The Saffman helicity invariant of Hosking & Schekochihin (Phys. Rev. X, vol. 11, issue 4, 2021, 041005), which we here call the Hosking integral, has emerged as an important quantity that may govern the decay properties of magnetically dominated non-helical turbulence. Using a range of different computational methods, we confirm that this quantity is indeed gauge invariant and nearly perfectly conserved in the limit of large Lundquist numbers. For direct numerical simulations with ordinary viscosity and magnetic diffusivity operators, we find that the solution develops in a nearly self-similar fashion. In a diagram quantifying the instantaneous decay coefficients of magnetic energy and integral scale, we find that the solution evolves along a line that is indeed suggestive of the governing role of the Hosking integral. The solution settles near a line in this diagram that is expected for a self-similar evolution of the magnetic energy spectrum. The solution will settle in a slightly different position when the magnetic diffusivity decreases with time, which would be compatible with the decay being governed by the reconnection time scale rather than the Alfvén time.

Effects of plasma non-uniformities and kinetic dispersiveness on the spontaneous excitation of geodesic acoustic mode (GAM) by reversed shear Alfvén eigenmode (RSAE) are investigated numerically. It is found that, due to the turning points induced by the shear Alfvén continuum structure, the nonlinear excitation of GAM is a quasiexponentially growing absolute instability. As the radial dependence of GAM frequency and pump RSAE mode structure are accounted for, the radially inward propagating GAM is preferentially excited, leading to core localized thermal plasma heating by GAM collisionless damping. Our work, thus, suggests that GAM excitation plays a crucial role in not only RSAE nonlinear saturation, but also anomalous fuel ion heating in future reactors.

In this paper we propose a methodology for the efficient implementation of machine learning (ML)-based methods in particle-in-cell (PIC) codes, with a focus on Monte Carlo or statistical extensions to the PIC algorithm. The presented approach allows for neural networks to be developed in a Python environment, where advanced ML tools are readily available to proficiently train and test them. Those models are then efficiently deployed within highly scalable and fully parallelized PIC simulations during runtime. We demonstrate this methodology with a proof-of-concept implementation within the PIC code OSIRIS, where a fully connected neural network is used to replace a section of a Compton scattering module. We demonstrate that the ML-based method reproduces the results obtained with the conventional method and achieves better computational performance. These results offer a promising avenue for future applications of ML-based methods in PIC, particularly for physics extensions where a ML-based approach can provide a higher performance increase.

Plasma jets are widely investigated both in the laboratory and in nature. Astrophysical objects such as black holes, active galactic nuclei and young stellar objects commonly emit plasma jets in various forms. With the availability of data from plasma jet experiments resembling astrophysical plasma jets, classification of such data would potentially aid in not only investigating the underlying physics of the experiments but also the study of astrophysical jets. In this work we use deep learning to process all of the laboratory plasma images from the Caltech Spheromak Experiment spanning two decades. We found that cosine similarity can aid in feature selection, classify images through comparison of feature vector direction and be used as a loss function for the training of AlexNet for plasma image classification. We also develop a simple vector direction comparison algorithm for binary and multi-class classification. Using our algorithm we demonstrate 93 % accurate binary classification to distinguish unstable columns from stable columns and 92 % accurate five-way classification of a small, labelled data set which includes three classes corresponding to varying levels of kink instability.

