Fusion Plasma Physics Annual Report 2015

KTH Royal Institute of Technology
Stockholm, December 2016
2
Fusion Plasma Physics School of Electrical Engineering KTH Royal Institute of Technology SE-100 44 Stockholm, Sweden
Fusion Plasma Physics Annual Report 2015
3
Contents 1. Introduction ............................................................................................................................ 5 2. Research projects .................................................................................................................... 6
2.1. Active MHD control ........................................................................................................ 6 2.1.1. EXTRAP T2R device ............................................................................................... 6 2.1.2. Improved Model Predictive Control by Error Field Estimator ................................ 8
2.2. Applied magnetic perturbations as a tool for tokamak-relevant physics studies .......... 10 2.2.1. Hysteresis in the locking – unlocking mechanism of a TM ................................... 10 2.2.2. TM locking to an external field in the MST reversed-field pinch ......................... 11 2.2.3 Non-disruptive technique for the identification of field errors ............................... 12
2.3. Plasma - wall interactions.............................................................................................. 13 2.3.1. Co-deposited layers in the divertor regions of JET with the ITER-Like Wall ...... 13 2.3.2. Tracer techniques for the assessment of material migration .................................. 18 2.3.3. Microanalysis of divertor surfaces in JET with ITER-like wall ............................ 20 2.3.4. A 10Be marker experiment on beryllium migration in JET-ILW ........................... 20
2.4. Theoretical fusion plasma physics ................................................................................ 22 2.4.1. Exploitation of JET ................................................................................................ 22 2.4.2. Contributions to the MST1 Work Programme ....................................................... 23 2.4.3. Integrated modelling within the EUROfusion-WPCD ........................................... 23 2.4.4. Non-linear wave-particle interactions .................................................................... 24 2.4.5. Iterative approach to spatial dispersion .................................................................. 24
2.5. Computational methods for fusion plasmas .................................................................. 25 2.6. Confinement physics ..................................................................................................... 27
2.6.1. Pedestal properties and confinement in JET .......................................................... 27 2.6.2. ELM behaviour in AUG ......................................................................................... 29
3. Education and research training ........................................................................................... 30 3.1. Basic and advanced level education .............................................................................. 30 3.2. Research training ........................................................................................................... 31
4. Personnel .............................................................................................................................. 35 5. Income & Expenditure ......................................................................................................... 36 Bibliography ............................................................................................................................. 37
4
5
1. Introduction The research at the Department of Fusion Plasma Physics is part of the European research program in fusion energy through involvement in the European programme EUROfusion. While the long-term goal of the activity is the realisation of fusion power as a new energy source, the near term aim of the EUROfusion activity is the support for ITER. The international project ITER, involving seven partners around the world – EU, US, Japan, China, Russia, India, South Korea – aims a building and operating a fusion test reactor able to produce around 500 MW of thermal power, thereby demonstrating the feasibility of fusion energy. The construction of this facility is now in progress in Cadarache, France. The European fusion research in the H2020 framework programme is based on a “European Joint Programme” Grant Agreement (GA) between Euratom, and a Consortium “EUROfusion”, formed by parties representing a majority of the countries in Europe. This agreement, and the corresponding Consortium Agreement (CA) between the Consortium parties, forms the framework for the European fusion research programme. The total Euratom grant amount for the five-year period 2014-18 is 458 million Euro. The overall activity of the Department in the EUROfusion Annual Work Programme during 2015 has been on a similar level as in the first year of the programme (2014). Areas of particular high involvement have been investigation of plasma-wall interaction and plasma facing components and material research, and participation in experimental campaigns at the JET and ASDEX-Upgrade research facilities - in some cases as scientific coordinators. Another key activity is the participation in the integrated modelling within EUROfusion, where the Department is coordinator for the heating and current drive project. During 2015 a main contribution in this area has been the development of the modeling infrastructure, involving also organizing educational activities in the form of “code camps”. Another area of work is the development of active MHD mode control carried out at the local EXTRAP T2R device. The high level of the research activity is documented by over fifty papers published by researchers at the Department in 2015. The Department’s involvement in education at the undergraduate level at KTH remains high, with major efforts being the course in “Vector Analysis” given to the second year students of the Electrical Engineering programme, the advanced level courses in the Electro Physics Master Programme, and the BSc and MSc thesis projects offered. At the graduate level, there are currently twelve active PhD students at the Department, an unusually high number compared to earlier years. The high number of students is indicative of the presently very lively and productive research environment at the Department. The EUROfusion funding for research training has been essential in order to provide financial support for the PhD students. Also contributing to the successful research training is the European summer schools in fusion physics and the European FuseNet meetings for PhD students in fusion. Per Brunsell Department Head Fusion Plasma Physics
6
2. Research projects 2.1. Active MHD control P. Brunsell, L. Frassinetti, A. C. Setiadi (PhD student), R. Fridström (PhD student), and J. R. Drake In collaboration with: W. Suttrop, V. Igochine, Max-Planck-Institut für Plasmaphysik, Garching T. Bolzonella, G. Marchiori, G. Manduchi, Consorzio RFX, Padova C. R. Rojas, H. Hjalmarsson, EES/Automatic Control, KTH
2.1.1. EXTRAP T2R device A main feature of the EXTRAP T2R reversed field pinch device is a comprehensive system for development and test of methods for active MHD control. One aim of this research is to develop control methods for future RWM control experiments at the ASDEX Upgrade tokamak, which is turn is motivated by the desire to incorporate the possibility for active control in ITER and future tokamak reactor designs. Another distinguishing feature is an all metal (stainless steel/molybdenum) wall. The EXTRAP T2R device includes a number of plasma diagnostic systems, such as magnetics, interferometer, Thomson scattering, SXR detectors, visible and ultra-violet light spectrometers, bolometers, Langmuir and collection probes.
The main components of the MHD mode control system at EXTRAP T2R is an array of active control coils placed outside the conducting shell, and a corresponding array of sensor coils placed inside the shell, as shown in Fig. 2.1-2.
Figure 2.1.1-1. EXTRAP T2R device at Alfvén Laboratory KTH
7
Figure 2.1.1-2. Active MHD control system installed at EXTRAP T2R. Active control coils outside the shell in red, sensor coils inside the shell shown in blue, and the conducting shell itself depicted in orange colour.
The main features of system are: • 128 magnetic flux loop sensors at 4 poloidal and 32 toroidal positions inside the shell. • 128 active saddle coils at 4 poloidal and 32 toroidal positions outside the shell. • Saddle coils and sensor flux loops pair-connected at each toroidal position to form 64
independent “m=1” coils and sensors. • 64 “m=1” coil current transducers. • 32 two-channel power amplifiers units providing at total of 64 independent channels. • Integrated digital controller unit including CPU board, ADCs and DACs. • Control algorithms implemented in software. Audio amplifiers with output power of 800-1200 Watt are used for driving the active coils. Control algorithms are implemented in a Linux based PC equipped with a 3.0 GHz 6-core processor and 8 GB memory. The kernel of the Linux operating system has been modified so that all interrupts to the CPU are suspended during the shot. The kernel modification guarantees the availability of the CPU for the real-time feedback control algorithm. The integrated digital control system consists of several boards installed in a compact PCI crate. The data acquisition modules transfer data to a shared buffer in the host PC by Direct Memory Access. There are two data acquisition modules providing 16-bit analog-to-digital (ADC) for a total of 128 simultaneously sampled channels, and one digital-to-analog (DAC) module with 64 output channels. The setup of the data acquisition boards is facilitated by an additional Ethernet connection between the host PC and the boards. The minimum cycle time of the system with this channel configuration is of the order of 20 microseconds.
Feedback control algorithms are written in C++. Various types of controllers have been implemented, such as the conventional Proportional-Integrating-Derivative (PID) controller, as well as modern model based controller such as the Model Predictive Controller (MPC).
8
2.1.2. Improved Model Predictive Control by Error Field Estimator The RWM response model obtained from system identification has been used as a base for a state-space controller of the Model Predictive Control (MPC) type. The MPC controller has been successfully implemented and tested on EXTRAP T2R. The MPC employs a predictive model to calculate the optimal control action. It is as an advanced control technique and one of its advantages is that the MPC is able to incorporate state and actuator constraints in the controller design.
Figure 2.1.2-1. Block diagram and signal routing of the Error Field Estimator
Recently, the empirical model has been used to estimate the intrinsic machine error field. Two different error field estimation methods have been studied and compared. The error field estimator is then combined with the MPC to yield better radial field suppression at the wall. Figure 2.1.2-1 illustrates the basic idea: It is assumed that the error field is caused by an unknown input acting on the plant . The Error Field (EF) Estimator provides an estimate of the error field used to perform compensation by negative feedback.
Figure 2.1.2-2. Block diagram of the Disturbance Observer (DOB).
The first method is the Disturbance Observer (DOB). The main principles of the DOB are illustrated in Figure 2.1.2-2: The approximate inverse of the plant is found (−1) and used to reconstruct the input entering the plant. This reconstructed input is then compared to the actual input entering the plant, and the difference of the two is used as the estimate of the error field . A filter () is required for physical implementation of the DOB. Typically a low-pass filter is used, assuming a low-frequency, quasi-static error field. The second method is the Unknown Input Observer (UIO), illustrated in Figure 2.1.2-3. Similar to the DOB, the UIO assumes that there are unknown inputs entering the system. The difference is that UIO
9
provides state estimation along with disturbance estimation. The main idea for the UIO is that we need to update the dynamics of our system such that is incorporates the effects of the disturbance. The UIO requires some information regarding the dynamics of the disturbance. The most common assumption is a constant disturbance.
