arXiv subjects

Frank Jenko

Publications and source records attributed to Frank Jenko.

At least 37 records · Page 2Linked to original sources

arXiv2024

Global gyrokinetic analysis of Wendelstein 7-X discharge: unveiling the importance of trapped-electron-mode and electron-temperature-gradient turbulence

We present the first nonlinear, gyrokinetic, radially global simulation of a discharge of the Wendelstein 7-X-like stellarator (W7-X), including kinetic electrons, an equilibrium radial electric field, as well as electromagnetic and collisional effects. By comparison against flux-tube and full-flux-surface simulations, we assess the impact of the equilibrium ExB-flow and flow shear on the stabilisation of turbulence. In contrast to the existing literature, we further provide substantial evidence for the turbulent electron heat flux being driven by trapped-electron-mode (TEM) and electron-temperature-gradient (ETG) turbulence in the core of the plasma. The former manifests as a hybrid together with ion-temperature-gradient (ITG) turbulence and is primarily driven by the finite electron temperature gradient, which has largely been neglected in nonlinear stellarator simulations presented in the existing literature.

physics.plasm-ph↗

arXiv2023

Physics-Preserving AI-Accelerated Simulations of Plasma Turbulence

Turbulence in fluids, gases, and plasmas remains an open problem of both practical and fundamental importance. Its irreducible complexity usually cannot be tackled computationally in a brute-force style. Here, we combine Large Eddy Simulation (LES) techniques with Machine Learning (ML) to retain only the largest dynamics explicitly, while small-scale dynamics are described by an ML-based sub-grid-scale model. Applying this novel approach to self-driven plasma turbulence allows us to remove large parts of the inertial range, reducing the computational effort by about three orders of magnitude, while retaining the statistical physical properties of the turbulent system.

physics.comp-ph↗

arXiv2023

Magnetic flutter effect on validated edge turbulence simulations

Small magnetic fluctuations ($B_1/B_0 \sim 10^{-4}$) are intrinsically present in a magnetic confinement plasma due to turbulent currents. While the perpendicular transport of particles and heat is typically dominated by fluctuations of the electric field, the parallel stream of plasma is affected by fluttering magnetic field lines. In particular through electrons, this indirectly impacts the turbulence dynamics. Even in low beta conditions, we find that $E\times B$ turbulent transport can be reduced by more than a factor 2 when magnetic flutter is included in our validated edge turbulence simulations of L-mode ASDEX Upgrade. The primary reason for this is the stabilization of drift-Alfvén-waves, which reduces the phase shifts of density and temperature fluctuations with respect to potential fluctuations. This stabilization can be qualitatively explained by linear analytical theory, and appreciably reinforced by the flutter nonlinearity. As a secondary effect, the steeper temperature gradients and thus higher $η_i$ increase the impact of the ion-temperature-gradient mode on overall turbulent transport. With increasing beta, the stabilizing effect on $E\times B$ turbulence increases, balancing the destabilization by induction, until direct electromagnetic perpendicular transport is triggered. We conclude that including flutter is crucial for predictive edge turbulence simulations.

physics.plasm-ph↗

arXiv2023

Reassessing the impact of megaelectronvolt ions in fusion plasmas via gyrokinetic simulations

Gyrokinetic simulations conducted by Mazzi et al. reveal the suppression of turbulence in fusion plasmas through the destabilization of Toroidal Alfvén Eigenmodes (TAEs) by megaelectronvolt ions. Our analysis demonstrates that the authors' numerical findings are strongly influenced by the selected simulation settings, calling into question their claim that the resulting heat conductivity aligns with the TRANSP power balance. Specifically, we assert that the numerical results presented by Mazzi et al. are a direct consequence of the inadequacy of resolution employed in their numerical simulations. Notably, there are three primary factors contributing to this issue: (i) the employed radial box size is insufficient, leading to an undesirable impact of the boundary conditions on the simulations; (ii) the adoption of a higher minimum toroidal mode number fails to accurately resolve the entire range of TAEs, resulting in an underestimation of TAE drive and introducing spurious effects on wave-particle resonances; (iii) an insufficient resolution in the magnetic moment direction exacerbates these challenges. Upon addressing the aforementioned numerical issues, a significant increase in heat conductivity for each plasma species was observed by more than tenfold, diverging from the expected values derived from the TRANSP power balance calculations. Consequently, our results raise serious doubts on the authors' assertion of enhanced performance in the presence of strongly unstable TAEs, emphasizing the need for a reevaluation of their claims.

physics.plasm-ph↗

arXiv2023

Epistemic and Aleatoric Uncertainty Quantification and Surrogate Modelling in High-Performance Multiscale Plasma Physics Simulations