An important problem in nuclear fusion plasmas is the prediction and control of turbulence which drives the cross-field transport, thus leading to energy loss from the system and deteriorating confinement. Turbulence, being a highly nonlinear and multiscale process, is challenging to theoretically describe and computationally model. Most advanced computational models fall into one of the two categories: fluid or gyro-kinetic. They both come at a high computational cost and cannot be applied for routine simulation of plasma discharge evolution and control. Development of reduced models based on (physics informed) artificial neural networks could potentially fulfil the need for affordable simulations of plasma turbulence. However, the training requires an extensive data base and the obtained models lack extrapolation capability to scenarios not originally encountered during training. This leads to reduced models of limited validity which may not prove adequate for predicting scenarios in future machines. In contrast, we explore a data-driven model discovery approach based on sparse regression to infer governing nonlinear partial differential equations directly from the data. Our input data are generated by simulations of drift-wave turbulence according to the Hasegawa–Wakatani and modified Hasegawa–Wakatani models. Balancing model accuracy and complexity enables the reconstruction of the systems of partial differential equations accurately describing the dynamics simulated in the input data sets. Sparse regression is not data hungry and can be extrapolated to unexplored parameter ranges. We explore and demonstrate the potential of this approach for fusion plasma turbulence modelling. The findings show that the methodology is promising for the development of reduced and computationally efficient turbulence models as well as for existing model cross-validation.

A new formalism for the derivation of the cnoidal wave solution is presented by introducing a new set of initial conditions and a subacoustic moving frame. The resulting set of solutions is also constructed in a way that ensures the conservation of number density. The solutions are illustrated in a variety of graphical representations, and the effect of the amplitude's magnitude on the cnoidal wave form is presented. Interestingly, it is shown that the wavelength of the solutions decreases with amplitude. In addition, it is shown that the small-amplitude cnoidal wave solutions converge to linear waves in the small-amplitude limit.

Raytracing/Fokker–Planck (FP) simulations used to model lower-hybrid current drive (LHCD) often fail to reproduce experimental results, particularly when LHCD is weakly damped. A proposed reason for this discrepancy is the lack of ‘full-wave’ effects, such as diffraction and interference, in raytracing simulations and the breakdown of the raytracing approximation. Previous studies of LHCD using non-Maxwellian full-wave/FP simulations have been performed, but these simulations were not self-consistent and enforced power conservation between the FP and full-wave code using a numerical rescaling factor. Here, we have created a fully self-consistent full-wave/FP model for LHCD that is automatically power conserving. This was accomplished by coupling an overhauled version of the non-Maxwellian TORLH full-wave solver and the CQL3D FP code using the Integrated Plasma Simulator. We performed converged full-wave/FP simulations of Alcator C-Mod discharges and compared them with raytracing. We found that excellent agreement in the power deposition profiles from raytracing and TORLH could be obtained, however, TORLH had somewhat lower current drive efficiency and broader power deposition profiles in some cases. This discrepancy appears to be a result of numerical limitations present in the TORLH model and a small amount of diffractional broadening of the TORLH wave spectrum. Our results suggest full-wave simulation of LHCD is likely not necessary as diffraction and interference represented only a small correction that could not account for the differences between simulations and experiment.

We investigate three-dimensional (3-D) bow shocks in a highly collisional magnetized aluminium plasma, generated during the ablation phase of an exploding wire array on the MAGPIE facility (1.4 MA, 240 ns). Ablation of plasma from the wire array generates radially diverging, supersonic (M_S \sim 7$M_S \sim 7$), super-Alfvénic (M_A > 1$M_A > 1$) magnetized flows with frozen-in magnetic flux (R_M \gg 1$R_M \gg 1$). These flows collide with an inductive probe placed in the flow, which serves both as the obstacle that generates the magnetized bow shock, and as a diagnostic of the advected magnetic field. Laser interferometry along two orthogonal lines of sight is used to measure the line-integrated electron density. A detached bow shock forms ahead of the probe, with a larger opening angle in the plane parallel to the magnetic field than in the plane normal to it. Since the resistive diffusion length of the plasma is comparable to the probe size, the magnetic field decouples from the ion fluid at the shock front and generates a hydrodynamic shock, whose structure is determined by the sonic Mach number, rather than the magnetosonic Mach number of the flow. The 3-D simulations performed using the resistive magnetohydrodynamic (MHD) code Gorgon confirm this picture, but under-predict the anisotropy observed in the shape of the experimental bow shock, suggesting that non-MHD mechanisms may be important for modifying the shock structure.

