinement
remains to be done to achieve fully integrated scenarios, i.e.
is expected to be achieved, e.g. in reversed-shear operation.
scenarios that are also compatible with the partially detached
Stationary operation has been obtained experimentally at q95

divertor condition of a burning physics device (the divertor
5, with the maximum performance just in line with ITER
plasma has to be partially detached to reduce heat loads on
requirements for steady-state operation at Q ∼ 5. Operation in
these regimes requires strong plasma shaping and simultaneous the divertor targets to acceptable levels (IPB chapter 4 [20])).
control of the current and pressure profiles (chapter 6 of this In particular, steady-state scenarios privilege operation at
issue [9]) and active control of RWMs and possibly NTMs relatively low density and high electron temperature to
(chapter 3 of this issue [17]). optimize the current drive efficiencies. Generally, steady-state
A dimensionless parameter G = βN H89 /q95 2
is considered and hybrid scenarios that can be considered for ITER have
to be the figure of merit for evaluating inductive, hybrid and made impressive progress since the IPB; the domain is in full
steady-state scenarios (chapter 6, section 6.2.1 [9]). H89 is progress and new scenarios allowing the remaining issues to
the confinement enhancement factor relative to an L-mode be progressively alleviated are being proposed.

S8
 Chapter 1: Overview and summary

Figure 10. Filamentary structure of ELMs displayed in MAST [23].

central heating [21]. A theory-based model that explains these

observations is still under development [22].
Impurity transport is a critical issue in burning plasma
regimes due to possible fuel dilution and radiative cooling in
the core. In ELMy H-mode plasmas, the impurity transport is
reduced compared with L-mode, but impurity accumulation at
the centre is usually not observed. With a relatively flat electron
density profile, hollow profiles of impurities are observed,
which is consistent with neoclassical temperature gradient
screening, which should have a favourable effect on higher Z
impurity profiles in low collosionality plasmas such as ITER
Figure 9. Density peaking, defined as ne (ρ = 0.4)/ne (ρ = 0.8) even when turbulent diffusivity dominates over neoclassical.
versus the line average density ∼in 1020 m−3 , panel (a), and versus However, impurity accumulation at the centre is observed in
νeff , panel (b), for the subset of stationary plasmas in the AUG enhanced confinement regimes with ITBs. These plasmas are
H-mode database, with total NBI heating power of 5 MW. Different
values of q95 are plotted with different symbols [21]. characterized by a steep density gradient at the ITB. The central
accumulation of high Z impurities tend to be more pronounced
3.3. Particle and impurity transport (chapter 2, section 3.4) than low Z impurities. These observations are consistent with
neoclassical theory. The application of ECH or ICH inside the
Progress in particle transport studies has been slower than ITB region has been shown to be effective in reducing the level
in heat transport studies because of the complication of the of density peaking and consequently impurity accumulation.
mixture of two sources (NBI in centre and gas-puffing and
recycling at the edge) and involvement of convective transport.
However, particle transport is very important in burning 3.4. Edge pedestal and ELMs (chapter 2, section 4)
plasmas, since the density profile affects fusion performance
At the time of IPB, H-mode pedestal investigation has already
and consequently plasma profile and stability. Global particle
revealed that the maximum pressure gradient is usually
transport studies separating core and edge transport have
consistent with ideal ballooning mode stability at the edge.
shown that the core particle confinement improves with
The parameter dependence of pedestal width was not clear.
density while edge particle confinement deteriorates with
density. Local particle transport studies show that the particle Recent investigation shows that the edge pedestal gradient is
flux is described with a summation of diffusive and inward consistent with peeling–ballooning mode theory. Figure 10
convective terms. In many cases these transport coefficients illustrates the filamentary structure of ELMs [23] (chapter 2,
are anomalous. Some experiments show strong variation of the section 4.8.3 [8]); this observation is in agreement with
D/χeff ratio, between 0.3 (at high density and low q95 ) and 2.0 the model theorizing that the ELMs are driven by medium-
(at low density and high q95 ), whereas other experiments show to-high n peeling–ballooning mode. We do not yet have
that this ratio is constant (the ratio varies from 0.15 to 0.25 in a quantitative theory on pedestal width validated against
one experiment and ∼1 in the other). A particle pinch at low experiments. Several empirical formulae have been proposed
collisionality has been observed in a number of devices and for the pedestal pressure to be used in the projection to ITER.
explained in the framework of the ITG/TEM transport theory Figure 11 shows an empirical scaling of pedestal pressure
or neoclassical theory (Ware pinch), but a clear experimental compared with experiments and extrapolated to ITER [24]
evidence for the existence of an anomalous inward pinch (chapter 2, section 4.2.3 [8]). This projection suggests that
was shown, where peaked density profiles without central a pedestal temperature of 5.3 keV can be achieved in ITER,
fuelling were observed with zero loop voltage. The density satisfying a requirement of the pedestal temperature (higher
profile is shown to peak with lower collisionalities (figure 9) than 4 keV) for Q ∼ 10 in inductive operation (chapter 2,
or edge heating but flattens with higher collisionalities or section 5 [8]).

