A5088210128- Magnetic confinement fusion research
- Nuclear reactor physics and engineering
- Fusion materials and technologies
- Superconducting Materials and Applications
- Ionosphere and magnetosphere dynamics
- Nuclear Physics and Applications
- Meteorological Phenomena and Simulations
- Nuclear Engineering Thermal-Hydraulics
- Radiation Therapy and Dosimetry
- Advanced Radiotherapy Techniques
- Fluid Dynamics and Turbulent Flows
- Particle accelerators and beam dynamics
- Traffic Prediction and Management Techniques
- Atmospheric and Environmental Gas Dynamics
- Gaussian Processes and Bayesian Inference
- Radiation Effects and Dosimetry
- Radiation Detection and Scintillator Technologies
- Tropical and Extratropical Cyclones Research
- Medical Imaging Techniques and Applications
- Distributed and Parallel Computing Systems
- Climate variability and models
- Ocean Waves and Remote Sensing
- Computational Fluid Dynamics and Aerodynamics
- Scientific Measurement and Uncertainty Evaluation
- Fault Detection and Control Systems
Salk Institute for Biological Studies
2025
Dutch Institute for Fundamental Energy Research
2018-2024
Scripps Institution of Oceanography
2024
United Kingdom Atomic Energy Authority
2022-2024
Culham Science Centre
2022-2024
Eindhoven University of Technology
2023
University of Lisbon
2022
Fusion Academy
2019-2021
Fusion (United States)
2019-2021
Atomic Energy (Canada)
2014
Abstract The JET hybrid scenario has been developed from low plasma current carbon wall discharges to the record-breaking Deuterium-Tritium plasmas obtained in 2021 with ITER-like Be/W wall. development started pure Deuterium refinement of current, and toroidal magnetic field choices succeeded solving heat load challenges arising 37 MW injected power ITER like environment, keeping radiation edge core controlled, avoiding MHD instabilities reaching high neutron rates. have re-run Tritium...
Abstract Model-based plasma scenario development lies at the heart of design and operation future fusion powerplants. Including turbulent transport in integrated models is essential for delivering a successful roadmap towards ITER DEMO-class devices. Given highly iterative nature models, fast machine-learning-based surrogates are fundamental to fulfil pressing need faster simulations opening up pulse design, optimization, flight simulator applications. A significant bottleneck generation...
This paper outlines an approach towards improved rigour in tokamak turbulence transport model validation within integrated modelling. Gaussian process regression (GPR) techniques were applied for profile fitting during the preparation of modelling simulations allowing rigourous sensitivity tests prescribed initial and boundary conditions as both fit derivative uncertainties are provided. was demonstrated by a JETTO simulation JET ITER-like-wall H-mode baseline discharge #92436 with QuaLiKiz...
Abstract The recent deuterium–tritium campaign in JET-ILW (DTE2) has provided a unique opportunity to study the isotope dependence of L-H power threshold an ITER-like wall environment (Be and W divertor). Here we present results from dedicated transition experiments at JET-ILW, documenting tritium plasmas, comparing them with matching deuterium hydrogen datasets. From earlier it is known that as plasma isotopic composition changes deuterium, through varying deuterium/hydrogen concentrations,...
Abstract JET experiments using the fuel mixture envisaged for fusion power plants, deuterium and tritium (D–T), provide a unique opportunity to validate existing D–T prediction capabilities in support of future device design operation preparation. The 2021 experimental campaign has achieved powers sustained over 5 s ITER-relevant conditions i.e. with baseline or hybrid scenario full metallic wall. In preparation campaign, extensive predictive modelling was carried out several assumptions...
Abstract A review is given on the highlights of a scatter-shot approach developing machine-learning methods and artificial neural networks based fast predictors for application to fusion exhaust. The aim enable facilitate optimized improved modeling allowing more flexible integration physics models in light extrapolations towards future devices. project encompasses various research objectives: (a) developments surrogate model power & particle exhaust plants; (b) assessments...
Abstract Deep learning sequence models trained on personalized genomics can improve variant effect prediction, however, applications of these are limited by computational requirements for storing and reading large datasets. We address this with GenVarLoader, which stores genomic data in new memory-mapped formats optimal locality to achieve ∼1,000x faster throughput ∼2,000x better compression compared existing alternatives.
Predictability of burning plasmas is a key issue for designing and building credible future fusion devices. In this context, an important effort physics understanding guidance being carried out in parallel to JET experimental campaigns H D by performing analyses modelling towards improvement the DT optimization JET-DT neutron yield born alpha particle physics. Extrapolations from recent experiments using maximum power available have been performed including some most sophisticated codes...
Within integrated tokamak plasma modeling, turbulent transport codes are typically the computational bottleneck limiting their routine use outside of post-discharge analysis. Neural network (NN) surrogates have been used to accelerate these calculations while retaining desired accuracy physics-based models. This paper extends a previous NN model, known as QLKNN-hyper-10D, by incorporating impact impurities, rotation, and magnetic equilibrium effects. is achieved adding light impurity...
