You are on page 1of 31

Simulation of supra-thermal ions

during plasma edge instabilities in tokamak


fusion plasmas

Work conducted at: Max Planck Institute for Plasma Physics

Scientific work to achieve the degree


B.Sc. in Physics
at the Physics Faculty of the Technical University of Munich.

Supervised by Hon.-Prof. Sibylle Günter


Dr. Matthias Hoelzl
Max-Planck-Institute for Plasmaphysics (IPP)

Submitted by Hamza Ouerfelli


Steinickeweg 7
80999 München
+49 89 123 456 89
Submitted in Munich, on Date
Attachement I

Statement

I hereby declare that I have independently written the thesis I submitted and have used no other sources
and resources than those specified.

Place, Date, Signature


Contents

1 Abstract 5

2 Introduction 5
2.1 Plasma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Properties of plasma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3 Fusion energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.4 Tokamaks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.5 Heating methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.6 Plasma instabilities in tokamaks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.7 Divertor plasma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.8 Motivation for this research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

3 Theoretical background 11
3.1 Safety factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.2 Separatrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.3 H-mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.4 Edge-localized modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.4.1 Types of ELMs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.4.2 Peeling and ballooning modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.4.3 Prevention and control techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.5 Suprathermal particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.5.1 Banana orbits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.5.2 Pitch angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.6 MagnetoHydroDynamics (MHD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

4 Numerical methods 19
4.1 Nonlinear MHD code JOREK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.1.1 Discretization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.2 Particle model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.3 Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

5 Experimental observations 19
6 Tests and results 20
6.1 Simulation of ELMs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
6.1.1 Energies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
6.1.2 Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
6.1.3 Field lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
6.1.4 Q-profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
6.1.5 Density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
6.1.6 Particle content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
6.2 Simple tests for particle motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
6.3 Suprathermal particle dynamics in simplified ELM fields . . . . . . . . . . . . . . . . . . . 20
6.4 Suprathermal particle dynamics during realistic ELM crash . . . . . . . . . . . . . . . . . . 20

7 Discussion 20

8 Conclusions and outlook 20


1 Abstract

A large number of supra-thermal ions will be generated during the fusion of atomic nuclei in future fusion
experiments like ITER. During plasma edge instabilities (ELMs), these particles are partly lost from the
plasma. In present experiments, this process can be studied by creating supra-thermal ions via the
injection of energetic neutral particles into the plasma. The magneto-hydrodynamic code JOREK allows to
simulate ELMs in realistic tokamak plasmas. In the present project, the orbits of supra-thermal particles are
studied during an ELM to assess particle losses and compare qualitatively to experimental observations.

2 Introduction

2.1 Plasma

Known as the fourth fundamental state of matter, plasma consists of a gas/fluid of ions and free electrons
along with a neutral background (unless fully ionized) which shows a collective behavior and interaction
with electromagnetic fields [1]. Being the most abundant form of visible matter in the universe (stars,
intracluster medium, ...) and being fundamental for many technological applications (e.g. fusion reactors,
nanotechnologies, ...), the study of the properties of plasma is of great importance. The sun in particular,
and stars in general are a prime example of a naturally occurring plasma state. Replicating the nuclear
fusion taking place inside it (with a different confinement), is the main motivation for high-temperature
plasma physics and fusion research.

2.2 Properties of plasma

In contrast to classical states, there is no distinct phase transition from gas to plasma. In this case, it is
convenient to define the degree of ionization of a gas α, which strongly depends on the temperature T and
the ionization energy Wion [2]:
ni T 3/2
α= ∝ exp−Wion /T (1)
nn ni
One can speak of a plasma instead of a gas, when α is equal or greater than a few percents, since the
plasma properties discussed below start to dominate.
The typical density of a magnetic fusion hydrogen plasma is about ne ≈ 1020 m−3 and it is already fully
ionized for T ≤ 100eV . By decreasing the temperature below Wion ≈ 10eV , recombination dominates
and the plasma falls back into the gaseous state.

Debye shiedling The Coulomb Potentials of the relatively slow moving ions is shielded when the density
of the much faster free electrons in the plasma is sufficient large. As a consequence, their effective
q
0 T
"Debye-Hückel" potentials decreases exponentially on a length scale of the Debye length λD ≈ 7.43 e2 ne
Because of that a plasma of volume L3  λ3D appears to be quasi neutral on a macroscopic scale.

Magnetized plasma Since it contains charged particles, a plasma interacts with electromagnetic fields. If
this interactions heavily influences the motion of the particles, the plasma is called a magnetized plasma.
While particles can still move freely parallel to (homogeneous) magnetic field lines, its perpendicular
motion is restricted by the Lorentz forces. In fact, particles rotate with the gyration frequency ωc =

2mα T
qα B/mα , where α ∈ e, i, around the field lines on circles of the Lamour radius ρL = |qα |B in average.
The centre of such a circular trajectory is called the guiding centre.

