annual progress report 2009 - kth/annual_report_2009.pdf · annual progress report 2009 ... the...

ASSOCIATION EURATOM - VR

Swedish Fusion Research Unit

Annual Progress Report 2009

Progress Report 2009 of the Fusion Association EURATOM-VR

Cover Pictures: Upper-left: Fig 2.5.1-8 Symmetric chirping pattern for a marginally unstable TAE mode, in the absence of fast particle collisions, in response to kinetic !-particle drive calculated using HAGIS Upper-right: Fig.3.1.1-5 Left: CAD model of the MAST neutron camera sitting on the rail in relation to MAST vacuum vessel. Lower: Fig. 2.2.1-5: (a) Time evolution of radial magnetic field harmonics m=1 for a plasma shot with RIS (revised intelligent shell) and reference signal for the harmonic (1,-12) set to 0.4mT and (b) corresponding spectrum at t=25ms. (c) Corresponding time evolution of the active coil current harmonics and (b) spectrum at t=25ms. The vertical continuous lines in frames (a) and (c) highlight the end of the discharge. Compiled from contributions from the research groups of the Swedish Fusion Research Unit James R Drake Head of the Swedish Research Unit Association EURATOM-VR Division of Fusion Plasma Physics Alfvén Laboratory School of Electrical Engineering Royal Institute of Technology KTH SE – 100 44 Stockholm, Sweden


i

ASSOCIATION EURATOM - VR

Swedish Fusion Research Unit

Annual Progress Report 2009

Table of Contents Preface…………………………………………………….. 1 1 Executive summary ……………………………………. 3

1.1 General introduction 3 1.2 Research Unit 4 1.3 Overview of research activities 5 1.4 Highlights of the research activity 7

2 Support advancement of the ITER physics base ……..10 2.1 Energy and particle confinement and transport 10 2.2 MHD stability and plasma control 20 2.2.1 Active MHD mode control experiments 20 2.2.2 Application of resonant magnetic perturbations 26 2.2.3 MHD stability 31 2.3 Power and particle exhaust, plasma-wall interaction 33 2.3.1 Plasma wall interaction 33 2.3.2 Accelerator-based analysis of materials 48 2.3.3 Studies of dust collection using aerogel collectors 49 2.4 Physics of plasma heating and current drive 51 2.5 Energetic particle physics 59 2.5.1 Physics of burning fusion plasmas 59 2.5.2 Runaway electrons in tokamaks 65

3 Plasma auxiliary systems – diagnostics ………………. 68 3.1 Neutron diagnostics 68 3.1.1 Neutron Diagnostic instrumental development for JET and MAST 69 3.1.2 Participation in the experimental programme at JET 77 3.1.3 Development of analysis tools based on neural networks 80 3.1.4 R&D for ITER within the EFDA Work Programme 84 3.1.5 Contracts and collaborations 90 3.2 Plasma spectroscopy 92


ii

4 Concept improvements …………………………………97 4.1 Computational methods with applications for the RFP 97

5 Emerging technology ………………………………….. 99 5.1 High purity ODS-tungsten materials 99

6 EFDA JET………………………………………………. 6.1 EFDA JET Orders 101 6.2 JET enhancements 101 6.3 JET campaign participation C20-C25 107 6.4 Fusion Technology at JET 108

7 EFDA Task Force and Topical Group activity …….. 109 7.1 Task Force Integrated Tokamak Modelling 109 7.2 Task Force Plasma Wall Interaction 110 7.3 Topical Group Materials 111 7.4 Topical Group Diagnostics 111 7.5 Topical Group MHD 112 7.6 Topical Group Transport 112 7.7 Goal Oriented Training in Theory (GOTiT) 113

8 ITPA activity …………………………………………..114 8.1 Overview of ITPA activity 114

9 Technology …………………………………………….. 115 9.1 Art. 5.1b actions 115

10 Other activities ………………………………………. 116 10.1 Training and education 116 10.2 Public information 116

Appendix I Fusion for Energy Grants ………………… 118 Appendix II: Contact Information …………………… 119


1

Preface On behalf of the Swedish Fusion Research Unit, I am pleased to present the Annual Progress Report for 2009 covering research carried out under the Contract of Association between the Swedish Research Council (VR) and the European Atomic Energy Community, EURATOM. James R Drake Head of the Swedish Fusion Research Unit Association EURATOM-VR


2


3

1 EXECUTIVE SUMMARY 1.1 General introduction Controlled thermonuclear fusion offers the prospect of an intrinsically safe, virtually inexhaustible energy source. It is seen as potentially having a key role in the long-term energy system, primarily for base load electricity production, provided it can be developed to become economically competitive. ITER (the international fusion energy test reactor) plays a vital role in the development of fusion as an energy source. After a series of pre-agreements, the final ITER International Agreement was signed by the seven international parties in November 2006 and is now under construction in Cadarache, France. The seven parties are the EU, Japan, USA, Russian Federation, China, South Korea and India. The mission of the ITER experiment is as follows:

• Demonstrate capability of steady state fusion power production. • Optimise burning plasma confinement under reactor conditions. • Have dimensions comparable to a power station and produce about 500 MW of fusion

power (10x more power than needed to run it). • Demonstrate or develop new technologies and materials required for fusion power

stations. • Have a construction period of 10 years and an operation period of 20 years.

The European Atomic Energy Community (EURATOM) is the Domestic Agency representing the European Union in the ITER International Organisation. A Joint Undertaking for ITER and the Development of Fusion Energy (Fusion for Energy or “F4E”) was approved by the Council of Ministers in 2006 and was formally launched in June 2007. F4E is charged with procurement and delivery of Europe’s contributions to ITER. The development of fusion power is a key action in the European Framework Programme and the research is co-ordinated and managed as a part of the EURATOM agreement. The procurement and delivery of the European contribution to ITER is the responsibility of the Joint Undertaking, Fusion for Energy. However a substantial support effort is required in the accompanying European fusion programme, which is co-coordinated by the European Commission under EURATOM auspices. The work is performed by various groups in the member states under Contracts of Association (CoA). The Contract of Association establishing the Swedish Research Unit (RU) is between the Swedish Science Research Council (VR) and EURATOM. The CoA defines the roles of the Association Steering Committee, the Research Unit and the leadership of the RU, the Head of Research Unit (HRU). The CoA also includes a Work Plan for the Association which normally spans several years. There are additional agreements between the Associations. The European Fusion Development Agreement (EFDA) provides a framework for coordinating activities which compliment the activities of the Joint Undertaking.

• Co-ordinated activities in physics and emerging technology.


4

• The collective use of the JET facilities, which is the largest fusion experiment now in operation and is located in the UK.

• Training and career development of researchers, promoting links to universities and carrying out support actions.

• European contributions to international collaborations that are outside the Joint Undertaking for ITER and the Development of Fusion Energy.

The long-term strategy of the European fusion programme is based on well-defined steps. Operation of present-generation experiments, in particular the JET experiment, has established a base for the design of ITER. These experiments will now be used to plan for the exploitation of ITER. The results of the ITER project together with efforts carried out in parallel should enable the next step, the construction of a demonstration reactor, DEMO. The focus is now on ITER, however progress to fusion power plants includes additional elements: (i) a test facility for materials, the International Fusion Materials Irradiation Facility, IFMIF, (ii) continued exploration of concept improvements that may, in the longer term, be attractive, and (iii) continued development of the technology required for DEMO and future power plants. The Swedish fusion research unit encompasses a range of competencies that are important for the European fusion programme and for the ITER project. The Swedish Association has as its basic goal to make important contributions to the ITER project and to the long term goal of a prototype fusion reactor. 1.2 Research Unit The formation of a Swedish Fusion Research Unit is enabled by the Contract of Association between EURATOM and the VR. Swedish fusion research activities are carried out at four universities and one industry, which together form the Swedish Research Unit. The following universities participate in the fusion research: the Royal Institute of Technology (KTH) in Stockholm, Chalmers University of Technology (CTH) in Göteborg, Uppsala University (UU) and Lund University (LU). A group at Studsvik Energy AB is also a part of the Research Unit. The activity of the Association EURATOM-VR is directed by the Steering Committee, which during 2009 included the following members:

Members: Yvan Capouet (EU RTD-J4), Ruggero Gianella (EU RTD-J4), Marc Cosyns (EU RTD-J5), Doug Bartlett (alternate EU RTD-J4), Lars Börjesson (VR), Johan Holmberg (VR), Goran Bogdanovic (Ministry of Education).

HRU: James R Drake (KTH) Secretary: Per Karlsson (VR).

The research activities are carried out within the areas of fusion plasma physics and fusion technology. The fusion plasma physics research is mainly carried out at universities and is concerned with transport, stability including active control of instabilities, plasma wall interaction, heating of the plasma, energetic particles, and diagnostic development and implementation, in particular neutron diagnostics. This research includes both experimental and theoretical work with a strong element of computer code simulation. Emerging Technology projects are in the area of tungsten and tungsten alloy development. Studsvik Energy AB has been involved in fusion technology and is now focusing on work for Fusion


5

for Energy, the European agency responsible for fulfilling the EU responsibilities to the ITER International Organisation. The research activities within the RU are organized in a number of research projects which are presented in the Work Plan of the RU. These activities are well integrated into the EURATOM fusion programme and the activity is a part of the accompanying programme which supports the ITER project. It includes substantial participation in the EFDA JET projects as well as collaboration with other Associations. A special feature of the Swedish fusion Research Unit is that it is university-based and involves doctoral student participation and education. 1.3 Overview of research activities The Contracts of Association and the EFDA Agreement provide the framework for the co-ordinated European fusion research activity. All the member states with fusion research units participate in EFDA. The EFDA leadership during 2009 includes the EFDA Leader, Mr. Jerome Pamela and the EFDA Associate Leader for JET, Mr. Francesco Romanelli. The leadership is aided by staff forming Close-Support Units. The EFDA Steering Committee, made up of representatives from the Associations that are members of EFDA, functions as a management board for EFDA. The instrument for co-ordination of the work is a Work Plan prepared by the EFDA leadership and approved by the EFDA Steering Committee. The work plan normally spans several years. An annual EFDA Work Programme, based on the Work Plan, is also prepared by the EFDA leadership and approved by the Steering Committee. The EFDA Work Programme together with the Work Plans for the Research Units is used as the basis for the annual Work Programmes prepared for each research unit. The Work Programme for the Swedish Research Unit is approved by the Association Steering Committee on an annual basis. The Work Programme for the Swedish Research Unit for 2009 is summarised in Table 1.1. The total volume of activity in terms of Professional Person Years is about 39 PPY (a PPY is based on salaried activity). The number of professionals involved is about 65, of which 16 are PhD students. The two major areas of activity are Provision of support to the advancement of the ITER Physics Basis and Development of plasma auxiliary systems, the latter being primarily neutron diagnostics and spectroscopy–based diagnostics. The university academic staffs are also involved in undergraduate teaching, which is not included in the PPY accounting. The training of PhD students is an important part of the activity of the academic professional staff and this activity is integrated in the research projects. The activity of the Swedish RU is well integrated into the EFDA Work Programmes. The EFDA Work Programme is organised into Task Forces (TF) and Topical Groups (TG). There is a Task Agreement for each TF and TG where the activity undertaken by each RU is specified. In addition there is an EFDA JET Work Programme which specifies the exploitation of the jointly operated JET experiment. The JET activity is based on Scientific and Technical (S/T) tasks carried out by the Research Units under the supervision of the EFDA JET Associate Leader. A summary of Swedish RU involvement in the Work Programme for the EFDA Task Forces and Topical Groups and the EFDA JET Work Programme for 2009 is provided in Table 1.2. This table also indicates the financial provisions for additional support as specified in the CoA (Art 8.2) and the EFDA Agreement


6

(Art 6.3). An indication of the level of integration of the Swedish RU into the EFDA work programmes is seen in the fact that about 63% of the Swedish RU activity is directly specified in EFDA and EFDA JET Task Agreements (39% receiving baseline support and 24% receiving additional support in some form).

Table 1.1. Overview of Swedish RU Work Programme activity for 2009. EFDA Work Programme area Approximate person

power years (PPY) Provision of support to the advancement of the ITER Physics Basis

25.9 PPY

Energy and particle confinement / transport CTH-4.9 KTH-0.7 UU-0.5

MHD stability and plasma control KTH-3.0 CTH-1.2 UU-0.8

Power and particle exhaust, and plasma wall interaction KTH-4.2 UU-0.1

Physics of plasma heating and current drive KTH-2.0 Energetic particle physics CTH-5.0

KTH-1.5 UU-0.7

Theory and modelling for ITER CTH-0.8 KTH-0.4 UU-0.1

Development of plasma auxiliary systems 8.9 PPY Neutron diagnostics UU-7.1 Spectroscopy diagnostics KTH-1.3

LU-0.5 CTH-0.15

Development of concept improvements and advances in fundamental understanding of fusion plasmas

0.8 PPY

Operational regimes and plasma characteristics for improved concepts; theory and modelling

KTH-0.8

Emerging technologies 1.3 PPY Material science and advanced materials for DEMO (tungsten and tungsten alloys)

KTH-1.3

Secondments 2.4 PPY CSU-JET KTH-1.0

CTH-1.0 JOC UU-0.4 TOTAL 2009 39.3 PPY Training and career development All universities are

involved and the personnel participation is included under the respective activity areas.


7

The principal areas where the Swedish Association focuses its activity that are included in Task Agreements of the EFDA and EFDA JET Work Programme are summarised in Table 1.2.

Table 1.2. Summary of Swedish RU involvement in EFDA Task Forces and Topical Groups and EFDA JET Work Programme for 2009. Baseline

support PPY

Additional support

PPY EFDA Work Programme area Art 8.1 Art 8.2a EFDA Task Force Integrated Tokamak Modelling 1.7 1.0 EFDA Task Force Plasma Wall Interaction 1.5 1.0 EFDA Topical Group Materials 1.3 EFDA Topical Group Diagnostics 2.2 1.2 EFDA Topical Group Transport 2.7 0.2 EFDA Topical Group Heating & Current Drive 0.2 EFDA Topical Group Magnetohydrodynamics 2.7 0.2 subtotal EFDA TF and TG 12.3 3.6 EFDA JET Work Programme area for scientific and technical tasks

Art 6.3 Notif

Art 8.2c Art 6.3 Order

Enhancement Project 0.2 0.3 JET Fusion Technology 1.3 Campaigns 1.7 3.1 EFDA Goal Oriented Training Art 8.2e EFDA GOTiT 0.5 CSU Secondment Art 8.2g EFDA JET CSU 2.0 TOTAL 15.5 9.5 Percent of total Research Unit PPY 39.4% 24.1%

1.4 Highlights of the research activity Support to the advancement of the ITER physics base

• Energy and particle confinement: The transport work during 2009 concerns effects of plasma flows, momentum transport including transport simulations, particle transport in fluid and quasilinear kinetic descriptions including collisions, fluid closure, edge plasma physics including resonant magnetic perturbations and detailed geometry effects.

• Energy and particle confinement: The stability of ion temperature gradient modes have been analyzed using a semi-analytical model based on an approximate solution of the gyrokinetic equation.


8

• Energy and particle confinement: Various reasons for the magnetic field threshold for runaway electron generation in tokamak disruptions were investigated. It was shown, that even in rapidly cooling plasmas, where hot-tail generation is expected to give rise to substantial runaway population, whistler waves can stop the runaway formation below a certain magnetic field unless the post-disruption temperature is very low.

• MHD control: By enabling output-tracking, the Intelligent Shell (IS) concept has been generalized to the Revised Intelligent Shell (RIS). The output-tracking capability allows a non-zero reference radial magnetic field, expressed in term of Fourier coefficients with time-dependent amplitude and phases.

• MHD control: Plasma breaking due to resonant and non-resonant magnetic perturbations has been studied experimentally with controlled magnetic perturbation spectra.

• Plasma wall interaction: An important contribution in the area of PWI is tritium retention studies and the assessment of fuel inventory in castellated PFCs, i.e. in structures similar to those foreseen in ITER.

• Plasma wall interaction: Results of the systematic dust survey in TEXTOR confirm several previous data. The new important finding is the evidence for the existence of carbon debris, 10 nm – 200 µm, in the collected material. These may be products of brittle destruction.

• Physics of plasma heating: A fast code for routine simulations of ICRH is being developed. The code is based on the formulation used for the PION code using a 1D –Fokker-Planck solver with a model for the parallel velocity. The upgrading of the code consists of using replacing the formula for the power deposition with direct solution of the wave equation.

• Physics of plasma heating: The wall loads due to fusion alphas as well as NBI- and ICRF-generated fast ions have been simulated for ITER Reference Scenario-2 and Scenario-4 including the effects of ferritic inserts (FI), Test Blanket Modules (TBM), and 3D wall with two limiter structures.

Development of plasma auxiliary systems

• Neutron diagnostics: The work in the field of neutron diagnostics at Uppsala University for 2009 included the following main research activities: • Continued characterization of the TOFOR and MPRu neutron spectrometers at JET • Participation in the experimental programme at JET during campaigns C26 – C27, in

particular experiments concerning fast ions • Development of modelling and data analysis programs for the JET spectrometers • Instrumental development and installation of a camera/spectrometer system for MAST • R&D for neutron spectrometry on ITER, within the EFDA Topical Group Diagnostics

and the ITPA Diagnostics • Spectroscopy:

• Observations and interpretations of the q-profile behavior with auxiliary heating systems at JET

• Modelling of the total radiative power. Applications to the Tore Supra plasma • Correction method development for alkali beam emission spectroscopy (BES)

measurements and implementation into the RENATE BES simulation code (Istvan Pusztai)


9

Development of concept improvements • An article featuring the semi-implicit root solver SIR, designed for use in a new PDE

solver method, became of the most successful (most downloaded) IMACS papers in 2009.

• The novel time-spectral generalized weighted residual method (GWRM) was successfully applied to problems in ideal and resistive MHD. The potential of the method is to study confinement on long time scales without the use of time steps and associated limiting criteria.

The major specialised equipment used by the Association includes the following: The EXTRAP T2R reversed-field pinch is located at KTH. The UU group has delivered neutron spectrometers to JET (MPRu and TOFOR).


10

2 Support advancement of the ITER physics base 2.1 Energy and particle confinement and transport J. Weiland, L. Frassinetti, T. Fülöp, H. Nordman, P. Strand, M. Tendler, H. Wilhelmsson, Ansar Mahmood, V. I. Pusztai, Andreas Skyman, Michal Stransky. Introduction This project aims at obtaining an understanding of the main transport features of today’s fusion experiments and to find ways of improving the performance of a reactor. This is done by participating in several international working groups: (JET task forces, ITPA, the Integrated Modelling Task Force, (ITM), and the Transport Topical Group (TTG). There are also several other international collaborations going on. Summary The transport work during 2009 concerns effects of plasma flows, momentum transport including transport simulations, particle transport in fluid and quasilinear kinetic descriptions including collisions, fluid closure, edge plasma physics including resonant magnetic perturbations and detailed geometry effects. Also a workshop “From Leonardo to ITER” was organized in Göteborg where many fundamental aspects of transport were discussed. EFDA Task Force and Topical Group highlights Integrated Tokamak Modelling Task Force (ITM-TF) The European Task Force on Integrated Tokamak Modelling has the mid- to long-term goal of providing a validated suite of software tools for ITER exploitation. P. Strand is as of October 2006 the Task Force Leader. The work within the Task Force has progressed rapidly in several areas and the ITM has been starting to have an impact in the broader fusion community and are actively taking part in the development on the broader modeling programme in Europe. In addition, VR is represented by T. Hellsten as the project leader in the Heating & Current Drive project (Integrated Modelling Project #5 - MP#5) – coordinating the physics and modeling development in that area. Active participation in the projects outside of the leadership functions above is in IMP#5, Non-linear MHD (IMP#2) and to a lesser extent in the Turbulence and Transport project (IMP#4) together with a somewhat smaller contribution to the IMP#3 (transport code and whole device project). During 2007 resources were provided to coordinate a proposal under the EU Seventh Framework Capacities programme. This has lead to the EUFORIA project being funded at 3.65 M Euro during the three year period starting January 2008 with Chalmers being acting as host. EFDA JET highlights Our drift wave transport model has continued to give good results in comparisons with experiment in connection with our work under the EFDA-JET 2009 programme. Momentum transport has during 2009 mainly been tested on the formation of transport barriers (JET69454) where self consistent simulations were made of four channels, ion and electron temperatures and poloidal and toroidal momenta with surprisingly good agreement with the


11

experiment. Also the poloidal spinup was reproduced to location and magnitude. However finite beta effects are required for these results and are now being studied in more detail. We recall the good high beta results with our model from 2008 (Laborde et. al. Phys.Plasmas 15, 102507 (2008)). Also the works on transient momentum transport and stiffness have continued and the results are now improving. ITPA highlights We have participated in the ITPA Transport and Confinement, topical group in Naka where results on momentum transport and transport barrier formation were presented. A remarkable new result was the simulation of the formation of a transport barrier in four channels, ion and electron temperature and poloidal and toroidal momentum. This included the spinup of poloidal momentum in ITB and the total success of this simulation further secures the fact that the poloidal spinup of momentum in JET ITB’s is due to zonal flows. Good agreement was also obtained in some cases where the ITG mode was stable in the transport barrier region while the trapped electron mode was governing the transport. We have now used also the off diagonal elements from the stress tensor (included also in Multi Mode Model (MMM) 2008), thus we are now using only fluid theory. A new numerical scheme (due to Pereverzev) has made the code more stable and we can now simulate the spinup of poloidal momentum dynamically, (see fig. 2.1-1). The toroidal effects increase both the diagonal and off diagonal transport leaving the steady state results almost unchanged but increasing the transient transport. An important result from discussions in Naka was that the fact that the opposite sense of rotation in the simulations as compared to the experiment is likely to be due to the fact that the simulations describe main ions while the experimental measurements are made on impurity ions.

Fig. 2.1-1. Poloidal spinup in ITB as simulated for JET 69454. The full line is the initial conditions. The dashed line is neoclassical with barrier and the dotted line is the simulated. The mechanism is nonlocal.

Transport simulations Our drift wave transport model has been developed to include the perturbed parallel ion heatflow and this has been tested on momentum and energy transport with minor differences to the 2008 version which has been included in the MMM 2008. The general features of MMM 2008 were tested in References 14 and 15 in this section giving better agreement with experiment than GLF23. Recent results were reported under JET and ITPA highlights. We have participated in the ITPA (International Tokamak Physics Activities) Expert group meetings the EFDA-JET programme, the Transport Topical Group (TTG) and the Integrated Tokamak Modelling Task force (ITM-TF).


12

Fundamental understanding of fusion plasmas-Theory-Effects of plasma flows on transport Modeling of the Edge Plasma of MAST in the Presence of Resonant Magnetic Perturbations It has been demonstrated on DIII-D [T.E. Evans et al., Nucl. Fusion 48 024002 (2008)] and later on JET [Y. Liang et al., Phys. Rev. Lett. 98 265004 (2007)] that edge localized modes (ELMs) can be suppressed or mitigated by applying resonant magnetic perturbations (RMP) to the high confinement mode (H-mode) of a tokamak. Resonant coils for RMP are installed or planned on almost all large tokamaks: DIII-D, JET, MAST, ASDEX-Upgrade (AUG) and ITER. The widely accepted mechanism for ELM suppression during RMP is the reduction of the pressure gradient in the pedestal region below the stability limit for type I ELMs. The main contribution to the pressure gradient decrease is the pedestal density drop – the so-called ‘pump-out effect’, while the pedestal temperature does not drop and might even increase. Up to now this effect was not completely understood. On the other hand, as was known from several earlier [X. Z. Yang et al., Phys. Fluids B3 3448 (1991)] and recent [R. A. Moyer et al., J. Nucl. Mater. (2008)] observations, inside the stochastic layer the radial electric field becomes less negative or even positive and co-current toroidal rotation is generated. An analytical model which can describe these effects has been described in [V. Rozhansky, M. Tendler et al Nuclear Fus. 50 034005 (2010)] and the first simulations made by the B2SOLPS5.2 transport code demonstrated that the results are consistent with the analytical predictions. The impact of RMP on the structure of the edge plasma of MAST has been studied using the B2SOLPS5.2 transport code. Results are compared with experimental data obtained on MAST in the shots with and without RMP. The effect of RMP is taken into account by adding an electron current along a stochastic magnetic field to the current continuity equation solved in the code. The expression for the electron current is of Rechester - Rosebluth type with the stochastic diffusion coefficient calculated numerically. From analytical models it is known that the radial ion current which is generated to compensate the electron current changes the radial electric field in the separatrix vicinity and accelerates the plasma toroidally in the co-current direction. The additional particle flux, which is proportional to the ion current, might lead to a density drop near the separatrix. Simulations were performed for three L-mode shots with different densities with and without RMP. It was found that the electric field at the core side of the separatrix becomes less negative and the plasma is accelerated in the co-current direction in all the cases. The variations in the electric field and in the toroidal rotation are consistent with those measured in these shots. An H-mode shot where more frequent type III ELMs are triggered by the application of the coils has been simulated as well. For this discharge it was demonstrated that the additional particle flux in the barrier region due to the stochastic field might explain the pump-out effect and the rise of the pedestal electron temperature observed. However, for the L-mode cases the additional particle flux due to the stochastic field is not sufficient to cause the significant pump-out which is observed. The simulations show that in order to match the observed pump out, the transport coefficients in the L-mode shots with RMP have to be significantly increased compared to the shots without RMP. The effect is more pronounced for the low density cases. This increase in the transport coefficients correlates with the observed increase in the amplitude of ion saturation current fluctuations. It was shown that the level of the magnetic field perturbations required to match the pump-out effect in the H-mode and the variation of the radial electric field and toroidal rotation in the L-mode is significantly smaller than that calculated using the vacuum magnetic field, i.e. significant screening of the perturbed magnetic field is predicted. The mechanism of


13

screening that had been suggested in [E. Kaveeva, V. Rozhansky, M. Tendler, Proc.37th EPS Conf. on Contr. Fus. and Plasma Phys. (2010) P-2.139 ] is now applied to analysis of the MAST L-mode shots with RMP. It was shown that for the parameters of experiment the screening factor 53!=" which corresponds to the order of magnitude reduction of the stochastic conductivity in plasma with respect to the stochastic conductivity calculated using vacuum magnetic field. This is consistent with the magnetic diffusivity used in the simulation mDst

8105 !"= to be order of magnitude smaller than that calculated using vacuum magnetic field mDst

7105 !"= . Electric field and toroidal rotation calculated for the L-mode with and without RMP is consistent with the values measured on MAST in the same shot with the reciprocating probe, Fig.2.1-2 and Fig.2.1-3.

Fig.2.1-2. Comparison of experimental and simulated radial electric fields at the outer midplane for L-mode shots with and without stochastisity.

Fig. 2.1-3 Comparison of experimental and simulated Mac Mach numbers at the outer midplane for L-mode shots with and without stochasticity.

As a result it is demonstrated both in H and L-regimes of MAST a strong screening of the RMP takes place. The stochastic diffusion coefficient is reduced by order of magnitude with respect to the vacuum magnetic field. The change of the radial electric field and toroidal rotation velocities during RMP calculated with the reduced values of perturbed magnetic field are similar to the experimental observations in the L-mode of MAST. Impurity particle transport in tokamak plasmas in neoclassical and fluid descriptions Impurity particle transport in tokamaks has been studied using an electrostatic fluid model for main ion and impurity temperature gradient ( ITG! ) mo de and t rapped elect ron (TE! ) mode turbulence in the collisionless limit and neoclassical theory. The impurity flux and impurity density peaking factor obtained from a self-consistent treatment of impurity transport are compared and contrasted with the results of the often used trace impurity approximation. Comparisons between trace and self-consistent turbulent impurity transport have been performed for ITER-like profiles. It was shown that for small impurity concentrations the trace impurity limit is adequate if the plasma is dominated by ITG turbulence. However, in case of TE mode dominated plasmas the contribution from impurity modes may be significant, and therefore a self-consistent treatment may be needed.


14

Collisional model of quasilinear transport driven by toroidal electrostatic ion temperature gradient modes The stability of ion temperature gradient modes and the quasilinear fluxes driven by them have been analyzed in weakly-collisional tokamak plasmas using a semi-analytical model based on an approximate solution of the gyrokinetic equation, where collisions are modeled by a Lorentz operator. The results (Fig. 2.1-4) show that although the frequencies and growth rates of ion temperature gradient modes far from threshold are only very weakly sensitive to the collisionality, the ion-temperature-gradient threshold for stability is affected significantly by electron-ion collisions. The decrease of collisionality destabilizes the ITG mode driving an inward particle flux, which leads to the steepening of the density profile. Closed analytical expressions for the electron and ion density and temperature responses have been derived without expansion in the smallness of the magnetic drift frequencies. The results have been compared with gyrokinetic simulations with GYRO and illustrated by showing the scalings of the eigenvalues and quasilinear fluxes with collisionality, temperature scale length and magnetic shear.

Fig 2.1-4 Stability boundaries of the ITG mode. The upper curves are the critical inverse ion-temperature-gradient scale length and the lower ones are the normalized real frequency of the mode. Left: Shear scaling of the stability boundary and real frequency for a/L_n=1. The solid black lines are from COMET for adiabatic electrons, the dash-dotted blue lines are the results of linear GYRO simulations. Right: a/L_n-scaling oft the stability boundary and real frequency for s=1. The collisionality increases from the thin to the thick lines. The stability boundary for adiabatic electrons is indicated with dotted lines.

Pedestal confinement in JET The hybrid confinement regime has been extensively explored in the 2008-2009 JET campaigns, obtaining confinement factors up to H98~1.4 and normalised pressure !N~3 in both low and high triangularity (") configurations. The aim of the present study was to examine the contribution of the pedestal to the total confinement in the hybrid regime in JET. The pedestal parameters are measured using the new High Resolution Thomson Scattering (HRTS) data measuring electron temperature and density combined with Charge Exchange data measuring the ion temperature and the effective charge. Ion density is estimated from Zeff assuming carbon as main impurity. To investigate the origin of the better confinement in the hybrid scenarios, a comparison with baseline H-mode scenarios is made.