Figure 2.1.2-3. Block diagram of the Unknown Input Observer (UIO)
The main parameters of the error field that need to be estimated are for each Fourier mode number the amplitude and phase.
Figure 2.1.2-4. MPC w/o and with error field compensation. Time evolution of the radial magnetic field at the wall for difference toroidal Fourier modes, n=-11, -8, -6, 4, 5, and 8.
The implementation of DOB and UIO to perform estimation and error field compensation alongside MPC feedback control is relatively simple. The experimental results are shown in Figure 2.1.2-4, comparing the radial field suppression for MPC w/o and with error field compensation, using DOB and UOI methods, respectively.
10
2.2. Applied magnetic perturbations as a tool for tokamak- relevant physics studies L. Frassinetti, P. Brunsell, R. Fridström (PhD student), M.W.M. Khan (PhD student), In collaboration with: S. Munaretto, B.E. Chapman, University of Wisconsin-Madison, USA R. Sweeney, F. A. Volpe, Dept. Applied Physics and Mathematics, Columbia University, USA External magnetic perturbations are an important tool in tokamaks to mitigate edge localized modes and/or to influence the neoclassical tearing mode island dynamics in order to optimize ECCD stabilization. On the other hand magnetic perturbations produce also undesired effects such as plasma flow braking. A clear study of the corresponding underlying physics is relatively complicated in tokamaks because of the limited number of active coils that inevitably produces a broad spectrum of side-band harmonics. The EXTRAP T2R reversed-field pinch is equipped with a set active and sensor coils that can produce external magnetic perturbations (MPs) with a specific harmonic in a controlled fashion (i.e. with defined amplitude and phase). The feedback system of EXTRAP T2R has the capability of suppressing the entire RWM and error-field spectrum and simultaneously producing a clean external MP. EXTRAP T2R is therefore a useful machine to investigate the magnetic perturbation effect on the plasma dynamics. During 2015 a series of studies aimed at the understanding of the mechanisms that lead to the tearing mode locking-unlocking mechanism have been conducted. The work has been extended studying the a similar mechanisms in the MST experiment, located at University of Wisconsin-Madison.
2.2.1. Hysteresis in the locking – unlocking mechanism of a TM The RMP interaction with the tearing modes (TMs) can reduce the plasma rotation and eventually cause wall-locking. The wall-locking has been observed to produce disruptions in tokamaks and shot terminations in reversed-field pinches (RFPs). Clearly, rotation is desirable for achieving high confinement and disruption free fusion plasmas. Therefore, understanding of the physical mechanisms behind TM wall-locking and - unlocking is desirable. The TM dynamics in the presence of an RMP have been theoretically studied in both tokamaks and RFPs. The TM locking and unlocking is characterized by a hysteresis behavior, where the locking threshold (the RMP amplitude required to lock the mode) is significantly higher than the threshold for unlocking. Theoretically, the hysteresis in the TM locking and unlocking process can be described by the magnetic island time evolution equations subject to an RMP. When an RMP is applied to a rotating plasma, it produces an electromagnetic (EM) torque that brakes the TM velocity. The velocity reduction spreads to rest of the plasma through the viscosity. The surrounding plasma produces a viscous torque that acts to restore the natural rotation and a lower plasma rotation is obtained. If the RMP is larger than the locking threshold, the TM goes through a transition to the wall-locked state. To unlock the TM, the EM torque must be reduced to a value lower than the viscous torque. However, soon after
11
the locking, the velocity reduction profile relaxes, producing a sudden drop in the viscous torque. Therefore, to unlock the mode, the RMP must be reduced to a value significantly lower than the locking threshold. In addition, the theory predicts that the EM torque increases after the locking, due to the growth of the TM island. This should lower the unlocking threshold even more, in other words deepening the hysteresis The work in EXTRAP T2R has investigated from the experimental point of view the above described mechanisms. The results show that after TM locking to an external RMP, a relaxation of the velocity profile with a subsequent drop in the viscous torque and an increase of the TM island size is observed. The relaxation of the velocity profile is shown in figure 2.2.1. The drop in the viscous torque causes hysteresis in the RMP locking/unlocking threshold or in other words, to unlock the TM, the RMP amplitude needs to become significantly lower than the RMP amplitude required for locking the TM. The increased island size results in an increased EM torque and thereby a deepening of the hysteresis. The experimental results are in qualitative agreement with the theoretical model.
2.2.2. TM locking to an external field in the MST reversed-field pinch Previous works in RFPs have mainly focused on the interaction between a single harmonic RMP and a TM of corresponding resonance. However in tokamaks, the number of RMP coils that can be placed outside the machine-wall is limited by other equipment, such as ports for neutral beam injection (NBI) and radio frequency (RF) heating. This leads to a broad spectrum of harmonics, which can potentially be resonant with more than one TM. For example, to describe the effect of magnetic perturbations on one TM island in ASDEX, the torques on all TM islands had to be included In addition to error field correction, the feedback system can be used to produce RMPs with a pre-set amplitude and poloidal mode number m. Due to the coils limited toroidal extent, they produce a broad spectrum in toroidal mode number n. MST has several TMs that are naturally rotating together with the plasma fluid. Hence, MST is a suitable device to study the simultaneous braking of several TMs, due to a multi-harmonic RMP. In this work, the RMP effect on the tearing mode dynamics is experimentally studied in MST. With the application of a multi-harmonic RMP, the work has experimentally investigated the TM dynamics and locking threshold in MST. Observations showed that the locking threshold was increased with increasing plasma density. A model, describing the TM interaction with the wall and an external multi-harmonic RMP, showed qualitative agreement with the experimental data. An example of the results is shown in figure 2.2.2.
Figure 2.2.1. Velocity reduction profile at wall- locking (blue unfilled circles) and before unlocking (red filled diamonds) for shot 24 770
12
It is shown that the EM torque acting on multiple resonant surfaces has to be modeled to describe the evolution of other TMs n > 6. Since the TM rotation is connected to the plasma rotation, the inclusion of multiple modes is also important to describe the viscous torque and estimate the kinematic viscosity.
2.2.3 Non-disruptive technique for the identification of field errors An error field (EF) detection technique using the amplitude modulation of a naturally rotating tearing mode (TM) is developed and validated in the EXTRAP T2R reversed field pinch. The technique was used to identify intrinsic EFs of m/n = 1/−12, where m and n are the poloidal and toroidal mode numbers. The effect of the EF and of a resonant magnetic perturbation (RMP) on the TM, in particular on amplitude modulation, is modeled with a first-order solution of the modified Rutherford equation. In the experiment, the TM amplitude is measured as a function of the toroidal angle as the TM rotates rapidly in the presence of an unknown EF and a known, deliberately applied RMP. The RMP amplitude is fixed while the toroidal phase is varied from one discharge to the other, completing a full toroidal scan. An example is shown in figure 2.2.3. Using three such scans with different RMP amplitudes, the EF amplitude and phase are inferred from the phases at which the TM amplitude maximizes. The estimated EF amplitude is consistent with other estimates (e.g. based on the best EF-cancelling RMP, resulting in the fastest TM rotation).
Figure 2.2.2. Experimental and modeled TM data for modes n = 6; 7 and 8, in blue, green and red color, respectively.
Figure 2.2.3. Time-varying amplitude of the field (blue) and the phase of the TM (red) in the presence of a Br = 2 G RMP during rotation period. The vertical dashed line intersects the point where the poloidal field is maximized (blue point) and the TM phase at which this maxima occurs (red point). The solid horizontal line shows the phase of the RMP.
13
2.3. Plasma - wall interactions M. Rubel, P. Petersson, H. Bergsåker, A. Garcia-Carrasco (PhD student), P. Ström (PhD student), A. Weckmann (Ph.D. student), Y. Zhou (Ph.D. student) In collaboration with G. Possnert and D. Primetzhofer (Uppsala University) Plasma-wall interactions (PWI) comprise all processes involved in the exchange of mass and energy between the plasma and the surrounding wall. Two inter-related aspects of fusion reactor operation - economy and safety - are the driving forces for studies of PWI. The major issues to be tackled are: (i) lifetime of plasma-facing materials (PFM) and components (PFC), (ii) accumulation of hydrogen isotopes in PFC, i.e. tritium inventory; (iii) carbon and metal (Be, W) dust formation. PWI is one of the primary areas where integration of the Physics and Technology programmes is being achieved. The work at KTH in the field of PWI and fusion- related material physics has been fully integrated with the international fusion programme: (i) EU Fusion Programme, (ii) International Tokamak Physics Activity (ITPA), (iii) International Atomic Energy Agency (IAEA), (iv) Implementing Agreements of International Energy Agency (IEA). It is demonstrated by the participation in: • EUROfusion Work Programme:
o Work package JET2: Analysis of Plasma-Facing Components from JET o Work package PFC: Plasma-Facing Components for ITER o Work package MAT: Diagnostic Materials for DEMO
• ITPA, IAEA and IEA activities. Experimental work is carried out at home laboratory, JET, ASDEX-Upgrade and Forschungszentrum Juelich. The research programme is concentrated on: • Material erosion, migration and re-deposition. • Fuel retention studies and fuel removal techniques. • Dust generation processes in fusion devices. • Characterization of plasma-facing materials including testing of high-Z metals. • Development and testing of diagnostic components: first mirror test at JET for ITER • Development of diagnostic methods for PWI studies.