~ perpendicular to the magnetic field drifts away the


Particle drifts The application of an external force F
position of the guiding centre perpendicular both to force and magnetic field. This drift velocity is given by:

1 F~ × B
~
~v = (2)
qα B 2
For an applied external electric field, the equation reduces to the charge independent "E-cross-B" drift:
~ ×B
E ~
~vE = (3)
B2
~ also leads to a drift:
An inhomogeneity of B

1 F~ × B~
~v = 2
(4)
qα B
The diamagnetic drift can be derived from the force due to a pressure gradient, the plasma pressure being
defined in terms of magnetohydrodynamics (see 3.9)

∇p × B~
~uD = − 2
(5)
qnB

The evolution of the guiding center is determined through the component of this equation along the
magnetic field:
dU0k dv~e
m = eEk − µ.∇.B − m~b. (6)
dt dt

The information shown in this section follows this reference [3].

2.3 Fusion energy

A fusion reaction, is a reaction in which multiple atomic nuclei are combined into larger, heavier ones. The
difference in the nuclear binding energy results in kinetic energy release for reaction products with atomic
number smaller or equal to iron. A fusion reaction occurs when two nuclei overcome the repelling Coulomb
barrier. The energy barrier is strongly reduced below the classical expectations due to quantum tunneling
effects [4].

The process occurs under influence of the dominant attractive force at sub-atomic length scales, the strong
interaction, once the spatial distance is sufficiently small. This close proximity can only be reached under
certain conditions, one of them is a sufficiently high velocity relative to each other, which directly relates to
the temperature for a collection of particles. The heating will result in a higher average particle velocity, or
even in the freeing of the electrons, creating a mixture of atomic nuclei and free electrons: a plasma.

The fusion probability is characterized by a highly temperature dependent function, the reaction rate,
derived from the cross section (or velocity averaged cross section). For an artificial fusion reactor on
earth, a reaction with the highest possible rate at the lowest temperature is looked for. The fusion reaction
of the hydrogen isotopes deuterium and tritium satisfies this requirement, and is therefore considered as
the primary option for generating fusion energy on earth.

The deuterium-tritium reaction releases an energy of Ef usion = 17.6M eV distributed over the products
(helium nucleus and a neutron) according to their mass ratio.

D + T → 4 He + n + 17.6M eV (7)

While Deuterium is abundant in nature, Tritium is extremely scarce due to radioactive decay, but can be
produced in a Lithium reaction with neutrons which is called Tritium breeding.

Although, at the atomic scale, the requirements of fusion are clear and feasible, many challenges arise
at the macroscopic scale. In fact, since no material can support such temperatures, an electromagnetic
confinement is needed. Although there are many concepts and designs available, this thesis focuses
on the most established confinement method and design: the tokamak, which uses magnetic fields in a
torus-like configuration and which is the design of the largest international fusion experiment ITER.

The expectations suppose that nuclear fusion is going to have several advantages over fission as a source
of energy: ample fuel supplies, increased safety since the there is no chain reaction and the required
conditions are external, reduced radioactivity in operation and less high-level nuclear waste. Nevertheless,
practical and economical problems arise when trying to reach the necessary combination of temperature,
pressure, and duration. Although research into fusion reactors began in the 1940s, until today, no design
has been able to produce more fusion power output than the electrical power input, defeating the purpose.
The management of the released neutrons, which over time degrade the wall materials of the apparatus,
is a second important issue affecting the process.
2.4 Tokamaks

Since the electrons and ions in the plasma are charged, they will gyrate because of the Lorentz force
around the field lines. Their perpendicular to the magnetic field motion is thus extremely restricted,
suppressing energy losses by many orders of magnitude, leading to a very effective confinement of
particles and their associated energy [5].

So what is needed is a configuration where magnetic field lines form nested surfaces, called flux surfaces.
The tokamak, which is one of several types of magnetic confinement devices to produce controlled thermonuclear
fusion power, operates according to this principle with a combination of externally generated toroidal
magnetic fields and poloidal magnetic fields generated by an electric current driven in the (conducting)
plasma. A schematic view of this configuration is shown in 20

Figure 1: Tokamak [6]

2.5 Heating methods

As explained in the previous parts, the plasma has to be heated to its operating temperature of greater than
10 keV (over 100 million degrees Celsius). After that, part of the generated energy will serve to maintain
the plasma temperature as fresh deuterium and tritium are introduced [7]. In current tokamak magnetic
fusion experiments, the produced fusion energy is still insufficient to maintain the plasma temperature.
That’s why external heating must constantly be supplied.