15

The correlation between the pedestal temperature and the pedestal electron density, Fig. 2.1-5, is useful to describe the four groups of shots. The baseline low ! shots are part of a current scan study. High current shots are characterized by high pressure. The baseline high ! shots are part of current scan and fuelling experiments. The pedestal pressure increases with increasing current. At fixed current the pedestal density is increased by gas fuelling while the pedestal temperature is decreased. As a result the pedestal pressure is kept fixed and the data follows approximately a constant pressure curve. For the hybrid low ! a large variation in the pedestal temperature is present, while the density is approximately constant. This behaviour is due to (i) variation of the heating power, (ii) current profile optimisation by variation of heating timing and current waveforms (iii) variation of the fuelling schemes that were employed in NTM avoidance schemes. The hybrid high ! are characterized by current scan and gas fuelling. In this regime the pedestal pressure is not maintained at increased fuelling levels. Figure 2.1-6(a) shows the dependence of the pedestal (Wped) and total (Wtot) stored energy calculated from kinetic profiles compared to stored energy according the scaling W98. For the hybrid plasmas Wtot rises well above W98 by up to 50% compared to the baseline plasmas, as expected from the improved confinement factor H98 in these plasmas. Moreover, also the pedestal energy for the hybrid plasma is approximately 50% higher than baseline Wped. The relative contribution of the pedestal to the total energy content versus H98 is shown in Fig. 2.1-6(b). A weak negative trend is present for the baselines. For the hybrids, the large scatter does not allow any conclusive claim, but the results seem to suggest the opposite (positive) trend for the high ! regime. A clear separation is visible between the low and high triangularity discharges, independent of the regime (hybrid or baseline). To further investigate the origin of the improved confinement, the role of electrons and ions has been studied separately by calculating the total stored energy for the two species. Fig. 2.1-6(c) shows the ratio between of Wions and Welec versus H98. This ratio, being Wions/Welec>1 for the hybrids, seems to show that the hybrid increased performances are driven more by the ions. Note that at the same H98 the ion contribution to the total energy is higher for the low ! hybrids than for the high !. For the baseline plasmas the opposite relation holds, Wions/Welec<1. In Fig. 2.1-6(d) the ratio between core Ti and Te is shown versus the electron density. While for the hybrid scenarios 1<Ti/Te<2, for the baseline scenarios Ti/Te ! 1 in the core for the entire density range.

Fig. 2.1-5. Pedestal temperature versus electron pedestal density. Dashed lines represent constant pressure curves. The green dots are low hybrids plasma with density control in which only the input power is varied.


16

Fig. 2.1-6 (a) Total energy (full symbols) and pedestal energy (open symbols) versus W98. The dashed lines are arbitrary lines to highlight the separation between hybrids and baseline plasmas and the numbers correspond to the slopes. (b) Pedestal energy normalized to total energy versus H98. (c) Ratio of ions energy and electron energy versus H98. (d) Ratio of Ti over Te in the core versus the core density.

Fig. 2.1-7 shows examples of profiles for different collisionality !eff. A negative correlation of the density peaking with !eff is seen (Fig. 2..1-7(a)). The peaking of the electron temperature increases slightly with !eff (Fig. 2.1-7(b)). As a result, the pressure profile peaking barely depends on collisionality (Fig. 2.1-7(c)). For !eff <0.25 however, corresponding to the low " hybrids with the highest Ti/Te and highest H98, the trend of the Ti peakedness is broken. In this regime the electrons and ions are decoupled. No strong ion ITB is seen in these plasmas, but the ion temperature is indeed steeper for the high Ti peaking shots (Fig. 2.1-7(d)).


17

RMP screening in stochastic magnetic fields It was demonstrated on many tokamaks that the edge localized modes (ELMs) could be suppressed or mitigated by applying resonant magnetic perturbations (RMP) to the high confinement regime (H-regime) of a tokamak. The resonant coils for RMP are installed or planned on almost all large tokamaks: DIII-D, JET, MAST, ASDEX-Upgrade (AUG), NSTX and ITER. The widely accepted mechanism of ELMs suppression during RMP is the reduction of the pressure gradient in the pedestal region below the stability limit for type I ELMs. The main contribution to the pressure gradient decrease is the pedestal density drop – the so-called ‘pump-out effect’, while the pedestal temperature does not drop and might even increase. The existing models for ‘pump-out effect’ require knowledge of the level of the

Fig. 2.1-7. Profiles normalized to the pedestal height for (a) electron density (b) electron temperature and (c) electron pressure, for !eff <0.15 (red) and !eff>1.9 (blue). (d) Normalized Ti profiles for low " hybrids with Ti

core/Tiped>3.8 (red)

and Ticore/Ti

ped<2.6 (blue).

!

!

!

!

(a)

(b)

(c)

(d)


18

magnetic field perturbations in plasma. However, there is some evidence that the vacuum magnetic field perturbations are strongly screened by the plasma so that the resulting RMP are significantly lower that the vacuum ones. The existing RMP screening models are based on the screening of a separate magnetic island. According to these models plasma inside the non-rotating magnetic island is at rest while outside there is both poloidal and toroidal rotations. Inertia force in the layer adjacent to the island separatrix causes radial currents which are closed by the electron parallel currents. They alter the vacuum magnetic perturbations. The screening effect is calculated in the cylindrical approximation, so that the account of toroidal effects and turbulent transport is bound to change the result. Moreover, the screening models for the separate island are not directly applicable to the case of RMP due to overlapping of the magnetic islands and stochastization of the magnetic field. Hence, the radial pressure gradient, the radial electric field, poloidal and toroidal flows remain finite inside a region of stochastization in contrast to the case of a separate island. Therefore, the impact of the RMP to the issue of screening is required. Hence, the project addresses the screening effect for the stochastic magnetic field. It is demonstrated that in the existing experiments the RMP screening might be rather large and should be taken into account in the simulations of the edge plasma parameters. The issue of RMP screening is extremely important for the success of ITER. Publications section 2.1 Peer reviewed journals

1. J. Weiland, R. Singh, H. Nordman, P. Kaw, A.G. Peeters and D. Strinzi, Symmetry breaking effects of toroidicity on toroidal momentum transport. Nuclear Fusion. 49 s. 065033 (2009).

2. M. Ansar Mahmood, T. Rafiq, M. Persson and J. Weiland, Collisionless trapped electron and Ion temperature gradient modes in an advanced tokamak equilibrium, Physics of Plasmas 16, 022503, (2009).

3. A. Zagorodny and J. Weiland, Non-Markovian renormalization of kinetic coefficients for drift type turbulence in magnetized plasmas. Physics of Plasmas. 16 (5) s. 052308. (2009).

4. V. Parail, P. Belo, P. Boerner, X. Bonnin, G. Corrigan, D. Coster, J. Ferreira, Foster, L. Garzotti, G.M.D. Hogeweij, W. Houlberg, F. Imbeaux, J. Johner, F. Kochl, V. Kotov, L. Lauro-Taroni, X. Litaudon, J. Lonnroth, G. Pereverzev, Y. Peysson, G. Saibene, R. Sartori, M. Schneider, G. Sips, P. Strand, G. Tardini, M. Valovic, S. Wiesen, M. Wischmeier, R. Zagorski, JET EFDA contributorsb and EU ITM Task Force. Integrated modelling of ITER reference scenarios, Nuclear Fusion 49, 075030 (2009).

5. Fülöp, T., Nordman, H. (2009). Turbulent and neoclassical impurity transport in tokamak plasmas. Physics of Plasmas. 16 (3) s. 032306 (2009).

6. Pusztai I, Fülöp T, Candy J, Hastie J, Collisional model of quasilinear transport driven by toroidal electrostatic ion temperature gradient modes, Physics of Plasmas, 16 p. 072305 (2009).

7. M. N. A. Beurskens T. H. Osborne L. D. Horton L. Frassinetti R. Groebner A. Leonard P. Lomas I. Nunes, S. Saarelma., "Pedestal width and ELM size identity


19

studies in JET and DIII-D; implications for ITER", Nuclear Fusion, vol. 51, (no. 7,) pp. 124051, 2009

8. M. Tendler, et al., “Measurements of the temporal evolution of plasma rotation in the TCABR tokamak”, Nucl. Fusion, vol. 49, 115026, 2009

Summer College on Plasma Physics, ICTP Trieste, 10 – 28 August 2009.

9. Weiland, J. Turbulent Transport in Fusion Plasmas, Effects of Toroidicity and Fluid Closure. American Institute of Physics Proceedings. 1188 s. 193 (2009).

From Leonardo to ITER, Gothenburg May 18-19 (2009)

10. Zagorodny, A., Weiland, J. Non-Markovian Effects in Drift-type Turbulence. American Institute of Physics Proceedings. From Leonardo to ITER s. 72 (2009)

11. Wilhelmsson, H. From Waves to Vortices: A Novel Approach to General Relativity. American Institute of Ph

Book Chapters

12. H. Wilhelmsson, (2009). Large-scale vortex dynamics. Recent Research Developments in Physics. 8 s. 127. Transword Research Network, Trivandrum, India (2009).

13. H. Wilhelmsson, (2009). Coupled Vortex Solutions Recent Research Developments in Physics. 8 s. 117. Transword Research Network, Trivandrum, India (2009).

14. J. Weiland, A. Zagorodny, V. Zasenko, Fluid and Kinetic Modelling on Timescales Longer than the Confinement Time in Bounded Systems. American Institute of Physics Proceedings. From Leonardo to ITER, s 96 (2009).

Presentations at EPS conferences

15. F.D. Halpern, A.H. Kritz, G. Bateman, R.V. Budny, J. Weiland, Multi-mode modeling of toroidal momentum confinement in tokamaks, 36th EPS Conference, Sofia, June 29-July 3 2009, ECA Vol. 33E, Paper P-2 199.

16. A.H. Kritz, F.D. Halpern, G. Bateman, D.C. McCune, R.V. Budny, A.Y. Pankin, T. Rafiq, J. Weiland, PTRANSP: Predictive Integrated Tokamak Modeling, 36th EPS Conference, Sofia, June 29-July 3 2009, ECA Vol. 33E, Paper P-2 200.

Talks at ITER and ITPA workshops

17. J. Weiland, P. Mantica, T. Tala, V. Naulin, K. Krombe, A.G. Peeters, D. Strinzi, Simulation of toroidal and poloidal momentum including symmetry breaking toroidal effects, ITPA workshop on Transport and Confinement, Naka, Japan March 31 – April 2 2009.

Other workshops and conferences

18. C. Giroud, C. Angioni, M. Baruzzo, H. Nordman, Impurity Characterization in JET operational scenarios, 12th INT. Workshop on H-mode Physics and Transport Barriers. Princeton, Sept. 30 (2009).

19. D. Coster, X. Bonnin, D. Reiter, .....P. Strand, Simulations of the Edge Plasma: the role of atomic, molecular and surface physics. ICAMDATA-2008, 1125, 113-122 (2009).

20. V.Zasenko, A. Zagorodny and J. Weiland, Permittivity of plasma in random field of moderate intensity, Bogolyobov Kyiv Conference, Modern Problems of Theoretical and Mathematical Physics, September 15 – 18 2009, Kyiv, Ukraine.


20

RUSA meeting Uppsala May 5-6 (2009). 21. J. Weiland, Transport research, invited overview of the international progress 2008 –

2009.

Licentiate Thesis 22. Pusztai I: Modelling and measuring transport in fusion plasmas, Chalmers University

of Technology, September 2009. 2.2 MHD stability and plasma control 2.2.1. Active MHD mode control experiments P. Brunsell, E. Olofsson (PhD student), L. Frassinetti, J. R. Drake In collaboration with: W. Suttrop, Max-Planck-Institut für Plasmaphysik, Garching, T. Bolzonella, R. Paccagnella, G. Manduchi, RFX Team, Consorzio RFX, E. Witrant, UJF-INPG/GIPSA-Lab, Grenoble, France, C. R. Rojas, H. Hjalmarsson, EES/Automatic Control, KTH EXTRAP T2R device Fusion Plasma Physics operates the EXTRAP T2R reversed field pinch device shown in Fig. 2.2.1-1. The device is equipped with a thin shell, which enables studies of resistive wall mode stability. An extensive array of external control coils provides excellent capabilities for research on active magnetic feedback control. The active control system was developed and installed on EXTRAP T2R during 2004-2006 in collaboration with Consorzio RFX. The main parameters of the device are shown in the table below: Table 2.2.1-1. EXTRAP T2R parameters

Parameter Notation Value Unit

Major radius R 1.24 m

Minor radius a 0.183 m

Wall diffusion time !v 6.3 ms

Plasma pulse length !d <100 ms

Plasma current Ip <150 kA

Plasma electron temperature (typical) Te 300 eV

Plasma electron density (typical) ne 1x1019 m-3


21

Fig. 2.2.1-1 The EXTRAP T2R device at the Alfvén Laboratory

A total of around 900 magnetic sensors are part of a comprehensive diagnostic system for studies of MHD instabilities and active MHD mode control. Other plasma diagnostics installed on EXTRAP T2R include, electric and magnetic probe array for turbulence studies, collector probes for plasma wall interaction studies, VUV and visible spectroscopy, Thomson scattering, interferometer, neutral particle time-of-flight diagnostic, bolometer array, SXR camera. The active MHD mode control system installed on EXTRAP T2R is based on extensive arrays of active coils and sensors distributed over the toroidal surface as shown in Fig. 2.2.1-2. The main features of the active control system are: • Magnetic flux loop sensors at 4 poloidal and 32 toroidal positions inside the thin shell, a

total of 128 sensors. • Active saddle coils at 4 poloidal and 32 toroidal positions outside the thin shell, a total of

128 coils. Saddle coils and sensor flux loops are pair-connected at each toroidal position to form 64 independent m=1 coils and sensors.

• Power amplifiers units providing at total of 64 independent channels. Audio amplifiers are used with output power of 800-1200 Watt and bandwidth 1 Hz to 25 kHz. Amplifier output currents are up to 20 A, providing 800 At in the power coils and a maximum radial magnetic field at the coil centre of about 3 mT.

• An integrated digital computer based controller unit. The controller unit is contained in a VME bus crate and includes ADCs for analog input of 64 magnetic sensor signals and 64 coil current signals, Board with PPC CPU, DACs for analog output of 64 amplifier control voltages. Controller algorithms are implemented in software.


22

The research program on MHD stability and control aims at the development of physics understanding and practical methods for tokamak and RFP. Control algorithms are implemented and tested utilizing the control system installed at EXTRAP T2R. A specific task is the development of a controller design for the planned experiments on active Resistive Wall Mode (RWM) stabilization at ASDEX Upgrade, which is carried out in collaboration with Max-Planck-Institute für Plasmaphysik and Consorzio RFX. The revised intelligent shell The basic idea underlying the Intelligent Shell (IS) concept is that the active coil must respond to the radial magnetic field measured by the corresponding sensor coil by producing a magnetic field with opposite direction. Ideally, the final result of the IS feedback would be to obtain a zero radial magnetic field at the sensor coils in order to simulate a perfect conductor surface. Indeed, the IS has proven to be a useful tool for RWM stabilization in several machines. By enabling output-tracking, the IS concept has been generalized to the Revised Intelligent Shell (RIS). The output-tracking capability allows a non-zero reference radial magnetic field, expressed in term of Fourier coefficients with time-dependent amplitude and phases.

Fig. 2.2.1-3: Model plant for Revised Intelligent Shell (RIS) feedback.

Fig. 2.2.1-2 Two-dimensional arrays of sensor flux loops and active saddle coils installed at EXTRAP T2R.


23

The model for the realization of the RIS feedback scheme is shown in Fig. 2.2.1-3. It is composed of three main parts: (i) the actuators (power amplifiers and active coils) that produce coil currents Ii(t), (ii) the machine with its passive structures (vacuum vessel and resistive shell) and the sensor coils that measure bi(t), (iii) the digital controller that determines the feedback law and calculates the voltage input vi(t) to the actuators. To implement the RIS algorithm, the identification of the matrices Gact, GT2R and FRIS, is necessary. The identification of the matrix Gact can be considered as a calibration of the actuators: given an input signal to a power amplifier it is necessary to know the corresponding coil current produced. From a practical point of view, an analytic expression dependent on free parameters is obtained by modeling the actuators as a system composed of a low pass filter, a high pass filter, a delay and a static gain. The free parameters are calculated by measuring the coil current produced by using input voltages perturbed with a Pseudo-Random-Binary-Sequence (PRBS) and then comparing experimental results with the model. The identification of the matrix GT2R can be considered as a calibration of the machine passive structures (shell, vessel, sensor coils and other mechanical structures): for given active coil currents in the actuators, the signal measured by any sensor coil is required. Note that this formulation takes into consideration the fact that the field measured by a sensor coil is affected by the coil currents produced by all the actuators. Same as for the actuator identification, an analytic expression is obtained and the free parameters are determined by applying the PRBS method to vacuum shots. The RIS algorithm is developed using two main simplifications: (i) the system identification performed in vacuum shots is valid also in plasma shots and (ii) the controller gain tuning is performed assuming a nominal instability growth rate. The identification of the matrix FRIS, consists in the determination of its two parts Meq and FPID. Meq compensates for the coupling of the active coil to neighbouring sensor coils by inducing small currents in neighbouring active coils so that total measured field in the neighbour sensors is close to zero. FPID is a diagonal matrix containing the gains of a Proportional-plus-Integral-plus-Derivative (PID) controller for each channel. The gains are determined by comparing the EXTRAP T2R model with a known second order unstable process for which the optimal gains are known. This last step implies the assumption of a nominal instability growth rate of the order of the shell diffusion time. An experimental application of the RIS with all references set to zero is shown in Fig. 2.2.1-4. Panel (a) shows the time evolution the radial component of the magnetic field harmonics as measured by the sensor coils for a shot without feedback, panel (b) with the standard IS and panel (c) with RIS and reference signals set to zero for all harmonics. The IS and the RIS achieve similar suppression of the radial field. A RIS shot with reference signal bref = 0.4mT for mode m=1, n=-12 is shown in Fig. 2.2.1-5. Panels (a) and (b) show the radial magnetic field. Panels (c) and (d) show the coil current harmonics induced to generate the magnetic field shown in Fig. 2.2.1-5. Clearly, the coil current pattern is quite complicated. This is explained by the fact that, in addition to generating the desired field harmonic, the controller is simultaneously generating coil currents for active feedback compensation of magnetic field errors present in EXTRAP T2R device.


24

Fig. 2.2.1-4: Time evolution of radial magnetic field harmonics m=1 for a plasma shot (a) without feedback, (b) with intelligent shell and (c) with revised intelligent shell.

Fig. 2.2.1-5: (a) Time evolution of radial magnetic field harmonics m=1 for a plasma shot with RIS and reference signal for the harmonic (1,-12) set to 0.4mT and (b) corresponding spectrum at t=25ms. (c) Corresponding time evolution of the active coil current harmonics and (b) spectrum at t=25ms. The vertical continuous lines in frames (a) and (c) highlight the end of the discharge.


25

The Fourier mode-decoupled control system The RIS feedback control is not tuned on the experimental plasma response but a common nominal growth rate is assumed for all modes. Closed-loop identification of the plasma parameters has been done by applying the PRBS method to plasma shots. The identified mode growth rates and the shell diffusion times are plotted in Fig. 2.2.1-6. These datasets are then the fundament for the design of an Fourier mode-decoupled controller. Output (total radial field) tracking and m=1 mode tracking are both considered in this controller scheme.

Fig. 2.2.1-6. Growth rate and diffusion time determined with the closed loop identification experiments as function of the toroidal mode number n.

Development of MHD mode control at ASDEX Upgrade The ASDEX Upgrade enhancement project for active MHD control is carried out in collaboration between Max-Planck-Institut für Plasmaphysik, Forschungszentrum Jülich, Consorzio RFX and Royal Institute of Technology (KTH). ASDEX Upgrade is currently being enhanced with 24 in-vessel saddle coils for (i) Edge Localised Modes (ELM) mitigation by Resonant Magnetic Perturbation (RMP), (ii) locking of plasma modes to rotating magnetic error fields for disruption avoidance and (iii) feedback control of Resistive Wall Modes (RWM). These coils are arranged as three toroidal rings of eight coils each at different poloidal positions at the low field side inside the ASDEX Upgrade vacuum vessel. The coils will be used to produce stationary or time-dependent perturbation fields with toroidal mode numbers n=1-4. The phase and amplitude can be specified as programmed waveforms or obtained as result of feedback control based on detected growing plasma instabilities. The coils are driven by 1 to 24 switched converters. If fewer converters than coils are available, two or more coils are connected in series for each converter. KTH involvement in the project is mainly in the design of the local controller that produces the request waveforms for the coil power supplies based on supervising commands from the ASDEX Upgrade control system and measurements of a set of poloidal field sensors in the


26

ASDEX Upgrade vacuum vessel. Furthermore, it performs signal conditioning operations to the sensor signals and the signal processing algorithms to ensure the desired controller response. For RWM feedback the local controller derives mode signals from the sensor signals (several different modes are controlled simultaneously) and determines the best-fitting amplifier (coil) current distributions. 2.2.2 Application of resonant magnetic perturbations L. Frassinetti, M.W.M. Khan (PhD student), P. Brunsell, J. R. Drake, External magnetic perturbations are now an essential tool in fusion plasma physics to address severe problems related to Edge Localized Modes (ELMs), magnetic islands and Neoclassical Tearing Modes (NTMs). Resonant Magnetic Perturbations (RMPs), by interacting with the corresponding plasma Tearing Modes (TM), can (i) produce stochastic magnetic field in the pedestal region mitigating ELMs , (ii) induce rotation for NTM-stabilization or (iii) move the magnetic island to an optimal position for Electron Cyclotron Current Drive (ECCD) stabilization. While theoretical studies on the interaction between RMPs and TMs are relatively developed, recent experimental works are focused mainly on the RMP effect on global plasma parameters. From an experimental point of view, the effect of an RMP on the TM dynamics and on the plasma flow is not yet studied in a controlled fashion. Effect on tearing mode dynamics In EXTRAP T2R, the most internal TM corresponds to the helicity (m,n)=(1,-12) and it is characterized by a helical angular velocity in the range 20-40 krad/s, in agreement with plasma flow measurements. The typical TM dynamics in EXTRAP T2R is characterized by sawtooth crashes that occur with a period around 0.3-0.4ms. When no external perturbation is present, the TM amplitude is relatively steady in between crashes and the TM velocity has no significant variations. The application of an external perturbation can drastically affect this behaviour. In Fig. 2.2.2-1, the time evolution of the most internal TM (corresponding to helicity m=1, n=-12) is shown when an RMP with amplitude 0.3mT and helicity (1,-12) is applied. A relatively clear sinusoidal modulation is present. This modulation is well related to the mode phase. Fig. 2.2.2-1(b) shows the correlation between the TM amplitude and the phase difference between the TM and the applied RMP. Approximately, amplitude amplification occurs when the TM is in phase with the RMP and amplitude suppression when the TM is out of phase. A similar correlation is found for the TM velocity, but acceleration and deceleration phases have a !/2 shift respect to the TM amplitude amplification and suppression. This behaviour is in good agreement with a model describing the evolution equations for magnetic islands chains in toroidal pinch plasmas subject to applied RMPs. According to the model, the RMP modifies the Rutherford equation and the torque balance equation with a perturbation proportional to the RMP field. A simulation obtained by adapting the model to EXTRAP T2R geometry and equilibrium gives a reasonable agreement with the experimental results, Fig. 2.2.2-1(b).


27

Figure 2.2.2-1. (a) TM amplitude time evolution when a 0.3mT RMP with helicity (1,-12) is applied. (b) Correlation between TM amplitude and phase difference for experimental data (full dots) and modeled data (open dots).

A significantly different TM dynamics is obtained when the RMP amplitude is relatively large. In Fig. 2.2.2-2, the TM amplitude time evolution is shown for an RMP of amplitude 0.5 mT. In this case the modulation is much stronger and when TM and RMP are out of phase, the TM amplitude can be suppressed to zero. After the suppression, the TM has a ! phase flip and re-appears approximately in phase with the RMP; see Fig. 2.2.2-2(b). A reasonable agreement between model and experiment is found. Finally, it is worth mentioning that the two dynamics described above can occur during the same sawtooth cycle. For example, the dynamics characterized by the weak TM amplitude oscillation can be followed by the phase jumps and eventually by the TM wall locking.


28

Fig. 2.2.2-2. (a) TM amplitude time evolution when a 0.5mT RMP with helicity (1,-12) is applied. (b) Correlation between TM amplitude and phase difference for experimental data (full dots) and modeled data (open dots).

Effect on plasma flow and tearing mode velocity The RMP produces a perturbation on the TM velocity. The perturbation is dependent on the phase difference between TM and RMP, and the therefore the velocity is characterized by phases of acceleration and deceleration; see Fig. 2.2.2-3(a). Due to the plasma viscosity and to the mode amplitude modulation, the phase difference is not linear in time; the TM spends more time in the deceleration phase than in the acceleration phase. Therefore, an overall TM braking effect is observed; see the continuous line in Fig. 2.2.2-3(b). It is different time scales for the two phenomena: the velocity modulation occurs on the fast time scale of around 10 µs, while the velocity braking is much slower with a slow time scale of around 10ms. In Fig. 2.2.2-3(b), the TM velocity has been smoothed with a 1 ms window to remove the fast oscillations and highlight the braking. Also the plasma flow is modified by the RMP application. In Fig. 2.2.2-3(b), the dots correspond to the toroidal velocity of oxygen impurity ions from spectroscopic measurement of the Doppler line shift. The OV impurity ions are mainly concentrated in the EXTRAP T2R plasma core and assuming that impurities and plasma co-rotate, the measured velocity is representative of the plasma flow in the region where the TM is resonant. Since the ion velocity is a line integrated measurement while the TM velocity can be considered as a local


29

measurement, differences are unavoidable. Nevertheless, TM and OV impurity velocities show a reasonable agreement.

Fig. 2.2.2-3. (a) TM velocity modulation on a short time scale when a 0.5mT RMP with helicity (1,-12) is applied. (b) TM velocity (continuous line) and OV velocity (dots) on a long time scale. (c) TM velocity profile variation for three different shots. Red: RMP (1,-12) is applied; blue: RMP (1,-13) and green: RMP (1,-15).

The deceleration is not limited to the plasma core; the whole velocity profile is affected. This result is obtained both by measuring the toroidal velocity of impurities concentrated not in the core and by analyzing the velocity of TMs that are not resonant in the core. Fig. 2.3.3(c) shows the variation of TM velocity profile when an RMP is applied. The three lines correspond to three different shots. For the red line, an RMP with helicity (1,-12) is applied, for the blue the helicity (1,-13) is used and for the green the helicity (1,-15). In all three cases the RMP amplitude is 0.3 mT and the profiles are calculated when the TM velocity has


30

reached a steady state. Despite the RMP amplitude is the same, the TM velocity variation is different both in the amplitude and in the profile shape: the maximum variation is obtained approximately at the resonant surface corresponding to the helicity of the applied RMP. Plasma rotation braking by non-resonant fields Non-resonant m=1 modes in EXTRAP T2R typically correspond to weakly stable or unstable RWMs. Experiments have shown that if the non-resonant RWMs are not suppressed by feedback control, this leads to slowing down of the plasma fluid rotation and may also result in wall locking of other core-resonant TMs. A controlled series of experiments to study possible plasma rotation braking by non-resonant external fields has been carried out. The array of external coils installed on EXTRAP T2R is utilized to produce helical fields with poloidal mode number m=1 and toroidal mode number n that correspond to non-resonant modes. Changes in the plasma rotation occurring with the applied field are estimated using impurity ion Doppler spectroscopy. In addition, resonant TM phase velocities are measured in order to obtain information on the radial profile of the plasma rotation. Preliminary results with an external (m=1, n=-10) helical field are quite interesting. The initial plasma velocity is 50-55 km/s, in the same direction as the toroidal plasma current and central toroidal magnetic field. With the external field applied, a reduction of the plasma velocity is observed, which is stronger with increasing external field amplitude. At the highest external field around 0.6 mT, the plasma rotation slows down to 35-40 km/s. Measurements of TM phase velocities suggests that there is an overall decrease of the plasma rotation profile, indicating a non-local braking mechanism. A first conclusion from the experiments is that the retardation observed is non-localised. Neoclassical Toroidal Viscosity (NTV) due to the applied asymmetric external field is one possible process that results in a non-local braking effect. However, further investigations are needed in order to make more definite conclusions about the origin of the braking process observed in EXTRAP T2R experiments. Publications sections 2.2.1 and 2.2.2 Peer reviewed journals

1. Lorenzo Frassinetti, Per R Brunsell, Marco Cecconello, James R Drake, "Heat transport modeling in EXTRAP T2R", Nuclear Fusion, vol. 49 pp. 025002, 2009.

2. Lorenzo Frassinetti, Per R Brunsell, James R Drake, "Heat transport in the quasi-single-helicity islands of EXTRAP T2R", Physics of Plasmas, vol. 16 pp. 0325503, 2009.