S9
M. Shimada et al

development of a resistive wall mode (RWM), whose growth

time is set by the wall resistive time (typically a few ms
in present tokamaks and ∼0.3 s in ITER). However, with
sufficient plasma rotation velocity, growth of the RWM is
suppressed and plasma operation well above the no-wall beta
limit can be obtained. This behaviour has been well known
since before the time of the writing of the IPB. The ability of
these plasmas to exceed the ideal MHD no-wall beta limit for
many wall times is now, in retrospect, well understood in terms
of the effect of rotational stabilization of the RWM growth.
The failure of rotational stabilization and onset of a beta-
limit disruption with application of increasing levels of non-
axisymmetric error field was also reported, but not understood,
in the IPB. Since then, this apparent dependence of the resistive
wall beta limit on error-field level has been conclusively
elucidated in terms of resonant field amplification (RFA), a
Figure 11. Comparison of pedestal pressure values in PDB3V2 collateral effect of RWM onset, wherein plasma operation
database (chapter 2, section 4.3.2. [8]) with the model given in near or slightly above the no-wall ideal limit results in an
section 4.2.3 [24]. This model predicts Tped = 5.3 keV for ITER at RWM-mediated amplification of natural or externally applied
nped = 7.0 × 1019 m−3 . error fields. If the level of the amplified field becomes high
enough, the resulting rotational drag slows the plasma rotation
3.5. Stabilization of neoclassical tearing modes (NTMs) below the critical frequency needed to suppress RWM growth
(chapter 3, section 3.2.2) and a plasma energy confinement collapse and/or disruption
ensues. The close agreement of predicted and observed RFA
NTMs, internal MHD instabilities driven by a deficit of
values in itself constitutes a rather elegant and conclusive
bootstrap current at rational q = m/n surfaces, are frequently
confirmation of the underlying basic RWM theory. Correction
observed in present-day positive shear plasmas, i.e. during
(nulling out) of external error fields avoids the rotation braking
ELMy H-mode operation (the basis for the ITER inductive
caused by RFA and allows sustained plasma operation—with
scenario). Unconstrained growth of NTM modes can lead to
sufficient plasma rotation drive—at βN values well above the
deterioration of energy confinement and sometimes eventual
corresponding no-wall limit.
disruption. The threat of NTMs to ITER operation in both
The error-field correction capability in ITER should
the inductively driven ELMy H-mode scenario and in the
enable stabilization of the RWM by rotation especially when
similar hybrid scenario has been recognized since the time
the rotation is large. However, given the relatively low external
of the writing of the IPB. In the IPB, the possibility (and
necessity) of controlling NTM growth through the use of local momentum input in ITER, feedback stabilization of the RWM
electron cyclotron current drive (ECCD) was highlighted as an by saddle coils is implemented.
inductive-scenario enabling element. Experiments conducted The DIII-D tokamak is equipped with a set of 12 internal
in a number of tokamaks have conclusively demonstrated the single turn feedback coils to provide fast feedback control of
feasibility of using well-controlled ECCD to suppress NTM the RWM in those plasmas in which the rotation is below the
growth and/or prevent onset of the instability. Figure 12 threshold. Figure 13 shows that direct feedback has sustained
shows the first demonstration of m/n = 3/2 NTM suppression a plasma with βN almost 4 well over the no-wall limit for over
using ECCD [25]. Active control techniques to track 1 s [26].
the position of the rational q surface and target the ECCD at the The suppression of RWMs has been investigated for ITER
desired location have been demonstrated. Assessments of the steady-state scenarios. A multi-input, multi-output linear
ECCD requirements and ECCD deposition control needed for quadratic Gaussian (LQG) controller has been produced on
the control of NTMs (m/n = 3/2 and 2/1) in ITER inductive the basis of a semi-analytical model of RWM with a code
and hybrid scenarios have been developed. Present estimates treating the configuration of the plasma, saddle coils and
suggest that NTM can be suppressed with an ECCD power of vacuum vessel. Analyses show that this controller, without
10–30 MW in ITER. The overall prognosis based on successful using the second time derivative of the measured perturbed
suppression of NTMs in ITER is found to be well in hand. field, is able to suppress highly unstable modes with Cβ ∼ 0.6–
0.8 (Cβ = (βN − βN (no wall))/(βN (ideal wall) − βN (no wall)))
3.6. Feasibility of sustained operation above the no-wall with coil voltages of about 300 V/turn, which is implemented
ideal MHD beta limit (chapter 3, section 3.2.3) in ITER.