2.3.1. Co-deposited layers in the divertor regions of JET with the ITER-Like Wall The determination of fuel retention and material migration leading to the buildup of co- deposits – especially beryllium erosion and transport – belong to key research areas in operation of the JET tokamak with the ITER-Like Wall (JET-ILW). The analysis of PFC from a campaign which started in 2011 with clean tile surfaces has given a unique opportunity of looking for a material migration and deposition history. This also includes tracing of the plasma operation and the history of heating power increase throughout the campaign: starting from limiter plasmas and then followed by the divertor operation with a gradually increased heating. During the latter operation period impurity species (e.g. nitrogen) were seeded to cool the plasma edge. The retention of nitrogen is to be addressed. The focus of this work was on a correlation between the operation and the deposition accompanied by material mixing.
The study was carried out for the tungsten-coated carbon fibre composite (W-CFC) divertor plates and for erosion-deposition probes (EDP) retrieved from the JET vessel after the 2011-
14
2012 campaigns which lasted totally for about 18.9 h including 13 h of X-point operation. Specimens were obtained by coring the inner (Tile 1, 3 and 4) and the outer divertor (Tile 6, 7, 8) plates. In total nineteen samples have been examined. The study of wall probes was done for test mirrors made of polycrystalline molybdenum and for a stainless steel cover of the quartz microbalance (QMB); details about the structure and location of wall probes are in. It has already been known that the thickness of co-deposited layers in most positions does not exceed 1 µm. To study the impurity concentrations in these layers time-of-flight heavy ion elastic recoil detection analysis (ToF-HIERDA) was used. Its main advantages are: (a) high sensitivity, (b) good separation of light isotopes up to neon, (c) depth profiling with high resolution for about 0.5µm (i.e. information depth) from the sample surface and (d) quantitative measurements for different elements. Under normal conditions a sensitivity lower than 10-3 is normal. However, as both the impinging ion and the recoiled target ion are leaving the target at shallow angles the depth profiles are very sensitive to the sample roughness which consequently deteriorates depth resolution. For samples where the roughness is larger than the analyzed depth it is not possible to determine a quantitative depth profile. Instead the ratio of different elements in the analyzed layer was calculated. This feature was used in the examination of W-CFC cored samples from the divertor.
A representative HIERDA spectrum for a divertor target (Tile 4 in this case) is shown in Figure 2.3.1-1. Hydrogen, deuterium, beryllium, carbon, nitrogen and oxygen along with tungsten, steel or Inconel components are the main species detected in most locations. The mass resolution is not sufficient to resolve the different steel elements Fe, Ni, Cr accurately and they are therefore treated as a single element. In addition, molybdenum has been found in several locations. It originates both from the Mo-W marker layers on some tiles and the erosion of Inconel 625 parts of the radio frequency antenna. An important feature is a clear trace of nitrogen thus proving that gases for plasma edge cooling are efficiently retained in co- deposits on PFC. This confirms results of our works dedicated to the determination of nitrogen retention in TEXTOR and ASDEX-Upgrade.
Figure 2.3.1-1: Example of ToF-HIERDA spectrum from Tile 4.
An overview of the poloidal deposition pattern of Be, C, N and O is shown Figure 2.3.1-2. The divertor cross-section with the so-called S-coordinate is also shown. The data are given in the form of concentration ratios with respect to beryllium, deposited as a result of erosion from limiters in the main chamber: O/Be, N/Be and C/Be. In this notation low values correspond to points with high relative beryllium content. This particularly applies to the top part (apron) on Tile 1 where thick Be deposits were formed, as demonstrated in Figure 2.3.1-3, from which presents distribution and absolute Be and D contents along the tile. A clear maximum occurs on the flat apron. Except that Be-rich region, in the majority of other analyzed areas, a strong signal from the tungsten, or molybdenum for Tile 3 that has a special marker coating, substrate could be detected thus indicating that the entire deposited layer was probed. For a
15
uniform layer of Be on top of W, under the current analysis conditions, a clear tungsten signal should be seen as long as the layer thickness was below 2x1019 Be cm-2. For most of the points the Be signal that was detected should correspond to less than 10 % of this as the W signal is not strongly affected. For the high field side of Tile 4 the layer is possibly thicker but should be less than 50%. Also on the upper part of Tile 6 the W signal is affected but should be less than 25%. In the total amount of Be was found to be below 1018 at./cm2 on Tile 4 and 6 with higher amounts found locally in valleys formed by the CFC structure, this indicates indeed that the estimates above is an upper limit.
Figure 2.3.1-2. Amount of light impurities related to the concentration of Be are given for different positions in the diveror . On top of the plot the corresponding tile numbers are given. Positions of the tiles are given in the insert. Note that the scale for N is different from the C and O scales.
Figure 2.3.1-3. Be and D concentration on Tile 1. Att: scales are different for different parts of the tile.
Using the data presented in Fig. 2.3.1-2 and 2.3.1-3 and the estimate above it is possible to estimate how much carbon, oxygen and nitrogen is found in the layer. From the histogram in Figure 2.3.1-2 one infers that the atomic concentration of nitrogen is between 3% on top of Tile 1 and 15 % on Tile 3 in the inner divertor. However, when absolute numbers are considered, one perceives that the retention of nitrogen is the largest in the deposit on Tile 1. The same is true for oxygen and carbon. The obtained pattern is a result of the last period of operation with 151 pulses with the same position of strike points. Carbon migrates by erosion
16
and re-deposition and it is transported also to Tile 4 where ratio changes from 0.49 on the sloping part to 0.93 in the corner under the protection plate of Tile 5. It is a result of a multi- step process, as determined earlier in marker experiments with 13C-labelled methane. However, the most important point is that the absolute carbon concentrations are small reaching only a level (1019 cm-2), even in the thickest layers.
The chemical form of deposited species could not be studied. The formation of BeO by gettering of oxygen impurities can be assumed with high probability. Indeed, oxygen is the most common impurity but in most locations O/Be <1. Therefore, the presence of other forms of Be in the divertor should be considered. While the formation of a certain amount of stoichiometric compounds, e.g. beryllium nitride (Be3N2) or carbide (Be2C), under tokamak operation cannot be fully eliminated, the dynamic conditions of erosion and deposition suggest rather a mixture of compounds containing also chemically bound nitrogen. The formation of mixed oxides, nitrides, borides and carbides has been detected using X-ray methods on the very surface layer and in thicker films on components from TEXTOR and, also on test mirrors from JET.
On the divertor-inserted probes with smooth surfaces (quartz microbalance and test mirrors from the First Mirror Test) quantitative depth profiling could be accomplished. Plots in Figure Figure 2.3.1-4(a) reveal the deposition history on a molybdenum mirror from the outer divertor. The total deposit thickness is around 180 nm. The lack of a sharp mirror-deposit interface could be a result of both material mixing and variations in the layer thickness, other effects such as the mirror roughness should be smaller. Beryllium is a dominant constituent in the deposited layer and its profile is nearly constant with O/Be, N/Be and C/Be are 0.3; 0.04 and 0.06, respectively. Only at the interface there is relative increase of the oxygen content attributed to: (a) unavoidable presence of the oxide (MoO2, MoO3) layer on the original mirror surface and (b) larger amount of oxygen impurities gettered by Be at the very beginning of the JET-ILW operation. Another increase of oxygen content at the very surface corresponds to the uptake of atmospheric oxygen. Carbon also has two regions of the increased content. They correspond to the initial phase of the ILW operation and exposure to air after the mirror retrieval from JET. In general, the C content is very low, on average 20 times less than Be, thus being in a very good agreement with spectroscopy measurements. There is also a clear feature of Inconel and steel components eroded from components such as the main chamber wall. The thickness of the layer where features of nitrogen and tungsten appear is smaller than for the elements described above. It reflects that fact of divertor operation started a few weeks after the initial ILW phase with limiter plasmas. The deposition of W is accompanied by the accumulation of nitrogen in the deposit. It is a strong indication that the nitrogen presence is directly related to the tokamak operation. Similar profiles have been also determined on the mirrors from the inner divertor where discrete structure of tungsten and Inconel deposition profiles could be traced.
Depth profiles in Figure 2.3.1-4(b) have been recorded for a cover of QMB located in the outer divertor. The features of beryllium, carbon and oxygen are as those on the mirror. The amount of Be on the cover is even more pronounced: O/Be, N/Be and C/Be ratios equal to 0.1; 0.03 and 0.01, respectively. The profile of steel components is different than that of Inconel on the mirrors because on the QMB there is combination of deposition of Ni+Cr+Fe from plasma and also mixing of those elements at the cover-deposit interface due to roughness and possibly local erosion re-deposition. Small traces of W were found but are not shown in the Figure. The nitrogen isotopes have different sources: 14N used during the entire divertor operation for a long period and, a rare isotope 15N (natural abundance 0.37%) used as a marker puffed locally from a single gas inlet module (GIM 14) in the divertor floor between
17
Tiles 6 and has only been found on the QMB cover so far. This was done only during four pulses which were still followed by two-weeks operation period (151 pulses) before the shutdown. The detection of 15N in the deposits leaves no doubts regarding the origin of nitrogen in all studied deposits. It also shows a great value of HIERDA in studies of plasma-facing materials.