Plasma supports collective modes and particle–wave interactions that lead to complex behaviour in, for example, inertial fusion energy applications. While plasma can sometimes be modelled as a charged fluid, a kinetic description is often crucial for studying nonlinear effects in the higher-dimensional momentum–position phase space that describes the full complexity of the plasma dynamics. We create a differentiable solver for the three-dimensional partial-differential equation describing the plasma kinetics and introduce a domain-specific objective function. Using this framework, we perform gradient-based optimization of neural networks that provide forcing function parameters to the differentiable solver given a set of initial conditions. We apply this to an inertial-fusion-relevant configuration and find that the optimization process exploits a novel physical effect.

The effect of sheared E × B flow on the blob dynamics in the scrape-off layer (SOL) of HL-2A tokamak has been studied during the plasma current ramp-up in ohmically heated deuterium plasmas by the combination of poloidal and radial Langmuir probe arrays. The experimental results indicate that the SOL sheared E × B flow is substantially enhanced as the plasma current exceeds a certain value and the strong sheared E × B flow has the ability to slow the blob radial motion via stretching its poloidal correlation length. The locally accumulated blobs are suggested to be responsible for the increase of plasma density just outside the Last Closed Flux Surface (LCFS) observed in this experiment. The results presented here reveal the significant role played by the strong sheared E × B flow on the blob dynamics, which provides a potential method to control the SOL width by modifying the sheared E × B flow in future tokamak plasmas.

We study the onset of electron heating in intense laser–solid interactions and its impact on the spectral quality of radiation pressure accelerated ions in both hole boring and light sail regimes. Two- and three-dimensional particle-in-cell (PIC) simulations are performed over a wide range of laser and target parameters and reveal how the pulse duration, profile, polarization and target surface stability control the electron heating, the dominant ion acceleration mechanisms and the ion spectra. We find that the onset of strong electron heating is associated with the growth of the Rayleigh–Taylor-like instability at the front surface and must be controlled to produce high-quality ion beams, even when circularly polarized lasers are employed. We define a threshold condition for the maximum duration of the laser pulse that allows mitigation of electron heating and radiation pressure acceleration of narrow energy spread ion beams. The model is validated by three-dimensional PIC simulations, and the few experimental studies that reported low energy spread radiation pressure accelerated ion beams appear to meet the derived criteria. The understanding provided by our work will be important in guiding future experimental developments, for example for the ultrashort laser pulses becoming available at state-of-the-art laser facilities, for which we predict that proton beams with \sim$\sim$150–250 MeV, \sim$\sim$30% energy spread, and a total laser-to-proton conversion efficiency of \sim$\sim$20% can be produced.

Power-transfer and fixed-point analysis of previous NIMROD simulations (Roberds et al., Phys. Plasmas, vol. 23, issue 9, 2016, 092513) improved the understanding of the effect of 3D (non-axisymmetric equilibrium) magnetic fields on sawtooth oscillations in the Compact Toroidal Hybrid (CTH) experiment. Computing the locations of order-1 fixed points, their Greene's residues, and local values for the rotational transform results in a description of CTH sawteeth consistent with Kadomtsev's model. A power-transfer analysis quantifies the distribution of energy among toroidal Fourier modes and their nonlinear interactions. The Lorentz power transfer drives sawtooth growth, and it is unambiguously interpreted as the flow of energy from toroidal mode n'$n'$ to mode n$n$, catalysed by \boldsymbol {B}_{n-n'}$\boldsymbol {B}_{n-n'}$. It has been reported previously that the CTH sawtooth frequency increases with the 3D field strength. This is attributed to an increased growth rate of the internal kink that drives sawtooth oscillations. Here, 3D fields remove energy from the kink, eliminating the possibility that these fields are an additional energy source that drives growth. Instead, 3D fields catalyse energy transfer from large-to-small scales, where magnetic reconnection is stronger. It is proposed that this energy transfer increases the reconnection rate at small scales, which is consistent with the increased growth rate observed at higher 3D field strengths.