The ITER steady-state scenarios require plasma operation 3.7. Feasibility of disruption mitigation using massive gas
with beta values close to or above the so-called no-wall ideal injection (chapter 3, section 3.3.6)
MHD beta limit (the no-wall limit). For the plasma current
profiles anticipated for ITER steady-state operation, the no- All tokamaks including ITER are inherently susceptible to
wall limit is calculated to be about βN =∼ 2.4, while the disruption, which can bring large heat and electromagnetic
required βN is
2.6. In the presence of a nearby resistive loads and nearly complete conversion of thermal plasma
wall, the manifestation of exceeding the no-wall limit is the current to relativistic (∼10 MeV) suprathermal runaway

S10
 Chapter 1: Overview and summary

Figure 12. Decreasing amplitude of m/n = 3/2 Mirnov coil signal shows that the NTM is suppressed during ECCD [25]. The high
frequency oscillations of the Mirnov coil signal are due to ELMs. The shift of the EC resonance position due to the variation of the toroidal
magnetic field is given, and βN of this discharge is compared with a (nearly) identical discharge without ELMs.

feasible to reach (or closely approach) the required electron

densities and that the resulting rapid radiative cooling of the
plasma also limits conduction of the plasma thermal energy to
the divertor targets and reduces the magnitude and toroidal
asymmetry of the halo currents that disruptions normally
cause. The technique of massive gas injection appears to offer
good promise for simultaneous mitigation of disruptions and
runaway conversion in all ITER plasma operation scenarios.
For the case of a vertical displacement event, which occurs
when the vertical position control is lost, the vertical movement
is slow (the time scale ∼0.5 s) in ITER, which ensures a
high reliability of detection and mitigation. For the case of
disruption, neural networks are being developed which trigger
the mitigation system. In the long term, a disruption avoidance
system should be developed.
Figure 13. Direct feedback in DIII-D enables sustained operation
well above the no-wall limit when plasma rotational stabilization is
insufficient [26]. 3.8. Digital plasma control systems (chapter 8)
Since the writing of the IPB, the development of very
electron current. In ITER, disruptions have no ‘single- sophisticated and comprehensive digital plasma control
event’ potential to damage the plasma-facing components systems has taken place in most major tokamak facilities.
to failure. However, repetitive worst-case disruptions are These systems, which typically comprise a number of generic
not acceptable, since disruptions would shorten the life time PC-type computers running asynchronously with real-time
of plasma-facing components and the deterioration of wall operating systems and interconnected with a fast GB/s
condition after disruption would reduce the availability of network, have allowed a variety of plasma control and real-
the machine. In Demo, very low probability of disruption time diagnostic data utilization tasks to be undertaken in a
(less than one disruption per year) and/or reliable active fully digital manner. The types of control performed in
measures of disruption avoidance is assumed due to its serious this manner range from routine plasma current, position and
consequences [27]. Thus disruption control constitutes an equilibrium control, plasma density control and beta control to
essential element for high availability of operation and long sophisticated multi-actuator control of the plasma safety factor
lifetime of plasma-facing components in ITER and elimination profile and internal transport barrier location. Capabilities
of disruptions in ITER would be essential for the high now possible include use of essentially real-time equilibrium
reliability of Demo. reconstruction data for plasma shape and position control
In the IPB, the injection of massive neutral gas, equivalent and divertor strike point control, incorporation of real-time
to increasing the plasma electron density by a factor of about MSE data in plasma current profile control algorithms and
200 (to a final density of about 2 × 1022 electrons m−3 ) was implementation of sophisticated algorithms to detect and track
identified as a possible disruption mitigation means. But at the location of NTMs for ECCD control of MHD stability.
that time there was no experimental or theoretical precedent The ability of the present generation of plasma control
to believe that this high degree of density increase could system to control the plasma parameters is now at a state
necessarily be achieved, or achieved without in itself causing that is basically limited only by the availability of diagnostic
a disruption. However, recent gas-injection experiments in measurements and the corresponding ‘actuators’ needed to
several medium-sized tokamaks, using massive quantities modify the parameters, even in a multiple-input, multiple-
of either neon or argon gas have demonstrated that it is output algorithm sense. The required control system capability