Figure 2.3.1-4. Depth profiles of species in layers on the deposition probes from the outer divertor.
In summary, both on the mirror and the QMB cover the amount of impurities is much lower compared to PFC surfaces of the divertor. It shows only limited transport of species to the shadowed areas. Even in the case of carbon only small quantities are measured on the probes. However, there are some differences between the probes. For instance, the cover has 30 times less deposition when compared to the louvre clip exposed for the same time in the outer divertor: 5x1016 cm-2 on the cover versus 170x1016 cm-2 on the clip, but this amount is nevertheless two orders of magnitude smaller found in flaking layers after the JET-C operation.
Concluding remarks The work has brought two important results to the better understanding and description of material migration processes in JET-ILW: (a) significant residence of nitrogen used for edge cooling and (b) transport of metals to the remote areas. On all surfaces some nitrogen was found although generally in lower concentrations that those of carbon and oxygen. There is no doubt that a fraction of the injected nitrogen is retained. The sticking mechanisms and the identification of a chemical form of nitrogen-containing compounds deserves further studies taking into account that the injected small amount of 15N was not desorbed or isotopically exchanged despite over 100 plasma pulses before the end of campaign. All types metals (Be, W, Inconel components) eroded from PFC are transported to locations with no plasma line-of-sight. This poses questions regarding their transfer mechanism whether this simple transport of neutrals sputtered from the target plates or chemical transport. While BeDx compounds can be responsible for the beryllium transport, the chemical transfer of tungsten would require oxides. The presence of W oxides on PFC surfaces has been detected on several occasions. A conclusive answer regarding the transport could probably be determined a dedicated marker experiment with the 18O2 rare isotope.
18
2.3.2. Tracer techniques for the assessment of material migration Tracer techniques are widely used in science and industry to determine for instance flows, reaction rates and mechanisms, i.e. to reveal decisive steps in studied processes. In the field PWI such techniques have been used to study material migration which is decisive for lifetime of PFC and fuel inventory. The term ‘‘tracer’’ denotes species introduced on purpose to plasma edge, either by puffing exotic (e.g. WF6) or rare isotope (e.g. 15N2, 18O2) gases, by ablating tracer material using lasers or, by exposing marker tiles coated with sandwich-type layers of heavy and light elements. Experimental approach to the determination of erosion phenomena involves the combined application of spectroscopy, mass spectrometry, surface probes for ex-situ studies and tracer materials. For many years a programme dedicated to testing of PFC and studies of erosion processes by tracer techniques have been carried out at the TEXTOR tokamak, especially using methane labeled with carbon-13. Isotope- based tracers are impractical in the case of heavy metals (e.g. W) both because of a tremendous cost of isotopic enrichment and the lack of quantitative surface analysis techniques capable of reliable distinguishing heavy isotopes of similar masses (resolution problem). Instead, another refractory metal being a proxy to tungsten can be deposited as a coating on a W component or a volatile compound may be used. The intention of this work is to provide a brief overview of results regarding: (a) the detection of nitrogen co-deposition after plasma edge cooling with a 15N marker; (a) mobility of heavy metal following the injection of volatile hexa-fluorides. MoF6 was injected to test whether such compound can be considered as a tracer for W migration studies in a device with a tungsten wall.
The experimental programme was carried out at TEXTOR In operation till December 2013, it was a mission-oriented tokamak with a focus on PWI processes and material tests. Intense tungsten testing done over the years in TEXTOR has led to a remarkable W background level on graphite PFC. As a consequence, molybdenum was selected as a proxy of tungsten to study high-Z migration. Molybdenum hexafluoride (MoF6) was puffed locally from the bottom of the machine through an inlet in a polished graphite plate installed on a test limiter. The entire procedure was similar to that used for the WF6 injection experiments. The puffing was done during 31 neutral beam injection (NBI) heated discharges: 14.2 x1020 Mo atoms in total, injection at 0.8–1.8 s into the discharge. The heating phase lasted 4.5 s in the period 0.8–5.3 s into the discharge. Nitrogen-15 was simultaneously (timing: 1–5 s) puffed into the plasma edge: 5.3x 1021 15N atoms in 22 pulses. The experiment was done on the last operation day of TEXTOR and followed by the dismantling of all in-vessel components making them available for ex-situ studies. Samples of dust and debris were also collected. Local and core spectroscopy measurements during the experiment with MoF6 were performed. Exposed probes and PFC tiles were analysed with a number of surface sensitive techniques.
Figure 2.3.2-1. Temporal evolution of MoI and FII lines during the MoF6 injection.
19
Figure 2.3.2-1 shows a typical evolution of the local spectroscopy signals of MoI and FII. The timing of injection marked in the graph corresponds to the opening and shutting of the valve located more than one meter from the vacuum vessel. As a result, the heavy gas enters the torus with a delay of approx. 0.45 s. This explains the shift of the corresponding Mo and F signals in comparison to the start/end of the puff. The shape of both traces is the same and the intensity decreases sharply after the end of injection. Also during the pre-injection phase no traces of both elements are detected locally.
Figure 2.3.2-2. Graphite plate coated by a deposit after the MoF6 puffing (a); concentration profiles of species in the deposit recorder with WDS along the line located 2 mm above the hole.
Figure 2.3.2-2(a) shows the graphite plate through which MoF6 was injected. A shiny metallic layer indicates local Mo deposition. The study with electron microprobe was done along more than 20 lines both near the injection hole and at distant locations. Concentrations of Mo, F, N and O along the line located 2 mm above the hole are plotted in Figure 2.3.2-2(b). Mo distribution is very broad, while fluorine is detected in a narrower band near the injection hole. The content of both elements in that region reaches 5x1022 m-2 but around 100 µm from the hole edge Mo formed a 7 µm thick deposit, as determined with surface profilometry. This local content may reach even 3–4x1023 m-2 Mo atoms. Electron microprobe measurements with a 30 keV electron beam have identified around 2.5x1023 Mo atoms m-2, while the F content in that location is 5x1022 m-2. Such massive Mo deposition around the hole is related to an immediate decomposition of MoF6. The proof of decomposition and Mo deposition was also found in the gas inlet system. The integrated amount of Mo detected on the plate corresponds to less than 2% of the amount which left the calibrated volume in the gas matrix. Other elements on the plate are in smaller quantities: 1x1022 m-2 O atoms and 1.5x1022 m-2 N atoms. It is not possible to determine how much oxygen has been co-deposited but the result certainly shows that the Mo layer is not fully oxidised (MoO2 and MoO3) when exposed to atmospheric air. High nitrogen content suggests direct codeposition of the species from plasma. Electron microprobe does not distinguish 14N and 15N isotopes. Relevant measurements were performed with ToF-HIERDA on the graphite and titanium catcher plates attached vertically to the side of the test limiter. Carbon, as expected, is the main constituent which is accompanied by two nitrogen isotopes, 14N (1.2x1021 m-2) and 15N (0.97x1021 m-2), distributed uniformly through the layer indicating their deposition from plasma. The amounts of 15N on the ALT-II limiter tiles were in the range: 0.9–4.5x1020 m-2 with the maximum on the tiles near the MoF6 injection. Oxygen profile is strongly peaked at the surface most probably due the ad-/absorption from air. There is also boron and a clear trace of helium. Fluorine content is small (around 4x1019 m-2) showing that the retention of that reactive element (with metals and hydrogen) does occur but fortunately it is not high.
20
2.3.3. Microanalysis of divertor surfaces in JET with ITER-like wall The erosion and migration of first wall material gives rise to several critical plasma-surface interactions issues for ITER and for other future big and high duty cycle fusion devices. One of them is how deuterium is trapped at surfaces in the divertor. It has been found that the
surface roughness and microscopic inhomogeneity is of considerable importance. This is being studied in JET using post mortem ion beam analysis methods with microscopic resolution (μ-IBA), in combination with other microscopic techniques and analysis methods. Figure 2.3.3-1 shows a poloidal cross section of the JET-ILW divertor indicating positions where surface samples were taken during the first shut down following operations with ILW.
Figure 2.3.3-2 shows an example of microanalysis of the W coated surface at divertor position 6/3. SEM images are overlaid with elemental maps if Be and D. The inhomogeneous spatial distributions are related to surface roughness features and complicate the interpretation of any analysis methods with poor spatial resolution. Taking advantage of the
capabilities of spatial resolution however, ambiguities can be resolved. An example is shown in figure 2.3.3-3. The bottom panels show elemental depth profiles from secondary ion mass spectrometry (SIMS) without microscopic resolution, where it is unclear what the effects of surface roughness and lateral inhomogeneity may be. E.g. it is not clear if the beryllium is confined within the deposited layer at the surface, or if it has penetrated somehow into the W substrate. By carefully selecting μ-IBA data from representative, microscopically flat regions, it was possible to construct the quantitative depth profiles in the top panels, to show firstly that the Be was confined to the deposited layer and did not penetrate significantly into the substrate, secondly that the deuterium had penetrated far into the W substrate and had been preferentially trapped at the W/Mo and Mo/W interfaces in the substrate coatings.