3. Lorenzo Frassinetti Erik Olofsson Per Brunsell, James Drake., "Resonant magnetic perturbation effect on the tearing mode dynamics in EXTRAP T2R: experimental results and modeling", 14th Workshop on MHD stability control, 2009


4. W Suttrop, D Hahn, J Hartmann, M Rott, B Streibl, W Treutterer, T Vierle, D Yadikin, I Zammuto, E Gaio, V Toigo, P Brunsell, E Olofsson, "Physical description of external circuitry for resistive wall mode control in ASDEX Upgrade", Proc 36th EPS Conference on Plasma Physics, pp. P-1.165, 2009


31


5. Erik Olofsson, Per Brunsell, James Drake, "Closed-loop direct parametric identification of magnetohydrodynamic normal modes spectra in EXTRAP -T2R reversed-field pinch", Proc 18th IEEE International Conference on Control Applications, pp. 1449, 2009

6. Erik Olofsson, Per Brunsell, "Closed-loop identification and controller reconfiguration for EXTRAP T2R reversed-field pinch", Proceedings of 23rd symposium on fusion engineering, 2009

7. Lorenzo Frassinetti Erik Olofsson Per Brunsell, James Drake., "Resonant magnetic perturbation effect on the tearing mode dynamics in EXTRAP T2R", Swedish Fusion Research Unit Meeting, 5-6 May 2009, Ångström Laboratory, Uppsala University, Sweden, 2009

Master thesis

8. Dong Li, Data analysis tool for study of MHD modes in active control experiments at the EXTRAP T2R, 2009

2.2.3 MHD stability C. Wahlberg Collaboration with: Jonathan Graves CRPP Lausanne and Ian Chapman UKAEA Culham The main focus of this work is on the GAMs and global MHD oscillations and instabilities existing in the core region of tokamaks and, in particular, the influence of toroidal plasma flows on such modes. The activity during 2009 has dealt both with global MHD aspects of axisymmetric and non-axisymmetric GAMs in static and rotating plasmas as well as the effects of toroidal plasma flows on the internal kink mode. Low-frequency MHD in toroidally rotating plasmas In Ref. [1], a generic equation is derived, describing, within a unified framework, i) both the Alfvén and the acoustic continua in the plasma, ii) Alfvén-acoustic global modes, and iii) infernal, ideal MHD instabilities in low-shear tokamaks including, in all three cases, the effects of toroidal plasma flows. This paper also gives an analytical description of the large-scale, global MHD properties of the axisymmetric GAMs existing both in static and rotating plasmas [3], including an explanation for the m = ±2 halo surrounding the GAM surface previously seen only in numerical computations. The theory developed provides a suitable starting point and framework for analysing rotational effects as well as other aspects of the infernal, ideal modes currently observed in several tokamaks (especially STs) with low magnetic shear in the core (q ! 1 Effects of toroidal rotation on the sawtooth instability Considerable progress in the understanding of rotational effects on the sawtooth instability in tokamaks has been achieved in a combined analytical and numerical study in collaboration with CRPP Lausanne and UKAEA Culham. In particular, the different roles played by the centrifugal effects on the equilibrium and by the radial profiles of the plasma density and rotation velocity for the stability of the internal kink mode have been clarified in detail in this work. The importance of selfconsistently including the centrifugal effects in the equilibrium


32

for a correct prediction of the growth rate of this mode is demonstrated in Fig. 2.2.3-1 for a plasma with parameters !a = 0.1, q0 = 0.938, r1/a = 0.3, "p = 0.3 and "0 = 0.66 %.

Fig. 2.2.3-1. Growth rate of the internal kink mode vs rotation velocity at the axis for different profiles of plasma density and rotation frequency calculated for a non-selfconsistent equilibrium in (a) and for a selfconsistent equilibrium in (b) The symbols show the numerical results and the dashed curves the analytical predictions.

In the numerical calculations, two different MHD stability codes have been used: CASTOR-FLOW, where the centrifugal effects are selfconsistently included in the equilibrium, and MISHKA-F, where such effects are not taken into account. The distingtly different results obtained with these two codes, as well as a comparison with the analytical theory, also developed in Ref. [2], are illustrated in the figure. From the analytical theory, the reason for the difference between the two models can be traced to the stabilising effect of the GAM induced by the plasma rotation, a phenomenon previously discovered within this project. This GAM exists solely as a consequence of the nonuniform plasma density and pressure created by the centrifugal acceleration of the plasma on the q = 1 surface [1], and a stabilising coupling of the internal kink instability to this mode cannot therefore take place if the centrifugal effects are not included in the equilibrium. In addition to the GAM stabilisation, the effects of the radial profiles of the plasma density and rotation velocity are also found to be significant. While such effects are rather small in the self-consistent calculation in Fig. 2.2.2-1b (due to the small !a = 0.1 used there), it is shown in Ref. [2] that the importance of the profile effects increases with decreasing aspect ratio. Publications section 2.2.3 Peer reviewed journals

1. Wahlberg, C., Low-frequency magnetohydrodynamics and geodesic acoustic modes in toroidally rotating tokamak plasmas, Plasma Physics and Controlled Fusion 51, 085006 (2009).

2. Wahlberg, C., Chapman, I. T., Graves, J. P., Importance of centrifugal effects for the internal kink mode stability in toroidally rotating tokamak plasmas, Physics of Plasmas 16, 112512 (2009)


33

Presentations at EPS conferences 3. Wahlberg, C., Radial structure of the two magnetohydrodynamic GAMs existing in

toroidally rotating tokamak plasmas, Proceedings of the 36th European Physical Society Conference on Plasma Physics, Sofia, P1.131 (2009).

2.3 Power and particle exhaust, Plasma-wall interaction 2.3.1 Plasma-wall interaction M. Rubel, B. Emmoth†, D. Ivanova (PhD student), Per Petersson Survey of dust formed in the TEXTOR tokamak In fusion context the term “dust” denotes all erosion products resulting from plasma-wall interaction processes and covers a range of particle dimensions from some nanometers to millimeters. Dust formation in carbon-wall machines is mainly related to sputtering and chemical erosion of carbon by fuel species followed by long-range migration and re-deposition of hydrogenated carbonaceous species (CxHyDz) in the form of co-deposited layers on plasma-facing components (PFC) and in remote areas with no direct line-of-sight to the plasma. Such layers often flake and peel-off from the substrate, thus forming dust agglomerates. There is no critical layer thickness when such process starts. Dust release occurs also due to arcing; the effect has frequently been observed during the start-up phase. Another mechanism of carbon dust production may be related to brittle destruction of carbon under localized high power loads. Melting, melt layer motion and eventually splashing of droplets would be main mechanism for the formation of metal dust. The formation and accumulation of dust may have a serious impact on economy and safety in operation of a reactor-class device. The primary concern is related to consequences of oxygen and/or water contact with hot dust in the case of air and/or water leak during the plasma operation: high risk of pressure rise and explosion. A water leak would lead to the gasification of carbon by the water - gas shift reaction (H2 and CO as products) and, in the case of metal dust, hydrogen release, e.g. Be + H2O ! BeO + H2. A risk of mobilization and release of radioactively contaminated (neutron-activated and tritiated) products by explosion is also taken into account in safety assessments. Another category is the operational risk connected with degraded performance of diagnostic (e.g. first mirrors) or pumping components such as cryopanels. The impact of levitation of charged tritiated dust on plasma performance has also been addressed. As a consequence, there are administrative and safety limits of dust content in the ITER vessel. The access to the reactor in-vessel components will be extremely limited. Therefore, in the current tokamak physics program one has to determine as accurately as possible: the impact of dust on plasma operation, the amount of dust formed (loose material and co-deposits), its distribution in the machine, physical and chemical properties such as size distribution, structure and composition with particular emphasis on the fuel content. The importance of such studies was recognized at TEXTOR already in the mid nineties. Since then several comprehensive dust surveys have been carried out on the occasion of major shut-downs which gave access to all in-vessel components. Studies have also been done in ASDEX-U, Tore Supra, JT-60 and in JET, where a large amount (approx. 100 g) of loose carbon flakes have been retrieved from one octant of the machine.


34

Dust particles moving in the plasma have been reported on many occasions. In present-day machines, the release of dust into the plasma is not considered as a serious operational or safety issue. Disruptions have only been noted in the case of large amount of material being instantly released to the plasma The aim of this paper is to provide an overview of a broad dust survey and morphology studies in connection with the opening of the TEXTOR vacuum vessel in 2008. Recent data are also briefly compared with results obtained during previous surveys. Dust collection and analysis The collection of dust was performed during the first hours after opening of TEXTOR. Images in Fig. 2.3.1-1 (a-c) show in-vessel components of TEXTOR. The locations from where the material for further analysis was sampled are indicated. In Fig. 2.3.1-1 b one notices on some tiles patches of flaking layers whereas vast areas of the liner (Fig. 2.3.1-1 c) are covered with blackish soot-like rough deposits. A dust collector with a cascade series of meshes and filtering paper connected to a vacuum pump was used. It was done in order to separate debris of different size (0.2 - 0.5 and 0.5 – 1 mm) from fine dust which was collected on a filtering paper. The material (loose particles) was separately taken from the bottom and the low-field side of the Inconel® liner, from the main graphite limiters pump toroidal belt ALT-II, poloidal and inner bumper and from the bottom shield of the dynamic ergodic divertor (DED). Magnetic and non-magnetic fractions were separated prior to further studies. Sampling of deposits from the liner and limiters was done using adhesive carbon stickers and ultra-fine copper nets for microscopy and by direct scraping with a plastic knife. Samples were examined with high-resolution scanning and transmission electron microscopy (SEM and TEM) combined with energy dispersive X-ray spectroscopy (EDX), Rutherford backscattering spectroscopy (RBS) and nuclear reaction analysis (NRA) using a 1.4 MeV 3He+ beam. The total fuel content was determined by means of thermal desorption spectrometry (TDS). In the latter case, the temporal evolution of fuel-related species versus temperature was measured by monitoring masses 2 (H2), 3 (HD), 4 (D2), 18-20, i.e. water group and various hydrocarbons CxHyDz. Calibration of the quadrupole mass analyzer was based on the consecutive injection gases H2, D2 and CH4 into the TDS chamber. A calibration factor for HD was approximated with arithmetical mean of the calibration factors for H2 and D2. The desorption was done at temperatures of up to 1273 K, the heating rate was 0.5 K/s on most occasions. The efficiency of fuel removal by long-term outgassing at 623 K, was also determined. The background desorption of the empty TDS chamber was performed for the both experiments. The sensitivity is on the level of 6 ! 1011 at"s-1 at the temperature of 1273 K.


35

Fig. 2.3.1-1. Plasma-facing components of TEXTOR: (a) Bottom part of the torus; (b) details of the inner wall and (c) the liner on the low field side. 1. ALT-II toroidal belt pump limiter: (a) erosion zone and (b) deposition zone; 2. main poloidal limiters; 3. bottom of the Inconel liner; 4. bottom DED shield; 5. inner wall bumper limiter; 6. blackish dust on the Inconel liner.

Fig. 2.3.1-2 SEM images showing the topography of flaking co-deposits from the toroidal ALT-II (a) and poloidal (b) limiters and corresponding TDS characteristics for release of deuterium-containing species from the two types of limiters (c) and (d), respectively.


36

Deposit structure and fuel content All SEM and TEM studies have been accompanied by EDX in order to identify the elemental composition arising from material mixing. Particular interest was focused on the identification of objects consisting essentially of pure carbon or only metal species. SEM images in Fig. 2.3.1-2 show topography of flakes isolated from the deposition zone of the ALT-II tile (Fig. 2.3.1-2 a) and deposits from the main poloidal limiter (Fig. 2.3.1-2 b). Thermal release of fuel from the two types of deposits is shown in Fig. 2.3.1-2 c and d, respectively. The results of TDS measurements for samples from several locations are summarized in Table 2.3.1-1 which provides data for the total amount of desorbed hydrogen isotopes (M=3 and M=4) and deuterium-to-carbon (D/C) ratio. Masses 19 and 20 are not shown because their intensity was two orders of magnitude smaller than the intensity of HD and D2 signals. Despite the rather small amount of each dust sample (milligrams) the background level for the studied masses did not exceed 1% of the total intensity. All data presented in the paper have been normalized for the specimen mass of 1 g.

Table 2.3.1-1. Retention of H and D in dust and co-deposites from various locations. The atomic density of carbon is 4.5!1022 g-1.

The largest amount of deuterium and the D/C ratio (~0.08) is found in stratified layers from the deposition zone on ALT. The layers contain also hydrogen (D/H in the range 0.7 – 0.8) which origin is most probably attributed to the adsorption of water vapor from ambient atmosphere when the flakes are stored after retrieval from the vacuum vessel of TEXTOR. Rough granular deposit on the poloidal limiter contains minute amount of deuterium (D/C = 0.0003). It is related to high surface temperature of limiters during plasma operation. Hydrogen content in this case could not be exactly quantified because the pressure rise caused by the release of hydrogen species (H2 and H2O) at the very beginning of desorption was greater then the operational range of the quadrupole mass analyzer. This observation supports the statement that the hydrogen presence is associated with water adsorption in the porosity of deposits. It may occur efficiently in deposits of high roughness and large amount of open pores. Unfortunately, proper porosity measurements could not be performed in this case because of too small quantity of material (mg range) available for studies. The fuel content in deposits on the ion cyclotron resonance heating (ICRH) antenna grill and loose dust from the liner’s bottom is significantly smaller than on ALT-II. The reason for a reduced fuel retention in flakes from the antenna grill has been discussed earlier: effective release of carbon and deuterium by energetic particles generated by ICRH pulses. The small fuel content in loose dust hovered from the bottom of the liner should be interpreted with care.


37

The material collected (less then 2 grams in total) was composed of species with highly diversified size and composition: from nano-size carbon dust to larger debris (0.1 – 1 mm) of graphite, ceramics and metal. As a consequence, the TDS result (D/C = 0.001) may not exactly reflect carbon-related fuel content. The survey has confirmed that fuel retention in TEXTOR is highly localized: the majority of deuterium is accumulated in deposition zones on ALT-II. Therefore, flaking layers from that area have been selected for long-term outgassing at constant temperature of 623 K. The aim was to assess the efficacy of fuel removal at the maximum bake-out temperature of the ITER divertor. The desorption was done in two stages: long-term annealing at 623 K for 70 hours followed by a steady temperature rise (~ 0.1 K/s) to 1273 K in order to complete the desorption. The results presented in Table 2.3.1-2 show that only about 10 % of deuterium has been removed by long-term treatment at 623 K. The efficient removal of remaining 90% could only be achieved at greater temperatures. It clearly indicates a limited use of outgassing at moderate temperatures as a method for fuel removal from carbon-based co-deposits in ITER.

Table 2.3.1-2. Efficiency of deuterium release by long-term outgassing of co-deposits on ALT-II.

Fig. 2.3.1-3 a shows a debris collected from the bottom of the liner, whereas TEM image (Fig. 2.3.1-3 b) presents an amorphous deposit with embedded nano-size particles of crystalline matter, as confirmed by electron diffraction (see pattern in Fig. 2.3.1-3 c). In all cases the elemental composition determined by EDX revealed that the crystalline material was composed of carbon. The origin of larger debris (0.2-0.8 mm) in Fig. 2.3.1-3 may be attributed to the damage of the graphite tile during the installation. However, the presence of fine crystalline carbon matter points to another possible mechanism of material disintegration and production of such objects: brittle destruction. One may tentatively suggest that small crystalline graphite objects in the loose material could be associated with brittle destruction of carbon PFC caused by highly localized power loads due to disruptions. The final confirmation of that phenomena taking place in TEXTOR is still needed by systematic combined studies using cameras and spectroscopy methods for dust tracing in the plasma and, by detailed dust survey during the next opening of the TEXTOR vacuum vessel. It should be stressed that until now no direct evidence of brittle destruction has been found in present day tokamaks, but it was detected in a reversed field pinch as a result of highly localized wall-locked modes. Therefore, the effect cannot be excluded in future devices in the case of giant edge localized modes (ELM), vertical displacement events (VDE) or disruptions. Fig. 2.3.1-4 shows metal species retrieved from the bottom of the liner (Fig. 2.3.1-4 a) and a splash from the surface of ALT-II (Fig. 2.3.1-4 b, c). The objects are composed predominantly of nickel, chromium and iron thus suggesting that these are erosion products from the liner or ICRF antenna grill. A careful analysis of the splash shape indicates that the original molten metal droplet did not re-solidify instantly after heating the graphite substrate.


38

It was still rotating when solidifying, as inferred from the twisted layers (left side in Fig. 2.3.1-4 b). This “layered” structure is also visible in Fig. 2.3.1-4 c which additionally allows for determination of the grain size: 5-15 µm. At the tips of the protruding grains on finds carbon co-deposits, as proven by secondary electron images (atomic number contrast) and EDX. Further analysis of splashes and ion tracks present on the surface is under way in order to determine the droplet velocity with respect to the magnetic field direction. Dust conversion factor Dust is not a safety or operational problem in present-day devices. Therefore, studies are carried out from the ITER perspective. The most intriguing and important issue is the so-called conversion factor, i.e. the ratio of the total amount of eroded and transported material to the amount of loose matter found in the vacuum vessel: “erosion-to-dust”. First of all, the determination of the total erosion meets difficulties because a part of eroded material, especially in the case of carbon, is pumped out. The second difficulty arises from the fact that loose dust contains a variety of species including ceramics from broken diagnostics, wall heating elements, tiny pieces of mechanically damaged tiles during the PFC installation, etc. As a consequence, the meaning of the conversion factor is narrowed to: “the amount of eroded and re-deposited matter to the amount of loose dust of operational origin”. The latter is somewhat difficult to define. In the following, the consideration of dust will be limited to carbon-based co-deposits referred to as the total erosion and the amount of loose carbon species. The total amount collected by vacuum cleaning of limiters and the liner (bottom and low-field side) was about 2.0 grams. The dominant part of this material is constituted by larger (up 1 mm) debris which could be related to the in-vessel work, i.e. installation of tiles. More exact number cannot be given because ultra-fine dust could not be weighted: (a) it sticks to the filtering paper; (b) some fine matter could not be removed from PFC because of strong adhesion to the substrate, probably due to Van der Waals forces. After subtracting the mass of above mentioned larger debris and ceramics the amount of carbon matter is 0.1-0.2 g. The other side of the equation for the conversion factor includes the amount of material eroded and re-deposited on PFC. Results of recent studies are consistent with previous findings that co-deposits in TEXTOR are predominantly formed in the deposition zones of the ALT-II tiles. The total amount of carbon in these 20-40 µm layers of density around 1.4 g cm-3 is assessed at the level of approximately 40 g. It leads to the conversion factor of 200-400. The process that may significantly influence the conversion factor is splitting, flaking and eventually peeling-off of the layer. The process is enhanced by exposure of the layers to air. SEM image in Fig. 2.3.1-5 show layers on the liner after contact with air. These stratified co-deposits split into 20-30 nm thin strata that easily disintegrate into very fine matter. As already discussed in previous papers, accumulation of water vapor in porous dust pieces and deposits causes splitting and flaking of the layers. The effect is enhanced when the layer is subjected again to vacuum conditions either in the SEM chamber or in a tokamak after a shut down period. The issue is still to be examined in more detail because the same may be expected in beryllium wall machine, because beryllium easily reacts with water and oxygen in air.


39

Fig. 2.3.1-3. (a) SEM images of graphite debris collected from the bottom of the liner (a); dust from the inner bumper limiter: amorphous carbon co-deposit with embedded nano-size particles of crystalline carbon (b) and corresponding diffraction pattern (c).

Fig. 2.3.1-5. Stratified dust collected from the Inconel liner: splitting of flakes into 30 nm strata after one year in air.

Fig. 2.3.1-4. Metal (Ni–Cr–Fe) objects: (a) dust from the liner and (b) a splash with (c) carbon deposits on edges.


40

Concluding remarks As already stated, the results of the systematic dust survey in TEXTOR confirm several previous data. The new important finding is the evidence for the existence of carbon debris, 10 nm – 200 µm, in the collected material. These may be products of brittle destruction. In such case it would be the first documented example in a tokamak. To clarify this point, more studies are needed during the TEXTOR operation (spectroscopy and cameras) and the next inspection of the vessel. From the ITER perspective the issue of brittle destruction and dust formation in present-day machines deserves and requires comprehensive approach. It is expected that a very detailed survey of dust in JET before the installation of the ITER-Like Wall (ILW) will shed light on the real extent of carbon and beryllium dust production that a reactor-class device may face. The issue of tungsten and beryllium metal dust will be one of important subjects in the operation of JET with ILW. Fuel inventory and co-deposition in castellated beryllium structures The assessment of fuel retention is of extreme importance for economy and safety in operation of next-step fusion devices. There are several pathways leading to the accumulation of hydrogen-containing species in the wall. In carbon surrounding the major mechanism is co-deposition. It is known that eroded material is transported and co-deposited together with fuel species in areas shadowed from the direct plasma line-of-sight. Grooves of castellation and side surfaces of plasma-facing components (PFC), i.e. surfaces located in gaps separating tiles, are considered as a potential trap for vast amount of co-deposited fuel atoms. All PFC in ITER and all beryllium limiters in JET during the ITER-Like Wall (ILW) Project will be castellated, i.e. composed of small blocks separated by narrow grooves (less than 1 mm) in order to reduce thermally-induced stress. The main question is whether fuel retention in such places may significantly contribute or even be decisive for the overall tritium inventory in ITER where the total number of castellated grooves will exceed 2 millions and the area of surfaces in the castellated grooves in the main chamber wall and divertor will be greater (by at least factor of 2) than the area of plasma facing-surfaces.. This calls for detailed studies of castellated structures and gaps between PFC tiles retrieved from present-day tokamaks. Until now, JET has been the only tokamak operated in the past with a carbon wall and large scale castellated metal structures: beryllium limiters and divertor tiles. Fig. 2.3.1-5a shows the in-vessel components in JET with belt limiters and Mk-0 divertor tiles made of beryllium. Other main chamber components (inner column cladding and upper plates) were made of carbon. Therefore, studies of Be limiters have given an unique chance to recognise the material transport to metal components during the long-term operation in the carbon-containing surrounding. Analysis of deposition in the narrow castellated grooves and in tile gaps can contribute to the assessment of material migration in environment containing both C and Be. It helps to distinguish the influence of chemical and physical processes in wall erosion and to draw conclusions regarding their impact on fuel inventory. The study was carried out for beryllium belt limiters protecting the main chamber wall (castellation 15-16 mm deep and 1 mm wide) operated for approximately 56000 s during campaigns in 1989-1992. There were approximately 2000 tiles installed in the machine; each 379 mm long, 106 high and 22 mm wide. They were separated by 4-5 mm wide gaps with inconel plates as spacers. All castellated grooves on the limiter were toroidal, whereas the gaps between tiles were in poloidal direction. Two of the tiles were retrieved from storage for detailed studies. In the Beryllium Handling Facility at JET each tile was cleaved in three places along the castellation in order to expose surfaces in the groove thus enabling analysis of co-deposits formed on both sides of six grooves. The preparation of samples by cleaving


41

minimised the risk of contamination of freshly open surfaces by deuterium from surfaces exposed to the plasma and the production of Be dust which would inevitably occur in case of sawing the tiles. Ex-situ examination of the tiles was done by means of nuclear reaction analysis (NRA) with a 3He+ beam to quantify and map surface contents of deuterium, beryllium and carbon. Most of the D, Be and C analyses were carried out with a 2.5 MeV 3He+ beam. Enhanced proton scattering (EPS) with a 2.5 MeV H+ beam was used to determine oxygen and carbon contents on the tiles. Morphology of plasma-facing and side surfaces of limiter tiles Images in Fig. 2.3.1-5b show the appearance of plasma-facing side for three consecutive segments (S-1 to S-3) of the cleaved tile. The corresponding deposition profiles of deuterium and carbon are plotted in Fig. 2.3.1-5c. The central area of the tile (Segments 2 and 3) was molten during the plasma operation but the remaining parts at both ends of the tile have fairly smooth shiny appearance (Segments 1 and 4, the latter is not shown here). In the melt zone, the surface is very rough and some gap bridging by molten Be is noted. However, the melt layer remained on the surface and castellated grooves were not filled with liquid metal; see Fig. 2a. Results indicate some variations in the fuel content along the tile. In absolute numbers (1-7x1017 cm-2), the amount of deuterium is small in comparison to over 1x1019 D cm-2 as measured in shadowed areas of the JET divertors. The integrated content on the plasma-facing side of one tile is around 2.1x1019 D atoms. The greatest content is found in the melt zone on Segment 2. Also the amount of carbon increases in that region but the quantity of carbon is quite small, maximum 1.8x1018 cm-2. Both C and D have been detected only in the very surface layer.

0

5

10

15

20

0 50 100 150 200 250 300Distance along Tile Surface [mm]

Deu

teriu

m a

nd C

arbo

n [e

17/c

m2]

D C

S-1 S-2 S-3c

S11

S22

S3

b a

Fig. 2.3.1-5 View inside the JET vessel with beryllium components: divertor floor and belt limiters, position of the studied tile marked with X (a); surface topography of plasma-facing surfaces (b) and distribution of deuterium and carbon on consecutive segments of the cleaved limiter tile (c).


42

Side surfaces located in gaps separating the tiles of several segments were analysed. Figures 2.3.1-6a and b show respectively the appearance of a segment with the molten surface and the distribution of deuterium and carbon from the near-plasma region down to the bottom of the tile. The results clearly show that the amounts are small and the deposition profiles are steep with the e-folding length, !, of around 2 mm. At the depth of approximately 10 mm into the gap the deuterium profile flattens and remains constant to the very bottom of the tile. A striking feature is the presence of a BeO layer, approximately 3-3.5 µm thick, found only in the region located deep in the gap (from 20 mm down) where the carbon content falls to the level below 1x1017 cm-2. Fuel retention is still detected in that oxide film. It is not possible to conclude when the formation of the oxide took place, i.e. at the machine or during the long-term storage in air. However, the result may indicate that the carbon-containing film at the upper part of tile prevented beryllium oxidation. This hypothesis is partly supported by the fact that no BeO has been found on the plasma-facing surfaces containing some carbon.

Fig 2.3.1-6. Side surface (a) and deposition profiles of D and C in the gap separating tiles.

Morphology of castellated grooves Images in Fig. 2.3.1-7a show two sides of the castellated groove created by cleaving the tile. Bright areas at far ends are shiny surfaces of metallic beryllium in the freshly fractured part. Deposition profiles of D and C inside the groove are plotted in Fig. 2.3.1-7b. The results obtained for both sides of six castellated grooves allow a statement that, despite some differences in the appearance of surfaces, all profiles are very similar. Therefore, the most important features are summarised by the following. In all cases the presence of deposit is observed just below the entrance to the gap. Smaller amounts of species found very close to the entrance may be attributed to thermal desorption of gases (D2 and CxDy molecules) due to high temperature of the limiter surface. The profiles reach their maximum at the depth of about 2-3 mm and then decay to a level below 1x1017 cm2 already at 4-5 mm. In neither case the D content exceeds 7x1017cm-2, which is considered to be a very small amount in comparison to that found on side surfaces of carbon tiles. The e-folding length, !, for deposition is 1.2-1.6 mm. This value agrees very well with that determined for the castellated Be Mk-I divertor and also found in some short-term experiments. The presence of D and C is detected down to the bottom of the groove. In some cases there is an increase in concentration at the bottom. This may indicate that the groove bottom is a final trap for C-D neutrals entering the gap at right angle to the tile surface. As inferred from EPS spectra, some oxygen

0

2

4

6

8

10

0 10 20 30 40 50 60 70Distance from Plasma [mm]

Deu

teriu

m a

nd C

arbo

n [e

17/c

m2]

D C

BeO layer on the surface

a

b


43

(up to 6x1017 cm-2) is present in co-deposits but there is no evidence of bulk BeO formation. The presence of oxygen is probably related to penetration of some quantities of oxygen-containing gases (water vapour, O2, CO2) into the film during the tile storage for 14 years. However, the amount of gases was limited because the tiles were stored in tightly sealed bags.

Fig. 2.3.1-7. Deposition pattern (a) and deposition profiles of deuterium and carbon (b) inside the castellated groove of the beryllium belt limiter tile.

When the composition of deposits in all grooves of the analysed tile is considered, the integrated amount of fuel in castellation is estimated at 6-10x1018 D atoms, which corresponds to about 30-50% of deuterium detected on the plasma-facing surfaces (compare to numbers given in the previous section), thus showing that the amount of retention and carbon deposition in castellated Be grooves is small. The most important fact coming from tile analysis is that the fuel presence is associated with co-deposited carbon. These results are highly consistent with data obtained for the castellated Be tile of the Mk-I divertor. Secondly, on beryllium surfaces freshly opened by cleaving no oxygen was detected with the methods applied in the study, detection limit 1x1017 cm2. It means that during six weeks after fracturing the tile only small quantity of oxygen the oxide formation was limited to the very surface layer.

