- Magnetic confinement fusion research
- Ionosphere and magnetosphere dynamics
- Plasma Diagnostics and Applications
- Solar and Space Plasma Dynamics
- Nuclear reactor physics and engineering
- Fusion materials and technologies
- Magnetic Field Sensors Techniques
- Magnetic Properties and Applications
- Geomagnetism and Paleomagnetism Studies
- Physics of Superconductivity and Magnetism
- Laser-Plasma Interactions and Diagnostics
- Characterization and Applications of Magnetic Nanoparticles
- Induction Heating and Inverter Technology
- Natural Language Processing Techniques
- Atomic and Subatomic Physics Research
- Superconducting Materials and Applications
- Advanced Numerical Analysis Techniques
- Electrohydrodynamics and Fluid Dynamics
- Advancements in Semiconductor Devices and Circuit Design
- Advanced DC-DC Converters
- Particle accelerators and beam dynamics
- Electric Motor Design and Analysis
- Electrostatic Discharge in Electronics
- Magnetic Bearings and Levitation Dynamics
- Electric Power Systems and Control
General Atomics (United States)
2021-2024
Princeton University
2020-2023
Los Alamos National Laboratory
2014-2021
Tibbar Plasma Technologies (United States)
2017-2021
Fermi Research Alliance
2021
Fermi National Accelerator Laboratory
2021
Columbia University
2020
Goddard Space Flight Center
2016
U.S. National Science Foundation
2016
University of Washington
2006-2015
Abstract Recent progress in the application of machine learning (ML)/artificial intelligence (AI) algorithms to improve Equilibrium Fitting (EFIT) code equilibrium reconstruction for fusion data analysis applications is presented. A device-independent portable core solver capable computing or reconstructing different tokamaks has been created facilitate adaptation ML/AI algorithms. large EFIT database comprising DIII-D magnetic, motional Stark effect, and kinetic generated developments...
Accurate plasma state reconstruction will be crucial for the success of ITER and future fusion plants, but harsh conditions a burning make diagnostic operation more challenging than in current machines. Integrated data analysis (IDA) based on Bayesian inference allows improved information gain by combining many diagnostics into single step using sophisticated forward models. It also provides framework to seamlessly combine predictive modeling data, which can invaluable data-poor environment....
Abstract Nonlinear 3D MHD simulations of shattered-pellet injection (SPI) in JET show prototypical SPI-driven disruptions using the M3D-C1 and NIMROD extended-MHD codes. Initially, radiation-driven thermal quenches are accelerated by activity as pellet crosses rational surfaces, leading to a radiation spike, global stochasticization magnetic field, complete quench. Eventually, current quenches, preceded spike seen Ohmic heating becomes equal radiative cooling. The results qualitatively...
Abstract Kinetic equilibrium reconstructions make use of profile information such as particle density and temperature measurements in addition to magnetics data compute a self-consistent equilibrium. They are used multitude physics-based modeling. This work develops multi-layer perceptron (MLP) neural network (NN) model surrogate for kinetic Equilibrium Fitting (EFITs) trains on the 2019 DIII-D discharge campaign database reconstructions. We investigate impact including various diagnostic...
Abstract A mechanism for steady inductive helicity injection (SIHI) current drive has been discovered where the driving fluctuations are not generated by plasma but rather imposed injectors. Sheared flow of electron fluid distorts to current. The model accurately predicts time dependent toroidal current, injector impedance scaling, and profile produced in HIT-SI experiment. These results show that a stable equilibrium can be efficiently sustained with can, principle, controlled. Both large...
Bayesian statistics offers a powerful technique for plasma physicists to infer knowledge from the heterogeneous data types encountered. To explain this power, simple example, Gaussian Process Regression, and application of inverse problems are explained. The likelihood is key distribution because it contains model, or theoretic predictions, desired quantities. By using prior knowledge, inferred quantities interest based on given can be inferred. Because not single prediction, uncertainty...
This work presents a method for predicting plasma equilibria in tokamak fusion experiments and reactors. The approach involves representing the current as linear combination of basis functions using principal component analysis toroidal densities (Jt) from EFIT-AI equilibrium database. Then utilizing EFIT's Green's function tables, are created poloidal flux (ψ) diagnostics generated (Jt). Similar to idea physics-informed neural network (NN), this physically enforces consistency between ψ,...
Increasing the helicity injector drive frequency up to 68.5 kHz on Helicity Injected Torus-Steady Inductive (HIT-SI) experiment has produced spheromaks with current amplifications of 3.8, ideal n = 1 kink stability, improved toroidal symmetry and pressure confinement. Current centroid calculations from surface magnetic probes show an outward shift in field at frequencies above 50 kHz. Grad-Shafranov equilibria indicate confinement higher operating frequencies. The minimum characteristic...
We present a comparison study of 3-D pressureless resistive MHD (rMHD) and presureless two-fluid models the Helicity Injected Torus with Steady Inductive helicity injection (HIT-SI). HIT-SI is current drive experiment that uses two geometrically asymmetric injectors to generate sustain toroidal plasmas. The comparable size collisionless ion skin depth di predicates importance Hall term for HIT-SI. simulations are run NIMROD, an initial-value, extended code. modeled plasma density temperature...