2.3.4. A 10Be marker experiment on beryllium migration in JET-ILW An isotopic marker experiment has been designed to study the migration of beryllium from the main chamber to the divertor in JET with ITER-like wall. One of the beryllium tiles at the inner wall in JET has been enriched with 10Be through irradiation with thermal neutrons in a
21
fission reactor. The tile was installed in JET in 2011 and exposed to the plasma throughout the first two periods of operations with ITER-like wall.
Figure 2.3.4-1
Using the extremely sensitive accelerator mass spectrometry (AMS) method, the 10Be content in re-deposited beryllium layers all over the JET has been investigated after the first JET shut down in summer 2012. Several numerical models exist for the materials migration and mixing problem in JET with ITER-like wall and the marker experiment is designed for comparison with the numerical models. Figure 2.3.4-1 shows a representation of the inboard wall in JET, with the 16 vertical beams of inner wall guard limiters (IWGL). The marker tile position is indicated with a cross and the red line shows the direction of a typical magnetic field line at the plasma boundary. Figure 2.3.4-2 shows the distribution of Be deposited in the divertor, and the marker fraction in the deposit. Interesting results are firstly the poloidal marker distribution, which is more uniform than predicted for materials migration in the diverted phase of the plasma discharges, secondly the high absolute marker level. One interpretation is that there may be previously unexpected channels for transporting Be to the divertor during the limiter phases of the plasma discharges. The analysis continues, with data from the second shut down with ILW.
Figure 2.3.4-2.
22
2.4. Theoretical fusion plasma physics T. Johnson, T. Hellsten, E. Tholerus (PhD student), P. Vallejos (PhD student) In collaboration with the EUROfusion member CCFE, CEA, CIEMAT, ENEA, EPFL, IPP, IST, LPP-ERM-EMS, VTT, Wigner RCP, with PPPL, F4E and ITER-IO. The theoretical fusion plasma physics group 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 is particularly active in developing numerical models and codes for studies of ion cyclotron resonance heating, ICRH, and in validating them against experiments. This work is well integrated into the European fusion program through participation in: the Integrated Tokamak Modelling Task Force, and the exploitation of the JET facility. The main codes developed by the group are PION, FIDO, SELFO, SELFO-light, RFOF and FOXTAIL. PION was the first self-consistent code for modeling ICRH and NBI heating using simplified models and is used routinely at JET. For more advanced modelling the Monte Carlo code FIDO was developed to calculate the distributions of resonant ions taking into account effects of finite orbit width, RF-induced spatial transport and interaction between MHD waves and fast ions. By coupling the FIDO and wave code LION the self-consistent ICRH code SELFO was developed. Recently the SELFO-light code has been developed which is similar to LION, but with a more advanced wave solver.
2.4.1. Exploitation of JET Analysis has been performed of JET experiments on “Fusion Product Studies”, which were conducted under the scientific coordination of Sergei Sharapov (CCFE) and Torbjörn Hellsten (KTH). In these experiments third harmonic ICRF were used to accelerate deuterons to MeV energies, which strongly enhanced the D-D fusion yield and also produced unusually high intensities of several other reactions that is of interest for diagnosing a high performance D-T plasma. The experiments were very successful. They provided e.g. a test of recent upgrades to the neutron camera detector system and the new gamma detectors. In addition, the FLR cut- off in the ICRF accelerated fast deuteron population was measured with high precision by the TOFOR neutron spectrometer and the gamma diagnostic BGO, which enabled detailed benchmarking with ICRF modelling. The results are illustrated in figure 2.4.1, comparing measured and modelled the distribution functions of ICRF accelerated ions. These experiments have resulted in a number of publications.
23
Figure 2.4.1: Comparison of measurements using TOFOR and BGO and numerical results using SPOT/RFOF and ASCOT/RFOF.
During 2015 experiments were performed on the optimization of ICRF heating in JET with the ITER-like wall. Considerable improvement in the coupled power was achieved by optimizing the gas puffing near the antenna. In addition, the KTH group has participated in the work on JET experiments M15-24 on “Target discharge for TAEs in DTE2 and fast particle physics in all scenarios” and M15-27 on “ICRH scenarios for DT”, as well as in the modelling activities “T15-01” on “DT scenario extrapolation”. In each of these projects, significant efforts have been made modelling several discharges and presenting the results during internal JET meeting.
2.4.2. Contributions to the MST1 Work Programme During 2015 experiments have been performed at AUG, under MST1, on the effect of ECRH on the stability of ICRF driven TAE modes. The results have shown that a dramatic change in the TAE stability is obtained when ECRH is switched on. The KTH group has performed extensive ICRF modelling of these discharges and the next step is to use the ICRF results in an analysis of the TAE stability. MST1 experiments were also performed on the 3rd harmonics ICRF acceleration of deuterium beam ions at AUG. Similar experiments have already been performed at JET, but never before at AUG. During the experiments a clear acceleration of deuterium beam ions were observed using the NPA. Modelling has been performed of these experiments using both PION and SELFO.
2.4.3. Integrated modelling within the EUROfusion-WPCD The group participates in integrated modelling within EUROfusion, where Thomas Johnson is Task Coordinator for the Heating and Current Drive activities. During 2015 the main contributions have been to work on the installation of the ITM infrastructure at JET, within the ETS4JET project, including the development and testing of new ITM standard to ensure platform independence and to enable public releases of Kepler-actors based on physics codes. In addition, a new implementation of the NBISIM code has been developed to facilitate a simplified transition to IMAS. Also ICRF wave codes have been benchmarked using the WPCD infrastructure.
24
2.4.4. Non-linear wave-particle interactions The 1D code for studying the interactions between Alfven eigenmodes and fast ions that was developed during 2014 has been extended to include a quasilinear model. The new code has been used to study the validity of the quasilinear approximation for bump-on-tail systems with phase decorrelation. In addition, a new code, FOXTAIL, for the interactions of fast ions and MHD modes in full tokamak geometry has been developed. The novelty of this code is the use of canonical angles from an action-angle coordinate system of the unperturbed motion. In this system the time stepping for modelling MHD interactions is limited only by the time- scale of the toroidal precession. The first paper published on the code focusses on the mathematical formulation and the numerical algorithms employed.
Figure 2.4.2: Bounce and precesion frequencies in ITER for particles with the magnetic moment 100 keV/T, as calculated using the new FOXTAIL code.
2.4.5. Iterative approach to spatial dispersion The wave equation for ICRF waves include spatial dispersion that in general should be modelled by an integral equation. To reduce the computational effort, while retaining the complete integral operator, an iterative method has been developed in which an elliptic differential equation is solved multiple times. To separate the different wave length within the wave field a wave-let transform is used, fig 2.4.3. Localised spectral representations could provide a significant speed up of the evaluation of the dielectric tensor and is particularly well suited when a quasi-homogeneous approach to spatially dispersive responses are applicable.
Figure 2.4.3. Wavelet spectrum of solution wave equation in a dispersive media, where the dispersion is treated iteratively.
25
2.5. Computational methods for fusion plasmas J. Scheffel In collaboration with: PhD student K. Lindvall, KTH (since November 2015) Prof H. Nordman, Chalmers University of Technology In light of the need for efficient multiple time and spatial scale simulations in physics, it is surprising that time-spectral methods for partial differential equations have not yet been studied and explored systematically. For example, turbulence in fusion plasmas at high Reynolds or Lundquist numbers are presently addressed by gyrokinetic codes that are allocated millions of CPU hours for parallel processing on supercomputers. If there exists a new avenue that may alleviate the requirements on computer power for these crucial problems, it is certainly of great importance to explore it. The present project concerns a time-spectral approach named the Generalized Weighted Residual Method (GWRM). In this method, traditional finite time differencing is replaced by a spectral representation of the time domain by use of Chebyshev polynomials. The computed solutions are truncated, approximate semi-analytical Chebyshev polynomial series valid for all time, spatial and, optionally, physical parameter domains and are immediately tractable for mathematical analysis. Scalings with physical parameters are thus obtainable in a single computation. The GWRM has so far been successfully employed for solving an extensive set of problems, and time is found ripe for application to two central problems in magnetic confinement fusion physics. These are drift wave turbulence in tokamaks and kinetic effects on resistive, pressure driven instabilities in reversed-field pinch (RFP) plasmas. A short background follows. Drift wave turbulence is driven by density or temperature gradients. This universal phenomenon is inherently limiting performance in the tokamak. It can be studied in a host of models, reaching from two-fluid to advanced, nonlinear gyrokinetic models. Turbulence driven by pure ITG modes in 2D geometry, neglecting trapped electrons, is to be studied using the present GWRM model and compared to results obtained with the fluid turbulence code developed at Chalmers University. Particular emphasis will be on accuracy and efficiency and subsequent development of more advanced turbulence models within the GWRM. As a next step, extension to kinetic drift wave turbulence will be attempted. Kinetic (finite Larmor radius) effects on the stability of resistive g-modes in the RFP is a long standing crucial problem for the reactor potential of the concept in the sense that the energy confinement required depends on the non-existence or suppression of these modes. In recent studies, it has been shown that resistive g-modes can be stabilised by thermal conduction effects, but only at low beta. At high beta and Lundquist numbers, however, it remains to be proven that the RFP is stable against these modes. No RFP modelling has so far taken into account the (stabilising) effect of finite Larmor radius on this phenomenon. Using our previous experience from the Vlasov-fluid model, being well suited for the present problem, a GWRM initial-value code is to be designed for determining the scaling of resistive g-mode growth rates in hot, high beta RFP plasmas. These results are relevant for the experimental RFP device Extrap T2R at KTH in Stockholm. The purpose and aim of the project is a continuing effort in addressing the above two problems, together with the following related computational plasma physics problem.