The most widely used heating methods are:

Ohmic heating or inductive mode Analogical to the heating in an electric light bulb or an electric heater
(hair dryer, fan heater, etc.), plasma can be heated by inducing a current through it, since it is also an
electrical conductor. The induced current providing most of the poloidal field is also a major source of initial
heating. The heat output depends on the resistance of the plasma and the voltage. As the temperature
rises, the electrical resistance of the plasma decreases and Ohmic heating becomes less effective. The
maximum temperature achievable by Ohmic heating in a tokamak is around 20-30 million degrees Celsius.
In order to reach higher temperatures, additional heating methods must be used.

Neutral-beam injection (NBI) Neutral particle injection means the introduction of fast atoms or molecules
into the magnetically confined plasma. The neutral beam is rapidly ionized once it enters the plasma,
its atoms are then caught by the magnetic field, transferring part of their energy to the plasma particles
by repeatedly colliding with them and thus raising the plasma temperature. Deuterium and tritium atoms
are particularly suitable as neutral particles, so that this plasma heating also contributes to fuel refilling. In
contrast to the Ohmic method, this process has no inherent energy (temperature) limitation, but the current
in the injectors limits its rate. NBI causes a comparably strong fast ion population and therefore also allows
to study effects relevant for future devices, where fusion-born fast particles will play an important role.

Radio-frequency heating High-frequency electromagnetic waves of suitable frequency (usually microwaves)


and polarization are generated by oscillators (often gyrotrons or klystrons) outside the torus, allowing their
energy to be transferred to the charged particles in the plasma, which in turn collide with other plasma
particles and thus increase the temperature. There are different techniques used depending on whether
the energy should first be transferred to the electrons (electron cyclotron resonance heating (ECRH)) or
the ions of the plasma (ion cyclotron resonance heating).

2.6 Plasma instabilities in tokamaks

Like all developed magnetic confinement concepts, tokamaks are also prone to various plasma instabilities
which limit their performance. For a plasma in equilibrium, a self-reinforcing perturbation or "mode" caused
by an arbitrary small force acting on it is called an instability [8]. The magnetohydrodynamics (MHD) model
is able to describe many instabilities classified as MHD instabilities. Similar to minimization of the potential
energy of systems in classical mechanics, small perturbations can be linearised and treated using that
energy principle. Numerical time integration of the equations is however often necessary for understanding
the non-linear saturation and dynamics of the instabilities. In the present work, focused on the effect of
Edge Localized Modes (ELMs) on supra-thermal ions, simulations are done using the JOREK [9] code.
In addition, large pressure gradient and large intrinsic current density in the pedestal( boottraps current)
gives rise to turbulence.
2.7 Divertor plasma

Divertors are devices used in ring-shaped tokamak and stellarator fusion reactors allowing the online
removal of the helium-4 fusion product (the so-called "fusion ash"), contaminants and waste material from
the fusion plasma while the reactor is operating. Thus controlling the buildup of fusion products in the
fuel, and removing impurities in the plasma that have entered into it from the vessel lining and which are
responsible for a loss of heat and a disturbance of the reaction. The divertor was already introduced during
the earliest studies of fusion power systems in the 1950s as a solution to those problems.

Figure 2: Divertor [10]

2.8 Motivation for this research

A large number of supra-thermal ions will be generated during the fusion of atomic nuclei in future fusion
experiments like ITER. During plasma edge instabilities (ELMs), these particles are partly lost from the
plasma. In present experiments, this process can be studied by creating supra-thermal ions via the
injection of energetic neutral particles into the plasma. The magneto-hydrodynamic code JOREK allows
to simulate ELMs in realistic tokamak plasmas. In the present project, the orbits of supra-thermal particles
are studied during an ELM to assess particle losses and compare to experimental observations. Supra-
thermal particles originate from heating methods (NBI) or from fusion processes. (e.g. ASDEX Upgrade
has essentially only ions from NBI, ITER/DEMO will have many fusion-born fast particles)

The confinement of these particles is important since the strong accumulation of Helium atoms in the
plasma and losses from edge region during ELMs are not desired.
3 Theoretical background

3.1 Safety factor

The safety factor q of a toroidal fusion power reactor is the ratio of the times a particular magnetic field
line travels around a toroidal confinement area’s "long way" (toroidally) to the "short way" (poloidally). This
number is an important factor for plasma stability. When a field line is traced, the safety factor q is given
by:
numberof toroidalturns m
q= = (8)
numberof poloidalturns n

Martin David Kruskal and Vitaly Shafranov, who first developed the concept [11], noticed that the plasma
in pinch effect reactors would be stable if q > 1. In general, a plasma with a high q at the boundary is more
stable. The q-profile depends on the equilibrium conditions and the form of the plasma. Flux surfaces with
rational values of q are prone to plasma instabilities.