The paper presents experimental results from the SMOLA device on the testing of the helical mirror confinement hypothesis. Helical mirror confinement is the technique of an active control of axial plasma losses from a confinement zone by multiple magnetic mirrors that move along the axis in the reference frame of the plasma that experiences \boldsymbol{E} \times \boldsymbol{B}$\boldsymbol{E} \times \boldsymbol{B}$ rotation due to an applied radial electric field. Theory predicts that a helical mirror will provide an axial force that modifies the plasma flow and, simultaneously, density pinching to the axis. The force direction depends on the plasma rotation direction. Experimental data on the axial plasma losses at different direction of the magnetic mirror movement are presented. If the trapped ions move in the direction opposite to the direction of the axial losses, then the particle flux reduces in the broad range of the plasma density. The confinement improves with the increase of the fraction of the trapped particles (effective mirror ratio was up to R_{{\rm eff}}=5.8\pm 1.4$R_{{\rm eff}}=5.8\pm 1.4$). If the trapped ions move in the same direction as the axial losses, then the flux depends on density. At intermediate densities, the integral flux through the transport section rises compared to the plasma flowing through the straight magnetic field. The effective mirror ratio is lower and does not significantly depend on the fraction of the trapped particles (effective mirror ratio at intermediate density was R_{{\rm eff}}=3.3\pm 0.8$R_{{\rm eff}}=3.3\pm 0.8$).

The kinetic–magnetohydrodynamic (MHD) hybrid simulation approach for macroscopic instabilities in plasmas can be extended to include the kinetic effects of both thermal ions and energetic ions. The new coupling scheme includes synchronization of the density and parallel velocity between thermal ions and MHD, in addition to pressure coupling, to ensure the quasineutrality condition and avoid numerical errors. The new approach has been implemented in the kinetic-MHD code M3D-C1-K, and was used to study the thermal ion kinetic effects and Landau damping in fishbone modes in both DIII-D and NSTX. It is found that the thermal ion kinetic effects can cause an increase of the frequencies of the non-resonant n=1$n=1$ fishbone modes driven by energetic particles for q_\mathrm {min}>1$q_\mathrm {min}>1$, and Landau damping can provide additional stabilization effects. A nonlinear simulation for n=1$n=1$ fishbone mode in NSTX is also performed, and the perturbation on magnetic flux surfaces and the transport of energetic particles are calculated.

Understanding generation and mitigation of runaway electrons in disruptions is important for the safe operation of future tokamaks. In this paper we investigate the runaway dynamics in reactor-scale spherical tokamaks, focusing on a compact nominal design with a plasma current of 21 megaamperes (MA), 1.8 T magnetic field on axis and major radius of approximately 3 m. We study both the severity of runaway generation during unmitigated disruptions, and the effect that typical mitigation schemes based on massive material injection have on runaway production. The study is conducted using the numerical framework dream (Disruption Runaway Electron Analysis Model). We find that, in many cases, mitigation strategies are necessary to prevent the runaway current from reaching multi-MA levels. Our results indicate that, with a suitably chosen deuterium–neon mixture for mitigation, it is possible to achieve a tolerable runaway current and ohmic current evolution. However, this does not account for the runaway source due to wall activation, which has been found to severely limit successful mitigation at conventional aspect ratios, but whose definition requires a more complete wall specification. Furthermore, the majority of the thermal energy loss is found to happen through radial transport rather than radiation, which poses a risk of unacceptable localised heat loads.