S11
M. Shimada et al

level acceptable for divertor targets is possible with such ELM-

pacemaking at a frequency of 4 Hz or above. The ITER design
includes pellet injectors on the high field side for fuelling and
on the low field side for ELM-pacemaking. However, confine-
ment deterioration is observed at frequent pellet injection and
is a concern for ITER.
Magnetic perturbation induced by saddle coils can make
edge magnetic islands overlap with each other, which enhances
the transport in the pedestal, keeping the pressure gradient
below the critical level to induce ELMs at safety factors ∼3.7
[30], (chapter 2, section 4.9.2 [8]). The saddle coils located
inside the vacuum vessel are under study, but such in-vessel
coils are associated with engineering difficulties in ITER. ELM
suppression with out-vessel coils should be investigated. ELM
suppression at a safety factor ∼3 is yet to be demonstrated.
Possible triggering of NTM and consequences of elimination
of toroidal rotation also need to be investigated.
Furthermore, improved confinement regimes free of type-I
ELMs have been explored. For example, figure 15 shows
that highly shaped, moderate q (q95 ∼ 4) and high βp
Figure 14. Comparison of ELM sizes obtained in this study for (∼1.6) discharges are associated with benign ELMs (chapter 2,
19 Hz pellet triggered (P in filled black square) and 3 Hz (R in open section 4.7.5 [8]) [31]. The enhanced D-alpha (EDA) regime
black diamond) and 20 Hz (R in filled black diamond) intrinsic (chapter 2, section 4.7.2 [8]) has been observed along with its
ELMs to values from a type-I ELM scaling for ASDEX Upgrade
and JET [28].
quasi-coherent modes. However, the viability of this regime
is unlikely in ITER because of the low edge temperature
that is required. The quiescent H (QH) mode (chapter 2,
(and many representative algorithms) is already available for
section 4.7.3 [8]) shows ELM-free discharges with good
ITER plasmas.
confinement. However, extrapolation to ITER is uncertain
without a reliable physics-based model. High impurity levels
3.9. Particle control and power dispersal (chapter 4) of QH-plasmas are also a concern.
The heating power that the fusion plasma produces and The plasma-facing components for the initial operation
supplied externally needs to be eventually exhausted at the consist of the following; the divertor targets are covered
edge. Since the wetted surface area is expected to be ∼4 m2 at by carbon-fiber-composite (CFC) graphite; tungsten is used
the target in ITER, 80 MW heat loss through the separatrix of at the dome and baffle (upper target) regions for its low
ITER would result in an average power density of 20 MW m−2 yield of physical sputtering by neutral particles; beryllium
at the target. In order to keep the peak target power density is used for the first wall for its small impact on the plasma
below ∼10 MW m−2 , a substantial reduction of the heat load performance and high oxygen gettering. CFC targets are
is required. Also particle exhaust by divertor is required to commonly used as plasma-facing components in present-
control the density and purity of the core plasma. At the day experiments due to its compatibility with a wide range
time of IPB writing, divertor code calculations suggested the of plasma parameters. However, tritium retention control,
feasibility of radiative cooling enhanced by partial detachment required with CFC targets, remains a key issue and more
with ITER parameters. This provided optimism that inter- efforts are called for to investigate retention mechanism and
ELM heat load handling and particle exhaust would not be a develop efficient removal techniques. Early experiments with
problem for ITER. ELMs were expected to cause a serious heat a horizontal target and with a limiter configuration suggested
load on the divertor but a quantitative analysis was yet to be that ∼30% of tritium injected is retained in the vessel with C
made and an ELM mitigation scheme was yet to be developed. walls and divertor targets. If we simply extrapolate this data to
Quantitative estimates based on experimental database ITER, the T retention reaches a project guideline (limit on the
have shown that power fluxes on the ITER divertor targets amount of T) rather quickly. However, recent experiments with
associated with type-I ELMs could be close to or above vertical target show that ∼3% of injected tritium is retained in
marginal for an acceptable divertor lifetime, which has the vacuum vessel: one order of magnitude reduction of the
motivated development of back-up scenarios. Potential build-up rate. Furthermore, it was shown that the build-up of
scenarios include those with more frequent, smaller ELMs tritium retention could be significantly reduced by a factor of
triggered by frequent pellet injection, edge ergodizaiton and ∼1/5 by the coverage of carbon surface by beryllium. The
regimes with benign or no ELMs (type II or grassy ELMs). level of tritium retention predicted for ITER is thus reduced
The idea of ELM-pacemaking (chapter 4, sec- but the uncertainties are large and the methods for its removal
tion 2.7.3 [29]) stems from the experimental observation that need more development. The primary challenge is to remove
the amplitude of intrinsic ELM decreases with ELM frequency tritium from shadow areas, not accessible by plasma, as well as
(figure 14) [28]. The ELMs induced by small pellets show sim- from surfaces with mixed material deposition (e.g. Be/C/W).
ilar amplitudes as the intrinsic ELMs. Analysis with ITER Elimination of graphite from the vessel would be attractive
parameters suggests that reduction of ELM amplitude to a from the viewpoint of tritium retention control. High

