arXiv subjects

Ralph Kube

Publications and source records attributed to Ralph Kube.

7 recordsLinked to original sources

arXiv2022

Machine-Learning enabled analysis of ELM filament dynamics in KSTAR

The emergence and dynamics of filamentary structures associated with edge-localized modes (ELMs) inside tokamak plasmas during high-confinement mode is regularly studied using Electron Cyclotron Emission Imaging (ECEI) diagnostic systems. ECEI allows inference of electron temperature variations, often across a poloidal cross-section. Previously, detailed analyses of filamentary dynamics and classification of the precursors to ELM crashes have been done manually. We present a machine-learning-based model, capable of automatically identifying the position, spatial extent, and amplitude of ELM filaments. The model is a deep convolutional neural network that has been trained and optimized on an extensive set of manually labeled ECEI data from the KSTAR tokamak. Once trained, the model achieves a 93.7% precision and allows to robustly identify plasma filaments in unseen ECEI data. The trained model is used to characterize ELM filament dynamics in a single H-mode plasma shot. We identify quasi-periodic oscillations of the filaments' size, total heat content, and radial velocity. The detailed dynamics of these quantities appear strongly correlated with each other and appear qualitatively different during the pre-crash and ELM crash phases.

physics.plasm-ph↗

arXiv2021

Near real-time streaming analysis of big fusion data

While experiments on fusion plasmas produce high-dimensional data time series with ever increasing magnitude and velocity, data analysis has been lagging behind this development. For example, many data analysis tasks are often performed in a manual, ad-hoc manner some time after an experiment. In this article we introduce the DELTA framework that facilitates near real-time streaming analysis of big and fast fusion data. By streaming measurement data from fusion experiments to a high-performance compute center, DELTA allows to perform demanding data analysis tasks in between plasma pulses. This article describe the modular and expandable software architecture of DELTA and presents performance benchmarks of its individual components as well as of entire workflows. Our focus is on the streaming analysis of ECEi data measured at KSTAR on NERSCs supercomputers and we routinely achieve data transfer rates of about 500 Megabyte per second. We show that a demanding turbulence analysis workload can be distributed among multiple GPUs and executes in under 5 minutes. We further discuss how DELTA uses modern database systems and container orchestration services to provide web-based real-time data visualization. For the case of ECEi data we demonstrate how data visualizations can be augmented with outputs from machine learning models. By providing session leaders and physics operators results of higher order data analysis using live visualization they may monitor the evolution of a long-pulse discharge in near real-time and may make more informed decision on how to configure the machine for the next shot.

physics.plasm-ph↗

arXiv2018

Non-Oberbeck-Boussinesq zonal flow generation

Novel mechanisms for zonal flow (ZF) generation for both large relative density fluctuations and background density gradients are presented. In this non-Oberbeck-Boussinesq (NOB) regime ZFs are driven by the Favre stress, the large fluctuation extension of the Reynolds stress, and by background density gradient and radial particle flux dominated terms. Simulations of a nonlinear full-F gyro-fluid model confirm the predicted mechanism for radial ZF propagation and show the significance of the NOB ZF terms for either large relative density fluctuation levels or steep background density gradients.

physics.plasm-ph↗

arXiv2018

Intermittent electron density and temperature fluctuations and associated fluxes in the Alcator C-Mod scrape-off layer

The Alcator C-Mod mirror Langmuir probe system has been used to sample data time series of fluctuating plasma parameters in the outboard mid-plane far scrape-off layer. We present a statistical analysis of one second long time series of electron density, temperature, radial electric drift velocity and the corresponding particle and electron heat fluxes. These are sampled during stationary plasma conditions in an ohmically heated, lower single null diverted discharge. The electron density and temperature are strongly correlated and feature fluctuation statistics similar to the ion saturation current. Both electron density and temperature time series are dominated by intermittent, large-amplitude burst with an exponential distribution of both burst amplitudes and waiting times between them. The characteristic time scale of the large-amplitude bursts is approximately 15μs. Large-amplitude velocity fluctuations feature a slightly faster characteristic time scale and appear at a faster rate than electron density and temperature fluctuations. Describing these time series as a superposition of uncorrelated exponential pulses, we find that probability distribution functions, power spectral densities as well as auto-correlation functions of the data time series agree well with predictions from the stochastic model. The electron particle and heat fluxes present large-amplitude fluctuations. For this low-density plasma, the radial electron heat flux is dominated by convection, that is, correlations of fluctuations in the electron density and radial velocity. Hot and dense blobs contribute approximately 6% of the total fluctuation driven heat flux.