This work suggests several methods of uncertainty treatment in multiscale modelling and describes their application to a system of coupled turbulent transport simulations of a tokamak plasma. We propose a method to quantify the usually aleatoric uncertainty of a system in a quasi-stationary state, estimating the mean values and their errors for quantities of interest, which is average heat fluxes in the case of turbulence simulations. The method defines the stationarity of the system and suggests a way to balance the computational cost of simulation and the accuracy of estimation. This allows, contrary to many approaches, to incorporate aleatoric uncertainties in the analysis of the model and to have a quantifiable decision for simulation runtime. Furthermore, the paper describes methods for quantifying the epistemic uncertainty of a model and the results of such a procedure for turbulence simulations, identifying the model's sensitivity to particular input parameters and sensitivity to uncertainties in total. Finally, we introduce a surrogate model approach based on Gaussian Process Regression and present a preliminary result of training and analysing the performance of such a model based on turbulence simulation data. Such an approach shows a potential to significantly decrease the computational cost of the uncertainty propagation for the given model, making it feasible on current HPC systems.

physics.plasm-ph↗

arXiv2023

Verification of the Fourier-enhanced 3D finite element Poisson solver of the gyrokinetic full-f code PICLS

We introduce and derive the Fourier-enhanced 3D electrostatic field solver of the gyrokinetic full-f PIC code PICLS. The solver makes use of a Fourier representation in one periodic direction of the domain to make the solving of the system easily parallelizable and thus save run time. The presented solver is then verified using two different approaches of manufactured solutions. The test setup used for this effort is a pinch geometry with ITG-like electric potential, containing one non-periodic and two periodic directions, one of which will be discrete Fourier transformed. The results of these tests show that in all three dimensions the L2-error decreases with a constant rate close to the ideal prediction, depending on the degree of the chosen basis functions.

physics.plasm-ph↗

arXiv2023

Zonal Flow Excitation in Electron-Scale Tokamak Turbulence

The derivation of an intermediate-scale gyrokinetic-electron theory in nonuniform tokamak plasmas [Chen H. et al 2021 Nucl. Fusion 61 066017] has shown that a Navier-Stokes type nonlinearity couples electron-temperature-gradient (ETG) modes and zonal flow (ZF) modes with wavelengths much shorter than the ion gyroradius but much longer than the electron gyroradius. This intermediate-scale ETG-ZF coupling is typically stronger than the Hasegawa-Mima type nonlinearity characteristic of the fluid approximation and is predicted to lead to relevant zonal flow generation and ETG mode regulation. Electron-scale, continuum, gyrokinetic simulation results are presented here which include both single-mode ETG and full-spectrum ETG turbulence. The zonal flow generation due to single ETG modes is investigated and the single-mode intermediate-scale results are found to be in agreement with theory. The full-spectrum results are then presented and explained qualitatively in terms of the single-mode results. It is found that the ETG-driven zonal flows regulate intermediate-scale electron heat flux transport to levels in the predicted range.

physics.plasm-ph↗

arXiv2022

Context-aware learning of hierarchies of low-fidelity models for multi-fidelity uncertainty quantification

Multi-fidelity Monte Carlo methods leverage low-fidelity and surrogate models for variance reduction to make tractable uncertainty quantification even when numerically simulating the physical systems of interest with high-fidelity models is computationally expensive. This work proposes a context-aware multi-fidelity Monte Carlo method that optimally balances the costs of training low-fidelity models with the costs of Monte Carlo sampling. It generalizes the previously developed context-aware bi-fidelity Monte Carlo method to hierarchies of multiple models and to more general types of low-fidelity models. When training low-fidelity models, the proposed approach takes into account the context in which the learned low-fidelity models will be used, namely for variance reduction in Monte Carlo estimation, which allows it to find optimal trade-offs between training and sampling to minimize upper bounds of the mean-squared errors of the estimators for given computational budgets. This is in stark contrast to traditional surrogate modeling and model reduction techniques that construct low-fidelity models with the primary goal of approximating well the high-fidelity model outputs and typically ignore the context in which the learned models will be used in upstream tasks. The proposed context-aware multi-fidelity Monte Carlo method applies to hierarchies of a wide range of types of low-fidelity models such as sparse-grid and deep-network models. Numerical experiments with the gyrokinetic simulation code \textsc{Gene} show speedups of up to two orders of magnitude compared to standard estimators when quantifying uncertainties in small-scale fluctuations in confined plasma in fusion reactors. This corresponds to a runtime reduction from 72 days to about four hours on one node of the Lonestar6 supercomputer at the Texas Advanced Computing Center.