S12
 Chapter 1: Overview and summary

Figure 15. Change of ELM activity (divertor Dα signal) with δ, q95 and βp at Ip = 1 MA ( t = 0.2 s in all cases). Giant ELMs are replaced
by grassy ELMs as q95 increases with fixed δ and βp ((a) to (b) to (c)); as δ increases with fixed q95 and βp ((e) to (d) to (c)); and as βp
increases with fixed δ and q95 ((h) to (c)). Compared with (b) (δ = 0.47 and q95 = 5.1), an almost grassy ELMy phase appears at a higher
δ = 0.54 even at a low q95 = 4 (g) [31].

Z materials such as tungsten are attractive for their long

lifetime and are deemed most applicable to a fusion reactor.
However, experiments with tungsten [32] show that tungsten
can accumulate in the centre in discharges with an ITB
and/or with a peaked density profile (figure 16) in the
absence of sawtooth/fishbones and frequent ELMs (chapter 4,
section 2.5.4.1 [29]). It has also been demonstrated that high
central heating could suppress the accumulation (figure 16).
Further experimental studies with all high Z plasma-facing
components in large and medium-sized tokamaks are required
to investigate the specific restrictions that tungsten may impose
on ITER operation. Surface melting after disruption and
subsequent formation of irregular surfaces are of serious
concern, since heat loads during disruption in ITER are
estimated to be one or two orders of magnitude higher than
in the present machines and these irregular surfaces can easily
melt or evaporate at normal operation. Therefore further efforts
should be made in the development of disruption prediction, Figure 16. Peaking of the W concentration (cW ) as a function of
peaking of background density [32]. Shown are the ratios of cw at
mitigation and avoidance. Since adoption of tungsten might the location where the electron temperature is 3 keV to cw at the
limit plasma operational flexibility, the tungsten target is not location where the electron temperature is 1 keV. Discharges with
suitable for the initial operation, in which a wide operational pure NBI heating (black circles) show the strongest peaking,
space is fundamental. After establishing reliable operation whereas central ECRH reduces the cW peaking significantly already
modes, it would be better to replace CFC targets with tungsten at low additional heating power.
to facilitate tritium retention control and to demonstrate a
target suitable for a next-generation reactor and it may be 3.10. Energetic particle physics (chapter 5)
better to replace the beryllium first wall with tungsten at One of the major scientific goals for ITER is to demonstrate
least partially. Development and validation of theory-based and investigate burning plasmas, in which a significant power
models of impurity transport is also useful for the assessment of is produced by the DT fusion reactions. Such plasmas
plasma performance with tungsten. The current combination are characterized by a large isotropic population of fusion
of CFC/tungsten targets and beryllium walls is expected to alphas, providing the dominant heating of the plasma. In DT
minimize T retention and W erosion with a wide range of plasmas, self-heating is provided by the α-particles generated
parameters, which is essential for the initial operation. at 3.5 MeV by the D-T fusion reactions. Other fast or energetic
Dust of sizes ranging from 10 to 0.1 mm is observed ions with energies in the MeV range, well above the thermal
in tokamaks. Dust has recently attracted attention in distribution of the plasma bulk, are generated by ion cyclotron
light of its potential impact on safety in ITER, due to its resonant heating (ICRH) and NBI. The behaviour of energetic
high chemical reactivity and possible activation (chapter 4, ions is a key subject of burning plasma studies. Transport and
section 2.6.5 [29]). Understanding of its formation confinement of fusion α-particles not only impact machine
mechanisms and measurement and removal methods need to performance by affecting the fusion yield but can also damage
be developed. the first wall.

S13
M. Shimada et al

Figure 17. Time evolution of the EPM radial structure, decomposed in poloidal Fourier harmonics. Here τA0 = R0 /vA (r = 0) (R0 is the
major radius and vA is Alfven speed). The toroidal mode number is n = 4. The nonlinear modification of αEo − R0 q 2 → (dβE /dr) (change
in alpha particle pressure) is also shown [33].