26
General and efficient GWRM modules need be developed and benchmarked. Solution of the two advanced problems mentioned above will necessarily advance the GWRM to higher flexibility and capacity. This will lead to a versatile formulation that could be applied to a large set of problems. The present study should thus provide valuable information on limits and potential of the GWRM. Now we turn to results obtained during 2015. Linear drift waves Presently, the basis for a 2D GWRM code is being developed to determine the temporal evolution of nonlinear ITG and TE modes, using the physical model developed at Chalmers. For benchmarking, all linear terms have so far been implemented, using the parameters of the standard Cyclone base case and a 2D box, the sides of which are 50 ion Larmor radii. Preliminary results indicate very good agreement between theoretical and GWRM linear growth rates. GWRM: Breakthrough in spatial subdomain computations During 2015 we have found a method to drastically reduce the number of computational operations and the memory requirements when solving the algebraic equations for the Chebyshev coefficients that represent the linear or nonlinear system of partial differential equations relevant to the problem at hand. For guaranteed convergence, the GWRM requires simultaneous information from all spatial domains of the computational domain. A primitive algorithm would then simultaneously involve all the Chebyshev coefficient equations of all the subdomains when the corresponding Jacobian is inverted in each iteration. This would be costly; say that the number of Chebyshev modes in total is N – then memory requirements scale as N2 and computational complexity scales as N3, which would be prohibitive for advanced problems. In the new method, all physics equations are solved locally in each spatial subdomain for each iteration, whereas only the boundary condition equations, linking the domains, are solved globally. This procedure not only drastically reduces the number of global equations to be solved; it enables parallelization of the computations of each subdomain and reduces the algebraic complexity of the equations, with a strong reduction in required memory space and even more dramatic reduction in computational operations. In 2015 we also found that sparse matrix methods have strong potential for further reduction of the matrix operations. GWRM: Application to numerical weather prediction During a visit in 2015 to SMHI, The Swedish Meteorological and Hydrological Institute in Norrköping, Sweden, some interesting contacts for future collaboration in the field of numerical weather prediction (NWP) was made. Since meteorologists solve very similar partial differential equations as plasma physicists (essentially advanced Navier-Stokes equations) they face the same problems relating to efficiency. This is particularly critical for NWP since detailed, very advanced prognoses are to be computed and interpreted on time scales of a few hours. Of special interest is when the weather is such that it requires the solution of so-called chaotic equations, being extremely sensitive to initial values. Using a set of chaotic equations, being suggested by Lorenz in 1984, the GWRM has been compared in terms of convergence, accuracy and efficiency with standard methods that use finite differences for time-stepping. It is found that for these problems the GWRM solutions are optimal for time intervals being approximately a factor 100 longer than those used in the finite difference methods. We have during 2015 developed advanced GWRM solvers that employ automatic time interval optimization techniques. The accuracy and efficiency of the GWRM is found to be highly competitive, for example when repeated perturbations of an initial state need be computed for aggregating reliable scenarios. Results will be published in 2016.
27
2.6. Confinement physics L. Frassinetti, E. Stefanikova, M. Tendler in collaboration with JET and AUG researchers
2.6.1. Pedestal properties and confinement in JET Operations in JET have been resumed in autumn 2011 after the shutdown necessary to install the ITER-like wall (hereafter called ILW). Type I ELMy H-mode operation in JET with the ITER-like Be/W wall (JET-ILW) generally occurs at lower pedestal pressures compared to those with the full carbon wall (JET-C). In 2015, the study of the pedestal properties in JET with the new ILW has continued. The activity in this area has focused on investigating the role of collisionality and of the plasma current. Role of collisionality on the confinement in JET-ILW The baseline type I ELMy H-mode scenario with H=1 has been re-established in JET with the new tungsten divertor and beryllium main wall (JET-ILW) in 2011. Comparing carbon wall (JET-C) discharges in similar conditions, a degradation of the confinement has been observed in JET- ILW, with the reduction mainly driven by a lower pedestal pressure. The JET-ILW at low collisionality seems to reach a confinement comparable to the JET-C reaching H98≈1.0, as shown in figure 2.6.1. The present work has studied the confinement and the pedestal in a dimensionless collisionality scan in order to investigate the reason for the confinement improvement. The pedestal structure is significantly affected by the collisionality. Specifically, a reduction of the width with decreasing collisionality is observed. At low collisionality, the pedestal width is in god agreement with the expectations from the KBM prediction which expects wpe=0.076(βθ
ped)0.5. The high stored energy at low collisionality is achieved by both the increased pedestal pressure (50%) and by the increased core pressure (75%). The pedestal stability show an improvement with decreasing collisionality. The reason is not clear yet and it will subject of further investigation in 2016. It was not possible to find a perfect match between the present data set and JET-C, so a quantitative comparison cannot be done, but, qualitatively, JET-C and JET-ILW have similar trends with collisionality.
Figure 2.6.1 Confinement factor versus βN (a) and versus collisionality (b) in JET-ILW
28
Dependence of confinement and pedestal structure on plasma current in JET-ILW and comparison with JET-C High plasma current and field are necessary to develop high performance scenarios with high thermal energy (Wth>10MJ) and temperature. During the 2014 campaign, JET-ILW has reached 4.0MA in the baseline scenarios in quasi stationary-conditions, avoiding W accumulation and controlling the divertor heat loads. The stored energy in JET-ILW tends to be lower than in JET-C. However, comparable Wth is obtained at 2.5MA (but using different strike point positions).
This part of the work studies the role of the pedestal in the JET-ILW confinement of a current scan and discusses the differences with the JET-C. An increase of the pedestal energy with Ip is observed in both JET-ILW and JET-C. In JET-ILW, the increase of the pedestal stored energy is related to the pedestal density, while the pedestal temperature does not scale significantly with Ip. In JET-C, both density and temperature increase with Ip. This difference could be related to the large gas fuelling in JET-ILW to control W accumulation. In the core, the low JET-ILW pedestal temperature (compared to JET-C plasmas obtained at similar current) produces low core temperature due to the profile stiffness. Moreover, the low density peaking due to the increased collisionality in JET-ILW reduces further the core contribution to stored energy. The pedestal width analysis has been performed at each Ip level for the shots with the highest H98, both in JET-C and JET-ILW. Figure 2.6.2(a) shows pedestal pe widths versus Ip. No clear trend with Ip is observed, but JET-C tends to have a lower pedestal width than JET-ILW. This might be related to the higher gas level used in JET-ILW [3]. Figure 2.6.2(b) shows the pedestal pe versus Ip and figure 2.6.2(c) shows the correlation of the
Figure 2.6.2. (a) pe pedestal widths vs Ip; (b) pe pedestal heights vs Ip; (c) grad pe vs Ip. Open symbols – JET-C, full symbols – JET-ILW.