Side A Side B

a Surfaces in the castellated groove

0

3

6

9

12

15

18

0 3 6 9 12 15Depth inside castellated groove [mm]

Deu

teriu

m a

nd C

arbo

n [e

17/c

m2] D, Side A

D, Side BC, Side AC, Side B

b


44

Summary of deposition in castellated limiters An important contribution of this work to tritium retention studies is the assessment of fuel inventory in castellated PFCs, i.e. in structures similar to those foreseen in ITER. As shown previously, much more pronounced fuel accumulation is measured in the presence of the CFC structure than in the case of Be tiles. Only small amount of fuel has been found in all studied regions of the castellated tiles: maximum 7x1017 cm-2. One may suggest that during the long-term tile storage at least some fraction of the originally retained deuterium desorbed due to isotope exchange with hydrogen from water vapour. However, the small amount co-deposits and steep profiles of the co-deposited carbon indicate that the original deposition was not pronounced. The results from studies of co-deposits on carbon substrates from TEXTOR have shown that in 5 years the D content decreased by approximately 25-30%. With this scenario for the Be tiles with C-D deposits, the deuterium content would have decreased by 50-55% after 14 years. There is no difference in the amount and profiles of deuterium and carbon between narrow toroidal castellated grooves (1mm) and wider poloidal gaps (4-5 mm) separating tiles. This might suggest that operation with beryllium limiters and carbon on the central column and upper plates (as it was in this particular case) does not result in the formation of thick deuterated carbon films. It does not exclude, however, the formation of such films in other locations, as discussed previously. These results from JET operated with the carbon wall and relatively small quantities of beryllium should not be immediately translated into quantitative predictions and far reaching conclusions regarding the material migration and fuel inventory in ITER because the planned material configuration in the divertor (W and CFC) and on the main chamber wall (Be) will be different than in any device operated to date. This will change both the scenario of material erosion and will influence fuel co-deposition. The clarification of this issue will be one of the main targets in the science program of JET when it is operated in the future with an ITER-like wall comprised of castellated beryllium limiters and tungsten lamellae-structured load bearing tiles in the divertor. It is unreasonable to expect, however, that the deposition in the castellation grooves will be completely eliminated. The development of efficient techniques of fuel removal from all in-vessel components remains, therefore, crucial for the operation of a reactor-class device. Publications section 2.3.1 Peer reviewed journals

1. M. Rubel, J.P. Coad, J. Likonen and V. Philipps, “Analysis of fuel retention in plasma-facing components from controlled fusion devices”, Nucl. Instrum. Meth. B267 (2009) 711-717

2. M. Psoda, M.J. Rubel, G. Sergienko, P. Sundelin and A. Pospieszczyk, “Material mixing on plasma –facing components: Compound formation”, J. Nucl. Mater. 386-388 (2009) 740-743

3. M. Rubel, P. Coad and D. Hole, “Overview of long-term fuel inventory and co-deposition in castellated beryllium limiters at JET”, J. Nucl. Mater. 386-388 (2009) 729-732

4. M. Rubel, G. De Temmerman, P. Sundelin, P. Coad, A. Widdowson, D. Hole, F. Le Guern, M. Stamp and J. Vince, “An overview of comprehensive First Mirror Test for ITER at JET”, J. Nucl. Mater. 390-391 (2009) 1066-1069


45

5. A. Widdowson, M.J. Baldwin, J. P. Coad, R.P. Doerner, J. Hanna, D.E. Hole, G.F. Matthews, M. Rubel, R. Seraydarian and H. Xu, “Testing of beryllium marker coatings in PISCES-B for the JET ITER-like wall”, J. Nucl. Mater. 390-391 (2009) 988-991

6. B. Emmoth, A. Kreter, A. Hallen, M. Jakubowski, M. Lehnen, A. Litnovsky, P. Pettersson, V. Philipps, G. Possnert, M. Rubel, B. Schweer, P. Sundelin, B. Unterberg and P. Wienhold, “In-situ measurements of carbon and deuterium deposition under conditions with external magnetic perturbations in TEXTOR”, J. Nucl. Mater. 390-391 (2009) 179-182

7. P. G!sior, F. Irrek, P. Petersson, H.J. Penkalla, M. Rubel, B. Schweer, P. Sundelin, E. Wessel, J. Linke, V. Philipps, B. Emmoth, J. Wo"owski and T. Hirai, “Laser- induced removal of co-deposits from graphitic plasma-facing components: Characterisation of irradiated surfaces and dust particles”, J. Nucl. Mater. 390-391 (2009) 585-588

8. P. Sundelin, C. Schulz, V. Philipps, M. Rubel, G. Sergienko and L. Marot, “Nitrogen-assisted removal of deuterated carbon layers”, J. Nucl. Mater. 390-391 (2009) 647-650

9. J. Likonen, J.P. Coad, D.E. Hole, S. Koivuranta, M. Rubel, T. Renvall and E. Vainonen-Ahlgren, “Post-mortem measurements of fuel retention at JET with MkII-SRP divertor”, J. Nucl. Mater. 390-391 (2009) 631-634

10. G. F. Matthews, P.Edwards, H.Greuner, A.Loving, H.Maier, Ph.Mertens, V.Philipps, V.Riccardo, M.Rubel, C.Ruset, A.Scmidt, E.Villedieu and JET-EFDA contributors to the ITER-like Wall Project, “Current status of the JET ITER-like Wall Project”, Phys. Scr. T138 (2009) 014030.

11. G. De Temmerman, S, Lisgo, R.P. Doerner, P. John, A. Litnovsky, L. Marot, S. Porro, P. Petersson, M.J. Rubel, D.L. Rudakov, G. Van Rooij, I. Villalpando, J. Westerhout and J. Wilson, “Interactions of diamond surfaces with fusion relevant plasmas”, Phys. Scr. T138 (2009) 014013.

12. M.I. Airila, L.K. Aho-Mantila, S. Brezinsek, J.P. Coad, A. Kirschner, J. Likonen, D. Matveev, M. Rubel, J.D. Strachan, A. Widdowson and S. Wiesen, “ERO modeling of local deposition of injected 13C tracer at the outer divertor of JET”, Phys. Scr. T138 (2009) 014021.

13. A. Widdowson, S. Brezinsek, J.P. Coad, D.E. Hole, J. Likonen, V Philipps, M. Rubel, M Stamp and JET-EFDA contributors, “An overview of recent erosion - deposition studies at JET”, Phys. Scr. T138 (2009) 014005.

14. J.P. Coad, D.E. Hole, M. Rubel, A. Widdowson and J. Vince, “Deposition results from rotating collector diagnostics in JET”, Phys. Scr. T138 (2009) 014023.

15. D. Ivanova, M. Rubel, V. Philipps, M. Freisinger, Z. Huang, H. Penkalla, B. Schweer, G. Sergienko, P. Sundelin and E. Wessel, “Survey of dust formed in in the TEXTOR tokamak: structure and fuel retention”, Phys. Scr. T138 (2009) 014025.

16. A. Kreter, M.J. Baldwin, R.P. Doerner, D. Nishijima, P. Petersson, A. Pospieszczyk, M. Rubel and K. Umstadter, “Fuel retention in carbon materials under ITER-relevant mixed species plasma conditions”, Phys. Scr. T138 (2009) 014012.


17. M. Rubel, D. Ivanova, V. Philipps, M. Freisinger, J. Linke, O. Neubauer, H. Penkalla, B. Schweer, G. Sergienko and E. Wessel, “Dust particles in controlled fusion devices: Generation mechanism and analysis”, 36th EPS Conf. on Plasma Physics and Controlled Fusion, Sofia, Bulgaria, June 2009. Europhys. Conf. abstracts ECA 33E (2009) O-4.051


46

Invited talks at international conferences and workshops. 18. M. Rubel, “Fusion Reactor Materials: Aspects related to safety and radioactivity”,

Presented at the 12th Int. Workshop on Plasma-Facing Materials and Components (PFMC-12), Jülich, Germany, May 2009. Invited tutorial lecture

19. A. Widdowson, J.P. Coad, H.G. Esser , D.E. Hole, A. Kreter , J. Likonen, V Philipps, M. Rubel, M Stamp and JET-EFDA contributors, “An overview of recent erosion - deposition studies at JET”, Presented at the 12th Int. Workshop on Plasma-Facing Materials and Components (PFMC-12), Jülich, Germany, May 2009. Invited talk

20. E. Tsitrone, C. Brosset, J. Bucalossi, T. Dittmar, E. Gauthier, F. Linez, T. Loarer , B. Pégourié, H. Khodja, C. Martin, P. Roubin, K. Krieger, M. Mayer, R. Neu, V. Rohde, J. Roth, A. Kirschner, A. Kreter , A. Litnovsky, V. Philipps, M. Rubel, P. Wienhold, “Recent fuel retention studies in Tore Supra, ASDEX Upgrade and TEXTOR”, Presented at the 12th Int. Workshop on Plasma-Facing Materials and Components (PFMC-12), Jülich, Germany, May 2009. Invited talk

21. M. Rubel, “Fusion reactor materials”, Presented at the 9th Carolus Magnus Summer School on Plasma and Fusion Energy Physics, Herbeumont-sur-Semois, Belgium, September 2009. Invited lecture

22. V. Philipps, G. Matthews, H. Maier, Ph. Mertens, P. Edwards, A. Loving, V. Riccardo, H. Greuner, R. Neu, A. Schmidt, M. Rubel, C. Ruset and E. Villedieu, “Status of the ITER-Like Wall Project at JET”, Presented at the 9th Int. Symposium on Fusion Nuclear Technology (ISFNT-9), 01-047, PO3-046, Dalian, China, October 2009. Plenary lecture

23. G. Matthews, H. Altmann, H. Greuner, E. Grigore, S. Jachnich, M. Knaup, P. Lomas, A. Loving, H. Maier, Ph. Mertens, M. Norman, I. Nunes, N. Pace, V. Philipps, V. Riccardo, M. Rubel, C. Ruset, V. Thomson, I. Uytdenhouwen, E. Villedieu, P. de Vries and the ITER-Like Wall Team, “Tungsten elements of the ITER-Like Wall”, Presented at the 17th European Fusion Physics Workshop, Velence, Hungary, December 2009. Invited talk

Talks at ITER and ITPA workshops

24. M. Rubel, G. De Temmerman, P. Sundelin, P. Coad, A. Widdowson, D. Hole and J. Vince, “Comprehensive First Mirror Test for ITER at JET” Presented at the 12th ITPA Meeting on Diagnostic, Moscow, Russia, April 2009.

25. M. Rubel, D. Ivanova, M. Freisinger, V. Philipps, B. Schweer and E. Wessel, “Survey of dust in the TEXTOR tokamak”, Presented at the 12th ITPA Meeting on SOL and Divertor Physics, Amsterdam, May 2009.

26. M. Rubel, R. Doerner, S. Mukherjee and S. Suzuki, “Tritium retention and removal studies – recent progress”, Presented at the Executive Committee Meeting of IEA Implementing Agreement on Fusion Reactor Nuclear Technology, Dalian, China, October 2009.

27. V. Philipps, D. Ivanova , M. Freisinger, M. Rubel, “Fuel release from from dust and deposits in TEXTOR”, Presented at the 13th ITPA Meeting on SOL and Divertor Physics, San Diego, California, USA, December 2009,


28. J. Likonen, A. Hakola, J.P. Coad, A. Widdowson, J. Strachan, S. Koivuranta, D.E. Hole and M. Rubel, “Deposition of C-13 tracer at the JET MKII-HD divertor”, Presented at the 12th Int. Workshop on Plasma-Facing Materials and Components (PFMC-12), Jülich, Germany, May 2009.


47

29. G. De Temmerman, S, Lisgo, R.P. Doerner, P. John, A. Litnovsky,L. Marot, S. Porro, P. Petersson, M.J. Rubel, D.L. Rudakov, G. Van Rooij, I. Villalpando, J. Westerhout and J. Wilson, “Diamond as a plasma-facing material for fusion?”, Presented at the 12th Int. Workshop on Plasma-Facing Materials and Components (PFMC-12), Jülich, Germany, May 2009.

30. G.F. Matthews, P.E. Edwards, A. Loving, V. Philipps, V. Riccardo, M. Rubel, E. Villedieu for the ILW Project Team, “Current status of the JET ITER-Like wall Project”, Presented at the 12th Int. Workshop on Plasma-Facing Materials and Components (PFMC-12), Jülich, Germany, May 2009.

31. M.I. Airila, L.K. Aho-Mantila, S. Brezinsek, J.P. Coad, A. Kirschner, J. Likonen, D.

Matveev, M. Rubel, J.D. Strachan, A. Widdowson and S. Wiesen, “ERO modeling of local deposition of 13C tracer at the outer divertor of JET”, Presented at the 12th Int. Workshop on Plasma-Facing Materials and Components (PFMC-12), Jülich, Germany, May 2009.

32. V. Philipps, F. Nachtrodt, A. Pospieszczyk, P. Wienhold. A. Huber, U. Samm, H. Reimer, S. Richter, M. Rubel and D. Ivanova, “In-situ deposition of tungsten layers by local injection of WF6 through TEXTOR test limiters”, Presented at the 12th Int. Workshop on Plasma-Facing Materials and Components (PFMC-12), Jülich, Germany, May 2009.

33. J.P. Coad, D.E. Hole, M. Rubel, A. Widdowson and J. Vince, “Deposition results from rotating collector diagnostics in JET”, Presented at the 12th Int. Workshop on Plasma-Facing Materials and Components (PFMC-12), Jülich, Germany, May 2009.

34. D. Ivanova, M. Rubel, V. Philipps, M. Freisinger, B. Schweer, M. !"obi#ski, P. Sundelin, H. Penkalla and E. Wessel, “Survey of dust formed in TEXTOR and of debris generated by fuel and co-deposit removal methods”, Presented at the 12th Int. Workshop on Plasma-Facing Materials and Components (PFMC-12), Jülich, Germany, May 2009.

35. A. Kreter, M.J. Baldwin, R.P. Doerner, D. Nishijima, P. Petersson, A. Pospieszczyk, M. Rubel and K. Umstadter, “Fuel retention in carbon materials under ITER-relevant mixed species plasma conditions”, Presented at the12th Int. Workshop on Plasma-Facing Materials and Components (PFMC-12), Jülich, Germany, May 2009.

36. M. Rubel, I. Uytdenhouwen, J. P. Coad, G. De Temmerman, A. Hakola, D. Hole, J. Likonen, A. Widdowson and JET- EFDA Contributors, “Surface structure and composition of First Mirrors tested in JET for ITER”, Presented at the 12th Int. Workshop on Plasma-Facing Materials and Components (PFMC-12), Jülich, Germany, May 2009.

37. P. Petersson, A. Kreter, G. Possnert and M. Rubel, “Nuclear Microbeam Analysis of Deuterium Distribution on Plasma-Facing Materials from Controlled Fusion Devices”, Presented at the 19th Int. Conf. on Ion Beam Analysis, Abstract Book, p. 113, Cambridge, United Kingdom, September 2009.

38. M. Rubel, I. Uytdenhouwen, J. P. Coad, G. De Temmerman, A. Hakola, D. Hole, J. Likonen, A. Widdowson and JET- EFDA Contributors, “Surface structure and composition of First Mirrors tested in JET for ITER”, Presented at the 9th Int. Symposium on Fusion Nuclear Technology (ISFNT-9), 01-036, PO3-035, Dalian, China, October 2009.

39. D. Ivanova, M. Rubel, V. Philipps, M. Freisinger, B. Schweer, M. !"obi#ski, P. Sundelin, H. Penkalla and E. Wessel, “Survey of dust formed in a tokamak and of debris generated during removal of fuel and co-deposits”, Presented at the 9th Int.


48

Symposium on Fusion Nuclear Technology (ISFNT-9), 01-047, PO3-046, Dalian, China, October 2009.

40. M. Rubel, G. De Temmerman, I. Uytdenhouwen, J. P. Coad, D. Hole, J. Likonen, A. Widdowson, “First mirrors test in JET for ITER: An overview of optical performance and surface morphology”, Presented at the 1st Int. Conf. on Frontiers in Diagnostic Technology, Frascati (Rome), Italy, November 2009

2.3.2 Accelerator-based analysis of materials H. Bergsåker, B. Emmoth† , G. Possnert and J. Åström

Materials migration studies with the micro-beam A beam line for nuclear reaction analysis with a microbeam has been set up at the Ångstrom Laboratory in Uppsala, as shown in Fig. 2.3.2-1. Spatial resolution of a few micrometers has been achieved and the analysis samples from a number of fusion devices in Europe have been made. The analysis includes i) 2D surface images showing the deuterium distribution in deposited films and ii) depth profiles across the layers cross section showing the sediment layers.

Fig. 2.3.2-1. The microbeam experimental area of the 5 MeV Tandem Pelletron Accelerator Laboratory at The Ångström Laboratory, Uppsala University.

Studies of fuel retention include:

• Quantitative measurement of level of Hydrogen isotope retention. • Surface concentrations and depth profiles.

Cross sections of deposited layers at the JET divertor surfaces have been investigated with microscopy nuclear microbeam analysis methods. The elemental composition of the archeologically layered deposits is determined with the aim to draw conclusions about materials migration in the divertor region. The nuclear methods are easy to make quantitative compared to alternative methods and particularly sensitive to hydrogen isotopes and light elements like carbon and beryllium. The spatial resolution of a few micrometers is ideally


49

suited for the layers deposited in JET, which are hundreds of micrometers thick. An Example of the results is shown in figures 2.3.2-2 (2D map of the Beryllium layers). The influence of different polishing techniques in preparing the cross section samples has been investigated and detailed depth profiles as well as lateral elemental distributions have been determined for layers that are 100-800 µm thick, the thickest layers corresponding to JET operation 1998-2007

Fig. 2.3.2-2. 2D map of beryllium in a JET sample showing the archeologically layered deposits.


• J R Drake, H. Bergsåker, G. Possnert, M Rubel, Per Petersson, Swedish Association VR-EURATOM: Capabilities in View of PWI TF Work Programme and ITER Construction, Presentation at the EFDA Technical Meeting on European facilities for Plasma Surface Interaction, Garching, 18 June 2009..

2.3.3 Studies of dust collection using aerogel collectors H. Bergsåker, S. Ratynskaia

Understanding of dust dynamics is crucial and its importance is twofold. First, it makes it possible to understand and predict dust mobilization and transport; dust particles penetrating beyond the scrape-off layer (SOL) will sublimate and contribute to the impurities while those in the SOL plasmas can survive and even grow under certain conditions. Second, dust moving in the SOL will collide with plasma facing components (PFC) and, depending on velocity of


50

such impacts, it may contribute not only to PFC (e.g. wall) erosion and destruction but also to the dust production. Novel diagnostics based on collection if dust by very porous light targets- silica aerogels- have been recently proposed. Aerogel, a highly porous material with a density of a few tens of kg per cubic meter, allows capturing of dust particles present during the discharge without destroying them. First exposures in TEXTOR scrape-off layer plasmas have demonstrated the feasibility and potential of the method. The experiment showed that such targets are able to capture both slow and fast particles with sizes in the range from submicrometre to about 100 µm. The technique provides information on dust velocity and size distribution as well as dust flux estimates. The composition and texture of the captured dust can also be studied in detail to shed light on dust formation processes. Time resolved measurements revealed that largest numbers were found on the samples exposed during the flat-top phase, where the density of the collected particles was in correspondence with a particle flux density of 20–50 particles cm!2 s!1, for particles that were sufficiently big to be visible optically (>10 µm) and sufficiently fast to stick to the surface. Faster particles penetrate deeper in the samples and from the impact tracks, shown in Fig. 2.3.3-1 one can deduce velocity (of the order 2-3 km/s for a 10 micron dust).

Fig.2.3.3-1 Optical micrograph showing impacts tracks with particles residing at the bottom.

Publications section 2.3.3 Peer reviewed journals

1. S. Ratynskaia, H. Bergsåker, B. Emmoth, A. Litnovsky, A. Kreter and V. Phillips Capture by aerogel-characterization of mobile dust in tokamak scrape-off layer plasmas, Nuclear Fusion 49, 122001 (2009).

Presentations at EPS conferences 2. S. Ratynskaia, H. Bergsåker, B. Emmoth, A. Litnovsky, A. Kreter and V. Phillips

Capture by aerogel-characterization of mobile dust in tokamak scrape-off layer plasmas, 36th EPS Sofia Bulgaria, July 2009

Invited talks at international conferences and workshops.


51

3. S. Ratynskaia, Mobile dust in tokamak SOL plasmas, Presented at the Summer College on Plasma Physics at International Center for Theoretical Physics (ITCP), Trieste, Italy, August 25, 2009. Invited lecture.

2.4 Physics of plasma heating and current drive T. Hellsten, T. Johnson, M. Laxåback, A. Hannan (PhD student), K. Holmström (PhD student), J. Höök (PhD Student), Q. Mukhtar (PhD student) In collaboration with CEA, IPFN, TEKES, UKAEA, CRPP, ENEA, Hellenic associations, Courant Institute New York The research is focused on studying wave-particle interactions relevant for fusion experiments, in particular for heating, current drive and excitation of waves by fast particles. The group develops codes for predicting the effects of ICRH, and validates them against experiments. The program is well integrated into the European fusion program through participation in: the Integrated Tokamak Modeling Task Force, the exploitation of the JET facility and the EU training programme GOTiT. The three main codes developed by the group are PION, FIDO and SELFO. PION was the first self-consistent code for modeling ICRH and NBI heating using a model for the power deposition and solves a simplified Fokker-Planck equation for the distribution function. PION has become the standard code for routine simulation at JET. The Monte Carlo code FIDO calculates the distribution functions of the resonant ion species taking into account effects caused by finite orbit width and RF-induced spatial transport due to absorption of the momentum of the wave. The SELFO code calculates the wave field, using the LION code, and the distribution function, using the FIDO code, self-consistency is obtained by means of iterations. The FIDO code is being upgraded to include interaction with MHD waves allowing self-consistent studies MHD modes during ICRH; at the moment by using simple models of the MHD-modes. Exploitation of JET Thomas Johnson and Martin Laxåback have been long term seconded at JET. Thomas Johnson has been involved in analysing the experiments at JET and Martin Laxåback in the planning, coordination and monitoring of the JET programme. ITM task force The group participates in the Integrated Tokamak Modeling Task Force. Torbjörn Hellsten is project leader for IMP5, the project for developing models for heating, current drive and fast particle effects for ITER. Our group participates in the development of codes for simulating heating and current drive in the ion cyclotron frequency range and with interactions of fast particles with low frequency Alfvén waves. The group is involved in developing two new codes for ICRH, one for routine analysis, replacing PION, and one for advanced simulation including finite orbit effects to replace FIDO. The work on the advanced code based on Monte Carlo methods for solving a 3D distribution function have been focused on developing methods to improve the Monte Carlo method comprising of an adaptive !f-method and a higher order method to improve the convergence near boundaries where the diffusion coefficients vanish.


52

Development of codes for modelling of ICRH A fast code for routine simulations of ICRH is being developed. The code is based on the formulation used for the PION code using a 1D –Fokker-Planck solver with a model for the parallel velocity. The upgrading of the code consists of using replacing the formula for the power deposition with direct solution of the wave equation. This enables calculation of the power deposition with several cyclotron resonances or harmonics of it at different locations and to calculate electron current driven by the magnetosonic waves, which in the past have to be done with more advanced codes. Adaptive Monte Carlo !f -method Calculations of the distribution functions, which because of the complicated geometry are done with Monte Carlo methods, are time consuming. In order to speed up the calculation and /or to make them more accurate an adaptive !f-method has been developed and tested for one dimension, with promising results. The draw back with conventional !f-methods is either that they only allow a small local deviation of the distribution function or that new particles have to be continuously added. These problems are solved by resampling the distribution function and improving the approximation of the distribution function in when it is resampled in order to minimize the source term and the number of simulated particles. Stochastic differential equations with singular diffusion coefficients Monte Carlo operators When calculating the distribution function one experience in inhomogeneous plasmas that interacting regions and phase space are limited. At the boundaries the diffusion coefficients vanishes or become discontinuous resulting in poor convergence. To speed up the calculations and/or to improve the convergence higher order methods are developed. Promising results have been obtained for 1D-models. Coulomb collision operator for neoclassical calculations A method to include collisions with background plasmas consistent with neoclassical transport in the banana regime has been developed; suitable for orbit averaged Monte Carlo codes. By shifting the parallel velocity of the test particles, the collision operators can be expressed in terms of standard Coulomb collision operator for isotropic Maxwellian background distribution functions. The shift is determined from quasi-neutrality and can be expressed in terms of flux functions !if(")R0, which depend on the density gradients, ion temperature gradients and fluxes caused by wave-particle interactions or particle losses. As the shift, which determines the radial electric field, increases and becomes comparable to the thermal velocity of the ions the relation between this shift and the temperature gradients and the losses become strongly non-linear with the possibility of bifurcated solutions. It is found that at large gradients the relationship between radial electric field, parallel velocity, temperature and density gradient in the neoclassical theory is modified such that coefficient in front of the logarithmic ion temperature gradient, which in the standard neoclassical theory is small and counteracts the electric field caused by the density gradient, now changes sign and contributes to the built up of the radial electric field. Wave-particle interaction in toroidal plasmas A comprehensive treatment of wave-particle interactions in toroidal plasmas including collisional relaxation, applicable to heating or anomalous wave induced transport, has been obtained by using Monte Carlo operators satisfying quasi-neutrality. This approach enables a self-consistent treatment of wave-particle interactions applicable to the banana regime in the neoclassical theory. The possibility to drive current by absorbing the waves on trapped


53

particles has been studied and how the wave-particle interactions affect the bootstrap current. Three mechanisms appear: detrapping into co- and counter-passing orbits; changes of the bootstrap current due to wave induced particle transport; broadening of the trapped orbits producing dipolar like current. Wave-particle interactions by directed waves occurring only at one of the legs of trapped particles result in a selective detrapping into co- or counter-passing orbits, depending on the propagation direction of the waves. A factor of order unity, depending on the ratio of the effective detrapping by wave-particle interactions and collisions, can be recovered of the momentum absorbed on the trapped particles for current drive. This new current drive mechanism by selective detrapping of orbits during ELD/TTMP may partly explain the observed enhanced current drive in recent fast wave current drive. Upgrade of the PION code to treat the ITER-like antenna in JET The PION code has widely used at JET to study ICRF heating. However, the code was built to handle only one type of ICRF antenna at a time. The installation of the ITER.like ICRF antenna in JET has therefore require an upgrade of the code to treat scenarios where the plasma is heated by several antennas with different geometry, and frequency. This upgrade has been completed. RF-Induced rotation Rotation in plasmas can have beneficial effects. For instance, it can enhance the stabilizing effect of a resistive wall. Shear in the rotation is also believed to be an important factor for transport barriers. In ITER and future reactors the ratio between beam momentum and energy will be low because of the high energy required for central heating the induced rotation by the beam is not expected to give rise to strong plasma rotation, it is interesting to consider other mechanisms with a potential to produce rotation. Intriguing observations of rotation in plasmas heated by Radio Frequency (RF) waves with little or no external momentum input have been made in several machines. There is as yet no complete understanding of the mechanisms behind the observed rotation. Effects due to fast ions, MHD and transport have been proposed, but reliable theoretical predictions of rotation in ITER or a reactor with low momentum are not yet available. In this respect, measurements of rotation profiles are crucial. Experiments in JET have been carried out to study the effect of ICRH on the toroidal rotation of the plasma. In the outer part of the plasma a co-current rotation appears correlated with the magnitude of heating but independent of the cyclotron resonance position (frequency) or antenna phasing. This is in contrast to the central plasma rotation, which was affected by the frequency and wave spectrum. When applying the heating at the high field side, a central counter torque was found producing hollow rotation profiles most pronounced for waves propagating counter to the plasma current. The changes in the rotation profile depend on antenna phasing, frequency, plasma current and current profile. Hollow rotation profiles were seen either for hollow current profiles produced with LH or for low current with peaked current profiles. Intrinsic rotation experiments with ripple Recently experiments have shown that the level of toroidal field ripple can have a major impact on plasma performance. In the light of this work, the design of the ferritic inserts in ITER has been re-evaluated. One issue that has not been previously investigated is the role of ripple in plasmas with little momentum input, e.g. Ohmic of RF heated plasmas. Although these plasmas have low performance in present day machines the rotation is believed to be comparable to that in ITER, where the momentum injection will be significantly lower and the intrinsic rotation is expected to play an important role. Furthermore, the ripple tends to


54

drive rotation counter to the neutral beam driven rotation, thus reducing the already small rotation. We have therefore performed experiments at JET to study Ohmic and RF heated plasmas with ripple levels ranging from 0.08% up to 1.5%; compared to expected levels in ITER of 0.5-1%. Rotation analysis of charge exchange measurements (from short NBI pulses used for diagnostic purpose only) as well as MHD mode numbers and frequency data, shows that ripple indeed changes the rotation of both Ohmic and RF heated plasmas. In both cases ripple drives counter rotation. In Ohmic plasmas rotation is changed roughly rigidly by the same amount in the core and at the edge suggesting the presence of an edge localised torque source. Observation of increased counter-rotation in Ohmic plasmas indicates a strong torque due to non-ambipolar transport of thermal ions. In ICRH plasmas the rotation change in the plasma core is larger (see figure) indicating that the torque source in this case would be less edge localised and that fast-ion as well as thermal ion effects may be involved. To test this hypothesis numerical modelling has been performed with the Monte Carlo codes ASCOT and SELFO to calculate the toroidal torques and the plasma rotation, driven by non-ambipolar transport of ions. In Ohmic discharges these torques appear when thermal ions are transported by the ripple in the outer parts of the plasma, while in ICRF heated plasmas the ripple-induced transport of fast ions provides also a source that extends further towards the plasma core.

Figure 2.4-1. Toroidal rotation profiles from charge exchange measurements for ICRF heated pulses with Ip=1.5 MA and <BT>=2.2 T, PICRF~3 MW for two ripple levels. Top: pulse # 74688 with 0% ripple and PICRF=3.1 MW; bottom: pulse # 74686 with PICRF= 2.9MW bottom). The plasma centre is at R0=3.02 m, the ICRF resonance is off-axis on the high-field side at Rres= 2.71 m.


55

Figure 2.4-2. Toroidal rotation as a function of ripple for ICRF pulses with 1.5 MA, off-axis Rres=2.71 m, for the centre (3.1 m, solid symbols) and the edge (R=3.8 m, open symbols). Blue circles are measurements in H-mode and red triangles are L-mode.

Fast ion losses in ITER due to non-axisymmetric magnetic fields The wall loads due to fusion alphas as well as NBI- and ICRF-generated fast ions have been simulated for ITER Reference Scenario-2 and Scenario-4 including the effects of ferritic inserts (FI), Test Blanket Modules (TBM), and 3D wall with two limiter structures. The simulations were carried out using the Monte Carlo codes ASCOT and SELFO. The ferritic inserts were found very effective in ameliorating the detrimental effects of the toroidal ripple: the fast ion wall loads are reduced practically to their negligible axisymmetric level. The thermonuclear alpha particles overwhelmingly dominate the wall power flux. In Scenario-4 practically all the power goes to the limiters, while in Scenario-2 the load is fairly evenly divided between the divertor and the limiter, with hardly any power flux to other components in the first wall. This is opposite to earlier results, where hot spots were observed with 2D wall. In contrast, uncompensated ripple leads to unacceptable peak power fluxes of 0.5 MW/m2 in Scenario-2 and 1 MW/m2 in Scenario-4, with practically all power hitting the limiters and substantial flux arriving even at the unprotected first wall components. The local TBM structures were found to perturb the magnetic field structure globally and lead to increased wall loads. However, the TBM simulation results overestimate the TBM contribution due to an over-simplification in the vacuum field. Therefore the TBM results should be considered as an upper limit.