3.2 Separatrix

In a dynamical system, a separatrix is a boundary between regions with distinct dynamical behaviours. In a
fusion reactor, the term refers to the surface seperating closed and open field lines, separating the toroidally
confined region from the region where field lines connect to material surfaces in divertor tokamaks [12] (see
Fig.3).

If the current in one (or more) shaping coils is increased to a high enough degree, one (or more) ’X-points’
can appear (see Fig. 4). It is a point in space at which the poloidal field has zero magnitude. The separatrix
intersects at the X-point, and defines the last closed flux surface (LCFS), since all flux surfaces external to
this surface are unconfined. By establishing an X-point and separatrix, the plasma edge is uncoupled from
the vessel walls, and exhausted heat and plasma particles are preferentially diverted towards a known
region of the vessel near the X-point, the divertor.

3.3 H-mode

Discovered by Friedrich Wagner in 1982 during neutral beam heating of the plasma at the German tokamak
ASDEX [13], the high-confinement mode, or H-mode, is an operating mode where the energy and particle
confinement is significantly increased compared to the standard regime, also called L-regime or L-mode
[14]. The H-mode features an Edge Transport Barrier (etb) in the plasma, just inside the separatrix,
where cross-flux-surface transport of energy and particles is reduced by a significant factor. The energy
Figure 3: Sketch of tokamak geometry, including separatrix

confinement time is thus strongly improved, and the achievable values of the triple product increased.

Prior to the H-mode’s discovery, all tokamaks operated in the low-confinement mode, which is characterized
by relatively large amounts of turbulence allowing the energy to escape the confined plasma. Moreover,
it was observed that as the heating power applied to an L-mode plasma increased, the confinement
decreased. However, it was discovered in 1982 on the ASDEX tokamak that if the heating power applied
using neutral beams was increased beyond a certain critical value, then the plasma spontaneously transitioned
into a higher confinement state, the H-mode, characterized by a significant reduction of edge plasma
turbulence and a high edge pressure gradient. Figure 5 shows a comparison of the pressure proles in H-
and L-mode plasmas.

Due to its improved confinement properties, H-mode quickly became the desired operating regime for most
tokamak reactor designs. Even if various aspects of the phenomenon have been successfully explained,
up to now, a self-consistent model explaining the L-H transition has not been found.
Figure 4: Schematic view of an X-point tokamak plasma

Nevertheless, the H-mode still faces some problems. Firstly, the plasma could easily fall back to the L-
mode because of an uncontrolled increase in density. Secondly, the improvement of the confinement also
includes the impurities, such as heavy ionized atoms from the walls, leading to a larger loss of energy by
radiation. Moreover, the steep density and temperature gradients at the edge of an H-mode plasma may
cause an additional kind of instability as discussed in the following section.

3.4 Edge-localized modes

One of the issues for tokamak plasmas like ITER are called the Edge Localized Modes (ELMs) [16] , on
which this thesis will focus. Edge localized modes are understood to refer to plasma instabilities that occur
at the edge of tokamak H-mode plasmas, which results in a massive emission of particles and energy.
The reasons for this are the steep pressure gradients in the peripheral region of the plasma, typical for
H-mode, and large current densities caused by the bootstrap current [17]. ELMs can periodically reduce
the plasma energy by up to 20% on a very short time scale. The sudden expulsion of such a large amount
of high energy particles can cause serious damage to the plasma facing components.The components of
the current devices can bear the heat, but an extrapolation to ITER shows that the energy would surpass
the 0.5M J/m2 energy limit on the divertor [18]. In this section, the phenomenology of the different ELM
Figure 5: Schematic comparison of the pressure profiles in L- and H-mode plasmas. The transition from
L-mode (black) to H-mode (red) leads to the formation of a transport barrier, where turbulence is
suppressed. Consequently, the edge pressure gradient increases, while leaving the core profile
mostly unchanged (pedestal). [15]

types is described, afterwards the physics behind them is explained. Finally, different ELM control and
suppression techniques are discussed.

3.4.1 Types of ELMs

ELMs can be assigned to different types, which differ in their cause and frequency [19]:

The type I ELMs can cause a pedestal energy loss of up to 20%. It describes an ELM phase during which
large amounts of energy burst out of the plasma at a moderate frequency of 10-100 Hz, which increases
with increasing heating power. The type II ELMs occur only at specific conditions that are hard to maintain,
at strongly shaped plasmas at high density. The energy per ELM is lower than during a type I phase which
still guarantees good confinement while controlling the impurities in the confined region. They are also
called grassy ELMs The type III ELMs occur at low amplitudes and high frequencies, which decrease with
increasing heating power. The quantity of energy is transferred through the edge transport barrier is rather
high leading to a lower confinement. They are also called small ELMs.