A large effort was dedicated to the development of 3.11. Diagnostics (chapter 7)

methods to simulate fusion born alphas in plasmas without
ITER will require an extensive set of plasma and first
significant fusion reactivity. Ripple losses are relatively well
wall measurements for machine protection, plasma control
understood. The synergy between new observations and
and physics evaluation. Because of the harsh radiation
advanced modelling has led to an optimization of ferritic environment, diagnostic system selection and design involves
inserts in ITER to reduce ripple-induced α-particle losses with a range of challenges not previously encountered: for
reversed-shear configurations to a negligible level. example, radiation induced conductivity and radiation
The interaction of fast ions generated by additional heating induced EMF in magnetic sensors mounted in the vacuum
with low frequency MHD has been investigated in a variety of vessel, enhanced erosion of diagnostic first mirrors due to
experiments. The linear theory of kinetic ballooning modes energetic particle bombardment and enhanced absorption and
and localized interchange modes is well advanced, and the photoluminescence in windows and optical fibres. The
behaviour with sawteeth is well described. diagnostic designs also have to satisfy stringent requirements
The field of linear stability thresholds for collective on tritium confinement, vacuum integrity, remote handling
instabilities was advanced through a large number of maintainability and reliability. Access will be restricted
experimental results and significant progress in theoretical and must maintain neutron streaming below allowable limits.
simulations. Damping and drive mechanisms are qualitatively Taken together, these aspects mean that the provision of
understood, although quantitative predictions for specific diagnostics for ITER is arguably the most challenging
modes are still to be ameliorated, especially in regimes in which undertaken thus far in the history of plasma diagnostics.
fluid and kinetic models give significantly different results. It is expected that there will be a phased introduction
of powerful operation and advanced scenarios. The
The understanding of the nonlinear phase of the interaction
demands on the measurements will rise accordingly with
between waves and fast ions was significantly improved,
more parameters being brought under real-time control.
particularly in the weakly nonlinear regime. Figure 17 shows The parameters necessary to support the different planned
nonlinear interaction between energetic particle modes and operating scenarios—inductive ohmic L-mode, inductive
alphas, suggesting potentially significant effect on alphas [33]. ELMy H-mode, hybrid operation, and steady-state operation,
Measurements of the modes are used to extract information as the operation advances through the H, D and DT
about the background plasma and/or the fast ion population. phases—have been identified. For each parameter, detailed
Although nonlinear modelling appears to be in qualitative requirements (ranges, time and space resolutions, accuracies)
agreement with experiments, limited information is available have been developed and the full specifications are included
on the fast ion redistribution and losses, due to the difficulty in tabular form. In addition, for each parameter a justification
in achieving large amplitude modes in present devices and in has been prepared.
having sufficiently sensitive diagnostic tools to measure the R&D is in progress to resolve the remaining issues. These
energy and radial distribution of the fast ions. include specific component testing, for example on prototype

S14
 Chapter 1: Overview and summary

Table 4. Initial set-up and possible upgrade scenarios of heating and current drive systems.
Start-up Scenario 1 Scenario 2 Scenario 3 Scenario 4
Power (MW) Power (MW) Power (MW) Power (MW) Power (MW)
NB (1 MeV) 33 33 50 50 50
IC (40–50 MHz) 20 40 20 40 20
EC (170 GHz) 20a 40b 40b 40b 20b
LH (5 GHz) 0 20 20 0 40
Total 73 133 130 130 130
a
20 MW of EC will be used either in two upper ports to control neoclassical tearing modes, or in one
equatorial port for main heating or current drive.
b
EC will allow use of four allocated top ports for the power upgrade. No additional equatorial ports
are therefore foreseen for this system. The total installed power is given in the table and the total
maximum power into the torus is limited to 110 MW.