56

Figure 2.4-3. Wall load in perspective projection in ITER scenario-2 with uncompensated ripple. A view from inside ITER in the perspective projection, looking counterclockwise.

Sawtooth destabilisation by kinetic effects The restricted frequency range of the planned ICRH antennas in ITER are such that minority He3 is likely to be employed. Due to the negligible or reverse current drive contributions from minority He3, it was thought [M. Laxaback and T. Hellsten, Nucl. Fusion, 45, 1510 (2005)] that MHD control with toroidally propagating waves would not be viable. In contrast, the new explanation recently been given in Ref. [J P. Graves et al, Phys. Rev. Lett. 102, 065005 (2009)] for the sawtooth control mechanism does not rely on net driven current, and was therefore predicted to function even with minority He3. Consequently, minority He3 experiments in JET have been devised and carried out and interpreted with simulations in order to demonstrate the viability of sawtooth control using He3 minority in ITER, and to conclusively show that the previously assumed classical mechanism [V.P. Bhatnagar, et al, Nucl. Fusion 34, 1579 (1994)] cannot explain ICRF sawtooth control experiments in JET. The experiments demonstrate the viability of sawtooth control using ITER relevant He3 minority at low concentration. Simulation of RF induced currents have been confirmed to be negligible, as expected, but the recently developed fast ion mechanism [J P. Graves et al, Phys. Rev. Lett. 102, 065005 (2009)] has been analysed for these discharges, and shown to be responsible for sawtooth control (see figure). The success of these experiments in controlling sawteeth in JET greatly improves the prospect of using the planned ICRH system in ITER to shorten or lengthen sawteeth.


57

Figure 2.4-4. Comparison between measurements (black squares) and simulations (red curves) of the sawtooth period as a function of the distance between the sawtooth inversion radius and the ICRF resonance position.

ITG turbulence, 2D electric fields, fishbone induced losses ITG turbulence threshold Experiments were carried out in the JET tokamak to determine the critical ion temperature inverse gradient length for the onset of Ion Temperature Gradient modes and the stiffness of Ti profiles with respect to deviations from the critical value [P. Mantica, et al, Physical Review Letters, 175002, 2009]. Threshold and stiffness have been compared with linear and non-linear predictions of the gyro-kinetic code GS2. Plasmas with higher values of toroidal rotation show a significant increase in R/LTi, mainly due to a decrease of the stiffness level. This finding has implications on the extrapolation to future machines of present day results on the role of rotation on confinement. Neoclassical particle losses with static 2D electric fields It has been proposed that the so called “convective cells“ or poloidally localised 2D electric fields could be at least partly responsible for the loss of particle confinement (i.e. density pump out) observed in experiments where non-ambipolar losses are expected to be present. These include both toroidal magnetic field ripple experiments (ion loss channel) and resonant magnetic perturbation experiments (electron loss channel). In such experiments the poloidally localised losses could lead to the creation of poloidal electric fields and thus to localised 2D potential. We have been studying, both numerically and analytically, Neo-Classical (NC) losses in the presence of such an electrostatic and axisymmetric 2D potential. Analytical theory here covers presently non-collisional effects only. As a simulation tool for this analysis we use XGC-0 which is a guiding centre following Monte Carlo code developed for resolving NC transport. XGC-0 is able to follow both ions and electrons and it solves the radial electric field self-consistently from the radial current balance. On top of the existing functionality a simple model for a static 2D potential field was added either near X-point or at outer midplane. In collisionless simulations without the self-consistent radial electric field we found that losses are not much affected by the 2D static potential unless the potential is located in X-point region. This is in line with our analytical understanding showing that losses increase towards the inner target due to the modification of the loss cone. When we turn on the self-consistent radial electric field and collisions the significance of location of the 2D potential becomes much smaller as for both outboard- and X-point potential losses become similar. For


58

the case where we use 500 V negative potential (with plasma pedestal temperature 1 keV) the NC loss rate is roughly doubled. Our findings would, in principle, suggest that a 2D potential could be a candidate for explaining the density pump out. To get quantitative results one would, however, have to solve the 2D electric field self-consistently. Fishbone induced losses of ICRH accelerated fast ion A new collaboration has been started with the UKAEA modelling group to study interactions between ICRF accelerated fast ions MHD modes by coupling the SELFO code that calculates the ICRF ions and the HAGIS that describes the interaction of fast ions with MHD modes. Initial studies have shown that core localized Fishbone perturbations can drive losses of fast ions to plasma facing components. Publications section 2.4 Peer reviewed journals 1. V.G. Kiptily, C.P. Perez von Thun, S.D. Pinches, S.E. Sharapov, D. Borba, F.E. Cecil, D.

Darrow, V. Goloborod'ko, T. Craciunescu, T. Johnson, F. Nabais, M. Reich, A. Salmi, V. Yavorskij, M. Cecconello, G. Gorini, P. Lomas, A. Murari, V. Parail, S. Popovichev, G. Saibene, R. Sartori, D.B. Syme, M. Tardocchi, P. de Vries, V.L. Zoita, JET-EFDA Contributors, "Recent progress in fast ion studies on JET", Nuclear Fusion, vol. 49 pp. 065030, 2009

2. P. Mantica, D. Strintzi, T. Tala, C. Giroud, T. Johnson, H. Leggate, E. Lerche, T. Loarer, A. G. Peeters, A. Salmi, S. Sharapov, D. Van Eester, P. C. de Vries, L. Zabeo, K.-D. Zastrow, "Experimental Study of the Ion Critical-Gradient Length and Stiffness Level and the Impact of Rotation in the JET Tokamak", Physical Review Letters, vol. 102 pp.

3. J. P. Graves, I. Chapman, S. Coda, L.-G. Eriksson, and T. Johnson, "Sawtooth-Control Mechanism using Toroidally Propagating Ion-Cyclotron-Resonance Waves in Tokamaks", Physical Review Letters, vol. 102 pp. 065005, 2009 175002, 2009

4. L.-G. Eriksson, T. Hellsten, M. F. F. Nave, J. Brzozowski, K. Holmström, T. Johnson, J. Ongena, K.-D. Zastrow, JET-EFDA Contributors, "Toroidal rotation in RF heated JET plasmas", Plasma Physics and Controlled Fusion, vol. 51 pp. 044008, 2009

Other workshops and conferences 5. M.-L. Mayoral, J. Ongena, A. Argouarch, Yu. Baranov, T. Blackman, V. Bobkov, R.

Budny, L. Colas, A. Czarnecka, L. Delpech, F. Durodié, A. Ekedahl, M.Gauthier, M.Goniche, R. Goulding, M. Graham, J. Hillairet, S. Huygen, Ph. Jacquet, T. Johnson, V. Kiptily, K. Kirov, M. Laxåback, E. Lerche, J. Mailloux, I. Monakhov, M.F.F Nave, M. Nightingale, V. Plyusnin, V. Petrzilka, F. Rimini, D. Van Eester, A.Whitehurst, E. Wooldridge, M.Vrancken, JET-EFDA Task Force H, JET EFDA contributors, "Overview of Recent Results on Heating and Current Drive in the JET Tokamak", Proceedings of the 18:th Topical Conference on Radio Frequency Power in Plasmas, 2009

Presentations at EPS conferences 6. I. Chapman, J. P. Graves, T. C. Hender, S. D. Pinches, S. Saarelma, R. J. Akers, L. C.

Appel, R. V. Budny, R. J. Buttery, S. Coda, J. C. Connor, C. G. Gimblett, R.J. Hastie, G. T. A. Huysmans, V. G. Igochine, I. Jenkins, T. Johnson, H. R. Koslowski, Y. Liang, M. Maraschek, A. B. Mikhailovskii, M. F. F. Nave, S. E. Sharapov, G. Tardini, M. Turnyanskiy, I. Voitsekhovitch, C. Wahlberg, the ASDEX Upgrade MAST & TEXTOR


59

Teams, JET-EFDA Contributors, "Modelling the stability of the n=1 internal kink mode in tokamaks", Proceedings of the 36th EPS Conference on Plasma Physics, 2009

2.5 Energetic particle physics 2.5.1 Physics of burning fusion plasmas D. Anderson, M. Lisak, M. K. Lilley, R. Nyqvist (PhD student) Introduction The research programme is concentrated on the study of stability and transport in magnetically confined, impure plasmas with a special emphasis given to transport driven by microinstabilities and fast particle effects in fusion plasmas. The aim is to develop improved theoretical models within this area, to be able to explain experimentally observed phenomena in existing magnetic confinement devices and to make predictions on the implications for the design of future ones. One of the main objectives of tokamak devices such as JET and the planned ITER project is the study of alpha particle production, confinement and consequent alpha particle heating of D-T plasmas. Up to now the fusion experiments were in a sub- critical zone with the product of the plasma density n, plasma temperature T and the energy confinement time !e being of the order of nT!e< 8.3x1020 m-3keVs and without burn or with a small burn, e.g. with Q=0.61 achieved at JET. In a burning DT plasma with Q>10, the energetic alpha particle population will contribute a considerable part of the total plasma pressure (typically 10-20%) and the alphas can be expected to give rise to fundamentally new physics phenomena which may have a substantial impact on achieving and maintaining high temperatures, thus modifying the beta limits and energy confinement times of tokamak reactors. Furthermore, the alpha particle population itself may have a considerable impact on plasma stability, possibly affecting the alpha thermalisation as well as the transport of high-energy alphas and thermal ions. From a practical point of view, both the confinement capability, wall loading and fusion power density of a tokamak may be affected by the presence of the alpha particles. For a number of years the Chalmers group has carried out theoretical research concerning the physics of burning fusion plasmas. The research has mainly been devoted to the following questions:

• What fraction of the energy associated with the alpha particles will be transferred to the plasma and what are the direct losses of the alphas to the wall?

• Which collective effects and associated macroscopic processes will result from the alpha particles and what is the impact of alpha particles on plasma stability and transport?

• Is it possible to draw conclusions about fast ion behaviour already from existing experimental data?

• Is (quasi-) steady state thermonuclear burn possible and which methods are required for burn control?

Both the design features and the operation regime of a fusion reactor depend on the answers to these questions. The confinement properties of energetic alpha particles are of fundamental importance for alpha heating, burn control and alpha particle diagnostics. The next generation of tokamaks will produce a large amount of thermonuclear alpha particles, which may excite


60

wave instabilities. The basic reason for these instabilities is the deviation of the alpha distribution function from thermodynamic equilibrium, and the instability drive may thus be spatial as well as velocity space gradients, trapped as well as passing particles. The presence of thermonuclear instabilities may in turn cause anomalous losses of plasma energy and high-energy particles. Many aspects of fast ion collective effects are well understood, including excitation of MHD modes such as fishbone oscillations and Alfvén gap modes in toroidal devices. However, a "hot" topic is the role of kinetic and nonlinear Alfvén instabilities. Promising topics for future advances include explanations for rapid frequency chirping, unambiguous identification of energetic particle modes and a quantitative understanding of the radial structure of Alfvén modes. Generally, to achieve an improved understanding of nonlinear dynamics of the thermonuclear wave instabilities driven by energetic ions and to determine the wave saturation mechanism and its implications fast ion confinement, requires a further development of the theory of the nonlinear evolution of kinetic systems by H. Berk and B. Breizman. Nonlinear evolution of beam driven modes on MAST M. K. Lilley, M. Lisak, B. N. Breizman, S. E. Sharapov, S. D. Pinches Experiments on Alfvénic instabilities driven by super-Alfvénic beams in the spherical tokamak MAST have exhibited a variety of modes excited in a broad range of frequencies from Alfvén Cascade (AC) eigenmodes (Fig 2.5.1-1), Toroidal Alfvén Eigenmodes (TAE) (Fig 2.5.1-2), and chirping modes in the frequency range 50-150 kHz, to compressional Alfvén eigenmodes (CAE) in the frequency range 0.4-3.8 MHz (Fig 2.5.1-3), which is approaching the cyclotron frequency of plasma ions, (0.1 1) ci! !" ÷ .

Fig 2.5.1-1 Magnetic spectro-gram showing AC and TAE

Fig 2.5.1-2 Modes sweeping up in frequency over the TAE frequency range 90-130 kHz

Fig 2.5.1-3 Magnetic spectro-gram showing bursting CAE (B=0.5 T, I=750 kA, PNBI=3 MW)

In spite of the differences in the nature of the modes (shear Alfvén or CAE), and in the corresponding excitation mechanisms, the nonlinear evolution of these modes, as seen in their spectrograms, show similarities. The aim of the collaborative work with MAST is to demonstrate that the nonlinear evolution of these modes, when driven just above their excitation threshold, is determined by a competition between the electromagnetic field of the modes and the relaxation processes of the fast particles driving these modes. In particular, we have validated the recent theoretical finding that drag encourages beam-driven waves near marginal stability to exhibit an explosive scenario. An efficient fully nonlinear numerical tool BOT has now been created for this task, using a bump-on-tail model, the results of which show characteristic features typical of any near threshold energetic particle driven instability, examples of which are asymmetric and hooked frequency sweeping modes (Figs 2.5.1-4 and


61

2.5.1-5). It is noteworthy that the hooks of Fig 2.5.1-5 resemble experimentally observed chirping patterns. A reduced analytical model has been created which captures the affect of drag and diffusion on the mode evolution.

( )

2 ( 1)

(1 )

yx y

xy y yax

!

! !

"= +

"" "

+ = +" "

(0.1)

where x and y are effectively the electric field of the wave and the deviation of the frequency from the plasma frequency respectively. A numerical solution of Eqs 1.1 (shown in Fig 2.5.1-6) for the case shown in Fig 2.5.1-5 reproduces the evolution of a single hook. The frequency and time scales are in good quantitative agreement with the well isolated hooks seen in Fig 2.5.1-5 as shown in Fig 2.5.1-7.

Fig 2.5.1-4 Asymmetric frequency sweeping results from the BOT code including the effect of drag and velocity space diffusion

Fig 2.5.1-5 Hooked frequency spectrum with drag and an increase in diffusion from that in Fig 4.

Fig 2.5.1-6 Analytical hooked chirping, the solid line effectively shows the electric field and the dashed line shows deviation of the frequency from the plasma frequency

Fig 2.5.1-7 Quantitative agreement between hooks seen in Fig 5 and Fig 6 (white line) is shown.

These universal asymmetric features are observed in recent and past MAST experiments, indicating the central role that the drag might play in the evolution of beam or alpha particle driven waves in tokamaks. This has galvanised the current effort to perform quantitative modelling of beam driven Alfvénic instabilities on MAST in the presence of drag using the


62

HAGIS code, results of which agree with the main trends of the bump-on-tail analysis as seen in Figs 2.5.1-8 and 2.5.1-9.

Fig 2.5.1-8 Symmetric chirping pattern for a marginally unstable TAE mode, in the absence of fast particle collisions, in response to kinetic !-particle drive calculated using HAGIS

Fig 2.5.1-9 Hooked frequency spectrum for a TAE with drag and Krook collisions (to simulation diffusion), calculated using HAGIS

Alfvénic Eigenmodes within the q = 1 Radius in Sawtoothing Plasmas R. Nyqvist, B. N. Breizman, M. Lisak, S. E. Sharapov In JET discharges with high power ICRH and a central safety below unity, a hot ion population inside the q = 1 radius can stabilize the sawtooth crash during periods of up to 1 s or longer. During these periods, the central electron temperature builds up to significantly larger values than during ordinary sawtooth oscillations, finally saturating or even decreasing. The sawtooth free periods end with giant sawtooth crashes, preceded by magnetic activity in the toroidal Alfvén eigenmode (TAE) and sub-TAE frequency range (Fig 2.5.1-10), exhibiting a characteristic frequency evolution where cascading modes are transformed into “tornado modes” [P. Sandquist, S. E. Sharapov, M. Lisak and T. Johnson, Physics of Plasmas 14, 122506 (2007)].

Fig 2.5.1-10 Magnetic activity in JET discharge 66203 just prior to giant sawtooth crash.

Fig 2.5.1-11 Spectrum of numerically identified modes.

In this contribution, ideal and non-ideal Alfvén eigenmodes (AEs) within the q = 1 surface are investigated analytically and numerically with a suite of MHD codes. It is found that in


63

typical JET equilibria, the cascading modes can be identified as core-localized Alfvén cascades (AEs supported by reversed or extremely low shear) achieving their maximum amplitude close to, but below, the frequency range of a collection of low shear toroidal AEs, exhibiting “tornado” behavior due to the proximity of the magnetic axis (Fig 2.5.1-11).

Radiative Damping of Low Shear Toroidal Alfvén Eigenmodes R. Nyqvist, S. E. Sharapov, M. Lisak Radiative damping due to finite ion Larmor radius effects was considered analytically for core-localized low shear toroidal Alfvén eigenmodes (TAEs) in the center of a large aspect ratio tokamak. The introduction of the short ion Larmor radius length scale admits for wave tunneling into extremely short perpendicular wavelengths. The tunneled wave amplitude is propagated radially away from the low-shear region and considered lost. It was found, using a ballooning space formulation corresponding to a radial Fourier transformation, that radiative damping of the upshifted TAE mode in the toroidally induced Alfvén gap was negligible, while the downshifted mode could experience some damping in the very low shear limit (due to a somewhat more narrow potential barrier). These results are currently being reproduced using a real space representation. Wave-Particle Interaction in Toroidal Axisymmetric Systems R. Nyqvist, M. Lisak, D. Anderson, J. Zalesny The interaction between charged particles and plane electromagnetic waves in a large aspect ratio tokamak is investigated analytically by calculating the induced change in the constants of motion per poloidal transit period due to the presence of the wave. The wave-particle resonances are treated locally using a stationary phase method, and by forming a system of difference equations, a condition for stochastic particle motion is derived in terms of the wave amplitude. If the motion is stochastic, either due to large wave amplitude or other effects such as e.g. magnetic field ripples, an appropriate Fokker-Planck equation can be derived with the effect of the wave-particle interaction treated as a diffusive process. The associated diffusion coefficient can be calculated in terms of the changes in the constants of motion of the unperturbed particle orbits due to the wave. Mechanical analogy of the nonlinear dynamics of a driven unstable mode near marginal stability J. Zale!ny,1 S. Marczy"ski, M. Lisak, D. Anderson, A. Ga#kowski, 1 P. Berczy"ski,, S. Berczy"ski and R. Rogowski The universal integro-differential model equation derived by Berk et al. for studying the nonlinear evolution of unstable modes driven by kinetic wave particle resonances near the instability threshold is reduced to a differential equation and next as a further simplification to a nonlinear oscillator equation. This mechanical analogy properly reproduces most of the essential physics of the system and allows an understanding of the qualitative features of the theory of Berk et al.


64

EFDA JET highlights

• Sweeping up and tornado type modes in the toroidal Alfv´en eigenmode frequency range are observed prior to monster sawtooth crashes in JET and Alcator C-Mod discharges with a significant hot ion population. The possible types of modes are explained in terms of kinetic and MHD theory. The acquired mode structure is be used in future modelling of hot ion transport through the q = 1 magnetic surface during monster sawtooth crashes.

ITPA highlights

• Experiments on Alfvénic instabilities driven by super-Alfvénic beams in the spherical tokamak MAST have exhibited a variety of modes excited in a broad range of frequencies from Alfvén Cascade (AC) eigenmodes, Toroidal Alfvén Eigenmodes (TAE), and chirping modes in the frequency range 50-150 kHz, to compressional Alfvén eigenmodes (CAE) in the frequency range 0.4-3.8 MHz, which is approaching the cyclotron frequency of plasma ions, (0.1 1) Bi! !" ÷ . The Chalmers group has demonstrated that, in spite of the significant differences in the nature of the modes (shear Alfvén or CAE), and in the corresponding excitation mechanisms, the nonlinear evolution of these modes is determined by the type and strength of relaxation processes of the fast particles driving the waves. In particular, the recent theoretical finding that drag encourages the beam-driven waves near marginal stability to follow an explosive scenario has been validated. An efficient numerical tool has now been created for this task using a simplified bump on tail model, the results of which show similarities with features that are observed in recent and past experiments, indicating the central role that the drag might play in the evolution of beam or alpha particle driven waves in tokamaks. This has galvanised the current effort to perform quantitative modelling of Alfvénic instabilities in the presence of drag using the HAGIS code, preliminary results of which agree with the qualitative trends of the bump on tail analysis.

Publications section 2.5.1 Peer reviewed journals

1. J. Zale!ny,1 S. Marczy"ski,2 M. Lisak,3 D. Anderson,3 A. Ga#kowski,1 P. Berczy"ski,2, S. Berczy"ski and R. Rogowski, Mechanical analogy of the nonlinear dynamics of a driven unstable mode near marginal stability, Physics of Plasmas 16, 022110 (2009).

2. M. K. Lilley, B. N. Breizman, and S. E. Sharapov, Effect of dynamical friction on nonlinear energetic particle modes, Accepted for publication in Physics of Plasmas (Sept 2010)


3. R. Nyqvist, B. N. Breizman, M. Lisak, S. E. Sharapov, Fast Ion Driven Alfvén Eigenmodes within the q = 1 Radius. Presented at the 36th EPS Conference on Plasma Physics in Sofia, Bulgary, 2009.

4. R. Nyqvist, S. E. Sharapov, M. Lisak, Radiative Damping of Low Shear Toroidal Alfvén Eigenmodes. Presented at the 37th EPS Conference on Plasma Physics in Dublin, Ireland, 2010.


65

Other workshops and conferences 5. M. K. Lilley, B. N. Breizman, and S. E. Sharapov Destabilising effect of dynamical

friction on fast particle driven waves, Presented at the 11th IAEA TCM on Energetic Particles in Magnetic Confinement Systems, Kiev, 2009

6. M. K. Lilley, B. N. Breizman, and S. E. Sharapov, Effect of drag on nonlinear scenarios for fast particle driven waves, U.S. Transport Task Force Workshop, Annapolis, 2010

7. R. Nyqvist, B. N. Breizman, M. Lisak, S. E. Sharapov and JET-EFDA contributors, Alfvénic Eigenmodes within the q = 1 Radius in Sawtoothing Plasmas, presented at the 11th IAEA TCM on Energetic Particles in Magnetic Confinement Systems, Kiev, 2009.

2.5.2 Runaway electrons in tokamak disruptions T. Fülöp, T. Fehér (PhD student), G. Papp (PhD student), M Jakobs (M Sc student) Introduction Due to a sudden cooling of the plasma in tokamak disruptions a beam of relativistic runaway electrons is sometimes generated, which can cause damage on plasma facing components due to highly localized energy deposition. This problem becomes more serious in larger tokamaks with higher plasma currents and understanding of the processes that may limit or eliminate runaway electron generation is very important for future tokamaks, such as ITER. During 2009, our model for runaway dynamics in the presence of impurity injection have been refined and extended with the hot-tail generation mechanism. Furthermore, we have explored the magnetic field dependence of the runaway beam formation and we started the development of a tool to assess the effect of resonant magnetic perturbations on runaway losses. Runaway electron generation in tokamak disruptions Runaway electrons can be generated in disruptions by the Dreicer, hot-tail and avalanche mechanisms. Analytical and numerical results for hot tail runaway generation have been included in a 1-dimensional model of electric field, temperature and runaway current, which has been applied to simulate disruptions and fast shutdown. We have shown that the peaked shape of the runaway current density profile may cause tearing modes to become unstable. Fast shutdown has been studied by prescribing varying amounts of injected impurities. Our results show that large argon content suppresses runaways in JET simulations but causes hot tail generation in ITER. A pellet code has been coupled to the runaway model, and it is extended to enable simulations of carbon doped deuterium pellet injection. Such pellets are seen not to give enough cooling for a fast current quench.


66

Fig.2.5.2-1 (a) Temperature and (b) currents in a simulation with 10% argon in JET. (c) a parameter scan for argon injection in JET. Circles (left axis): fraction of extra deuterium required for runaway suppression. Crosses (right axis): resulting current quench time. Magnetic field threshold for runaway generation in tokamak disruptions Experimental observations show that there is a magnetic field threshold for runaway electron generation in tokamak disruptions. We have studied two possible reasons for this threshold. The first possible explanation for these observations is that the runaway beam excites whistler waves that scatter the electrons in velocity space prevents the beam from growing. The growth rates of the most unstable whistler waves are inversely proportional to the magnetic field strength. Taking into account the collisional and convective damping of the waves it was possible to derive a magnetic field threshold below which no runaways are expected. The second possible explanation is the magnetic field dependence of the criterion for substantial runaway production obtained by calculating how many runaway electrons can be produced before the induced toroidal electric field diffuses out of the plasma. We have shown, that even in rapidly cooling plasmas, where hot-tail generation is expected to give rise to substantial runaway population, the whistler waves can stop the runaway formation below a certain magnetic field unless the post-disruption temperature is very low. EFDA Task Force and Topical Group highlights As part of task WP09-MHD-02-04 we have started to develop a model for 3D runaway drift orbits in 3D magnetostatic perturbed fields to assess the effect of RMP coils on loss enhancement. The relativistic drift equations for the runaway electrons have been solved with the ANTS (plasmA simulatioN with drifT and collisionS) code developed by M Drevlak at IPP Greifswald, modified for our purposes to include the effect of synchrotron radiation losses. The code calculates the drift motion of particles in 3D fields and takes into account collisions with background particle distributions. As test cases we used TEXTOR-like configurations, chosen to be similar to the ones where the runaways were shown to be suppressed by resonant magnetic perturbations created by dynamic ergodic divertor (DED) coils. The reason for this choice is that the TEXTOR experimental observations on both runaway losses and synchrotron radiation measurements allow us to benchmark our calculations. The magnetic equilibrium was calculated by VMEC [S.P. Hirshman et. al., Comput. Phys. Commun. 43, 143 (1986)] for the TEXTOR coils and the magnetic field perturbations have been modelled to be similar to the ones produced by DED-coils at TEXTOR. The results of this work will be reviewed in detail in the Annual report 2010. Publications section 2.5.2 Peer reviewed journals

1. Fülöp T, Smith H M, Pokol G, Magnetic field threshold for runaway generation in tokamak disruptions, Physics of Plasmas, 16 p. 022502 (2009).


67

2. Smith H M, Fehér T, Fülöp T, Gál K, Verwichte E, Runaway electron generation in tokamak disruptions, Plasma Physics and Controlled Fusion, 51 p. 124008, (2009).

Oral presentations at international conferences and workshops.

3. Fülöp T, Pokol G, Smith H M, Helander P, Magnetic field threshold for runaway generation in tokamak disruptions, Presented at the International Sherwood Fusion Theory Conference, Denver, USA, May (2009).

4. Fehér T, Smith H M, Fülöp T, Gál K, Simulation of runaway electron generation during plasma shutdown by impurity injection, Presented at the 11th IAEA Technical Meeting on Energetic Particles in Magnetic Confinement Systems, Kiev, Ukraine, September (2009).

M Sc Thesis

5. Jakobs M, Runaway electron dynamics in tokamak disruptions, Chalmers University of Technology, November 2009.


68

3 Development of plasma auxiliary systems - diagnostics 3.1 Neutron diagnostics Uppsala University: M.Cecconello, S.Conroy, G.Ericsson, H.Sjöstrand, M. Weiszflog, E.Andersson Sundén (PhD student), M.Gatu Johnson (PhD student), C.Hellesen (PhD student), E.Ronchi (PhD student), H.Hellberg (dipl. student), P.Kårén (dipl. student), B.Molander (dipl. student), M.Skiba (undergraduate student), J.Eriksson (undergraduate student). Introduction At JET, instrumental development has focused on the two neutron spectrometers TOFOR (time-of-flight optimized for rate) and MPRu (upgraded magnetic proton recoil spectrome-ter). For TOFOR, detailed work on the problem of back scattered neutrons has been carried out. Careful MCNP modeling of the JET scattering environment was done and the results were used in the analysis of experimental data. It was seen that scattered neutrons make a sig-nificant contribution to certain parts of the time-of-flight spectrum. In an interesting applica-tion of the broad-band capability of TOFOR measurements in the full range 1.5 < En < 20 MeV were performed, showing clear signatures of the dominating 2.5-MeV and 14-MeV neutron emission and also showing that the modeling of scattered neutrons was an important component in understanding the details of the recorded neutron spectrum. For the MPRu, work was carried out to extend the method of neutron yield measurements to include also the 2.5-MeV emission in D plasmas. The analysis of neutron emission spectrometry data from JET has mainly been focused on analysis of TOFOR results from 2008. These data concern several JET experiments aimed at studying fast ions and how they interact with impurities and magneto hydrodynamic (MHD) activities. The analysis effort was in both cases concluded and the results presented in papers. Interesting data were also obtained from a set of high-performance experiments run at the end of the 2009 JET experimental campaign. Here it was shown that JET can achieve plasma operations with a very high component of thermo-nuclear fusion reactions, up to 75%. Code development for analysis and modeling of in support of the JET spectrometers has also continued during 2009. A code for calculating neutron energy spectra from JET discharges was developed, where the exact orbits of the fuel ions are taken into account. Here it could be shown that orbit effects (finite Larmor radii) of energetic ions can indeed have a large impact on what is observed in collimated neutron measurements. Comparisons with TOFOR results are encouraging, but more work needs to be done for a more complete understanding. The development of a neutron camera/spectrometer system for MAST continued during 2009. The design of the instrument was finalized and the procurement and manufacturing of the hardware components of the camera started. The main components are the neutron shielding and its support frame, the internal magnetic and gamma rays shields for the detec-tors, the detectors themselves, their power supplies and the data acquisition system. Initial characterization of the detectors, in terms of energy resolution and response function, was carried out with gamma-ray and neutron sources. Magnetic measurements were carried out to test the performance of the magnetic shielding.