On the one hand, ELMs represent a way of freeing the plasma from impurities (ions, ...) through the
transport barrier, which is the case for type III ELM. On the other hand, the components involved, in
particular the divertor, are exposed to very high and spatially very focused power densities during an ELM
specially type I ELMs, which have to be avoided. Thus, it is necessary to either suppress ELMs completely
or to maintain a type III ELM regime. Type I ELMs are at the forefront of the investigations because they
have to be taken into special consideration when planning future fusion reactors.

By looking at the conditions under which ELMs occur, it is possible to learn more about them. ELMs occur
at the peeling-ballooning stability limit of the MHD theory and depend mainly on the edge current density
and the pressure gradient at the edge.

3.4.2 Peeling and ballooning modes

As a consequence of the operation in H-mode, short periodic plasma instabilities occur, which reduce
the plasma energy and density in the outer plasma regions. This is a consequence of the improved
confinement by the edge transport barrier in the H-mode. The formation of a pedestal, where the density
and temperature, and thus the pressure, have a steep radial gradient results from this insulating region
(etb). This is accompanied by a large plasma current density [20].

This pedestal provides the energy required for two plasma instabilities called peeling and ballooning
modes, their mode structures are shown in Figure 7. While ballooning modes are pressure gradient-driven
with high mode numbers and located at the outboard side of the plasma, peeling modes are current-driven
with a low mode number. These theory of MagnetoHydroDynamics (MHD), which is a fluid description of
the plasma (explained in the next subsection), is able to describe well those modes.

Figure 6: Example peeling (left) and ballooning (right) modes, as simulated with JOREK. The colors
indicate the perturbation of poloidal flux. Figures reproduced from [21]
Figure 7: The stable regions of the peeling and ballooning modes are shown. Peeling modes occur at high
edge current densities ≈ jedge , while ballooning modes are unstable for high pressure gradients
∇p. They can couple together to peeling-ballooning modes which decreases the area of the
stable region. Especially when operated in H-Mode the plasma parameters are close to the
peeling-ballooning stability limit [15].

3.4.3 Prevention and control techniques

For large-scale projects like ITER, it is assumed that type I ELMs would greatly reduce the lifespan of the
divertor. A main motivation in studying ELMs is therefore the suppression or control of ELMs. Research
involving prevention of edge localized mode formation is underway. A possible solution is the injection of
static magnetic noisy energy into the containment field as a containment-stabilization regime, which may
decrease ELM amplitude [22].

3.5 Suprathermal particles

Supra-thermal particles are particles with an energy beyond the thermal Maxwell distribution. They can
originate from heating methods (NBI) or products of fusion processes or wall atoms. For ASDEX Upgrade
for example are they essentially only from NBI, but ITER/DEMO will have many fusion-born fast particles.
They are heavy and therefore typically non relativistic.
3.5.1 Banana orbits

Banana orbits describe the back-and-forth movement of plasma particles, which are trapped in the outside
region of a tokamak plasma [23]. The field inside the tokamak and similar toroidal arrangements is stronger
on the inboard than the outboard side simply due to the magnets being closer together in that area. The
particles approximately following magnetic field lines lose parallel (to the field) magnetic energy when
moving towards the inboard side and can there be reflected (magnetic mirror [24]). Together with drifts,
the orbits of these particles are called banana orbits due to the banana-shape visible on the projection 8.

Banana orbits are very important for particles confinement in Tokamaks. The two most important parameters
in banana orbits are the width and the amplitude.

Figure 8: Schematic view of banana orbits [euro-fusion.org]

3.5.2 Pitch angle

The pitch angle of a charged particle is the angle between the particle’s velocity vector and the local
magnetic field.

3.6 MagnetoHydroDynamics (MHD)

MHD theory can be thought of as a combination of the Navier-Stokes equations for fluid dynamics with
Maxwell’s equations for electromagnetism. It is based on the useful simplification of considering the plasma
as an electrically conducting fluid [25] and obtained from the first three moments of the Boltzmann equation
[26], which describe mass, momentum and energy conservation, and which are analogous to the Euler
equations for fluid motion. One gets the equations of resistive MagnetoHydroDynamics (MHD) when those
equations are combined with Maxwell’s equations of electromagnetism.