The first sustainment of toroidal plasma current 50 kA at up to 3 times the injected currents, added in quadrature, using steady inductive helicity injection is described. Separatrix currents-currents not linking injectors-are sustained 40 kA. Decreases n=1 mode poloidal magnetic field higher amplifications indicate more quiescent, direct drive. Results are achieved HIT-SI (with a spheromak major radius 0.3 m) during deuterium operations immediately after helium operation. These results...
A time-dependent helicity balance model applied to a spheromak helicity-injection experiment enables the measurement of relaxation time during sustainment phase spheromak. The experiment, Helicity Injected Torus with Steady Inductive injection (HIT-SI), studies formation and through inductive injection. captures dominant plasma behavior seen in HIT-SI by using an empirical helicity-decay rate, composite Taylor state both content system calculate resulting current. During single-injector...
We present simulations of inductive helicity injection in the Helicity Injected Torus with Steady Inductive (HIT-SI) device that treats entire plasma volume a single dynamic MHD model. A new fully 3D numerical tool, PSI-center TETrahedral mesh code, was developed provides geometric flexibility required for this investigation. Implementation zero-β Hall model using PSI-TET will be presented including formulation self-consistent magnetic boundary condition wall HIT-SI device. Results from are...
Abstract A numerical database of over one thousand perturbed three-dimensional (3D) equilibria has been generated, constructed based on the MARS-F (Liu et al 2000 Phys. Plasmas 7 3681) computed plasma response to externally applied 3D field sources in multiple tokamak devices. Perturbed with n = 1–4 ( is toroidal mode number) periodicity are computed. Surrogate models created for equilibrium utilizing model order reduction (MOR) techniques. In particular, retaining first few eigenstates from...
New understanding and improved parameters have been achieved on the Helicity Injected Torus with Steady Inductive helicity injection current drive (HIT-SI) experiment. The experiment has a bowtie-shaped spheromak confinement region two injectors. inductive injectors are 180° segments of small, oval cross section toroidal pinch. Spheromaks currents up to 38 kA amplification 2 only 6 MW injector power. Taylor-state model is shown agree HIT-SI surface internal magnetic profile measurements....
The Helicity Injected Torus-Steady Inductive (HIT-SI) injectors have been operated experimentally and simulated with a 3D pressureless Hall magnetohydrodynamic (MHD) code (NIMROD) at frequencies of 14.5, 36.8, 53.5 68.5 kHz. Simulations are compared to experimental data using biorthogonal decomposition (BD) real synthetic surface magnetic probe signals. First, the time-averaged field squared (B2), measured by probes, is between experiment simulation. Next, spatial basis set from reference...
The Helicity Injected Torus-Steady Inductive (HIT-SI) experiment investigates steady inductive helicity injection with the aim of forming and sustaining a high-beta equilibrium in spheromak geometry using two semi-toroidal injectors. Results experiments unequal rates produced highest current (38 kA), amplification ( I tor / inj_quad ≈ 2) poloidal flux (ψ pol /ψ > 6) to date. Single-injector operations establish preferred direction generated for each injector depending on sign injected its...
Magnetic helicity balance and a fully relaxed Taylor-state model are shown to predict the magnitude of sustained equilibrium current in an inductively driven spheromak. The Helicity Injected Torus with Steady Inductive drive (HIT-SI) [T. R. Jarboe et al., Phys. Rev. Lett. 97, 115003 (2006)] forms sustains spheromaks using two injectors. By assuming is injected at rate 2VΨ, only decays through Spitzer resistivity Te measured Langmuir probe, predicted no fitting parameters. correctly predicts...
Kinetic simulations have demonstrated that three-dimensional reconnection in collisionless regimes proceeds through the formation and interaction of magnetic flux ropes, which are generated due to growth tearing instabilities at multiple resonance surfaces. Since kinetic intrinsically expensive, it is desirable explore feasibility reduced two-fluid models capture this complex evolution, particularly, strong guide field regime, where better justified. With goal mind, paper compares evolution...
The helicity injected torus with steady inductive injection device is a simply-connected toroidal that forms and sustains axisymmetric spheromak equilibria through the use of injection, which inject magnetic drive non-axisymmetric currents on edge plasma. Simulations high-gain scenarios, using NIMROD code based models validated low gain experimental operations, indicate formation closed flux surfaces at current amplifications six. Despite scheme adding perturbations order , these are able to...
A rotating tokamak plasma can interact resonantly with the external helical magnetic perturbations, also known as error fields. This lead to locking and then disruptions. We leverage machine learning (ML) methods predict events. use a coupled third-order nonlinear ordinary differential equation model represent interaction of perturbation rotation field. is sufficient describe qualitatively unlocking bifurcations. explore using ML algorithms simulation data experimental data, focusing on that...
We present nonlinear NIMROD resistive MHD simulations of the response a rotating plasma to an error field when has weakly damped linear tearing modes (TMs), stabilized by pressure gradient and favorable curvature. The curvature leads Glasser effect: occurrence real frequencies stabilization with positive stability index Δ′. A cylinder hollow is used model toroidal Linear rotation ψ̃w show that, in agreement analytical results, peak reconnected flux occurs for rate near TM phase velocity....
We present a framework for estimating the probability of locking to an error field in rotating tokamak plasma. This leverages machine learning methods trained on data from mode-locking model, including field, resistive magnetohydrodynamics modeling plasma, wall, and external vacuum region, leading fifth-order ordinary differential equation (ODE) system. It is extension model without wall introduced by Akçay et al. [Phys. Plasmas 28, 082106 (2021)]. Tearing mode saturation finite island width...