69

Artificial neural networks (NN) are a powerful analysis tool and are particularly useful for fast inversion of convoluted data. Developments of NN analysis applications for neutron mea-surements were undertaken in 2009. The applications were applied to such problems as neu-tron/gamma separation in liquid scintillators, unfolding of neutron spectra and reconstruction of neutron emissivity profiles. The results show that NN can provide results at a level consis-tent with more powerful analysis tools at a tiny fraction of the computational cost. Development of neutron measurement techniques for ITER was performed under the framework of the EFDA Topical Group for Diagnostics. Work on two contracts was con-cluded during 2009 (final reports in 2010). One concerned “neutron based measurements”, an effort coordinated by the UU-VR group; the UU contribution involved a simulation study of different techniques for high resolution neutron spectroscopy on ITER. It was found that so-called thin-foil techniques in most cases out-perform other techniques. A second contract was aimed at investigating different diagnostic techniques for measuring the fuel ion ratio, nT/nD. Here it was shown that neutron spectroscopy could provide (line integrated) results within ITER specifications for a large part of the required parameter space, although in this restricted study not for the optimal nT = nD scenario. The work on both these topics will continue in 2010. 3.1.1 Neutron Diagnostic instrumental development for JET and

MAST The work on instrument developments within the Uppsala-VR neutron diagnostic group has been focused on the Magnetic Proton Recoil neutron spectrometer (installed at JET in 1996, upgraded to MPRu in 2005), the TOFOR time-of-flight neutron spectrometer (installed at JET in 2005) and the new neutron camera for MAST (installation begun in 2009). The TOFOR instrument at JET TOFOR is a neutron time of flight spectrometer at JET. A collimated beam of neutrons escaping the fusion plasma scatters in a first detector and a fraction of the scattered neutrons interact with a second detector. From the time between these two scattering interactions, the energy of the incoming neutron is deduced. During 2009, detailed work on the problem of back scattered neutrons has been carried out. TOFOR, like all neutron spectrometers, does not only record neutrons emitted directly from the plasma, but also such that have scattered off the tokamak vessel into the line of sight. The scattered neutrons will in general have a lower energy than the direct ones: The value depends on the scattering angle and on the material composition in the vessel walls. shows a simulated spectrum from the scattering of 2.5 MeV neutrons on mainly Ni and C (see Fig. 3.1-1). The scattered neutron spectrum will interfere with the direct spectrum and any measurement will include both. Thus, a precise estimate of the scattered component is required for an accu-rate interpretation of the measured data. Figure 3.1-2 shows summed TOFOR data from several high-power discharges. The fit to the data contains a DD component with a dominant peak at 65 ns and a triton burn up component with a peak at 27 ns, corresponding to 14-MeV neutron energy. In addition, a scattered component has been included which accounts well for the background in the time-of-flight spectrum. It can clearly be seen that an omission or underestimate of the background would result in a misinterpretation of the incident neutron spectrum.


70

The MPRu instrument at JET The MPRu is ab initio calibrated, i.e. the flux on the foil of the spectrometer can be estimated from basic physical parameters. This ab initio calibration has been used, together with the JET neutron camera to measure the absolute neutron yield of DT plasmas. In this section, the initial work on extending this method to DD plasmas is presented. Information on the neutron emission profile is estimated from the data of the JET neutron camera system, and the stan-dard magnetic equilibrium code EFIT is used to get the flux surface configuration. The neu-tron emission profile, !, is modelled as being constant on a flux surface, using either a Gaus-sian or a peaked profile function.

Fig. 3.1-2 Measured neutron spectrum from a DD plasma. The fit (blue line) to the experimental data contains DD and TBN components (dashed black lines) as well as a scattered component (red line).

Fig. 3.1-1 Simulated scattered neutron spectrum from a monoenergetic 2.5 MeV source.


71

The optical detector response to the emission profile is simulated. An optical model is used, where JET structures are either opaque or transparent. The volume inside the vacuum vessel is split into voxels and the solid angle seen by a detector from each voxel is calculated. In Fel! Hittar inte referenskälla..1-3, the LOS matrices of the 19 neutron camera channels and the MPRu at JET are shown obtained from this simulation. The colour intensities in the figure correspond to the fractional solid angles (projected on a poloidal plane) that the corresponding detector sees from that voxel. ! is folded with the LOS matrices to calculate the modelled neutron flux at the camera detectors. The forward convolution method is used to fit the parameters of ! to the data of the neutron camera, minimising "2. Due to the ab initio calibration of the MPRu, the flux at the foil of the spectrometer can be calculated from the measured proton histogram.

However, in order to relate the measured neutron flux on the MPRu foil to that of the optical model described above we must correct the measurement for non-optical effects. The attenua-tion, transmission and the two scattering corrections have been estimated using two MCNPX models; one modelling the machine effects (attenuation, transmission and machine scattering) and one the collimator scattering. In the original work on DT yield determination, no correc-tion for anisotropy was deemed necessary due to the isotropic nature of the DT cross section. In the DD case, however, the cross section is highly peaked in the forward and backward directions. An anisotropy correction has been estimated from TRANSP simulations, using thermal, beam heated and supra-thermal plasmas as input. With the correction factors, the intensity of the neutron emission profile is renormalized and corrected. Finally the neutron yield, Yn can be calculated as the sum over the neutron yield in each voxel. All JET discharges which give a "2

red<3 in the model fit to the camera data were selected and yield results for full pulses are shown in Fig. 3.1-4, where the results from this method (Gaussian (blue) and peaked emissivity model (red)) are plotted against results obtained from the cross-calibrated fission chambers. A linear fit to the two data sets gives:

YMPRu,Peak = 1.052(±0.002)YFC and

YMPRu,Gauss = 1.077(±0.004)YFC

Fig. 3.1-3 Projected LOSs on the poloidal plane of (a) the MPRu, (b) the horizontal and (c) the vertical camera. For intensity scale, see text.


72

The uncertainties of the coefficients are solely due to the statistical uncertainties of the MPRu data. In the DT case, the systematic uncertainty of this method has been estimated to 7%.

Fig. 3.1-4 Neutron 2.5-MeV yield measured with the MPRu-camera system vs. fission chamber for selected JET pulses in the range 68000 - 73700 using a Gaussian (blue) and peaked (red) emissivity function. Lines are linear fits to the respective data set. Error bars are statistical uncertainties of the MPRu.

Neutron Diagnostics at MAST The collaboration with the Culham Centre for Fusion Energy has continued during 2009 for the development and installation of a proof-of-principle neutron camera with spectroscopy capabilities at the MAST experiment. On the basis of extensive design studies carried out in 2008, the design was finalized and the procurement and manufacturing of the neutron camera started in 2009. The main components of the neutron camera that were procured are the neu-tron shielding and its support frame, the magnetic and gamma rays shields, the detectors and their power supply and the data acquisition system. Initial characterization of the detectors, such as detector energy resolution and response function, was carried out with gamma-ray and neutron sources. Magnetic measurements were carried out to test the performance of the mag-netic shielding. Diagnostic design The final design of the proof-of-principle neutron camera consists of four lines of sight, two on the equatorial plan and two diagonal ones. At present only the collimators for the equato-rial lines of sight have been manufactured: the additional two lines of sight are planned to be installed during 2010. Figure 3.1-5, shows the final design of the neutron shielding.


73

Fig.3.1-5 Left: CAD model of the MAST neutron camera sitting on the rail in relation to MAST vacuum vessel. Right: assembled neutron shielding.

The overall weight of the shielding and of its support structure (custom made) is around 4.5 tons. The high purity polyethylene shielding material was purchased from John Caunt Scien-tific and manufactured by the Culham Centre for Fusion Energy. The design for the detectors holders and ancillary servicing system was also finalized. In Fig. 3.1-6 a model of the final assembly of the four detectors is shown: detector holders, which double as heat dissipation units via copper braids, can be seen. The cooling system installation is foreseen for 2010 when the camera will start operations. Figure 3.1-7 shows the area were the detectors will be installed on the back of the neutron shielding: clearly visible are the double soft iron magnetic shielding boxes (with a thin µ-metal thin foil on the inside of the inner box). Inside the boxes, lead blocks are installed to reduce the gamma rays background resulting from the thermal capture of neutrons in the surrounding materials.

Fig . 3.1-6. CAD model detail of the collimators (grey) and detectors (green) of the neutron camera. In dark red the detector holders that are also used as thermal contact for heat dissipation towards the heat exchanger via copper braid (brown).


74

Fig 3.1-7 Left: CAD model of the double magnetic shielding box with the penetration of the collimators. Right: assembled magnetic and gamma ray shielding structures before the installation of the detectors.

Neutron detectors Four liquid scintillator detectors were purchased from Scionix Ltd. The detectors have an active scintillator volume of 20 ! 15 ! 50 mm coupled via a short quartz optical light guide to a low profile fast photomultiplier R5611A from Hamamatsu (Fig. 3.1-8). An optical fibre is connected to the light guide and will be used for monitoring the stability of the PMT gain via a pulsed constant amplitude light source (LED). The LED source has been specifically designed and built to deliver light pulses in the wavelength region 450 – 480 nm to mimic the scintillation light from the detector with a 5 kHz frequency and pulse width of 100 ns.

Fig. 3.1-8 NE213 neutron detector for the MAST neutron camera. For calibration purposes, a 22Na radioactive source will be mounted on the back of the detector at the circular recess (top).

The detectors were tested and characterized with gamma ray (137Cs, 60Co and 22Na) and neu-tron (252Cf) sources. Four 22Na sources were purchased to be used for in-situ calibration of the detectors. The gamma sources were used in the determination of the detectors energy resolution, as shown in Fig.3.1-9, and for determining the detector response function. A


75

MCNP model of the detectors was developed and the predicted gamma ray pulse height spectrum was folded with the experimentally determined energy resolution function and then compared with the experimentally measured pulse height spectrum as shown in Fig.3.1-10.

0.00 0.25 0.50 0.75 1.00 1.250

5

10

15

20

25

30

35

Data: Calibration_DLModel: Ne213Weighting: y No weighting Chi^2/DoF = --R^2 = 1 a 10.03144 ±--b 17.27357 ±--c 3.60186 ±--

SEK 467

Res

olut

ion !L

/L (%

)

Relative Pulse Height L Fig. 3.1-9 Typical energy resolution of the neutron detector for three different gamma energies.

Fig. 3.1-10 Left: MCNP model of the neutron detector. Right: comparison between the experimentally measured and MCNP simulated pulse height spectrum for a 22Na source.

The maximum temperature increase of the detectors was measured in lab tests and found to be 7 °C per detector, in close agreement with an estimated increase of 30 °C for all four detectors installed in the diagnostic cavity. Four solid state temperature sensors have been purchased to be installed on the detector holders. The pulse shape discrimination capabilities of the detec-

0 0.5 1 1.50

0.5

1

1.5

2

2.5

3

3.5

4x 104

Energy MeV

Norm

aliz

ed c

ount

s

MeasuredFolding


76

tors were tested in the laboratory using a 252Cf neutron/gamma source. Individual pulses were recorded and separated in gamma and neutron events by comparing the charge under the long decay tail. An example of the achievable separation is given in Fig. 3.1-12 where is also shown the LED signature, clearly different from gamma ray and neutron pulses.

0 1800 3600 5400 7200 9000 108000

1

2

3

4

5

6

7

8 SEK467 SEK468 SEK469 SEK470

time (s)

T DE

T - T R

EF (o C

)

Fig. 3.1-11 Measured temperature increase for each of the four NE213 detectors at nominal operating high voltage. The detectors where switched off at about t=10000s.

Fig. 3.1-12 Neutron detector pulse shape discrimination results using a 252Cf source. Left: gamma (top band) and neutron (lower band) pulses clearly separated in two regions in a Qfast vs Qslow plot. Right: PSD results with a 252Cf source in combination with the LED light source (far right).

Data acquisition The data acquisition system consists of a fast two-channel waveform digitizer with 14 bit pulse height resolution, 250 MHz sampling frequency and 256 MB on board memory. The card is installed in a dedicated data acquisition computer for which a data acquisition system


77

was written in C and Python. Integration of the data acquisition system into the MAST data acquisition system has started and is expected to be completed during 2010. EFDA Task Agreement The partial funding and financial support for the MAST neutron camera, granted via a “Pro-posal for implementation of elements of the EFDA Work Programme under Priority support” EFDA (08) and more specifically with the task agreement TGS-01-01 “Development of a neutron spectrometer for fast particles studies at MAST”, was extended until the end of 2009. In response to the “Call for Participation in the Activities under the EFDA 2010 Work Pro-gramme” by the “Topical Group MHD activities” an application for baseline support was submitted under the heading “WP10-MHD-01-01 Experiments on fast particle instabilities” under the title of “Co-ordinate experiments on fast particle instabilities at MAST”. The aim of this activity is to carry out experimental measurements of the fast ions confinement, redistri-bution and losses in MAST following the excitation of collective instabilities by super-Alfvénic fast particles from deuterium neutral beam injection. In particular, the confinement, redistribution and losses of fast ions will be studied by measuring the neutron emissivity pro-file following the beam-beam (and beam-thermal) DD fusion reactions in different plasma scenarios (H-mode, on-axis and off-axis heating).

3.1.2 Participation in the experimental programme at JET The analysis of neutron emission spectrometry data from JET has mainly been focused on analysis of TOFOR results from 2008. The data from 2008 concern several experiments aimed at studying fast ions and how they interact with impurities and magneto hydrodynamic (MHD) activities. Preliminary results from these experiments were partly reported in the 2008 Annual report, and during 2009 the analysis was continued and presented in several papers. A discussion of high quality TOFOR data obtained from two such scenarios, both with deute-rium bulk plasmas, is offered below. Interesting data were also obtained from a set of high-performance experiments run at the end of the 2009 JET experimental campaign. Neutrons from plasmas with ICRF heated 3He and Be impurity In the first scenario, a minority of 3He was accelerated to MeV range energies with ion cyclo-tron resonance frequency (ICRF) heating, and the neutron emission from reactions between the energetic 3He and traces of 9Be was measured. In Fig.3.1-13 (a), the measured TOFOR data from one JET discharge are shown. The calculated tTOF spectrum of the neutron emission from energetic 3He reacting with thermal 9Be is shown in dashed black and the calculated contribution from thermo nuclear DD reactions is shown in solid red. In Fig.3.1-13 (b) the calculated spectral components are shown on the neutron energy scale. As can be seen there is a good agreement between calculated spectral components and the measured data. It was fur-ther found that the contribution to the total neutron rates from beryllium reactions scaled inversely with the 3He concentration. For low 3He concentrations, around 1%, the beryllium reactions made up almost 60% of the total neutron emission while for higher 3He concentra-tions, around 5%, the beryllium reactions contributed only about 10%. Details of the results are presented in1. During 2009, experiments were run at JET where the neutron emission from ICRF-accelerated 4He reacting with beryllium constituted a significant proportion of the total neutron rates.

1 M.Gatu Johnson, 2010 Nucl. Fusion 50 045005 doi: 10.1088/0029-5515/50/4/045005


78

Fig. 3.1-13(a) Measured TOFOR data during minority 3He acceleration. The expected tTOF spectrum from energetic 3He reacting with thermal 9Be is shown in dashed blue as well as expected

Further, the expected neutron emission from reactions between energetic 3He and 4He ions and plasma impurities during the non-activated phase of ITER operations was investigated. It was found that for some scenarios, mainly involving harmonic ICRF heating, such reactions could potentially be a problem as a too high neutron emission during this phase of ITER is not accepted. The results were presented at the IAEA technical meeting on fast particles in Kiev2. Neutron from 3rd harmonic ICRF heating A second set of experiments involved ions from deuterium neutral beam injection (NBI) heating that were accelerated with 3rd harmonic ICRF heating. This heating scheme is effi-cient in creating a population of high energy ions and very high neutron rates were observed using only modest ICRF heating powers. As reported in 2008, clear correlations between the energetic ions measured by TOFOR and the presence of MHD activity were seen. These cor-relations were further analyzed, and it was found that the orbits produced by the ICRF heating scheme have turning points very close to the edge of the TOFOR sight line (Fig. 3.1-14).

Fig.3.1-14. Typical orbit of a 1.1 MeV deuteron accelerated with 3rd harmonic ICRF heating. The edges of the TOFOR sight line are shown as the vertical dashed lines.

2 M.Gatu Johnson, 2010 Nucl. Fusion 50 084020 doi: 10.1088/0029-5515/50/8/084020


79

Thus, small shifts (around a few centimeters) of the ions gyro centers outwards along the major radius (R) could cause a significant reduction of the TOFOR signal. It was investigated if such shifts were compatible with the types of MHD activities present in these discharges. The results show that the toroidal Alfvén Eigenmodes that were seen could potentially cause such redistributions, and the details were presented at an invited talk during the IAEA tech-nical meeting on fast particles in Kiev. It was further seen that the entire gyro orbit of the energetic ions did not fit in the TOFOR sight line (Fig. 3.1-14). This has large consequences on the shape of the neutron spectrum from the energetic ions and was investigated in detail in a dedicated modeling effort during 2009 (see below). High performance NBI heated D plasmas At the end of campaign C27 (October 2009), a number of high performance pulses with high plasma current and high NBI power were run at JET. Preliminary analysis of TOFOR data from these experiments showed a very high level of thermo-nuclear emission, up to 75%, for several seconds. Extrapolating these results to DT plasma conditions indicate that record-breaking levels of fusion energy could be produced in future JET DT campaigns. Modeling work A MATLAB code for calculating neutron energy spectra from JET discharges was developed during the fall of 2009. The code uses the fuel ion distribution functions calculated by the ICRH code SELFO to generate the spectrum through a Monte-Carlo simulation. The calcu-lated spectra were then compared with experimental results from TOFOR.

Fig. 3.1-15 Calculated neutron spectra (blue lines) with and without line of sight (LOS) effects. The spectra are compared with experimental TOFOR data from JET (points with error bars).

In the calculations, the exact orbits of the fuel ions are taken into account, to investigate what effect this has on the spectrum. The reason is that, for certain plasma heating scenarios, large populations of fast fuel ions are formed. These fast ions may have Larmor radii of the order of decimeters, which is comparable to the width of the TOFOR viewing cone, and may therefore affect the recorded neutron spectrum. Data from JET discharge # 74937, with both NBI and 3rd harmonic radio frequency heating, was analyzed. The results show that the details of the line of sight (LOS) of the detector indeed affect the neutron spectrum (Fig. 3.1-15). This effect is probably important to other diagnostics techniques, such as gamma-ray spectroscopy

LOS effects not included

LOS effects included


80

and neutral particle analysis, as well. Good agreement with TOFOR data is observed, but not for the exact same time slice of the discharge, which leaves some questions yet to be investi-gated. Comparison of TOFOR and NPA results on distribution functions The first cross-validation of the energy distribution function (EDF) of fast deuterium ions derived from measurements with TOFOR and the high energy NPA (KF1) was carried out in plasmas with deuterium neutral beam injection and 2nd harmonic ICRF heating. The valida-tion of TOFOR and NPA measurements is important to form a solid basis for the fast ion diagnostic techniques to provide reliable information in the ITER relevant energy region ED <1 MeV. Figure 3.1-16 shows the absolute (i.e. without any normalization) energy distribution functions of deuterium from TOFOR and KF1 for the ten one-second time intervals for pulse 69247. From left to right, the first two panels are during NBI beam blip, the middle panel right after beam blip and the last two panels during periods of RF heating only.

Fig.3.1-16 Comparison of the deuterium energy distribution function derived from TOFOR and the high energy NPA of JET pulse 69247 with 1 s time resolution.

The development of the EDF over time was seen to follow the heating applied, with a build-up of a population of highly energetic deuterons during the first second of NBI injection, a stable high-energy population during the second, a decline period after the beam is switched off followed by a decreased but significant presence of high-energy deuterons during the RF-only phase. The agreement between the EDFs derived from TOFOR and NPA data is good both in the presence of combined NBI and ICRH as well as during the phase of IRCH heating only. Based on these preliminary results a proposal was submitted and accepted for a dedi-cated experiment on “Diagnosing the D distribution with ED < 1 MeV”. Substantial time and effort was put into preparations for this experiment. Unfortunately, due the technical problems with the JET machine, the experiment was cancelled only hours before its execution in Octo-ber 2009. The proposal will be re-issued for the JET experimental programme in 2011.

3.1.3 Development of analysis tools based on neural networks Artificial neural nets are a computational method used in numerous applications. Typically an artificial neural net will be trained on a dataset where the desired outcome is known then ap-plied to new data where the outcome is not known. Examples include pattern recognition, data mining, e-mail spam filters and many others. In the last year they have been applied to pulse


81

shape discrimination, profile unfolding and emissivity calculations in the group and these will be detailed here. Pulse shape discrimination A useful property of some scintillators is that the light pulses have different shapes when neutrons and gammas are detected. Numerous techniques have been applied to the separation of pulses into the two categories over the years with varying degrees of success. A detailed breakdown of how artificial neural nets were applied successfully is presented in Ref.3. Whilst the initial training of the net can take some considerable computation time, the final test is very fast, with over 300 kHz of samples being processed per second. The method is shown to have less false positives than other techniques and able to operate effectively at smaller pulse heights as well. A comparison of the performance of various PSD methods applied to neutron and gamma events from a 252Cf source is shown in Fig. 3.1-17.

Fig. 3.1-17 Fraction of incorrectly identified ! (top) and neutron (middle) events as a function of the deposited energy for all methods. (bottom) P as a function of the deposited energy for all methods.

3 E. Ronchi et al, Nucl. Instr. And Meth. A610 (2009) 534


82

Free unfolding A traditional problem in neutron spectroscopy is that whilst we would like to measure the neutron energy we typically measure another quantity instead. More formally, the spectrome-ter is said to have a detector response function which is convoluted with the neutron spectrum to give the measured quantity. Unfortunately there is no method to invert this process to give the exact neutron energy spectrum and numerous methods have been used to approximate the answer. Here we present a application where artificial neural networks were used to try to perform this function and the results are published. If no prior knowledge of the spectrum is assumed, the inversion method is classed as a free unfolder. The datasets that the network was trained with consist of random Gaussian shapes with between 1 and 6 such peaks being present in a spectrum. These were then convoluted with the detector response function for a liquid scintillator to give the measured dataset. Neural nets were then trained using these data sets. The trained nets were then applied to another measured dataset to determine the original neutron spectra. Another free unfolding code (MAXED) was also applied to this final dataset and the results compared. The results indicate that neural nets can provide unfolded spectra of similar quality to MAXED and con-siderably faster. Neutron emissivity tomography The so called inverse problem, encountered in the free unfolding work above, also appears when trying to determine the underlying emission profile of a plasma from a limited number of line integrals passing through it. KN3 is a neutron diagnostic at JET consisting of 19 lines of sight arranged in two fans (“cameras”). In order to provide an effective 2D image recon-struction from KN3 data a physics based model of the possible structures which may be viewed was created. This has the effect of reducing the effective number of free parameters to a more manageable quantity than for a true 2D unfolding. The details of the physics based model can be found in4. Components are created for the bulk thermal, NBI and RF heating components which can be added to different neural nets. These parametric forms were then used on data from the Trace Tritium Experiment (TTE) at JET. The data from the campaign was first examined using fuzzy clustering to assemble certain sets of discharges and to estab-lish the range of magnetic flux surfaces which would be used. Approximately 10000 time slices of KN3 data from the TTE were used to train the neural nets. Sample emissivity profiles are shown in Fig. 3.1-18 from a plasma discharge where tritium was puffed into the plasma at the outboard edge and diffuses inwards.

4 E. Ronchi et al, 2010 Nucl. Fusion 50, 035008; doi: 10.1088/0029-5515/50/3/035008


83

Fig. 3.1-18 Example of Neural Network tomographic reconstruction of a JET plasma with a Tritium puff (JET pulse 61138) using simplified models. Four different time slices of KN3 data are shown. Note that the intensity scale in the left column of figures is different for each plot.


84

The results from the neural networks are found to be in reasonable agreement with more com-plex tomographic inversion codes but orders of magnitude faster in processing5. This makes the technique suitable for use in real time plasma control where millisecond scale responses are necessary. Conclusion Artificial neural networks are a powerful analysis tool and are particularly useful for fast inversion of convoluted data. The results show that they can invert data at a level consistent with more powerful inversion tools at a tiny fraction of the computational cost.

3.1.4 R&D for ITER within the EFDA Work Programme During 2009, work on two contracts within the EFDA Topical Group for Diagnostics has been concluded, namely WP08/09-DIA-01-05, Fuel ion ratio, and WP08/09-DIA-01-06, Neutron based measurements. Work on both topics will be continued in 2010 under new con-tracts. Work on the contract WP08/09-TGS-01a, Measurements of confined Alpha Particles (subtask: neutron camera for MAST), has continued and will be extended into 2010. Evaluation of Neutron Spectrometry Techniques The work for the EFDA contract DIA-01-06 focused on an evaluation of different high-reso-lution neutron spectrometer (HRNS) techniques for ITER using realistic instrument response functions (IRF) and simulated (synthetic) neutron data for ITER relevant plasma conditions. The work was divided into four separate parts:

1. A specific simulation study of the thin-foil (non-magnetic) proton recoil (TPR) technique (reported in Annual report 2008).

2. Assembly of realistic IRFs for a number of HRNS techniques as candidates for ITER. 3. Setting up of an integrated work environment to assess the performance of different

HRNS techniques under realistic ITER neutron emission conditions. 4. Assessing the performance of five selected HRNS techniques using the work environ-

ment; assembly of results and reporting.

Relevant ITER conditions were defined, both in terms of plasma scenarios and interfacing of a spectrometer to the ITER machine, in order to determine the available neutron flux. Here we build on previous modeling work, where neutron flux calculations were done for different plasma scenarios at the ITER reference HRNS position in port cell 1, about 12m from the plasma centre. For a 500 MW ITER plasma (ITER standard H-mode scenario S2) and a first wall aperture radius RFW = 50 mm, the direct flux of 14-MeV neutrons at the spectrometer location is estimated to about 1·109 n/cm2/s. The flux scales linearly with fusion power, with first wall aperture area (or, equivalently, quadratically with RFW) and with spectrometer active area. Since the performance of some techniques is limited by the available neutron flux, inter-facing restrictions can limit their operational range (in fusion power, Pfus). The work concerned a single line-of-sight instrument installed in the ITER reference position, with the implication that all measurements will be line-integrated over the part of the plasma covered by the instrument’s viewing cone. Consequently, unless profile corrections are imposed, plasma parameters estimated from such data will be “effective” values, integrated over the viewing cone. There are a large number of plasma measurements to which neutron spectroscopy can contribute, such as neutron flux and emissivity, plasma rotation, fuel ion 5 E. Ronchi et al, Nucl. Instr. And Meth. A(2010); doi:10.1016/j.nima.2009.12.023


85

ratio in the core, ion temperature profile in the core, neutron fluence on first wall and thermal fraction in Q. The requirements on measurement accuracy differ for different measurements; for example, the ion temperature in the core is required with 10% accuracy at a time resolu-tion of 100 ms and a space resolution of a/10. The aim was to find an HRNS instrument for ITER that can contribute results on as many of these measurements as possible. HRNS techniques and instrumental response functions. We selected the following five techniques for evaluation: thin-foil proton recoil instruments of magnetic (MPR) and non-magnetic (TPR) types, a traditional doubled-scattering time-of-flight (ToF) system, chemical vapor deposit diamond (CVDD) and liquid scintillator (NE213). The latter two are examples of “in-beam” compact neutron spectrometers, while the others are examples of the more elaborate, designed systems in use in fusion research. Interfacing affects these techniques quite differently: the thin-foil techniques are of relatively low efficiency and are thus limited by the available neutron flux. To achieve reasonable energy resolution (FWHM/E !5%), conversion foil size and thickness must be kept small, making the interfacing in terms of first wall aperture the main limiting factor; for these tech-niques count rates of several MHz are possible. CVDD and NE213, on the other hand, are limited by count rate, where a maximum total rate in the detector of 1 MHz has been assumed here. This is a quite ambitious rate limit by today’s spectroscopy standards. For the ToF tech-nique, the presence of random coincidences constitutes a special problem at high count rates although advances in signal processing electronics could improve the rate capability consider-ably. The IRFs used in this work were structured as efficiency matrices, !(En, "), where " is the parameter measured by the specific HRNS technique (detector channel, tTOF, pulse-height etc.) and ! is the efficiency per incoming neutron to get a count in parameter bin " for neutron energy En. For the MPR technique the IRF from the present installation at JET was used. This IRF is based on careful characterization of the real instrument and it has been thoroughly veri-fied, tested and used with real experimental data. The IRF for the TPR was taken from the special study conducted in 2008. For the ToF technique an IRF given by a specially designed MCNPX model was used. Based on the experience with the TOFOR instrument at JET, a count-rate limit (due to random coincidences) of 170 kHz was imposed. For the NE213 we used a IRF based on careful measurements and modeling done at the Physikalisch-Technische Bundesanstalt (Braunschwieg, Germany). Considerable work was invested to construct a best-practice IRF for CVDD, which combined experimental data with a model of the main physi-cal processes. The simulation environment We set up an integrated simulation environment to carry out the calculations using the IRF selected above. The script performs the following tasks:

1. Read in the IRF for the HRNS technique to study. 2. Read in the parameters that determine the energy distribution of the neutron emission:

these are typically the bulk thermal temperature, Ti, and the intensities of the different emission components, ITH, INB, IRF (for thermal, neutral beam, radio-frequency, respec-tively). Construct the neutron emission intensity energy vector.

3. Perform the matrix multiplication between neutron energy vector and the IRF to obtain an ideal spectrometer response vector (binned in measured parameter ") for the selected HRNS technique.


86

4. Chose the time resolution, which gives number of counts (counting statistics) based on instrument efficiency. Scale the ideal spectrometer response vector to obtain the required total number of counts.

5. Use the value of each bin in the ideal response as the average in a Poisson distribution and sample a random integer from that distribution as the synthetic measurement data point.