physics.plasm-ph↗

arXiv2017

Unified transport scaling laws for plasma blobs and depletions

We study the dynamics of seeded plasma blobs and depletions in an (effective) gravitational field. For incompressible flows the radial center of mass velocity of blobs and depletions is proportional to the square root of their initial cross-field size and amplitude. If the flows are compressible, this scaling holds only for ratios of amplitude to size larger than a critical value. Otherwise, the maximum blob and depletion velocity depends linearly on the initial amplitude and is independent of size. In both cases the acceleration of blobs and depletions depends on their initial amplitude relative to the background plasma density, is proportional to gravity and independent of their cross-field size. Due to their reduced inertia plasma depletions accelerate more quickly than the corresponding blobs. These scaling laws are derived from the invariants of the governing drift-fluid equations and agree excellently with numerical simulations over five orders of magnitude. We suggest an empirical model that unifies and correctly captures the radial acceleration and maximum velocities of both blobs and depletions.

physics.plasm-ph↗

arXiv2016

Fluctuation statistics in the scrape-off layer of Alcator C-Mod

We study long time series of the ion saturation current and floating potential, sampled by Langmuir probes dwelled in the outboard mid-plane scrape off layer and embedded in the lower divertor baffle of Alcator C-Mod. A series of ohmically heated L-mode plasma discharges is investigated with line-averaged plasma density ranging from n_e/n_G = 0.15 to 0.42, where n_G is the Greenwald density. All ion saturation current time series that are sampled in the far scrape-off layer are characterized by large-amplitude burst events. Coefficients of skewness and excess kurtosis of the time series obey a quadratic relationship and their histograms coincide partially upon proper normalization. Histograms of the ion saturation current time series are found to agree well with a prediction of a stochastic model for the particle density fluctuations in scrape-off layer plasmas. The distribution of the waiting times between successive large-amplitude burst events and of the burst amplitudes are approximately described by exponential distributions. The average waiting time and burst amplitude are found to vary weakly with the line-averaged plasma density. Conditional averaging reveals that the radial blob velocity, estimated from floating potential measurements, increases with the normalized burst amplitude in the outboard mid-plane scrape-off layer. For low density discharges, the conditionally averaged waveform of the floating potential associated with large amplitude bursts at the divertor probes has a dipolar shape. In detached divertor conditions the average waveform is random, indicating electrical disconnection of blobs from the sheaths at the divertor targets.

physics.plasm-ph↗

arXiv2015

Convergence of statistical moments of particle density time series in scrape-off layer plasmas

Particle density fluctuations in the scrape-off layer of magnetically confined plasmas, as measured by gas-puff imaging or Langmuir probes, are modeled as the realization of a stochastic process in which a superposition of pulses with a fixed shape, an exponential distribution of waiting times and amplitudes represents the radial motion of blob-like structures. With an analytic formulation of the process at hand, we derive expressions for the mean-squared error on estimators of sample mean and sample variance as a function of sample length, sampling frequency, and the parameters of the stochastic process. % Employing that the probability distribution function of a particularly relevant shot noise process is given by the gamma distribution, we derive estimators for sample skewness and kurtosis, and expressions for the mean-squared error on these estimators. Numerically generated synthetic time series are used to verify the proposed estimators, the sample length dependency of their mean-squared errors, and their performance. We find that estimators for sample skewness and kurtosis based on the gamma distribution are more precise and more accurate than common estimators based on the method of moments.

physics.plasm-ph↗