Figure 2.6.3. Stability diagram at Ip = 4MA, for JET-C (a) and JET-ILW (b)
(a) (b) (c)
29
pedestal pressure gradient with Ip. A significant reduction, by a factor 2 or higher, for JET- ILW is observed. Pedestal pe widths from the selected subset of shots do not follow any obvious trend with Ip. But JET-ILW reaches comparable or larger pedestal widths than JET- C. Also the pressure gradient at the pedestal is strongly reduced in JET-ILW with high Ip. This is in accordance with stability analysis from the MISHKA code which showed that at high Ip JET-ILW is far from the stability boundary. This is shown in figure 2.6.3
2.6.2. ELM behaviour in AUG The Type I ELM behaviour in ASDEX Upgrade with full W plasma facing components is studied in terms of time scales and energy losses for a large set of shots characterized by similar operational parameters but different nitrogen seeding rate and input power. ELMs with no nitrogen can have two typical behaviours that can be classified depending on their duration, the long and the short ELMs. The work shows that both short and long ELMs have a similar first phase, but the long ELMs are characterized by a second phase with further energy losses. This is shown in figure 2.6.4. The second phase disappears when nitrogen is seeded with a flux rate above 1022(e/s). The phenomenon is compatible with a threshold effect. The presence of the second phase is related to a high divertor/scrape-off layer (SOL) temperature and/or to a low pedestal temperature. The ELM energy losses of the two phases are regulated by different mechanisms. The energy losses of the first phase increase with nitrogen which, in turn, produces the increase of the pedestal temperature. So the energy losses of the first phase are regulated by the pedestal top parameters and the increase with nitrogen is due to the decreasing pedestal collisionality. The energy losses of the second phase are related to the divertor/SOL conditions. The long ELMs energy losses increase with increasing divertor temperature and with the number of the expelled filaments. In terms of the power lost by the plasma, the nitrogen seeding increases the power losses of the short ELMs. The long ELMs have a first phase with power losses comparable to the short ELMs losses. Assuming no major difference in the wetted area, these results suggest that (i) the nitrogen might increase the divertor heat fluxes during the short ELMs and that (ii) the long ELMs, despite the longer time scale, are not beneficial in terms of divertor heat loads
30
3. Education and research training 3.1. Basic and advanced level education The following courses were given in 2015: Basic level courses ED1100 Engineering Science (J. Scheffel) The progress of technology. Development in the physics, chemistry and computer sciences. About understanding and modelling nature. Units. Estimates. Graphical models. Mathematical models. Proportionality. Model fitting. Dimensional analysis. Simulation modelling. The computer tool MAPLE. The roles of the engineer and the technology user. ED1110 Vector Analysis (L. Frassinetti) Learning oriented course in vector calculus. The course is useful for further studies of electromagnetic theory, wave propagation, fluid mechanics, plasma physics, gas dynamics and the theory of relativity. EH1010 Project Course in Electrical Engineering Course in development of new technological systems. First year students are offered hands-on projects, primarily carried out at the lab. The students are also trained in project management and presentation techniques. Advanced level courses Fusion Plasma Physics provides Advanced level courses for the KTH Master Programme in Electro Physics. The programme is given in collaboration with the Space and Plasma Physics Lab and the Electromagnetic Engineering Departments at the EE School. The programme focuses on the foundations of electrical engineering such as electromagnetic fields and their interaction with matter. Physical principles, mathematical methods and numerical models make up the core of the programme, providing the tools and skills needed to describe electro technical processes and analyse complex systems and problems in the field. ED2200 Energy and Fusion Research (J. Scheffel, P. Brunsell) An introduction to fusion oriented plasma physics is given. The central areas of fusion research are emphasised. The progress of fusion research and its present state are discussed in the perspective of future power generation. ED2210 Electromagnetic waves in Dispersive Media (T. Hellsten, T. Jonsson) The course introduces students to methods of treating electromagnetic waves. The electromagnetic theory is described by Fourier transforms in space and time which is advantageous when treating propagation and emission of waves in dispersive, anisotropic media. ED2220 Experimental Fusion Plasma Physics (P. Brunsell) The course gives the student an opportunity to become familiar with basic experimental and diagnostic techniques used in magnetic confinement fusion plasma physics research. In
31
addition, the student will gain practical experience of using some diagnostics that are available at EXTRAP T2R and analysing real measurement data. ED2235 Atomic Physics for Fusion (H. Bergsåker) The purpose of this course is to make the student familiar with those aspects of atomic physics that are most important in fusion research. The focus of the course is on basic understanding of atomic collisions and applications in plasma modelling, plasma diagnostics and plasma surface interactions. Much of the course content is applicable also in other contexts in plasma processing and technology, ion implantation and radiation effects. ED2246 Project in Fusion Physics (P. Brunsell) The student will learn about practical experimental research work by carrying out a small research project. The projects are performed in a real research laboratory environment; the EXTRAP T2R fusion research facility at the Alfvén Laboratory in KTH. The student will engage in a project that also leads to a more in-depth understanding of some common fusion plasma diagnostics methods. Degree projects The following degree projects were completed in 2015: Bachelor Degree in Electrical Engineering (15 credits) Anders Enquist and Linus Härenstam Nielsen Radio frequency heating of tokamak plasma and the importance of chaos Master Degree in Electrical Engineering (30 credits) Tomas Lycken Modelling of collisionless alpha-particle confinement in tokamaks 3.2. Research training Since 2011, the research training in fusion plasma physics and technology at the department is part of the Plasma Physics track of the new E2DOC Doctoral Program of the School of Electrical Engineering, as one of five research studies tracks.
The Department of Fusion Plasma Physics is a member of FuseNet since a few years. This is an organization for increasing, enhancing and broadening fusion science and technology training in Europe. In particular, our PhD students are taking part in the FuseNet PhD Events, the aims of which are to enable students to disseminate their research, develop a network of contacts and learn from each other's experiences. Participation in two events allows the students to apply for a European Fusion Doctorate Certificate.
As of late 2015, there were altogether 12 PhD students at the Department.
32
Graduate level courses In E2DOC, courses for a PhD Degree should cover 75-120 credits whereas thesis work should cover 120-165 credits, altogether 240 credits. For a Licentiate degree, courses should cover 45-60 credits and thesis work 60-75 credits, adding up to 120 credits. Courses at the advanced undergraduate level may be included, insofar as they are not requirements for admission. For licentiate and PhD degrees at most 15 credits or 30 credits from undergraduate courses may be included, respectively. The main institutional work carried out by the PhD students is within teaching. All students enrolled into the new E2DOC Doctoral Programme should take four general skills courses, amounting to 10 credit points, in the topics of oral and written communication, pedagogics, research methodology and research ethics: • LH200V Basic Communication and Teaching • AK3014 The Theory and Methodology of Science – Minor Course • AK3015 The Sustainable Scientist • DS3103 Introduction to Scientific Writing for Doctoral Students The following basic graduate level courses in the subject area of fusion plasma physics are recommended (course responsible teacher in parentheses): • ED3220 Motion of Charged Particles, Collision Processes and Basis of Transport Theory, 8 credits (T. Jonsson) • ED3230 Magnetohydrodynamics, 8 credits (J. Scheffel) • ED3240 Plasma waves I, 8 credits (T. Jonsson) • ED3260 Fusion Plasma diagnostics, 8 credits (P. Brunsell) The following advanced courses are recommended: • ED3305 Magnetohydrodynamics, advanced course, 6 credits (J. Scheffel) • ED3320 Fusion research, 8 credits (T. Jonsson) Recent retirements of professors have led to a redistribution of responsibilities for the courses above. Some courses, like Transport Theory, need be reorganized in order to better match the present research of the department. EFDA Training programmes From 2012, Fusion Plasma Physics participates in the BeFirst program for the four-year period including 2015. The aim of the BeFirst training program is the education and training of early stage researchers in the field of "Plasma-Facing Components". It will be achieved by performing cooperative research and training programs in the area "Beryllium for the first wall". The program is practically oriented and will be realized through performing a set of practical exercises in laboratories experienced in beryllium handling and analysis, but will also include courses providing a theoretical background. The program is jointly undertaken by the Euratom Associations KIT, CEA, ENEA, FZJ, AEUL, and finally VR, represented by Department of Fusion Plasma Physics.
33
Licentiate degrees 2015 Agung Chris Setiadi Thesis title: Model predictive control of resistive wall modes in the reversed-field pinch Main supervisor: Per Brunsell Date of licentiate defence: 12 June 2015 Faculty opponent: Docent Cristian Rojas, KTH Abstract The reversed-field pinch (RFP) is a magnetic confinement fusion (MCF) device. It exhibits a variety of unstable modes that can be explained by magnetohydrodynamic (MHD) theory. A particular unstable mode that is treated in this work is the resistive wall mode (RWM), which occurs when the shell of the device has finite conductivity. Application of control engineering tools appears to be important for the operation of the RFP. A model-based approach is pursued to stabilize the RWM. The approach consists of experimental modeling of the RWM using a class of system identification techniques. The obtained model is then used as a basis for Mode Predictive Control (MPC) design. The MPC employs the model to build predictions of the system and find a control input that optimizes the predicted behavior of the system. It is shown that the formulation of the MPC allows the user to incorporate several physics relevant phenomena aside from the RWMs. The results are encouraging for MPC to be a useful tool for future MCF operation. Emmi Tholerus Thesis title: The dynamics of Alfvén eigenmodes excited by energetic ions in toroidal plasmas Main supervisor: Prof em. Torbjörn Hellsten Date of licentiate defence: 22 May 2015 Faculty opponent: Dr Lars-Göran Eriksson Abstract Experiments for the development of fusion power that are based on magnetic confinement deal with plasmas that inevitably contain energetic (nonthermal) particles. These particles come e.g. from fusion reactions or from external heating of the plasma. Ensembles of energetic ions can excite plasma waves in the Alfvén frequency range to such an extent that the resulting wave fields redistribute the energetic ions, and potentially eject them from the plasma. The redistribution of ions may cause a substantial reduction heating efficiency, and it may damage the inner walls and other components of the vessel. Understanding the dynamics of such instabilities is necessary to optimize the operation of fusion experiments and of future fusion power plants. A Monte Carlo model that describes the nonlinear wave-particle dynamics in toroidal plasma has been developed to study the excitation of the abovementioned instabilities. A decorrelation of the wave-particle phase is added in order to model stochasticity of the system (e.g. due to collisions between particles). Based on the nonlinear description with added phase decorrelation, a quasilinear version of the model has been developed, where the phase decorrelation has been replaced by a quasilinear diffusion coefficient in particle energy. When the characteristic time scale for macroscopic phase decorrelation becomes similar to or shorter than the time scales of nonlinear wave-particle dynamics, the two descriptions quantitatively agree on a macroscopic level. The quasilinear model is typically less computationally demanding than the nonlinear model, since it has a lower dimensionality of phase space. In the presented studies, several effects on the macroscopic wave-particle dynamics by the presence of phase decorrelation have been theoretically and numerically analysed, e.g. effects on the growth and saturation of the wave
34
amplitude, and on the so called frequency chirping events with associated hole-clump pair formation in particle phase space. Several effects coming from structures of the energy distribution of particles around the wave-particle resonance has also been studied.