magnetic coils or bolometers, generic testing of candidate further work is required to develop control schemes of
materials, for example in-vessel cabling, and testing of new disruption, ELM, impurity, NTM, RWM, dust and tritium
diagnostic techniques which have the potential to be relatively retention.
rugged in the ITER environment. All optical/IR, spectroscopic (4) ITER provides a unique opportunity of investigating
and microwave systems view the plasma with a mirror and a physical phenomena under reactor conditions, which are
critical issue is the lifetime of this component. Dedicated R&D impossible to achieve in present machines. Continued
is in progress on candidate first mirrors. physics R&D, focused on ITER-relevant topics, will
Finally, ITER will provide a unique opportunity of reduce uncertainties of projection and help develop control
measurements in a fusion reactor environment. That schemes that will be essential for handling problems that
experience will be useful for measurements of ITER may arise during ITER operation. Physics understanding,
plasma itself during operation with standard and enhanced projection and control methodologies developed and
performance and for diagnostics development for the next step validated in ITER and other machines will be important
device. It is probable that there will still be an extrapolation in for paving the path for Demo.
plasma performance from ITER to the next step device. The
experience gained on ITER will reduce the uncertainties in
this extrapolation and provide the basis on which an informed
Appendix A.
choice of diagnostics can be made and diagnostic systems
Appendix A.1. ITER
designed.
ITER is a long pulse tokamak with elongated plasma and
4. Summary single null poloidal divertor. The nominal inductive operation
produces a DT fusion power of 500 MW for a burn length of
(1) The physics R&D after IPB are highlighted by significant 400 s, with the injection of 50 MW of auxiliary power. Its
progress in understanding of turbulent transport, theory- main features and characteristics of its heating systems are
based modelling of core, H-mode confinement at n ∼ nG , summarized in tables 2 and 4. The major components of the
improved helium exhaust modelling, sustainment of tokamak are the superconducting coils which provide toroidal
hybrid- and SS-relevant scenarios at high beta and and poloidal fields which magnetically confine, shape and
confinement, ELM mitigation, NTM and RWM control, control the plasma inside the toroidal vacuum vessel. The
disruption mitigation, understanding of energetic particle magnet system comprises TF coils, the CS, external PF coils
modes and diagnostics development. and correction coils (CC). The vacuum vessel has a double-
(2) These have contributed to improved feasibility of ITER walled structure.
achieving its goals. Especially, validation of core transport The tokamak fuelling system is designed to inject gas
models has progressed and analysis with ITER parameters and solid pellets of hydrogenic isotopes (H2 , D2 , T2 or DT).
suggests that the achievement of Q > 10 in the inductive During plasma start-up, low-density gaseous fuel will be
operation is feasible. Improved confinement and beta introduced into the vacuum vessel chamber by the gas-injection
have been observed with low shear (= high βp = system. The plasma will progress from electron-cyclotron-
’hybrid’) operation scenarios in many tokamaks. If similar heating assisted initiation, in a circular configuration touching
normalized parameters were achieved in ITER, it would the outboard limiter, to an elongated divertor configuration as
provide an attractive scenario with high Q (>10), long the plasma current is ramped up. Once the current flat-top
pulse (>1000 s) operation with beta < no-wall limit and value (nominally 15 MA for inductive operation) is reached,
benign ELMs. subsequent plasma fuelling (gas or pellets) together with
(3) For improved physics understanding, more work remains additional heating for ∼100 s leads to a high Q DT burn
in the areas of transport of momentum and particles and with a fusion power of about 500 MW. With non-inductive
transport and stability in the edge pedestal and effects of current drive from the heating systems, the burn duration is
TAE modes. Understanding of core impurity levels should envisaged to be extended to 1 h. In inductive scenarios, before
also be developed. For reliable and high duty operation, the inductive flux available has been fully used, reducing the

S15
M. Shimada et al

Figure 18. Initial operation plan of ITER.

fuelling rate so as to slowly ramp down the fusion power following DT phase can be made based on the confinement
terminates the burn. This phase is followed by plasma current characteristics during the H2 -phase. Characteristics of
ramp-down and finally by plasma termination. The inductively transient phenomena, such as disruption, vertical displacement
driven pulse has a nominal burn duration of 400 s, with a pulse events and ELMs, will be investigated, and mitigation
repetition period as short as 1800 s. The integrated plasma measures (impurity gas injection, pellet injection) will be
control is provided by the PF, pumping, fuelling (D, T and tested. The neural network will be trained and tested for
impurities such as N2 , Ar) and heating systems all based on disruption prediction. Erosion and re-deposition of first wall
feedback from diagnostic sensors. and divertor materials will be investigated. Wall conditioning
procedures will be developed including the procedures for
Appendix A.2. Operation scenarios and phases tritium removal. Formation of dusts will be investigated and
the dust removal techniques will be developed.
As an experimental device, ITER is required to be able to cope During the DT phase, a reference DT scenario will be
with various operation scenarios and configurations. Variants developed by optimizing DT fuelling, fusion power, auxiliary
of the nominal scenario are therefore considered for extended heating power and burn pulse length. Exploration will be made
duration plasma operation, and/or steady-state modes with a in wide operation regimes to investigate burning plasmas and
lower plasma current operation, with H2 , D2 , DT (and He) steady-state plasmas and reliable scenarios will be developed
plasmas, potential operating regimes for different confinement for long pulse engineering tests without severe disruptions,
modes, and different fuelling and particle control modes. vertical displacement events and giant ELMs.
Flexible plasma control should allow the accommodation
of ‘advanced’ plasma operation based on active control of
plasma profiles by non-inductive current drive, heating or References
fuelling. [1] ITER Physics Basis Editors et al 1999 Nucl. Fusion 39 2137
An initial operation plan of the first 10 years is shown [2] ITER Final Design Report, Cost Review and Safety Analysis
in figure 18. The first operation phase will be conducted 1998 ITER Council Proc. (ITER Documentation Series
with hydrogen or helium plasmas. During this phase, No 15) (Vienna: IAEA) p 39
commissioning of various equipment will be completed, and [3] Technical Basis for the ITER Final Design Report, Cost
Review and Safety Analysis 1998 (ITER Documentation
a reference operation scenario will be developed. Adequacy Series No 16) (Vienna: IAEA)
of heating systems for L–H transition, divertor function will [4] ITER Technical Basis 2002 (ITER EDA Documentation Series
be confirmed and extrapolation of performance during the No 24)