math.NA↗

arXiv2022

A general framework for quantifying uncertainty at scale

In many fields of science, comprehensive and realistic computational models are available nowadays. Often, the respective numerical calculations call for the use of powerful supercomputers, and therefore only a limited number of cases can be investigated explicitly. This prevents straightforward approaches to important tasks like uncertainty quantification and sensitivity analysis. This challenge can be overcome via our recently developed sensitivity-driven dimension adaptive sparse grid interpolation strategy. The method exploits, via adaptivity, the structure of the underlying model (such as lower intrinsic dimensionality and anisotropic coupling of the uncertain inputs) to enable efficient and accurate uncertainty quantification and sensitivity analysis at scale. We demonstrate the efficiency of our approach in the context of fusion research, in a realistic, computationally expensive scenario of turbulent transport in a magnetic confinement tokamak device with eight uncertain parameters, reducing the effort by at least two orders of magnitude. In addition, we show that our method intrinsically provides an accurate surrogate model that is nine orders of magnitude cheaper than the high-fidelity model.

physics.comp-ph↗

arXiv2022

Leveraging Stochastic Predictions of Bayesian Neural Networks for Fluid Simulations

We investigate uncertainty estimation and multimodality via the non-deterministic predictions of Bayesian neural networks (BNNs) in fluid simulations. To this end, we deploy BNNs in three challenging experimental test-cases of increasing complexity: We show that BNNs, when used as surrogate models for steady-state fluid flow predictions, provide accurate physical predictions together with sensible estimates of uncertainty. Further, we experiment with perturbed temporal sequences from Navier-Stokes simulations and evaluate the capabilities of BNNs to capture multimodal evolutions. While our findings indicate that this is problematic for large perturbations, our results show that the networks learn to correctly predict high uncertainties in such situations. Finally, we study BNNs in the context of solver interactions with turbulent plasma flows. We find that BNN-based corrector networks can stabilize coarse-grained simulations and successfully create multimodal trajectories.

physics.flu-dyn↗

arXiv2022

Gyrokinetic modelling of anisotropic energetic particle driven instabilities in tokamak plasmas

Energetic particles produced by neutral beams are observed to excite energetic-particle-driven geodesic acoustic modes (EGAMs) in tokamaks. We study the effects of anisotropy of distribution function of the energetic particles on the excitation of such instabilities with ORB5, a gyrokinetic particle-in-cell code. Numerical results are shown for linear electrostatic simulations with ORB5. The growth rate is found to be sensitively dependent on the phase-space shape of the distribution function. The behavior of the instability is qualitatively compared to the theoretical analysis of dispersion relations. Realistic neutral beam energetic particle anisotropic distributions are obtained from the heating solver RABBIT and are introduced into ORB5 as input distribution function. Results show a dependence of the growth rate on the injection angle. A qualitative comparison to experimental measurements is presented and few disagreements between them are found, being the growth rate in the simulations much lower than that from experiments. An explanation for the difference is advanced.

physics.plasm-ph↗

arXiv2021

Data-driven low-fidelity models for multi-fidelity Monte Carlo sampling in plasma micro-turbulence analysis

The linear micro-instabilities driving turbulent transport in magnetized fusion plasmas (as well as the respective nonlinear saturation mechanisms) are known to be sensitive with respect to various physical parameters characterizing the background plasma and the magnetic equilibrium. Therefore, uncertainty quantification is essential for achieving predictive numerical simulations of plasma turbulence. However, the high computational costs of the required gyrokinetic simulations and the large number of parameters render standard Monte Carlo techniques intractable. To address this problem, we propose a multi-fidelity Monte Carlo approach in which we employ data-driven low-fidelity models that exploit the structure of the underlying problem such as low intrinsic dimension and anisotropic coupling of the stochastic inputs. The low-fidelity models are efficiently constructed via sensitivity-driven dimension-adaptive sparse grid interpolation using both the full set of uncertain inputs and subsets comprising only selected, important parameters. We illustrate the power of this method by applying it to two plasma turbulence problems with up to $14$ stochastic parameters, demonstrating that it is up to four orders of magnitude more efficient than standard Monte Carlo methods measured in single-core performance, which translates into a runtime reduction from around eight days to one hour on 240 cores on parallel machines.

physics.comp-ph↗

arXiv2021

Quasilinear gyrokinetic theory: A derivation of QuaLiKiz

In order to predict and analyze turbulent transport in tokamaks, it is important to model transport that arises from microinstabilities. For this task, quasilinear codes have been developed that seek to calculate particle, angular momentum, and heat fluxes both quickly and accurately. In this tutorial, we present a derivation of one such code known as QuaLiKiz, a quasilinear gyrokinetic transport code. The goal of this derivation is to provide a self-contained and complete description of the underlying physics and mathematics of QuaLiKiz from first principles. This work serves both as a comprehensive overview of QuaLiKiz specifically as well as an illustration for deriving quasilinear models in general.

physics.plasm-ph↗

arXiv2021

Turbulence suppression by energetic particles: A sensitivity-driven dimension-adaptive sparse grid framework for discharge optimization