Armin Weckmann Thesis title: Material migration in tokamaks: Studies of deposition processes and characterisation of dust particles Main supervisor: Marek Rubel Date of licentiate defence: 15 December 2015 Faculty opponent: Docent Daniel Primetzhofer, Uppsala University Abstract Thermonuclear fusion may become an attractive future power source. The most promising of all fusion machine concepts is the tokamak. Despite decades of active research, still huge tasks remain before a fusion power plant can go online. One of these important tasks deals with the interaction between the fusion plasma and the reactor wall. This work focuses on how eroded wall materials of different origin and mass are transported in a tokamak device. Element transport can be examined by injection of certain species of unique and predetermined origin, so called tracers. Tracer experiments were conducted at the TEXTOR tokamak before its final shutdown. This offered an unique opportunity for studies of the wall and other internal components: For the first time it was possible to completely dismantle such a machine and analyse every single part of reactor wall, obtaining a detailed pattern of material migration. Main focus of this work is on the high-Z metals tungsten and molybdenum, which were introduced by WF6 and MoF6 injection into the TEXTOR tokamak in several material migration experiments. It is shown that Mo and W migrate in a similar way around the tokamak and that Mo can be used as tracer for W transport. It is further shown how other materials - medium-Z (Ni), low-Z (N-15 and F), fuel species (D) - migrate and get deposited. Finally, the outcome of dust sampling studies is discussed. It is shown that dust appearance and composition depends on origin, formation conditions and that it can originate even from remote systems like the NBI system. Furthermore, metal splashes and droplets have been found, some of them clearly indicating boiling processes.
35
4. Personnel Professor Per Brunsell (Department Head) Marek Rubel Jan Scheffel Professor, emeritus James Drake
Torbjörn Hellsten Bo Lehnert
Michael Tendler Associate Professor Henric Bergsåker Lorenzo Frassinetti Thomas Johnson Researcher Per Petersson Administrator Emma Geira Engineer Håkan Ferm PhD student Richard Fridström
Alvaro Garcia Carrasco M. Waqas M. Khan
Kristoffer Lindvall Stefan Schmuck
Agung Chris Setiadi Estera Stefániková
Petter Ström Emmi Tholerus Pablo Vallejos Olivares Armin Weckmann Yushan Zhou
36
5. Income & Expenditure The accounts for Fusion Plasma Physics for the year 2015 are summarized in the table below: Income & Expenditure 2015 kSEK Income KTH 11 163 Undergraduate education (GRU) 1 419 Research and research training (FOFU) 9 744 External research grants 12 049 Swedish Research Council (VR) 3 916 European Framework Programmes (Euratom) 6 547 Swedish Energy Authority (STEM) 1 522 Other external income 64 Other external income 55 Total 23 267 Expenditure Salary 12 353 Travel 1 375 Equipment 317 Operation 726 Rent 4 069 KTH central costs 3 669 Total 22 509 Result 758
Bibliography
[1] M. I. Airila, A. Jarvinen, M. Groth, P. Belo, S. Wiesen, S. Brezinsek, K. Lawson, D. Borodin, A. Kirschner, J. P. Coad, K. Heinola, J. Likonen, Marek Rubel, and A. Widdowson. Preliminary Monte Carlo simulation of beryllium migration during JET ITER-like wall divertor operation . Journal of Nuclear Materials, 463:800–804, 2015.
[2] G. Arnoux, J. Loenen, B. Bazylev, Y. Corre, G. F. Matthews, I. Balboa, M. Clever, R. Dejarnac, S. Devaux, T. Eich, E. Gauthier, Lorenzo Frassinetti, J. Horacek, S. Jachmich, D. Kinna, S. Marsen, Ph. Mertens, R. A. Pitts, M. Rack, G. Sergienko, B. Sieglin, M. Stamp, and V. Thompson. Thermal analysis of an exposed tungsten edge in the JET divertor. Journal of Nuclear Materials, 463:415–419, 2015.
[3] A. Baron-Wiechec, E. Fortuna-Zalesna, J. Grzonka, Marek Rubel, A. Widdowson, C. Ayres, J. P. Coad, C. Hardie, K. Heinola, and G. F. Matthews. First dust study in JET with the ITER-like wall : sampling, analysis and classification. Nuclear Fusion, 55(11), 2015.
[4] A. Baron-Wiechec, A. Widdowson, E. Alves, C. F. Ayres, N. P. Barradas, S. Brezin- sek, J. P. Coad, N. Catarino, K. Heinola, J. Likonen, G. F. Matthews, M. Mayer, Per Petersson, Marek Rubel, W. van Renterghem, and I. Uytdenhouwen. Global erosion and deposition patterns in JET with the ITER-like wall. Journal of Nuclear Materials, 463:157–161, 2015.
[5] Henrik Bergsaker, Igor Bykov, Per Petersson, G. Possnert, J. Likonen, S. Koivu- ranta, J. P. Coad, W. Van Renterghem, I. Uytdenhouwen, and A. M. Widdowson. Microscopically nonuniform deposition and deuterium retention in the divertor in JET with ITER-like wall. Journal of Nuclear Materials, 463:956–960, 2015.
[6] R. Bilato, N. Bertelli, M. Brambilla, R. Dumont, E. F. Jaeger, Thomas Johnson, E. Lerche, O. Sauter, D. Van Eester, and L. Villard. Status of the benchmark activity of ICRF full-wave codes within EUROfusion WPCD and beyond. In RADIOFRE- QUENCY POWER IN PLASMAS :, number 1689 in AIP Conference Proceedings, 2015.
[7] S. Brezinsek, A. Widdowson, M. Mayer, V. Philipps, P. Baron-Wiechec, J. W. Co- enen, K. Heinola, A. Huber, J. Likonen, Per Petersson, Marek Rubel, M. F. Stamp, D. Borodin, J. P. Coad, Alvaro Garcia Carrasco, A. Kirschner, S. Krat, K. Krieger, B. Lipschultz, Ch. Linsmeier, G. F. Matthews, and K. Schmid. Beryllium migration in JET ITER-like wall plasmas. Nuclear Fusion, 55(6), 2015.
[8] Igor Bykov, Henric Bergsaker, Per Petersson, Jari Likonen, G. Possnert, and C. Wid- dowson. Combined ion micro probe and SEM analysis of strongly non uniform
deposits in fusion devices. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 342:19–28, 2015.
[9] Igor Bykov, Henrik Bergsaker, G. Possnert, K. Heinola, J. Miettunen, M. Groth, Per Petersson, A. Widdowson, and J. Likonen. Materials migration in JET with ITER-like wall traced with a Be-10 isotopic marker. Journal of Nuclear Materials, 463:773–776, 2015.
[10] C. D. Challis, J. Garcia, M. Beurskens, P. Buratti, E. Delabie, P. Drewelow, Lorenzo Frassinetti, C. Giroud, N. Hawkes, J. Hobirk, E. Joffrin, D. Keeling, D. B. King, C. F. Maggi, J. Mailloux, C. Marchetto, D. McDonald, I. Nunes, G. Pucella, S. Saarelma, and J. Simpson. Improved confinement in JET high beta plasmas with an ITER-like wall. Nuclear Fusion, 55(5), 2015.
[11] I. T. Chapman, J. P. Graves, M. Lennholm, J. Faustin, E. Lerche, Thomas Johnson, and Simon Tholerus. The merits of ion cyclotron resonance heating schemes for sawtooth control in tokamak plasmas. Journal of Plasma Physics, 81(06), 2015.
[12] L. Cherigier-Kovacic, Petter Strom, and F. Doveil. Electric field induced lyman- Emission (EFILE) diagnostic for electric field measurements. In Proceedings of Sci- ence :. Proceedings of Science (PoS), 2015.
[13] L. Cherigier-Kovacic, Petter Strom, A. Lejeune, and F. Doveil. Electric field induced Lyman-alpha emission of a hydrogen beam for electric field measurements. Review of Scientific Instruments, 86(6), 2015.
[14] J. Citrin, J. Garcia, T. Gorler, F. Jenko, P. Mantica, D. Told, C. Bourdelle, D. R. Hatch, G. M. D. Hogeweij, Thomas Johnson, M. J. Pueschel, and M. Schneider. Elec- tromagnetic stabilization of tokamak microturbulence in a high-beta regime. Plasma Physics and Controlled Fusion, 57(1):014032–, 2015.
[15] J. W. Coenen, G. Arnoux, B. Bazylev, G. F. Matthews, A. Autricque, I. Bal- boa, M. Clever, R. Dejarnac, I. Coffey, Y. Corre, S. Devaux, Lorenzo Frassinetti, E. Gauthier, J. Horacek, S. Jachmich, M. Komm, M. Knaup, K. Krieger, S. Marsen, A. Meigs, Ph. Mertens, R. A. Pitts, T. Puetterich, M. Rack, M. Stamp, G. Sergienko, P. Tamain, and V. Thompson. ELM-induced transient tungsten melting in the JET divertor. Nuclear Fusion, 55(2), 2015.
[16] J. W. Coenen, G. Arnoux, B. Bazylev, G. F. Matthews, S. Jachmich, I. Balboa, M. Clever, R. Dejarnac, I. Coffey, Y. Corre, S. Devaux, Lorenzo Frassinetti, E. Gau- thier, J. Horacek, M. Knaup, M. Komm, K. Krieger, S. Marsen, A. Meigs, Ph. Mertens, R. A. Pitts, T. Puetterich, M. Rack, M. Stamp, G. Sergienko, P. Tamain, and V. Thompson. ELM induced tungsten melting

Date post:	28-Mar-2022
Category:	Documents
Upload:	others
View:	3 times
Download:	0 times

Fusion Plasma Physics Annual Report 2015

Documents