+ ∇.(ρv) = 0
dt
d~v
ρ = J~ × B
~ − ∇p (9)
dt
dρ p
( )=0
dt ργ

~ + ~v × B
E ~ = η J~

~ =0
∇.B
∂B (10)
= −∇ × E~
∂t
µ0 J~ = ∇ × B
~

This set of equations is dubbed ideal MHD for a resistivity η = 0 (otherwise resistive MHD). Adding
a viscosity leads to the visco-resistive MHD equations. These differential equations must be solved
simultaneously, either analytically or numerically.
This resistivity leads to a breaking in the magnetic topology; there is possible reconnection of the
magnetic field lines when they collide.
In addition, it is typical that the particle diffusivity and heat conductivity are included to model particle and
heat transport respectively.
Extended MHD is used to describe a class of phenomena in plasma that are higher order than resistive
MHD, but which can adequately be treated with a single fluid description. The included effects are for
example: Hall physics, electron pressure gradients, finite Larmor Radii in the particle gyromotion, and
electron inertia.

Like mentioned above in 3.4, MHD can describe many of the large scale instabilities occurring in fusion
plasmas such as ballooning modes, peeling modes, internal and external kink modes, tearing modes ... as
well as the plasma equilibrium. In experiments, the plasma is constantly evolving, with MHD instabilities
playing a large role.

Limitations In the plasma, collisions lead to a transport of particle and energy. Neoclassical transport
theory can be derived from the collisions of particles during their motion through the magnetised plasma.
This description is adequate for a transport parallel to the field lines and in the edge, but in the core plasma,
perpendicular transport is dominated by turbulent transport, which can be found in gyrokinetic simulations
at large computational cost and is only approximated in MHD and transport simulations.
4 Numerical methods

4.1 Nonlinear MHD code JOREK

JOREK is a fully non-linear MHD code for solving extended MHD in realistic X-point plasma geometry
including the plasma core, the separatrix, scrape-off layer until the divertor capable of calculating the
temporal evolution of a fusion plasma in realistic configurations. It has been initially developed for simulations
to study the stability of Edge Localized Modes (ELMs) [9], but is nowadays also used to simulate a
wide range of disruption-related phenomena including tearing mode growth [27] and rapid cooling due
to impurity injections [28].

JOREK internally consists of a collection of different models, each solving a set of equations for the same
quantities, but also introducing different specialized additions to the code. In this project "model 303" is
used, which is a reduced MHD model that includes the velocity component along the magnetic field and
has various extensions.

4.1.1 Discretization

A brief overview of the discretization is given here. The poloidal plane is discretized in both the radial and
poloidal direction, as shown in figure 9, where the poloidal grid used in the JOREK simulations is shown.

More information about the JOREK discretization can be found in [29].

4.2 Particle model

model 307

4.3 Approach

A loop was used to study the losses of particles and their orbits in a perturbed and unperturbed case, to
make a comparison after that ... to be completed at the end -Code additions and modifications

5 Experimental observations

Experimental results are took from various experiments made at different tokamaks like ASDEX-Upgrade
in Munich.The results of one of those experiments [30] will be explained here.
6 Tests and results

6.1 Simulation of ELMs

6.1.1 Energies

6.1.2 Temperature

6.1.3 Field lines

6.1.4 Q-profile

6.1.5 Density

6.1.6 Particle content

6.2 Simple tests for particle motion

6.3 Suprathermal particle dynamics in simplified ELM fields

6.4 Suprathermal particle dynamics during realistic ELM crash

6.5

6.6

6.7

7 Discussion

8 Conclusions and outlook

References

[1] Robert J Goldston and Paul H Rutherford. INTRODUCTION TO PLASMA PHYSICS. IOP Publishing
Ltd 1995, 1995. ISBN 0 7503 0325 5. URL http://www.astrosen.unam.mx/~aceves/verano/
libros/goldstone_plasma.pdf.

[2] Wakatani M. Nishikawa K. Basic Properties of Plasma. Springer, Berlin, Heidelberg, 2000. ISBN
978-3-642-08465-2. URL https://link.springer.com/chapter/10.1007/978-3-662-04078-2_
2#citeas.

[3] R. Fitzpatrick. Plasma Physics: An Introduction. Taylor & Francis, 2014. ISBN 9781466594265. URL
https://books.google.de/books?id=0RwbBAAAQBAJ.
[4] Jeffrey P. Freidberg. Plasma Physics and Fusion Energy. Cambridge University Press, 2007.
doi:10.1017/CBO9780511755705.

[5] T S Taylor. Physics of advanced tokamaks. Plasma Physics and Controlled Fusion, 39
(12B):B47–B73, dec 1997. doi:10.1088/0741-3335/39/12b/005. URL https://doi.org/10.
10882F0741-33352F392F12b2F005.

[6] 2 H. Jiang 3 Z. Ren 2 S. Li, 1 and C. Xu. Optimal tracking for a divergent-type parabolic pde system
in current profile control. jun 2014. URL https://www.hindawi.com/journals/aaa/2014/940965/
#copyright.