A newly developed sensitivity-driven approach is employed to study the role of energetic particles in suppressing turbulence-inducing micro-instabilities for a set of realistic JET-like cases with NBI deuterium and ICRH $^3$He fast ions. First, the efficiency of the sensitivity-driven approach is showcased for scans in a $21$-dimensional parameter space, for which only $250$ simulations are necessary. The same scan performed with traditional Cartesian grids with only two points in each of the $21$ dimensions would require $2^{21} = 2,097,152$ simulations. Then, a $14$-dimensional parameter subspace is considered, using the sensitivity-driven approach to find an approximation of the parameter-to-growth rate map averaged over nine bi-normal wave-numbers, indicating pathways towards turbulence suppression. The respective turbulent fluxes, obtained via nonlinear simulations for the optimized set of parameters, are reduced by more than two order of magnitude compared to the reference results.

physics.plasm-ph↗

arXiv2021

Sub-grid-scale effects in magnetised plasma turbulence

In the present paper, we use a coarse-graining approach to investigate the nonlinear redistribution of free energy in both position and scale space for weakly collisional magnetised plasma turbulence. For this purpose, we use high-resolution numerical simulations of gyrokinetic (GK) turbulence that span the proton-electron range of scales, in a straight magnetic guide field geometry. Accounting for the averaged effect of the particles' fast gyro-motion on the slow plasma fluctuations, the GK approximation captures the dominant energy redistribution mechanisms in strongly magnetised plasma turbulence. Here, the GK system is coarse-grained with respect to a cut-off scale, separating in real space the contributions to the nonlinear interactions from the coarse-grid-scales and the sub-grid-scales (SGS). We concentrate on the analysis of nonlinear SGS effects. Not only that this allows us to investigate the flux of free energy across the scales, but also to now analyse its spatial density. We find that the net value of scale flux is an order of magnitude smaller than both the positive and negative flux density contributions. The dependence of the results on the filter type is also analysed. Moreover, we investigate the advection of energy in position space. This rather novel approach for GK turbulence can help in the development of SGS models that account for advective unstable structures for space and fusion plasmas, and with the analysis of the turbulent transport saturation.

physics.plasm-ph↗

arXiv2020

Plasma Confinement Mode Classification Using a Sequence-to-Sequence Neural Network With Attention

In a typical fusion experiment, the plasma can have several possible confinement modes. At the TCV tokamak, aside from the Low (L) and High (H) confinement modes, an additional mode, dithering (D), is frequently observed. Developing methods that automatically detect these modes is considered to be important for future tokamak operation. Previous work with deep learning methods, particularly convolutional recurrent neural networks (Conv-RNNs), indicates that they are a suitable approach. Nevertheless, those models are sensitive to noise in the temporal alignment of labels, and that model in particular is limited to making individual decisions taking into account only its own hidden state and its input at each time step. In this work, we propose an architecture for a sequence-to-sequence neural network model with attention which solves both of those issues. Using a carefully calibrated dataset, we compare the performance of a Conv-RNN with that of our proposed sequence-to-sequence model, and show two results: one, that the Conv-RNN can be improved upon with new data; two, that the sequence-to-sequence model can improve the results even further, achieving excellent scores on both train and test data.

physics.plasm-ph↗

arXiv2020

Nonlinear symmetry breaking in electron temperature gradient driven turbulence

Nonlinear symmetry breaking may occur in systems with two or more states whose linear dynamics displays certain symmetries, one of which is preferred nonlinearly. We have identified a regime of electron temperature gradient (ETG) instabilities in a tokamak plasma with circular concentric flux surfaces that has its largest growth rate at a finite ballooning angle, establishing a symmetry that is nonlinearly broken to favor one sign for the ballooning angle. This is the first example of nonlinear symmetry breaking in simulations of a drift instability in the absence of externally imposed flow shear or asymmetry in the plasma column.

physics.plasm-ph↗

arXiv2020

Deep learning for Gaussian process tomography model selection using the ASDEX Upgrade SXR system

Gaussian process tomography (GPT) is a method used for obtaining real-time tomographic reconstructions of the plasma emissivity profile in a tokamak, given some model for the underlying physical processes involved. GPT can also be used, thanks to Bayesian formalism, to perform model selection -- i.e., comparing different models and choosing the one with maximum evidence. However, the computations involved in this particular step may become slow for data with high dimensionality, especially when comparing the evidence for many different models. Using measurements collected by the ASDEX Upgrade Soft X-ray (SXR) diagnostic, we train a convolutional neural network (CNN) to map SXR tomographic projections to the corresponding GPT model whose evidence is highest. We then compare the network's results, and the time required to calculate them, with those obtained through analytical Bayesian formalism. In addition, we use the network's classifications to produce tomographic reconstructions of the plasma emissivity profile, whose quality we evaluate by comparing their projection into measurement space with the existing measurements themselves.

physics.data-an↗