Abstract This work applies the coupled JINTRAC and QuaLiKiz-neural-network (QLKNN) model on ohmic current ramp-up phase of a JET D discharge. The chosen scenario exhibits hollow <?CDATA $T_\mathrm{e}$?> <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:msub> <mml:mi>T</mml:mi> <mml:mrow> <mml:mi mathvariant="normal">e</mml:mi> </mml:mrow> </mml:msub> </mml:math> profile attributed to core impurity accumulation, which is observed worsen with increasing fuel ion...
We present an ultrafast neural network (NN) model, QLKNN, which predicts core tokamak transport heat and particle fluxes. QLKNN is a surrogate model based on database of 300 million flux calculations the quasilinear gyrokinetic QuaLiKiz. The covers wide range realistic parameters. Physical features such as existence critical gradient for onset turbulent were integrated into training methodology. have coupled to modelling framework JINTRAC rapid control-oriented solver RAPTOR. frameworks are...
Near-surface ocean wind measurements are important for weather forecasting, determining surface transports, and estimating air-sea interactions; however, in-situ observations often limited. Previous studies indicate that the equilibrium range of gravity wave spectrum can be used to estimate velocity. This approach is tested using spectral from Coastal Data Information Program (CDIP) buoy array off Monterey, California. Quality controlled vectors inferred spectra statistically compared a...
Abstract For the first time pellet cycle of a multiple-isotope plasma is successfully reproduced with reduced turbulent transport modelling, within an integrated simulation framework. Future nuclear fusion reactors are likely to be fuelled by cryogenic injection, due higher penetration and faster response times. Accurate modelling crucial assess fuelling efficiency burn control. In recent Joint European Torus tokamak experiments, deuterium pellets reactor-relevant deposition characteristics...
Abstract A novel combination of two widely‐used clustering algorithms is proposed here for the detection and reduction high data density regions. The density‐based spatial applications with noise (DBSCAN) algorithm used regions k ‐means reduction. iterates while successively decrementing DBSCAN search radius, allowing an adaptive factor based on effective density. demonstrated a physics simulation application, where surrogate model fusion reactor plasma turbulence generated neural networks....
Abstract For more than a decade, an unprecedented predict-first activity has been carried in order to predict the fusion power and provide guidance second Deuterium–Tritium (D–T) campaign performed at JET 2021 (DTE2). Such provided framework for broad model validation development towards D–T operation. It is shown that it necessary go beyond projections using scaling laws obtain detailed physics based predictions. Furthermore, mixing different modelling complexity promoting extended...
In this report, we explore the ability of language model agents to acquire resources, create copies themselves, and adapt novel challenges they encounter in wild. We refer cluster capabilities as "autonomous replication adaptation" or ARA. believe that systems capable ARA could have wide-reaching hard-to-anticipate consequences, measuring forecasting may be useful for informing measures around security, monitoring, alignment. Additionally, once a system is ARA, placing bounds on system's...
This work presents a fast and robust method for optimizing the stationary radial distribution of temperature, density parallel current in tokamak plasma its application to first-principle-based modeling ITER hybrid scenario. A new solver is implemented RAPTOR transport code, enabling direct evaluation solution which profiles evolve. Coupled neural network emulation quasi-linear gyrokinetic QuaLiKiz model (QLKNN-hyper-10D), estimate state core can be found at unprecedented computational speed...
Neoclassical and turbulent heavy impurity transport in tokamak core plasmas are determined by main ion temperature, density toroidal rotation profiles. Thus, order to understand prevent experimental behaviour of W accumulation, flux-driven integrated modelling heat particle over multiple confinement times is a vital prerequisite. For the first time, quasilinear gyrokinetic code QuaLiKiz has been applied for successful predictions kinetic profiles an ASDEX Upgrade H-mode discharge turbulence...
Core turbulent particle transport with multiple isotopes can display observable differences in behaviour between the electron and ion channels. Experimental observations at JET mixed H-D plasmas varying NBI gas-puff sources [M. Maslov et al., Nucl. Fusion 7 076022 (2018)] inferred source dominated peaking, but isotope peaking. In this work, we apply QuaLiKiz quasilinear gyrokinetic model within JINTRAC flux-driven integrated modelling, for core validation multiple-isotope regime. The...
Abstract Nuclear fusion reactor plasmas will need to exhaust a significant proportion of energy flux through radiative processes, enable acceptable divertor loads. This can be obtained by line radiation from impurities, injected the plasma edge. There are however limitations on sustainable impurity content, since core lead deleterious electron heat sink. Moreover, dilution main ions reduces available fuel. Simultaneously, impurities have an impact turbulent transport, both and changes in...
Some radiotherapy patients are treated with titanium surgical aneurysm clips in the radiation field. This is of particular importance for stereotactic radiosurgery brain treatments, where length blade clip may be comparable to size study seeks determine extent dosimetric effects caused by radiosurgery, using polyacrylamide gel phantoms and EBT type Gafchromic films. Using scanned magnetic resonance imaging scanner, dose enhancement around 20% was noted at distances less than 2 mm away from...