[7] Anatolij B Mineev. Plasma heating systems. In Fundamentals of Magnetic Thermonuclear Reactor
Design, pages 281–290. Elsevier, 2018.

[8] Marco Buzio. Structural effects of plasma instabilities on the JET tokamak. PhD thesis, Imperial
College London (University of London), 1999.

[9] G.T.A Huysmans and O Czarny. MHD stability in x-point geometry: simulation of ELMs. Nuclear
Fusion, 47(7):659–666, jun 2007. doi:10.1088/0029-5515/47/7/016. URL https://doi.org/10.
1088%2F0029-5515%2F47%2F7%2F016.

[10] K. Kim, K. Im, H.C. Kim, S. Oh, J.S. Park, S. Kwon, Y.S. Lee, J.H. Yeom, C. Lee, G-S. Lee, G. Neilson,
C. Kessel, T. Brown, P. Titus, D. Mikkelsen, and Y. Zhai. Design concept of k-DEMO for near-term
implementation. Nuclear Fusion, 55(5):053027, apr 2015. doi:10.1088/0029-5515/55/5/053027. URL
https://doi.org/10.1088%2F0029-5515%2F55%2F5%2F053027.

[11] D Wróblewski and LL Lao. Determination of the safety factor in sawtoothing discharges in diii-d.
Physics of Fluids B: Plasma Physics, 3(10):2877–2881, 1991.

[12] B LaBombard, JE Rice, AE Hubbard, JW Hughes, M Greenwald, J Irby, Y Lin, B Lipschultz,


ES Marmar, CS Pitcher, et al. Transport-driven scrape-off-layer flows and the boundary conditions
imposed at the magnetic separatrix in a tokamak plasma. Nuclear fusion, 44(10):1047, 2004.

[13] F. Wagner, G. Fussmann, T. Grave, M. Keilhacker, M. Kornherr, K. Lackner, K. McCormick, E. R.


Müller, A. Stäbler, G. Becker, K. Bernhardi, U. Ditte, A. Eberhagen, O. Gehre, J. Gernhardt, G. v.
Gierke, E. Glock, O. Gruber, G. Haas, M. Hesse, G. Janeschitz, F. Karger, S. Kissel, O. Klüber,
G. Lisitano, H. M. Mayer, D. Meisel, V. Mertens, H. Murmann, W. Poschenrieder, H. Rapp, H. Röhr,
F. Ryter, F. Schneider, G. Siller, P. Smeulders, F. Söldner, E. Speth, K. H. Steuer, Z. Szymanski, and
O. Vollmer. Development of an edge transport barrier at the h-mode transition of asdex. Phys. Rev.
Lett., 53:1453–1456, Oct 1984. doi:10.1103/PhysRevLett.53.1453. URL https://link.aps.org/
doi/10.1103/PhysRevLett.53.1453.
[14] F Wagner. A quarter-century of h-mode studies.

[15] Valentin Igochine. Active Control of Magneto-hydrodynamic Instabilities in Hot Plasmas. Springer-
Verlag Berlin Heidelberg, 2015. doi:10.1007/978-3-662-44222-7.

[16] Howard Wilson. Edge localized modes in tokamaks. Fusion Science and Technology, 57(2T):174–
182, 2010.

[17] N. Ferraro L. Sugiyama F. Waelbroeck X. Q. Xu A. Loarte PG. T. A. Huijsmans, C. S. Chang and


S. Futatani. Modelling of edge localised modes and edge localised mode control. Physics of Plasmas.
doi:10.1063/1.4905231. URL https://doi.org/10.1063/1.4905231.

[18] P.T. Lang, A. Loarte, G. Saibene, L.R. Baylor, M. Becoulet, M. Cavinato, S. Clement-Lorenzo, E. Daly,
T.E. Evans, M.E. Fenstermacher, Y. Gribov, L.D. Horton, C. Lowry, Y. Martin, O. Neubauer, N. Oyama,
M.J. Schaffer, D. Stork, W. Suttrop, P. Thomas, M. Tran, H.R. Wilson, A. Kavin, and O. Schmitz.
ELM control strategies and tools: status and potential for ITER. Nuclear Fusion, 53(4):043004, mar
2013. doi:10.1088/0029-5515/53/4/043004. URL https://doi.org/10.1088%2F0029-5515%2F53%
2F4%2F043004.

[19] A Huber, G Arnoux, MNA Beurskens, S Brezinsek, P Coad, T Eich, JC Fuchs, W Fundamenski,
S Jachmich, A Korotkov, et al. Impact of large type i elms on plasma radiation in jet. Bulletin of the
American Physical Society, 54, 2009.

[20] C Konz, LD Horton, E Strumberger, S Günter, and GTA Huysmans. The peeling-ballooning model
revisited.