6. Using a component-based forward analysis method based on minimum Cash statistic, determine a number of relevant plasma parameters, such as Ti, ITH, INB, IRF etc. and store the result.

7. Repeat steps 5-6 a number of pre-determined times to obtain a reasonably large set of extracted parameter results; normally 100 to 1000 iterations are performed.

8. Perform a statistical analysis on the set of results to obtain the accuracy and precision in determining the sought-after parameter(s). Store results.

9. Go back to step 4 to chose a new time base (level of statistics). 10. Go back to step 2 to select a new neutron energy distribution. 11. Go back to 1 to select a new HRNS technique.

HRNS technique evaluation The simulation environment was used to evaluate the five selected HRNS techniques. Based on the modeling of the standard ITER scenario S2, the available neutron flux at the HRNS position was determined. Relevant emission conditions for pure thermal, NBI heated and RF heated scenarios were assembled from pre-calculated emission components, where the ther-mal temperature Ti and the intensities of the bulk, NB-bulk and RF-bulk components were varied in a range relevant for ITER operations. For example, the thermal temperature was varied in 2 keV steps over 1 < Ti < 40 keV, while the NB and RF intensity levels were only tested for 4 and 3 different values each, namely INB/Itot = 1, 5, 10 and 50%, and IRF/ Itot ! 0.01, 0.1 and 0.5% (corresponding to 3, 1 and 0% 3He concentration, respectively). In all studied scenarios it was assumed that a 10% precision (relative standard deviation) in the determina-tion of the parameter was required; this is indeed the present ITER requirement for the de-termination of the fuel ion temperature, on a time resolution of 100 ms. The results of the study are multi-dimensional matrices of precision and accuracy as function of model parame-ter and time resolution (Fig. 3.1-19). The accuracy was in practically all cases found to be within the limits set by the statistics and we therefore only report the result on the precision as an overall indicator of the uncertainty. It should also be pointed out that the study was mainly aimed at comparing the relative merits of the techniques. Since the synthetic data used did not include effects due to background, noise, scattered neutrons, pile-up etc. the result presented should be considered as “best case” estimates.


87

Fig. 3.1-19 (a) Precision (1!) in the determination of the effective ion temperature as a function of effective plasma ion temperature and time resolution for the MPR technique. The red line indicates the 10% precision region. (b) Time resolution for 10% precision as a function of temperature of a purely thermal DT plasma for TPR (black), MPR (red), TOF (turquoise), NE213 (green) and diamond (blue). Purely thermal (Ti=20 keV) 500 MW scenario.

In the case of purely thermal plasma conditions, all techniques fulfilled the ITER require-ments (! " 10%, #t " 100ms) for Ti > 17 keV, which was the lower limit for the compact techniques. For a thermal ITER plasma with Ti = 20 keV and Pfus = 500 MW, the data collec-tion time to reach ! = 10% is 3, 5, 20, 70, 80 ms for TPR, MPR, ToF, NE213 and CVDD respectively (Fig. 3.1-19). The shorter the time, the larger is the dynamical range over which the ITER requirement can be fulfilled; for example, the TPR and MPR techniques can still reach the requirement for a 20 times lower Pfus, i.e., at 25 MW. This picture is further corro-borated by the results from the determination of the non-thermal contributions to the neutron emission, where the MPR and TPR techniques again offer the best performance. The general conclusion of the study was that the thin-foil techniques in almost all cases out-perform the other techniques. Determination of nT/nD ratio using neutron spectroscopy ITER requires the measurement of the tritium to deuterium fuel ion ratio, nT/nD with a preci-sion of 20% over the parameter range 0.01< nT/nD < 10 with a time resolution of 100 ms. Pre-vious work has shown that neutron spectroscopy is potentially capable of providing this by measuring the relative strengths of the DD (2.5 MeV) and DT (14 MeV) neutron peaks in mixed-fuel operations. In this study the region where this measurement can be successfully performed is assessed using a simple plasma and neutron scattering model combined with realistic detector simulations. The plasma is approximated by a toroidally symmetric neutron source with a rectangular shape in the poloidal plane. The neutron emissivity is constant across the source, implying a constant ion temperature, Ti, and a purely thermal plasma is assumed. The volume of the source was chosen so that at Ti = 20 keV, the total fusion power was 350 MW. Neutron trans-port calculations were carried out to determine the shape of the neutron spectrum observed at the proposed ITER detector location for both DD and DT neutrons. Figure 3.1-20 shows typical neutron spectra for three different nT/nD ratios, using a realistic material composition of the ITER wall. Backscattered DT neutrons clearly impact the visibility of the DD peak for high nT/nD ratios.


88

Fig. 3.1-20 Graph showing the magnitude of expected neutron fluxes at the reference ITER HRNS position for different fuel ion mixtures. The 1% DD component corresponds to an nT/nD ratio of approximately 0.3. For higher nT/nD ratios the DD peak becomes more difficult to determine.

By varying nT/nD and Ti the emissivity of the DD and DT components can be determined and hence the ideal neutron spectrum on the detector. By using the detector response function for candidate detector systems, the ideal detected signal on the spectrometer can be determined. Applying Poisson errors to this generates a possible measured signal. The DD and DT com-ponents are then determined from this signal as would be done in an actual detector system. From this the measured nT/nD ratio is determined. Typically 2000 such measured signals are generated for one set of nT/nD and Ti and used to determine the precision and accuracy of the determined nT/nD. For the results discussed here the detector response function used is of the MPR type, operat-ing at the ideal working point for determination of DD signals. The active area of the neutron spectrometer is 10 cm2. The aperture in the ITER first wall of ITER is 10-30 cm in diameter. In the step for determining the DD and DT components, a fitting function minimizing the Cash statistic was used. The resulting average measured nT/nD is very close to the modeled value even for very low numbers of detected neutrons (high accuracy). The precision of the measurement is determined from the sigma of the distribution of measured nT/nD relative to the average value of the determined nT/nD. Figure 3.1-21a shows the 2D map of this quantity as a function of Ti and nT/nD of the modeled plasma for a 10 cm diameter first wall aperture. A line is drawn on the figure to illustrate where the boundary of 20% precision lies (the ITER requirement). Figure 3.1-21b shows the same quantity for the 30 cm diameter first wall aperture.


89

Fig. 3.1-21 a) 2D plot showing the estimated precision of the determination of nT/nD ratio for a 10 cm diameter ITER first wall aperture as measured with an MPR type spectrometer at the reference HRNS position. The series of crosses denotes the 20% precision boundary as required by ITER (requirement fulfilled above and to the left). b) Similar to figure a but for a 30 cm diameter first wall aperture. The nine fold increase in flux at the spectrometer allows a 3 fold increase in the maximum nT/nD measurable with 20% precision.

The MPR type spectrometer can satisfy the ITER requirements over part of the operating space. Two limits arise from the calculation. At low Ti (<6 keV) insufficient neutrons are detected to give a spectrum analyzable in the required time interval. At high nT/nD the back-ground from DT neutrons drowns out the DD signal. A larger first wall aperture results in higher neutron fluxes at the detector and pushes the 20% precision boundary higher in nT/nD and lower in Ti. One other spectrometer type has been considered, a Time of Flight (ToF) system. Having a higher detection efficiency and similar energy resolution it is possible to access higher nT/nD ratios with this technique. However the system would be operating at an extremely high rate (100s of MHz) which leads to problems with system stability and uncor-related (random) coincidences. These would need to be better assessed to confirm the possible use of the ToF detection system. Two effects contribute to improve the prospects for using neutron spectroscopy in determining the nT/nD. First, the use of 1 MeV D beams will add a significant contribution of mainly DD but also some DT neutrons to the spectrum. Secondly, at high nT/nD a third neutron component starts to become significant, namely TT neutrons. These generate a peak from t+t ! 5He+n at 8.6 MeV and a broad peak from 0 to 8 MeV from 3-body (t+t ! "+n+n) phase space. Using the TT 8.6 MeV peak would allow access to the high nT/nD region but needs to be investigated further. The possibilities offered by the 1 MeV D beam will have to be assessed in future work. Using an MPR type detector, the nT/nD ratio can be determined in certain operational regions as shown in this study. Further work needs to be done on incorporating more realistic emis-sivity profiles, determination of line integration effects, effect of errors in certain assumed quantities and the possibilities offered by including TT and NBI components in the analysis.


90

3.1 Contracts and collaborations Contracts Table 1 Neutron-based contracts within the Work Program 2008-2009 of EFDA’s Topical Groups for Diagnostics (DIA) and the cross-disciplinary grouping “TGS”.

Contract num-

ber: EFDA WP08-09-

Title Baseline support

(20%)

Priority support

(40%)

Hardware support

(40%)

VR resp. officer

TGS-01a-01 Neutron spectro-meter for fast particle confine-ment

1.5 ppy 0.33 ppy 26 kEUR* Marco Cecconello, UU

DIA-01-05 Neutron Spectros-copy as Fuel ion diagn.

0.4 ppy Sean Conroy, UU

DIA-01-06 Neutron-based diagnostics

1.5 ppy 0.33 ppy Göran Ericsson, UU **

* UU share of a total 56 kEUR within the contract ** Task coordination Sean Conroy from the UU group took up a position as long-term (4 year) secondee to JET-JOC on August 3, 2009. Collaborations Below is a list of the persons involved in the collaborations that have been described in the diagnostic project areas above: University Bicocca, Milano, Italy: G.Gorini, M.Tardocchi, M.Nocente, F.Ognissanto ENEA-Frascati, Rome, Italy: B.Esposito, L.Petrizzi, D.Marocco IST University, Lisboa, Portugal: J.Sousa, A.Combo, N.Cruz and R.Costa Pereira CSU-JET, UK: A.Murari, T.Johnson, J.Ongena, M.Laxåback, M.Tsalas, E.Asp UKAEA-JET, UK: S.Popovichev, V.Kiptily, S.Sharapov, S.Pinches UKAEA-MAST, UK: R.Akers, M.Turnianski, and the MAST team Institute for Theoretical Physics, University of Innsbruck, Austria: V.Yavorskij Risö National Lab., Denmark: Sören Korsholm IPPLM, Poland: M.Schulz, Rafal Prokopowicz Publications section 3.1 Peer reviewed journals

1. Andersson Sunden E, Sjöstrand H, Conroy S, Ericsson G, Gatu Johnson M, et al. The thin-foil magnetic proton recoil neutron spectrometer MPRu at JET, Nuclear Instruments and Methods in Physics Research Section A610 (2009) 682

2. Baranov F, Jenkins I, Alper B, Challis D, Conroy S, Kiptily V, et al. (incl. M.Cecconello), Anomalous and classical neutral beam fast ion diffusion on JET, Plasma Physics and Controlled Fusion 51 (2009) 044004

3. van Eester D, Lerche E, Andrew Y, Biewer M, Casati A, Crombe K, et al. (incl S.Conroy), JET 3He–D scenarios relying on RF heating : survey of selected recent experiments, Plasma Physics and Controlled Fusion 51(4) (2009) 044007


91

4. Ericsson G, Pomp S, Sjöstrand H, Traneus E., Comment on "Piezonuclear decay of thorium" [Phys. Lett. A 373 (2009) 1956], Physics Letters A 373(41) (2009) 3795

5. Frassinetti L, Brunsell R, Cecconello M, Drake R., Heat transport modelling in EXTRAP T2R, Nuclear Fusion 49 (2009) 025002

6. Kiptily V, von Thun P, Pinches D, Sharapov E, Borba D, Cecil E, et al. (incl M.Cecconello), Recent progress in fast ion studies on JET , Nuclear Fusion 49 (2009) 065030

7. Krasilnikov V et al. (incl. G.Ericsson, C.Hellesen, A.Hjalmarsson), Fundamental ion cyclotron resonance heating of JET deuterium plasmas, Plasma Physics and Controlled Fusion 51 (2009) 044005

8. Ronchi E, Söderström P, Nyberg J, Andersson Sunden E, Conroy S, Ericsson G, et al., An artificial neural network based neutron-gamma discrimination and pile-up rejection framework for the BC-501 liquid scintillation detector, Nuclear Instruments and Methods in Physics Research Section A610 (2009) 534

9. Ronchi E, Andersson Sunden E, Conroy S, Ericsson G, Gatu Johnson M, Kallne J, et al., A bipolar LED drive technique for high performance, stability and power in the nanosecond time scale, Nuclear Instruments and Methods in Physics Research Section A599 (2009) 243

10. Sjöstrand H, Andersson Sundén E, Conroy S, Ericsson G, Gatu Johnson M, et al., Gain stabilization control system of the upgraded magnetic proton recoil neutron spectrometer at JET, Review of Scientific Instruments 80(6) (2009) 063505

11. Testa D, Cecconello M, Schlatter C, et al., The dependence of the proton+triton thermo-nuclear fusion reaction rate on the temperature and total energy content of the high-energy proton distribution function, Nucl. Fusion 49 (2009) 062004


1. Gatu Johnson M, Cecconello et al., Cross-validation of JET fast deuterium results from TOFOR and NPA , 36th EPS Conference on Plasma Physics, Sofia, June 29 - July 3, 2009 ECA Vol.33E. 2009. p. P-2.151-.

Contributions to ITPA meetings, JET and MAST TF meetings:

1. M.Cecconello, S. Conroy, G. Ericsson, M. Weiszflog, R. Akers and M. Turnyanskiy, “Proof of Principle Neutron Camera for MAST”, MAST presentation Jan. 28, 2009

2. M.Gatu Johnson et al., “Neutron backscatter study with TOFOR”, JET TF-D meeting, March 5, 2009

3. C.Hellesen et al., “Measurements of fast ions and their interactions with MHD activity using neutron emission spectroscopy”, JET TF-D meeting, Sept. 17, 2009

4. H.Sjöstrand, E.Andersson Sundén et al., “Neutron calibration at JET: a comparison between KN1, KN3 and KM9”, JET TF-D meeting, Sept. 24, 2009

5. G.Ericsson et al., “Last results from the TOFOR Neutron Spectrometer”, JET TF-D meeting, Dec. 4, 2009

6. M.Cecconello, S. Conroy, G. Ericsson, M. Weiszflog, S. Sangaroon, “Study of a neutron spectrometer for fast particle confinement: Installation and commissioning of a neutron spectrometer at MAST”, MAST presentation Dec. 14, 2009

Ph.D. and Ph.Lic. thesis:

• Emanuele Ronchi, Neural networks applications and electronics development for nuclear fusion neutron diagnostics, Ph.D. thesis, Digital Comprehensive summaries of Uppsala dissertations form the faculty of Science and technology 673, 2009


92

Bachelor’s and Master’s theses • Henrik Hellberg, Simulations of neutral particles in a neutral beam heated fusion

plasma, Master of Engineering Physics thesis, Uppsala University 2009 • Björn Molander, Methods and practice of characterization of liquid scintillator

detectors, Master of Engineering Physics thesis, Uppsala University 2009 • Per Kårén, Ion distribution box models for the NBI heating system at JET using

neutron emission spectroscopy, Master of Engineering Physics thesis, Uppsala University 2009

• Two Theses for Master of Science (J.Eriksson, M.Skiba) were initiated during 2009 and will be reported in 2010.

3.2 Plasma Spectroscopy E Rachlew, Y Ding (MSc student finished Jan 2009), S Ehlers (MSc finished June 2009), I Pusztai (PhD student)

Introduction The plasma spectroscopy programme is based at the Department of Physics, School of Engineering Science, KTH with the participation of both graduate and undergraduate students and research visitors. The research programme contains direct experiments at various fusion plasmas (mainly at JET in 2009) as well as basic experiments for achieving atomic and molecular data for input to the modeling of the radiation from the plasma. In addition beam emission spectroscopy (BES) diagnostic development work has been carried out at CTH. The spectroscopy group has contributed to the JET-EFDA work programme in all the campaigns during the year. The group has had active international collaborations within the EFDA programme with the following:

• UKAEA MAST project • CEA Tore Supra project • IPP Prague Compass project • Jülich TEXTOR project • And also with the MIT Plasma Fusion Center Alcator C-Mod group.

Modelling of the divertor radiation from the ToreSupra tokamak using ADAS The total radiation from the ToreSupra tokamak plasma has been modeled using ADAS data and the Onion Skin Collisional Radiative model and compared with measurements from the bolometer at ToreSupra. The main results can be summarized as

• With finite particle confinement time, the general trend in comparison with ionization

balance with infinite tp is that the ionization stages are pushed to higher energy regions. Zeff and Prad vary monotonically with particle confinement time; the longer tp results in larger Zeff and weaker Prad.

• Charge exchange recombination with neutral hydrogen suppresses highly charged ionization stages, and dominates over other recombination processes (radiative, dielectronic and three body recombination) at lower ionization stages like O2+ O3+ O4+ C1+ C2+ C3+ etc at high Te region. Zeff ecreases gradually with increasing fraction of


93

neutral hydrogen. Prad peaks with 5% of ne for neutral hydrogen concentration, at which stage strongly radiative ions i.e. O4+, C2+ etc. reach the highest fraction.

• The uncertainties of the edge temperature can cause maximum uncertainty of about 20% in Prad and 5% in Zeff in the region of 10eV to 100 eV.

X-ray spectroscopy from the Alcator C-Mod tokamak Spectra of hydrogenlike argon (Ar17+) have been obtained from Alcator C-Mod plasmas using a spatially imaging high resolution x-ray spectrometer system. The ratio of the Ly alfa2 to Lyman alfa1 (1S1/2 – 2P1/2,3/2) was found to be ca 0.59 regardless of plasma parameters which is considerably greater than the ratio of statistical weights, !. This difference has been modeled and found to be due to the effects of electron and ion collisional excitation of fine structure sub-levels. The experimental results have been compared with calculations from COLRAD, a collisional radiative model and the results agree well. Spectroscopy R&D The group is partner in the ADAS consortium for development of data and analysis programs for the modeling of spectroscopic data from the hot plasmas. Some of the atomic and molecular data is still lacking and the group is active in both new experiments (using the synchrotron radiation) and new modelling of the emitted radiation. Correction of alkali beam emission spectroscopy measurements In a beam emission spectroscopy (BES) density profile measurement, if the line of sight is far from tangential to the flux surfaces, and the beam width is comparable to the scale length on which the light profile varies, the observation might cause an undesired smoothing of the light profile, resulting in a non-negligible underestimation of the calculated density profile. We investigated the characteristics and the magnitude of this error, using the comprehensive alkali BES simulation code RENATE, and introduced a deconvolution-based correction method which we demonstrated in simulated BES measurements on the COMPASS and TEXTOR tokamaks. The error can be significantly reduced by our method which calculates the emissivity along the beam axis from a measured light profile, taking the finite beam width and the properties of the measurement into account in terms of the transfer function of the observation.

Fig. 3.2-1 Demonstration of the emission reconstruction method on simulated Lithium BES measurement on the COMPASS tokamak (out-borad midplane injection, observed from the in-board side top port). [a] light profiles, [b] density profiles, [c] differences from the ideal light profile and [d] differences from the original density profiles. Using the method reduces the error to the level of accuracy of the density calculation method.

a

c

b

d


94

EFDA JET highlights Experiments at JET-EFDA

Fig. 3.2-2 The left figure shows the diagnostic set-up of the motional Stark effect diagnostic at JET and the right figure shows the measured and derived q-profiles with an early reversal (from R De Angelis et al, Nucl. Instrum. Meth. in Phys. Res. A, 2010)

The measurement of the safety factor q in tokamaks, which describes the winding of the helical magnetic field lines, is very important especially for the achievement of advanced scenarios. The motional Stark effect diagnostic can provide a direct measurement of the magnetic field orientation but the derivation of the q-profile requires a simulation of the magnetic equilibrium taking into account input from several other diagnostics. This analysis can be affected by large errors. In order to validate the results, q-profiles are compared with the radii of MHD modes, which can be attributed to surfaces of known q. The figure shows, left, schematic of the setup of the MSE diagnostics and, right, derived q-profiles with a reversal at an early time. For these reversed profiles, uncertainties in q values near the axis are higher than those for monotonic profiles, although in some cases they are seen to correspond with experimental data from X-ray tomography and with the observation of Alfvén cascades.

Fig. 3.2-3 Long-lived rotating current filament observed in a magnetically confined plasma at the JET tokamak. (Illustration by Paul Carma). The figure shows q=4 spinning current ribbon that provides the best fit to magnetic signals from Mirnov coil set, with 100-300 A)- From E R Solano et al, Phys. Rev. Lett.104 (2010)185003


95

Identification of a localized current structure inside the JET plasma has been reported based on measurements and coordinations of many diagnostics, including the motional Stark effect spectroscopy. The structure is a field-aligned closed helical ribbon which carries current in the same direction as the background current profile (co-current), it rotates toroidally with the ion velocity (co-rotating). The structure appears spontaneously in low density, high rotation plasmas, and can last up to 1.4s, a time comparable to a local resistive time. It considerably delays the appearance of the first edge localized mode. Publications section 3.2 Peer reviewed journals

1. A Ekedahl, K Rantamäki, M Goniche et al incl E Rachlew:"Effect of gas injection during LH wave coupling at ITER-relevant plasma-wall distances in JET", Plasma Phys. Control. Fusion 51, 044001 (2009)

2. V Riccardo, G Arnoux, P Beaumont et al incl E Rachlew: "Progress in understanding halo current at JET", Nuclear fusion 49,055012 (2009)

3. A Kivimäki, G Vall-llosera, M Coreno, M A Huels, M Stankiewicz and E Rachlew: Fluorescence emission at core-to-Rydberg excitations in the N2 molecule, J.Phys.B: At.Mol.Opt.Phys. 42,185103 (2009)

4. A Kivimäki, G Vall-losera, M Coreno, M A Huels, M Stankiewicz and E. Rachlew:"Line shape narrowing in the ultraviolet yield at the N 1s- !* resonance of the N2 molecule", J. Phys. B: At. Mol. Opt. Phys. 42, 075102 (2009)

5. E R Solano, P J Lomas, B Alper et al incl E Rachlew:"Observation of confined current ribbon in JET plasmas", Phys. Rev. Lett. 104 2010, 185003

6. R De Angelis, M Baruzzo, P Buratti, B Alper, L Barrera, A Botrugno, M Brix, L Figini, A Fonseca, C Giroud, N Hawkes, D Howell, E De La Luna, F Orsitto, V Pericoli, E Rachlew, O Tudisco, JET- EFDA Contributors, Localisation of MHD modes and consistency with q-profiles in JET, Nucl. Instrum. Meth. Phys. Res. A (2010)

7. J E Rice, M L Reinke, J M A Ashbourn, A C Ince-Cushman, Y A Podpaly, M F Gu, M Bitter, K Hill and E Rachlew, "The Ar17+ Ly2/Ly1 ratio in Alcator C-Mod plasmas", submitted P R A, 2009

8. Pusztai I, Pokol G, Dunai D, Réfy D, Pór G, Anda G, Zoletnik S and Schweinzer J, Deconvolution-based correction of alkali beam emission spectroscopy density profile measurements, Rev. Sci. Instrum. 80, 083502 (2009).

Contributions at international conferences and other workshops and conferences

1. J Hobirk, F Imbeaux, F Crisanti et al incl E Rachlew:" Improved confinement in JET hybrid discharges", 36th EPS conference on plasma physics, Sofia, Bulgaria , June 28 - July 3, 2009

2. K Crombe, Y Andrew, T M Biewer et al, incl E Rachlew:" Influence of rotational shear on triggering and sustainment of internal transport barriers on JET." 36th EPS conference on plasma physics, Sofia, Bulgaria , June 28 - July 3, 2009

3. E R Solano, P J Lomas, B Alper et al incl. E Rachlew :" Study of high temperature pedestals in hot ion H-mode in JET",36th EPS conference on plasma physics, Sofia, Bulgaria , June 28 - July 3, 2009

4. R. De Angelis, F. Orsitto, M. Brix, N. Hawkes, E. Rachlew, A. Botrugno, P. Buratti, A. Fonseca, D. Howell, V. Pericoli, O. Tudisco , JET-EFDA:"Current profile


96

measurements in JET Advanced Tokamak scenarios",36th EPS conference on plasma physics, Sofia, Bulgaria , June 28 - July 3, 2009

5. J.E. Rice, M.L. Reinke, J.M.A. Ashbourn, A.C. Ince-Cushman, M. Bitter and E. Rachlew:" The Ar 17+ Ly-alpha ratio in Alcator C-Mod plasmas",17th topical conference High Temperature Plasma Diagnostics, MonteRey, Calif., US, May 2009

6. G Vall-llosera, R Sankari, S Sarabiour, E Rachlew, E Kukk and M A Huels: Radiation damage to genetic sugars: why nature chose DNA, 55th Annual meeting of the Radiation Research Society, Oct. 4-7, 2009, Savannah, Georgia, US

7. G Vall-llosera, S Lacombe, R Panajotovic, S Ptasinska, S Sarabipour, F Hennies, E Rachlew, M A Huels: The morphology and degradation of adenine films, 55th Annual meeting of the Radiation Research Society, Oct. 4-7, 2009, Savannah, Georgia, US

8. A.Kivimäki, G. Vall-llosera, M Coreno, M A Huels, M Stankiewicz and E Rachlew: Ultraviolet fluorescence yields at the 1s - !* resonance of N2: the observation of line shape narrowing effect, 14th International conference on X-ray absorption fine structure, XAFS-14, Cameron, Italy, July 26-31, 2009

9. Pusztai I, Pokol G, Dunai D, Réfy D, Pór G, Anda G, Zoletnik S and Schweinzer J, Deconvolution-based correction of alkali beam emission spectroscopy (BES) density profile measurements, RUSA Meeting of the Swedish Fusion Research Unit, Uppsala, 2009.

MSC thesis

1. Sandra Ehlers:"Temporal drifts of the magnetic field in superconductive undulators", 2009

2. Yuan Ding: "Modelling of the radiative power loss from the plasma of the ToreSupra tokamak", 2009


97

4 Concept improvements 4.1 Computational methods with applications to the RFP J. Scheffel, A. Mirza (PhD student) In collaboration with: D.D. Schnack, Univ. Wisconsin-Madison, USA Introduction In spite of recent year’s improved understanding of reversed-field pinch confinement, theoretical and numerical models need further refinement. Because of the complexity of the strongly nonlinear MHD phenomena and the strongly separated Alfvén and resistive time scales, relatively limited physical effects have so far been included in the numerical computations. As a result, there are as yet only approximate results on the operational limits of confinement in the RFP. The fusion potential of the RFP certainly is of significant interest, since an RFP reactor would feature high plasma beta (resulting in compact physical dimensions), using little or no beam or radio frequency heating and using normal magnetic coils, resulting in low capital costs and low cost of produced electricity. Our numerical simulations of RFP plasmas in the advanced regime, where current driven tearing modes were eliminated by current profile control, have provided favourable scalings of on-axis temperature and poloidal beta. To better understand and possibly find modifying effects to the limited energy confinement scaling, improved modelling is required. In the project, we extend our earlier work to more precise numerical studies of the effect of pressure driven modes in the high beta scenarios. It has been claimed that resistive-g modes are eliminated when heat conduction is accounted for in the energy equation. We test this conjecture, formulated in traditional delta-prime theory, by developing a GWRM (see below) model, which solves the resistive MHD equations in the entire plasma domain. In particular, it is essential to determine the effect at higher, reactor relevant beta values. Studies of two-fluid effects, in particular the Hall and diamagnetic contributions to Ohm’s law will be subsequently included. The effect of Larmor radius on the resistive g-modes in the RFP will, as a next step, addressed using the Vlasov-fluid model. The Generalized weighted residual method (GWRM) We have developed a new computational tool for initial-value problems that produce semi-analytical explicit solutions. Employing spectral residual methods, now generalized to represent the time dimension spectrally, solutions to initial-value partial differential equations are determined as finite, approximate Chebyshev polynomial expansions. The solutions thus represent not only space but also time and physical parameters explicitly. Time stepping is avoided completely. This eliminates numerical stability restrictions on the time domain and problems with vastly separated time scales can thus be efficiently solved. The efficiency of the Generalized weighted residual method (GWRM) is further improved through an optimized use of temporal and spatial subdomains. The project combines two vital areas of research in fusion plasma physics and computational physics.


98

A number of benchmarking tests of the GWRM has been carried out. For a set of pde’s with wavelike solutions on two separated time scales, the solution traces the slower dynamics using less computational time than both the explicit Lax-Wendroff and the implicit Crank-Nicholson schemes. We have further shown that the GWRM provides accurate solution of the nonlinear Burger equation, which features shock-like structure for weak dissipation. Also the temporal solution of the linearized, Fourier-decomposed, ideal and resistive (compressible) MHD equations for a cylinder has been computed, confirming the applicability of the GWRM to systems of MHD equations. Solver for systems of nonlinear algebraic equations (SIR) For solving the systems of algebraic equations that represent the initial-value pde’s in the GWRM, a globally convergent and highly efficient root solver has been developed. The Semi-implicit root solver (SIR) is shown to be efficient, robust and simple in comparison with the standard Newton method using line search and completely avoids landing on spurious roots. The method was soon recognized by numericists and the authors of the publication [1] were awarded IMACS Honor Members 2010 due to the paper being one of the most successful (most downloaded) IMACS papers in 2009 Publications section 4.1 Peer reviewed journals

1. J. Scheffel and C. Håkansson, Solution of Systems of Non-linear Equations – a Semi-implicit Approach, Appl. Numer. Math. 59(2009)2430.


2. J. Scheffel and A. Mirza, Resistive g-modes and RFP confinement, 51st Annual Meeting of the APS Division of Plasma Physics, Atlanta, Georgia, USA 2-6 November, 2009.