Plasma-terminating disruptions in future fusion reactors may result in conversion of the initial current to a relativistic runaway electron beam. Validated predictive tools are required to optimise the scenarios and mitigation actuators to avoid the excessive damage that can be caused by such events. Many of the simulation tools applied in fusion energy research require the user to specify input parameters that are not constrained by the available experimental information. The conventional approach, where an expert modeller calibrates these input parameters based on domain knowledge, is prone to lead to an intractable validation challenge without systematic uncertainty quantification. Bayesian inference algorithms offer a promising alternative approach that naturally includes uncertainty quantification and is less subject to user bias in choosing the input parameters. The main challenge in using these methods is the computational cost of simulating enough samples to construct the posterior distributions for the uncertain input parameters. This challenge can be overcome by combining probabilistic surrogate modelling, such as Gaussian process regression, with Bayesian optimisation, which can reduce the number of required simulations by several orders of magnitude. Here, we implement this type of Bayesian optimisation framework for a model for analysis of disruption runaway electrons, and explore for simulations of current quench in a JET plasma discharge with an argon induced disruption. We use this proof-of-principle framework to explore the optimum input parameters with uncertainties in optimisation tasks ranging from one to seven dimensions. The relevant Python codes that are used in the analysis are available via https://github.com/aejarvin/BO_FOR_RE_SIMULATIONS/.

A spontaneous heating process is found to arise in a system where a magnetic island is present due to a linearly unstable tearing mode. The parity, the relative phases and the structure of the fields determined linearly by the tearing mode cause the compression of the plasma in the direction parallel to the magnetic field to heat the plasma in the vicinity of the separatrix in the nonlinear phase. Using a six-field electromagnetic fluid model, the process is found to be present in both two-dimensional single-helicity and three-dimensional multi-helicity simulations with both symmetric and asymmetric magnetic equilibrium profiles. A noteworthy feature of the model is that the higher-order compression terms responsible for the heating process are retained in the equations. The process is believed to be linked to experimental observations of localized hot-spots on externally induced magnetic islands.

Resonant filament-assisted mode conversion (FAMC) scattering of high harmonic fast waves (HHFW) by cylindrical field-aligned density inhomogeneities can efficiently redirect a fraction of the launched HHFW power flux into the parallel direction. Within a simplified analytic approach, this contribution compares the parallel propagation, reflection and dissipation of nearly resonant FAMC modes for three magnetic field line geometries in the scrape-off layer, in the presence of radio-frequency (RF) sheaths at field line extremities and phenomenological wave damping in the plasma volume. When a FAMC mode, excited at the HHFW antenna parallel location and guided along the open magnetic field lines, impinges onto a boundary at normal incidence, we show that it can excite sheath RF oscillations, even toroidally far away from the HHFW launcher. The RF sheaths then dissipate part of the power flux carried by the incident mode, while another part reflects into the FAMC mode with the opposite wave vector parallel to the magnetic field. The reflected FAMC mode in turn propagates and can possibly interact with the sheath at the opposite field line boundary. The two counter-propagating modes then form in the bounded magnetic flux tube a lossy cavity excited by the HHFW scattering. We investigate how the presence of field line boundaries affects the total HHFW power redirected into the filament, and its splitting between sheath and volume losses, as a function of relevant parameters in the model.

Shock waves in plasma are usually dealt with using magnetohydrodynamics (MHD). Yet, MHD entails the assumption of a short mean free path, which is not fulfilled in a collisionless plasma. Recently, for pair plasmas, we devised a model allowing one to account for kinetic effects within a MHD-like formalism. Its relies on an estimate of the anisotropy generated when crossing the front, with a subsequent assessment of the stability of this anisotropy in the downstream. We solved our model for parallel, perpendicular and switch-on shocks. Here we bridge between all these cases by treating the problem of an arbitrarily, but coplanar, oriented magnetic field. Even though the formalism presented is valid for anisotropic upstream temperatures, only the case of a cold upstream is solved. We find extra solutions which are not part of the MHD catalogue, and a density jump that is notably less in the quasi-parallel, highly magnetized, regime. Given the complexity of the calculations, this work is mainly devoted to the presentation of the mathematical aspect of our model. A forthcoming article will be devoted to the physics of the shocks here defined.