[21] M. Hoelzl et al. Simulating tokamak edge instabilities: advances and challenges. URL
https://www.researchgate.net/profile/Matthias_Hoelzl/publication/326265673_
Simulating_tokamak_edge_instabilities_advances_and_challenges_Invited_Talk_at_
45th_EPS_Conference_on_Plasma_Physics_Prague/links/5b432d0c458515f71cb59479/
Simulating-tokamak-edge-instabilities-advances-and-challenges-Invited-Talk-at-45th-EPS-Conf
pdf.

[22] P.T Lang, G.D Conway, T Eich, L Fattorini, O Gruber, S Günter, L.D Horton, S Kalvin, A Kallenbach,
M Kaufmann, G Kocsis, A Lorenz, M.E Manso, M Maraschek, V Mertens, J Neuhauser, I Nunes,
W Schneider, W Suttrop, H Urano, and the ASDEX Upgrade Team. ELM pace making and mitigation
by pellet injection in ASDEX upgrade. Nuclear Fusion, 44(5):665–677, apr 2004. doi:10.1088/0029-
5515/44/5/010. URL https://doi.org/10.1088%2F0029-5515%2F44%2F5%2F010.

[23] Igor Pankratov and Volodymyr Bochko. Long-lived banana orbit formation of suprathermal electrons
during mhd spikes in runaway tokamak discharges. East European Journal of Physics, (3):4–9,
Oct. 2019. doi:10.26565/2312-4334-2019-3-01. URL https://periodicals.karazin.ua/eejp/
article/view/14355.

[24] R.F. Post. The magnetic mirror approach to fusion. Nuclear Fusion, 27(10):1579–1739, oct
1987. doi:10.1088/0029-5515/27/10/001. URL https://doi.org/10.1088%2F0029-5515%2F27%
2F10%2F001.

[25] H. R. Strauss. Nonlinear, three-dimensional magnetohydrodynamics of noncircular tokamaks.


Physics of Fluids, 19(1):134–140, January 1976. doi:10.1063/1.861310.

[26] J. P. Hans Goedbloed and Stefaan Poedts. Principles of Magnetohydrodynamics: With


Applications to Laboratory and Astrophysical Plasmas. Cambridge University Press, 2004.
doi:10.1017/CBO9780511616945.

[27] Dmytro Meshcheriakov, Matthias Hoelzl, Valentin Igochine, Sina Fietz, Francois Orain, Guido T.A.
Huijsmans, Marc Maraschek, Mike Dunne, Rachael McDermott, Hartmut Zohm, Karl Lackner, and
Sibylle Günter. Numerical study of tearing mode seeding in tokamak x-point plasma. Physics of
Plasmas, 26(4), 4 2019. ISSN 1070-664X. doi:10.1063/1.5086402.

[28] D. Hu, E. Nardon, M. Lehnen, G. T.A. Huijsmans, and D. C. van Vugt. 3d non-linear mhd simulation
of the mhd response and density increase as a result of shattered pellet injection. Nuclear Fusion, 58
(12), 10 2018. ISSN 0029-5515. doi:10.1088/1741-4326/aae614.

[29] O. Czarny and G. Huysmans. Bézier surfaces and finite elements for mhd simulations.
Journal of Computational Physics, 227(16):7423–7445, 8 2008. ISSN 0021-9991.
doi:10.1016/j.jcp.2008.04.001.

[30] M Garcia-Munoz, S Äkäslompolo, P de Marne, M G Dunne, R Dux, T E Evans, N M Ferraro,


S Fietz, C Fuchs, B Geiger, A Herrmann, M Hoelzl, B Kurzan, N Lazanyi, R M McDermott,
M Nocente, D C Pace, M Rodriguez-Ramos, K Shinohara, E Strumberger, W Suttrop, M A Van
Zeeland, E Viezzer, M Willensdorfer, and E Wolfrum. Fast-ion losses induced by ELMs and
externally applied magnetic perturbations in the ASDEX upgrade tokamak. Plasma Physics and
Controlled Fusion, 55(12):124014, nov 2013. doi:10.1088/0741-3335/55/12/124014. URL https:
//doi.org/10.1088%2F0741-3335%2F55%2F12%2F124014.
Figure 9: The poloidal grid used in the JOREK simulations
Figure 10

Figure 11
Figure 12
(a) Caption 1 (b) Caption 2

Figure 13: Overall caption


(a) Caption 1 (b) Caption 2

Figure 14: Overall caption

(a) Caption 1 (b) Caption 2

Figure 15: Overall caption


(a) Caption 1 (b) Caption 2

Figure 16: Overall caption

(a) Caption 1 (b) Caption 2

Figure 17: Overall caption


Figure 18

Figure 19

Figure 20
Figure 21: JOREK quantities and their normalisations

You might also like