99

5 Emerging technology 5.1 High purity ODS-tungsten materials M. Muhammed, S. Wahlberg

Tungsten as a first wall material in a reactor. Develop method for fabrication of highpurity ODS-tungsten based materials

Nanostructured materials can advantageously be fabricated using powder metallurgical methods, i.e., via sintering of nano-scale powders at relatively low temperatures to limit diffusion and grain growth. The optimization of powder fabrication methods and the characteristics of the nano powder are essential in order to maintain the nanosize in the finally processed materials. In general, a precursor powder is prepared (containing the required elemental composition in the final products) to be further processed to the required product materials. In the different processing steps, it is advantageous to have powders with a minimum of particle agglomeration, high purity and with a uniform (homogenous) composition. In this project we developed specific chemical powder metallurgical methods for the fabrication of nanoscale tungsten based powders. The processing is carried out in two steps: i) Fabrication of doped tungsten containing precursors by a solvent mediated process; and ii) Thermal decomposition and reduction of the obtained precursor into an tungsten alloy powder. In the first part of the study, we used commercially available ammonium paratungstate (APT) as a starting material that was allowed to react with salts of lanthanum or yttrium under controlled conditions. The produced precursor was then calcined and reduced. In this method we were able to prepare powders with controlled bulk composition. However, the produced materials was not highly homogeneous when analyzes on the nanometre scale. The particle size of the precursor was 1-20 !m. The alloyed tungsten powder obtained after reduction was nanostructured (50-200 nm), but the particles were agglomerated. The consolidation was carried out by sintering using spark plasma technique (SPS) which produced consolidated materials with grain size down to 0.5 !m, but the distribution of the yttrium or lanthanum oxides was not homogeneously distributed between the tungsten grains when examined by high resolution TEM. We therefore have undertaken a detailed study to identify the reason for the inhomogeneous distribution of the doping oxides: We could conclude that, the inhomogeneity of the distribution of the doping elements into the W-based matrix is determined by the chemical reactions taking place between the Y and La during the formation of the precursor. The reaction between APT and the salt of Y or La was found to take place at the surface of the APT grains and less with the interior of the grains (APT grains are of 20-50 !m in size at the start of the reaction). Recently, we have developed a second method for precursor fabrication also starting from APT and salts of La or Y, but carried out under acidic condition. The process involves a complete phase transformation in which the APT solid is converted into a tungstic acid based material. By control of the particle growth in solution, we are able to fabricate a nano-sized


100

(<50nm) powder precursor, which can be reduced under hydrogen to a nanostructured (<50 nm) tungsten alloy powder. The particle size is during the complete powder fabrication process reduced about 1000 times from the starting material, APT, to the final tungsten alloy powder. The uniformity is improved and the degree of particle agglomeration dramatically decreased, as compared to the previous method.


101

6 EFDA JET The Swedish Association is very heavily involved in the JET activity. Aspects of the scientific programme that support JET have been included in the earlier sections of this Annual Report covering the Work Programme of the Association. Here the participation in EFDA JET Orders and the JET Campaigns C26-C27 are summarized in table form. 6.1 EFDA JET S/T Orders. Involvement of the RU in the EFDA JET Orders is summarized in Table 6.1.

Table 6.1 Swedish participation in JET Orders.

Task agreement Order Title Responsible JW6-TA-EP2-EDP Enhancement Project

JW6-OEP-VR-26 Erosion deposition diagnostics for ITER-like wall project

KTH M. Rubel

JW6-TA-EP2-SIW-04 24 July 2009 version Enhancement Project

JW6-OEP-VR-21 JW7-OEP-VR-28 JW8-OEP-VR-30

Spectroscopy in support of ITER-like wall project

KTH T. Elevant

JW9-TA-EX-ITC-05 ITC task

JW9-O-VR-33 Integration of transport and MHD codes at JET

CTH P. Strand

JW8-TA-EXP-04 S/T Orders Campaign

JW8-O-VR-29B

JET Campaigns C20-C27

6.2 EFDA JET Enhancements. Erosion deposition diagnostics for ITER-like wall project M. Rubel, D. Ivanova Study of material migration, i.e. local and global erosion and deposition processes, is a very important element of the JET programme. Various types of in-vessel diagnostics have been developed and used over the years. The issue of erosion-deposition will also be in focus during the JET operation with a full metal wall: ITER-Like Wall Project. Therefore, a set of probes has been procured. A set of various monitors, the so-called wall probes, have been developed and installed in JET. These are: (i) wall inserts in the inner wall cladding, (ii) deposition monitors and louver clips in the divertor, (iii) quartz micro-balance devices, (iv) rotated collectors and (v) cassettes with mirrors located. Items (iv) and (v) are located both on the main chamber wall and in the divertor: inner and outer leg and under the load bearing plate. In addition to those monitors, marker tiles have been designed in order to determine beryllium and tungsten migration and fuel inventory. On the main chamber wall the probes are installed in the equatorial position in a few toroidal positions. Fig. 6.2-1 shows schematically the location of diagnostics in the divertor. Images and drawings in Fig. 6.2-2 (a-e) show details of several different diagnostic tools: (a) deposition monitor manufactured by IPP-Garching, (b) wall insert manufactured by IPP-Garching, (c) elements of a louver clip


102

produced vy VR, (d) mirror (produced by VR) for the First Mirror Test at JET for ITER and (e) results of optical pre-characterisation on mirrors. In Fig. 6.2-3 (a-c) there are marker tiles for studies of material erosion: (a-b) beryllium from the main chamber wall and (c) tungsten from the divertor.

VR is responsible for the procurement of mirrors and louver clips and for the coordination of the entire programme related to the wall probes.

Fig. 6.2-2: (a) deposition monitor: location in the divertor, scheme of the cage, steel cover; (b) inner wall cladding tile with an insert coated partly with 40 nm tungsten layer, partly with 3 micrometers of beryllium; (c) jaws of a louver clip.

Fig. 6.2-1. A poloidal cross-section of the JET divertor. The location of erosion -deposition diagnostics is marked with arrows.

a

c b


103

Fig. 6.2-2: (d) molybdenum mirror for the First Mirror Test and (e) reflectivity of several mirrors measured before installation in JET. In Fig. (d) M denotes the mirror surface and L denotes a leg for precise fixing of a mirror inside a carrier.

Beryllium Tile (3 cm)

Nickel interlayer (2-3 µm)

Beryllium marker coating (7-9 µm)

Ni

Be

OCNi

Be

OCNi

Be

OC

Epoxy resin Be

coating

Bulk Be

Ni interlayer

d

e

b

c

Fig. 6.2-3: (a) Schematic drawing a of beryllium marker tile for studies of Be erosion; (b) a cross-section of the Be marker tile; (c) tungsten load bearing tile in the JET divertor: yellow stripes indicate the positions of marker lamellae.

a


104

Spectroscopy in support of ITER-like wall project (SIW) T. Elevant and J.R. Drake Spectroscopy in Support of the ITER-like Wall (SIW) is part of a package of diagnostic enhancements, which is to be implemented and used at JET during operation with an ITER-like wall during Framework Programme 7. Background The primary material choice for ITER is a full beryllium main wall with CFC (Carbon Fibre Composite) at the strike points together with tungsten at divertor baffles and dome. Since this combination never has been tested in a tokamak previously, the ITER-like Wall project has been launched at JET implying replacement of the present first wall by an ITER-like wall together with an adequate set of diagnostics. Until now, the JET wall and divertor has consisted of CFC, while the RF and LH antennas have consisted of Be- and Cu-coated stainless steel respectively, with CFC protection tiles. Small amounts of Be have been introduced by evaporation, and noble gases (4He, 3He, Ne, Ar) have frequently been injected for transport studies, or as minority species for RF-Heating. C has served as the reference species for low Z elements for core and edge emission. The reference species for medium Z elements in the VUV, XUV and X-ray region has been Ni. Installation of the ITER-like wall and a new divertor on JET will result in qualitatively different spectroscopic needs and technical challenges. The metal surfaces can be subjected to large power density and thus high erosion rates. In case the full W divertor option is selected, it is conceivable that C levels will be so low that Be has to act as the reference species for low Z elements. Spectra from W are very rich, and may result in blending with spectral lines from other elements in all wavelength ranges. The spectral line best suited to study W erosion is the one at 401 nm because modelled and experimental results are available to convert photon fluxes into erosion rates. However at this wavelength transmission losses by optical fibred are high, quantum efficiency of detectors is low and absolute calibration more difficult than for the spectral lines that best characterise C. For Be, spectral lines in the normal visible wavelength range exist, but many other suitable lines are in the blue as well. It will also be necessary to quantify the amount of Be released from W surfaces, the amount of W released from Be surfaces, and later in the experimental campaigns the release of Be and W from Be/W alloys once these have formed on the various surfaces. The two elements Be and W and their alloys, are prone to melting; therefore specific spectral lines from several locations need to be provided in real time for possible use in machine protection. Core concentrations of Be and W and their relationship with erosion rates need to be studied in detail to allow a meaningful prediction of the same wall mix on ITER which is presently foreseen to be W/Be/C. Features of the enhanced systems To meet the new demands a number of spectroscopic diagnostic systems are modified and enhanced. All instrumentation that is not installed in the torus hall will be located in a dedicated laboratory, Spectrometer Room (SR) in J1D, with a number of optical fibres routed from the various locations around the torus. Fibre patch panels in the spectrometer room will make the instrumentation interchangeable between different lines of sight should such need arise. The core emission from W is spread from the VUV (10nm-13nm) and XUV (4nm-7nm) to the Xray (0.1nm-0.4nm), with each spectral region being representative of a different part of the plasma, depending on the degree of ionisation and hence electron temperature. This motivates enhancements to the diagnostics KT4, KT7 and KX1 to complement information from KT2 and KS6, which have not been modified since they are already adequate. The


105

diagnosis of wall sources is a multi-dimensional task: high time resolution, good spatial resolution, large spectral coverage with good spectral resolution cannot be achieved simultaneously by one single instrument. The enhancements to KL1, KT1 and KT3 address the need for spatial coverage and resolution for spectral lines characteristic of Be, C, and W with slow time resolution. The enhancements to KS3 provide both wide spectral coverage for more detailed physics studies of erosion processes and ELM resolved time resolution, in specific locations on the inner wall and in the divertor. The diagnostic consists of 10 horizontal and 20 vertical viewing lines, connected to instrumentation via optical fibres. Light from these fibres is split between spectrometers with ~50 ms time resolution, and a set of Polychromator Assembly (PA) units with filters for ELM resolved fast time resolution. The horizontal lines of sight will observe Be limiters. The vertical lines of sight provide a spatial resolution in the divertor of ~30 mm using 10 lines for the inner and 10 for the outer divertor. Light in each optical fibre needs to be split further into more channels than previously, to cover simultaneously BeI, CII, WI and D!. Thus additional light channels with filters and fast PM tubes are required.

Table 6.2-1 Overview of envisaged spatial-, spectral- and time resolution and coverage of the various visible instruments.

Species Location Performance Instrument Be, C, W, D Inner and Outer Wall ELM average KL1 Main Chamber

(Images) Be, C, W, D Inner and Outer Divertor ELM average KL11 Divertor

(Images) All One location (" and #)

Inner Wall (Be) Large spectral coverage, high spectral resolution, ELM average

KS3 (Spectroscopy)

Be, C, W, D One location (" and #) Inner Wall (Be), Divertor profile

Specific spectral lines, high time resolution

KS3 (PM-Tubes)

All One location (" and #) Inner Wall (W)

Large simultaneous spectral coverage, high spectral resolution, ELM average

KS8 (High Resolution Survey Spectrometer)

Be, C, W, D Three locations (") in same Octant. Three locations (#) Inner Wall (W)

Specific spectral lines, seven ELM averaged, one high time resolution

KS8 (Compact Spectrometers)

Be, C, W, D One location ("), Complete profile (#) Inner Wall (Be), Upper Strike Points, Divertor profile

Specific spectral lines, ELM average

KT1-Visible (Scanning)

All Outer divertor profile, plus load bearing septum replacement plate.

Large spectral coverage, high spectral resolution, ELM average

KT3 (Spectroscopy)


106

Role of Fusion Division at KTH The responsibilities of VR are dealt with by the Fusion Division at KTH and include technical specifications, purchasing, installation, test, calibration and commissioning of the following installations. 24 units of four light-beam Polychromator Assemblies (PA) with PM Tubes Briefly, each PA consists of a light fibre input, four beam splitters and four output channels, each one with a band pass filter and a fast PM-tube. Initial tests were made and all 96 beam line geometries were found to be in agreement with specifications after the prototype had been amended. Wavelength throughput were also measured for all channels using a calibrated tungsten-halogen lamp, a 0,6 mm optical fibre for light transmission and an Ocean Optics HR2000 detector.

PA 23

-20%

0%

20%

40%

60%

80%

100%

400 450 500 550 600 650 700 750

Beam 1 /Ref. Beam 2 /Ref.

Beam 3 /Ref. Beam 4 /Ref.

Fig. 6.2-4 Measurement of relative transmission in PA 23 (ex. Prototype). All 96 units were found to satisfy specifications.

Equipment in Spectrometer Room (SR) SR will act as a hub of information that can guide light from different lines of sight in the torus to various instruments located in SR on demand. Mounts and tables, optical components, fibre connectors, beam splitters and patch panels for spectrometer room have therefore been specified, purchased or made, tested and installed to be used jointly by KS3, KS8, KT1/4/7.


107

6.3 JET campaign participation C26-C27 Involvement of the RU in the EFDA JET Experimental Campaigns C26-C27 (EFDA JET Order EFDA CSU- JW8-O-VR-29B is summarized in Table 6.3-1. The scientific information for the activity is reported in the earlier sections covering the Association Work Programme.

Table 6.3-1 Swedish RU involvement in EFDA JET Campaigns C20-C25 Task Force in JET Work Programme Persons seconded to JET Diagnostics G. Ericsson, S. Conroy, M Gatu

Johnson, H. Sjöstrand, M. Weiszflog, C. Hellesen, E. Ronchi, E. Andersson Sunden, E. Rachlew, L. Frassinetti

Exhaust M. Rubel, D. Ivanova, E. Rachlew Heating T. Hellsten, E. Rachlew, M.

Cecconello, G. Ericsson, C. Hellesen, M. Gatu Johnson, S. Conroy

MHD M. Cecconello, G. Ericsson, C. Hellesen, M Gatu Johnson, D. Yadikin

S2 E. Rachlew, D. Yadikin Transport J. Weiland, H. Nordman, M.

Stransky, E. Rachlew, T. Hellsten TOTAL

Furthermore it can be pointed out that VR provided personnel to EFDA JET activity through secondments (Art 9.4 contracts) as follows: Seconded to the Operator on JOC contracts:

• Modelling Group (T. Johnson). • Diagnostics (S. Conroy)

Close support unit: • Responsible Officer M. Laxåback. • Responsible Officer E. Asp.

VR also provided personnel as Session Leader, Scientific Coordinators, Diagnostic Coordinators and Control Room Experts in neutron diagnostics, spectroscopy and other diagnostics, as well as in plasma theory and modelling.


108

6.4 Fusion Technology at JET Involvement of the RU in the Fusion Technology at JET is summarized in Table 6.4-1. The scientific information for the activity is reported in the earlier sections covering the Association Work Programme.

Table 6.4-1 Swedish RU involvement in Fusion Technology at JET

Task Agreement / Fusion Technology Topic

Description Responsible Person

JW9-FT-JET Plasma facing components 3.49

Surface morphology and characterization of first mirrors

M.Rubel

JW9-FT-JET Plasma facing components 3.52

Analysis of deposited layers and preparations for 10Be marker experiments

H. Bergsåker


109

7 EFDA Task Forces and Topical Groups The EFDA Work Programme for 2009 included activities specified in Task Agreements and Topical Groups. VR had activity in two EFDA Task Forces: Task Force Integrated Tokamak Modelling (TF-ITM) and EFDA Task Force Plasma Wall Interaction (TF-PWI). VR is active in both of these Task forces. In particular, The TF Leader for TF-ITM was Pär Strand and the Deputy Task Leader for ICRH Heating, Current Drive and Fast Particles was Torbjörn Hellsten. Marek Rubel was a task co-coordinator in the task area of dust generation. VR had activity in four Topical Groups: Materials, Diagnostics, MHD and Transport. The scientific information for the activity is reported in the earlier sections covering the Association Work Programme.

7.1 Task Force Integrated Tokamak Modelling Table 7.1-1 gives an overview of the activity in the Task Force Integrated Tokamak Modelling:

Table 7.1-1 Swedish RU involvement in the EFDA Task Force Integrated Tokamak Modelling

Activity Description Commitment (in ppy)

WP09-ITM-TFL1 Task Force Leader

Pär Strand 0.75 (eligible for preferential support)

WP09-ITM-TFL1 Deputy TF Leader IMP5 Heating, CD and Fast Particles

Torbjörn Hellsten 0.25 (eligible for preferential support)

WP09-ITM-IMP3 Transport code and discharge evolution

T2 standardized interfaces

0.25

WP09-ITM-IMP5 Heating, current drive and fast particles

T1 porting verification T2 evaluation datastructures T3 fast ICRH code T5 advanced 3D Fokker-Planck

0.2 0.2 0.6 0.5

TOTAL 2.75 of which 1.0 ppy pref sup


110

7.2 Task Force Plasma Wall Interaction Table 7.2-1 gives an overview of the activity in the EFDA Task Force Plasma Wall Interaction:

Table 7.2-1 Swedish RU involvement in the EFDA Task Force Plasma Wall Interaction

Activity Description Commitment

WP09-PWI-01 Post mortem analysis

01/VR/BS deposits 01/VR/PS deposits 01/VR/PS hardware

0.4 ppy bs 0.2 ppy ps 8 kEuro eligible ps

WP09-PWI-02 Fuel removal technologies

01/VR/BS surface analysis 01/VR/PS surface analysis 01/VR/PS hardware

0.2 ppy bs 0.2 ppy ps 10 kEuro eligible ps

WP09-PWI-03 Dust generation and characterisation

00/VR/PS task co-ord 01/VR/BS morphology mirrors 01/VR/PS fuel content 01/VR/PS hardware 02/VR/PS Aerogel samples

0.125 ppy ps 0.2 ppy bs 0.15 ppy ps 5 kEuro hardware 0.15 ppy bs

WP09-PWI-04 Erosion, transport, deposition

01/VR/BS studies of C13 01/VR/PS studies of C13 01/VR/PS hardware

0.2 ppy bs 0.1 ppy ps 8 kEuro hardware

WP09-PWI-05 High Z scenarios

01/VR/BS microscopy 01/VR/PS microscopy 01/VR/PS hardware


WP09-PWI-07 Alloys and compounds

01/VR/BS materials mixing 01/VR/PS materials mixing 01/VR/PS hardware


TOTAL 2.575 ppy of which 1.025 ppy pref sup 56 kEuro hardware pref sup


111

7.3 Topical Group Fusion Materials Table 7.3-1 gives an overview of the activity in the Topical Group Materials based on the draft of the work programme as of 13 November, 2008:

Table 7.3-1 Swedish RU involvement in the EFDA Topical Group Fusion Materials

Activity Description Commitment

Task Agreement Tungsten and Tungsten Alloys Development WP08-09-MAT-WWALLOY

Activity 2 Structure materials development. Development of chemical powder metallurgy routes for high purity and nano-structured tungsten ODAS alloys.

1.3 ppy basic support

7.4 Topical Group Diagnostics Table 7.4-1 gives an overview of the activity in the Topical Group Diagnostics based on the Diagnostic Development work programme as of 30 January, 2009 and the TGS work programme as of 6 April, 2009:

Table 7.4-1 Swedish RU involvement in the EFDA Topical Groups Diagnostics (DIA) and Diagnostics for Physics Studies (TGS).


WP09-DIA-01 Diagnostic development

05/VR/BS feasibility study fuel ion 06/VR/BS task coordinator feasibility high resolution ITER 06/VR/BS coordinator collab and calc high resolution ITER

0.2 ppy bs 1.0 ppy bs 0.2 ppy ps

WP08-09-TGS-01a Measurements of confined alphas

01/VR/BS Installation and commissioning of neutron spectrometers at MAST 01/VR/PS Installation and…

1.0 ppy bs 0.2 ppy ps 26 kEuro hardware ps

TOTAL 2.6 ppy of which 0.4 ppy ps and 26 kEuro hardware pref sup


112

7.5 Topical Group MHD Table 7.5-1 gives an overview of the activity in the Topical Group Diagnostics based on the Diagnostic Development work programme as of 30 January, 2009 and the TGS work programme as of 6 April, 2009:

Table 7.5-1 Swedish RU proposed involvement in the EFDA Topical Group MHD.


WP09-MHD-02 Disruptions

04/VR/BS calc runaway electron drift orbits 04/VR/PS calc runaway electron drift orbits

0.3 ppy bs 0.3 ppy ps

WP09-MHD-05 Toroidal viscosity breaking and RWM control

00/VR/BS task coordination 01/VR/BS Extrap rotation breaking and MIMO SISO 01/VR/PS Extrap rotation breaking and MIMO SISO

0.125 ppy bs 0.925 ppy bs 0.55 ppy ps

TOTAL 2.2 ppy of which 0.85 ppy pref sup

7.6 Topical Group Transport Table 7.6-1 gives an overview of the activity in the Topical Group Diagnostics based on the Diagnostic Development work programme as of 30 January, 2009 and the TGS work programme as of 6 April, 2009:

Table 7.6-1 Swedish RU involvement in the EFDA Topical Group Transport.


WP09-TGS-02a Core impurity transport

01/VR/BS convec impurity response RF heating

2.6 ppy bs

WP09-MHD-02b Physics of rotation

01/VR/BS devices poloidal toroidal rotation 01/VR/PS modeling momentum transport

0.1 ppy bs 0.2 ppy ps

TOTAL 2.9 ppy of which 0.2 ppy pref sup


113

7.7 Goal Oriented Training (GOTiT). Table 7.7-1 gives an overview of VR participation in Goal Oriented Training in Theory as included in definitive version of task agreement, 14 July, 2008.

Table 7.7-1 Swedish RU involvement in the GOTiT for 2008 and 2009 (start 1 October, 2008).


WP08-GOT-GOTiT Goal oriented training in theory

MHD, waves fast particles - ICRH

0.4 (mentoring) 2.0 (trainees)

TOTAL 2008 & 2009 2.4


114

8 ITPA 8.1 Overview of ITPA activities.

The International Tokamak Physics Activity (ITPA) provides a framework for internationally coordinated fusion research activities. The ITPA continues the tokamak physics R&D activities that have been conducted on an international level for many years resulting in the achievement of a broad physics basis essential for the ITER design and useful for all fusion programs and for progress toward fusion energy generally. The ITPA has been operating under the auspices of the IAEA International Fusion Research Council since its inception in July 2001. The ITPA from February 25, 2008 now operates under the auspices of ITER in order to provide the framework for coordinated physics research activities proposed by the ITER-IO. These co-ordinated physics research activities will help develop the physics basis for ITER operation, integrate the expertise of the international fusion community into ITER, provide a pathway to exploit the capabilities of existing fusion facilities and programmes in support of ITER, and integrate results of the fusion programmes of the ITER Members into planning of ITER operation. The Swedish Research Unit has participated in the following ITPA Topical Groups:

• Transport and Confinement

• Scrape Off Layer and Divertor

• Diagnostics The ITPA will provide support to ITER in the fulfilment of its mission by helping to create a common international research programme organized around scientific issues and will facilitate the participation of the ITER Members in the ITER scientific programme. The integration of fusion researchers from the ITER Members into the ITER physics research programme will help ITER become a world-class scientific research facility for the benefit of the international fusion community. The scientific information for the ITPA activity is reported in the earlier sections covering the Association Work Programme.


115

9 Technology 9.1 Art. 5.1b actions. The association has carried out work on the following EFDA Technology Art 5.1b Contracts during 2009. During 2009 all activity for the Art 5.1b actions has been completed. EFDA/07-1712 Annex 1 Effects of copper impurities (A. Molander): Concluded. EFDA07-1712 Annex 2 Resistive wall modes (Y. Liu): Concluded.

EFDA/06-1437 Diagnostic design for ITER (G. Ericsson): Final report concluded. EFDA/06-1527 Manufacture of Be coatings (M. Rubel): Running:


116

10 Other activities 10.1 Training and education PhD training The physics programme of the Swedish Fusion Research Unit is university based. There are PhD programmes at CTH, KTH and UU. During 2009 there were 16 PhD students included in the Research Unit. On the average there are about 3 to 4 PhD examinations per year. This is slightly less than in previous years, due to a decrease in funding for PhD students. However there is no difficulty in recruiting students when funding is available. Masters programmes In addition to the PhD programme the three universities have Master of Science programmes where students can select fusion plasma physics topics for their thesis work. Annually there are about 15 MSc thesis students. 10.2 Public information The Swedish Public Information Group (PIG) officer within EFDA is Jan Scheffel at the Division of Fusion Plasma Physics, KTH in Stockholm. The Swedish PIG activities are mainly carried out in the form of lobbying, concentrated towards contact with government, parliament and energy administrators with the intention of strengthening fusion funding. The national funding for fusion is unchanged since the early 1990’s, and the situation is quite difficult with regards to retaining competence and enrolling younger scientists. The Swedish Energy Authority has a mission to consider fusion but has sofar essentially not supported fusion research. The year 2009 has been an intense year with substantial contact also with the public. In particular, there is a strengthened interest in fusion among school pupils and university students, most likely because of the debate on global warming and the energy future. Performed PIG activities Lobbying

• Personal contact with several politicians and administrators in the energy sector • Articles on fusion, to Swedish newspapers or produced by independent journalists

Contacts with the public

• Interview of PIG officer in Swedish national television’s most viewed news program, where the lack of fusion support from the Swedish Energy Authority was questioned

• A stand with fusion information at the largest Swedish conference on energy, with over 2000 attendants gathered for two days

• Talks on fusion to school children, students, associations and at other universities • Tutorship for several groups of upper secondary pupils doing project work in fusion


117

• Cooperation with the science center “House of Science” on “science points” Contacts with students, taking university courses

• Students, to become upper secondary school teachers, now take a course on fusion and energy

• A recently started popular university summer course on energy includes talks on fusion. Lectures Presentations were made at the Swedish ”Energimyndigheten Energiting 2009.” held in March 2009.

• James R Drake, ITER-internationellt fusionsexperiment med svenskt deltagande. • Jan Scheffel, Fusionskraftens potentiella roll inom framtidens energisamhälle • Göran Ericsson, Deltagande i internationell fusionsforskning – neutronspektroskopi

vid Uppsala universitet. • Per Brunsell, Deltagande i internationell fusionsforskning - EXTRAP T2R och aktiv

plasmakontroll. • Stefan Wikman, Utveckling av fusion - industriperspektiv


118

Appendix I: Fusion for Energy Grants The Joint Undertaking for ITER and the Development of Fusion Energy, called fusion for Energy or F4E, has the responsibility for procuring the necessary research and development as well as the equipment for ITER. F4E can provide support to groups in the member states in the form of Grants, which cover 40% of the costs of the Research and Development. This F4E activity is not a part of the EURATOM fusion programme, which provides contributions to the Research Units as established in the Contract of association and the European Fusion development Agreement. However it is important that the EURATOM programme maintains an effective contact with F4E since the EFDA programme must have a programme where preparations for ITER have the highest priority. Therefore information of Association VR activity for F4E is provided here for information. During 2009, Studsvik Energy AB has had the following activity summarised in Table A.1. Table AI.1. Overview of Fusion for Energy Grants 2009.

Contract No. Title Type Comments

F4E-2008-GRT-02-01 (ES-SF)

Detailed analyses of LOCA reference event scenarios

Grant Finished

F4E-2008-GRT-01-02 (ES-FS).

New analyses of fire reference events scenarios

Grant Started

F4E-2008-GRT-021 (MS-VV)

Corrosion issues Grant Started

F4E-2009-GRT-031 (ES-FS)

Safety assessment of EU TBM

Grant Proposal


119

Appendix II: Contact Information European Commission Douglas Bartlett [email protected] European Commission DG Research J.4 - CDMA 05/115 B-1049 Brussels, Belgium Tel: +32 2 2954314 Swedish Research Council (VR) Lars Börjesson [email protected], Tel.: +46 8 456 44 109 Per Karlsson [email protected], Tel.: +46 8 456 44 177 Johan Holmberg [email protected] Research Infrastructures Swedish Research Council (VR) SE-103 78 Stockholm, Sweden Royal Institute of Technology * James R Drake [email protected], Tel: +46 8 790 7721 * Per Brunsell [email protected], Tel: +46 8 790 6246 * Jan Scheffel [email protected] Tel: +46 8 790 8939 **Elisabeth Rachlew [email protected], Tel.: +46 8 5537 8112 * Alfvén Laboratory Department of Fusion Plasma Physics School of Electrical Engineering KTH, SE-100 44 Stockholm, Sweden **AlbaNova University Center Department of Physics KTH, SE-10 691 Stockholm, Sweden


120

Chalmers University of Technology *Mietek Lisak mietek.lisak @chalmers.se, Tel: +46 31 772 15 65 *Jan Weiland [email protected], Tel.: +46 31 772 15 75 **Tunde Fülöp [email protected], Tel.: +46 31 772 3180 *Department for Radio- and Space Science CTH, SE-412 96 Göteborg, Sweden **Department of Applied Physics CTH, SE-412 96 Göteborg, Sweden Uppsala University Göran Ericsson gö[email protected], Tel: +46 18 471 3446 Marco Cecconello [email protected], Tel+46 18 471 3043 Christer Wahlberg [email protected], Tel: +46 18 471 3104 Department of Physics and Astronomy Division of Applied Nuclear Physics Ångstrom Laboratory Uppsala University, Box 525, SE-751 20 Uppsala, Sweden Lund University Christer Jupén [email protected], Tel.: +46 222 7735 Department of Astronomy Lund University, Box 117, SE-221 00 Lund, Sweden Studsvik Sture Nordlinder [email protected], Tel.: +46 155 22 12 52 Studsvik Nuclear, SE-611 82 Nyköping, Sweden

annual progress report 2009 - kth/annual_report_2009.pdf · annual progress report 2009 ... the...

Documents