S16
 Chapter 1: Overview and summary

[5] Mukhovatov V. et al 2007 Progress in the ITER Physics Basis CD-ROM IT/1-1 and http://www-naweb.iaea.org/napc/
Nucl. Fusion 47 S404–S413 physics/fec/fec2004/datasets/index.html
[6] 1998 ITER Council Proc. (ITER Documentation Series No 15) [17] Hender T.C. et al 2007 Progress in the ITER Physics Basis
(Vienna: IAEA) p 148 Nucl. Fusion 42 S128–S202
[7] Wesson J. et al 2003 Tokamaks (International Series of [18] Wade M.R. et al 2005 Nucl. Fusion 45 407
Monographs on Physics vol 118 (Oxford: Oxford [19] Gribov Y. et al 2007 Progress in the ITER Physics Basis
University Press) (3rd version) Nucl. Fusion 47 S385–S403
[8] Doyle E.J. et al 2007 Progress in the ITER Physics Basis Nucl. [20] ITER Physics Basis Editors et al 1999 Nucl. Fusion
Fusion 47 S18–S127 39 2391
[9] Gormezano C. et al 2007 Progress in the ITER Physics Basis [21] Angioni C. et al 2003 Phys. Plasma 10 3225
Nucl. Fusion 47 S285–S336 [22] Angioni C. et al 2005 Phys. Plasma 12 112310
[10] ITER Physics Basis Editors et al 1999 Nucl. Fusion [23] Kirk A. et al 2004 Proc. 20th Int. Conf. on Fusion Energy 2004
39 2175 (Chapter 2, section 6.4) (Vilamoura, Portugal, 2004) (Vienna: IAEA) CD-ROM
[11] Sartori R. et al Scaling study of ELMy H-mode global and EX/2-3 and http://www-naweb.iaea.org/napc/physics/
pedestal confinement at high triangularity in JET Proc. 20th fec/fec2004/datasets/index.html
Int. Conf. on Fusion Energy 2004 (Vilamoura, Portugal, [24] Sugihara M. et al 2003 Plasma Phys. Control. Fusion
2004) (Vienna: IAEA) CD-ROM EX/6-3 and 45 L55
http://www-naweb.iaea.org/napc/physics/fec/fec2004/ [25] Gantenbein G. et al 2000 Phys. Rev. Lett. 85 1242
datasets/index.html [26] Okabayashi M. et al 2005 Nucl. Fusion 45 1715
[12] McKee G.R. et al 2003 Phys. Plasmas 10 1712 [27] Najmabadi F. et al 1997 Fusion Eng. Des. 38 3
[13] Mukhovatov V. et al 2003 Plasma Phys. Control. Fusion [28] Lang P.T. et al 2003 Nucl. Fusion 43 1110
45 A235 [29] Loarte A. et al 2007 Progress in the ITER Physics Basis Nucl.
[14] Suzuki T. et al 2004 Proc. 20th Int. Conf. on Fusion Energy Fusion 47 S203–S263
2004 (Vilamoura, Portugal, 2004) (Vienna: IAEA) [30] Evans T.E. et al 2004 Phys. Rev. Lett. 92 235003
CD-ROM EX/1-3 and http://www-naweb.iaea.org/napc/ [31] Kamada Y. et al 2000 Plasma Phys. Control. Fusion
physics/fec/fec2004/datasets/index.html 42 A247
[15] Sips A.C.C. et al 2003 30th EPS Conf. on Controlled Fusion [32] Neu R. et al 2005 Nucl. Fusion 45 209
and Plasma Physics (St Petersburg, Russia, 2003) vol 27A [33] Zonca F. et al 2002 Proc. 19th Int. Conf. on Fusion Energy
(ECA) O-1.3A 2002 (Lyon, France, 2002) (Vienna: IAEA) CD-ROM
[16] Shimada M. et al 2004 Proc. 20th Int. Conf. on Fusion Energy TH/4-4 and http://www.iaea.org/programmes/ripc/
2004 (Vilamoura, Portugal, 2004) (Vienna: IAEA) physics/fec2002/html/fec2002.htm

S17