Search NASA⌕ Search

SEARCH · Search NASA

Results for “error mitigation”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 541 records · Page 30

The Fragment Molecular Orbital Method Based on Long-Range Corrected Density-Functional Tight-Binding

The presently available linear scaling approaches to density-functional tight-binding (DFTB) based on the fragment molecular orbital (FMO) method are severely impacted by the problem of artificial charge transfer due to the self-interaction error (SIE), which hampers the simulation of zwitterionic systems such as biopolymers or ionic liquids. Here we report an extension of FMO-DFTB where we included a long-range corrected (LC) functional designed to mitigate the DFTB SIE, called the FMO-LC-DFTB method, resulting in a robust method which succeeds in simulating zwitterionic systems. Both energy and analytic gradient are developed for the gas phase and the polarizable continuum model of solvation. The scaling of FMO-LC-DFTB with system size N is shown to be almost linear, O( N 1.13–1.28 ), and its numerical accuracy is established for a variety of representative systems including neutral and charged polypeptides. It is shown that pair interaction energies between fragments for two mini-proteins are in excellent agreement with results from long-range corrected density functional theory. Here the new method was employed in long time scale (1 ns) molecular dynamics simulations of the tryptophan cage protein (PDB: 1L2Y) in the gas phase for four different protonation states and in stochastic global minimum structure searches for 1-ethyl-3-methylimidazolium nitrate ionic liquid clusters containing up to 2300 atoms.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Characterizing and Mitigating Intraday Variability: Reconstructing Source Structure in Accreting Black Holes with mm-VLBI

The extraordinary physical resolution afforded by the Event Horizon Telescope has opened a window onto the astrophysical phenomena unfolding on horizon scales in two known black holes, M87* and Sgr A*. However, with this leap in resolution has come a new set of practical complications. Sgr A* exhibits intraday variability that violates the assumptions underlying Earth aperture synthesis, limiting traditional image reconstruction methods to short timescales and data sets with very sparse (u, v) coverage. We present a new set of tools to detect and mitigate this variability. We develop a data-driven, model-agnostic procedure to detect and characterize the spatial structure of intraday variability. This method is calibrated against a large set of mock data sets, producing an empirical estimator of the spatial power spectrum of the brightness fluctuations. We present a novel Bayesian noise modeling algorithm that simultaneously reconstructs an average image and statistical measure of the fluctuations about it using a parameterized form for the excess variance in the complex visibilities not otherwise explained by the statistical errors. These methods are validated using a variety of simulated data, including general relativistic magnetohydrodynamic simulations appropriate for Sgr A* and M87*. We find that the reconstructed source structure and variability are robust to changes in the underlying image model. We apply these methods to the 2017 EHT observations of M87*, finding evidence for variability across the EHT observing campaign. The variability mitigation strategies presented are widely applicable to very long baseline interferometry observations of variable sources generally, for which they provide a data-informed averaging procedure and natural characterization of inter-epoch image consistency.

79 ASTRONOMY AND ASTROPHYSICS↗

Mitigating the Sign Problem through Basis Rotations

Quantum Monte Carlo simulations of quantum many-body systems are plagued by the Fermion sign problem. The computational complexity of simulating Fermions scales exponentially in the projection time β and system size. The sign problem is basis dependent and an improved basis, for fixed errors, leads to exponentially quicker simulations. Herein, we show how to use sign-free quantum Monte Carlo simulations to optimize over the choice of basis on large two-dimensional systems. We numerically illustrate these techniques decreasing the “badness” of the sign problem by optimizing over single-particle basis rotations on one- and two-dimensional Hubbard systems. We find a generic rotation which improves the average sign of the Hubbard model for a wide range of U and densities for L×4 systems. In one example improvement, the average sign (and hence simulation cost at fixed accuracy) for the 16×4 Hubbard model at U/t=4 and n=0.75 increases by exp[8.64(6)β]. For typical projection times of β ≳ 100, this accelerates such simulation by many orders of magnitude.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Mastering HPC Runtime Prediction: From Observing Patterns to a Methodological Approach

The continual expansion of high-performance computing (HPC) brings with it an increasing need for efficiency. Heavy investment in energy, hardware, and software infrastructure to support peta- and exascale computing requires the optimization of existing systems and, wherever possible, the discernment and adoption of best-practices towards these goals. Such is the case for runtime prediction. When a job is submitted to an HPC system, an estimate of its runtime is provided by the user in the form of "requested wallclock". Error in this user-provided estimate can lead to jobs being prematurely killed by the scheduler, increased wait time on the queue, and decreased system utilization. More than fifteen years of research has been directed at mitigating these effects by using data-driven runtime predictions. Codified here is a set of commonalities and insights emerging from this body of work, which we present as recommendations and best practices. These practices are combined into a methodological approach described and evaluated on an 11-million-job dataset from the National Renewable Energy Laboratory's petascale HPC system, Eagle. This dataset and the accompanying codebase have been released to the public domain for the benefit of the wider HPC research community.

high performance computing↗

Fundamentals of Solar PV Bolted Joint Loosening and Prevention

The photovoltaic (PV) industry has long reported anecdotal accounts of systems exhibiting intermittent or chronic fastener loosening, including joints that fail to maintain preload despite multiple re-tightening attempts. These occurrences are frequently, and often incorrectly, attributed to installer errors, vibration, or loading beyond design expectations (e.g., extreme weather events). Loose fasteners have serious implications and can significantly impact solar PV systems’ performance, reliability, and safety. However, effective bolted joint design and proper assembly practices can mitigate or eliminate loosening. Until the US Department of Energy’s Solar Energy Technologies Office (SETO) funded research on fasteners and solar PV structures, there was a notable gap in understanding the causes of loosening in solar PV systems.

14 SOLAR ENERGY↗

Mastering HPC Runtime Prediction: From Observing Patterns to a Methodological Approach: Preprint

The continual expansion of high-performance computing (HPC) brings with it an increasing need for efficiency. Heavy investment in energy, hardware, and software infrastructure to support peta- and exascale computing requires the optimization of existing systems and, wherever possible, the discernment and adoption of best-practices towards these goals. Such is the case for runtime prediction. When a job is submitted to an HPC system, an estimate of its runtime is provided by the user in the form of "requested wallclock''. Error in this user-provided estimate can lead to jobs being prematurely killed by the scheduler, increased wait time on the queue, and decreased system utilization. More than fifteen years of research has been directed at mitigating these effects by using data-driven runtime predictions. Codified here is a set of commonalities and insights emerging from this body of work, which we present as recommendations and best practices. These practices are combined into a methodological approach described and evaluated on an 11-million-job dataset from the National Renewable Energy Laboratory's petascale HPC system, Eagle. This dataset and the accompanying codebase have been released to the public domain for the benefit of the wider HPC research community.

high performance computing↗

SPECTRAL RELATIVE ABSORPTION DIFFERENCE METHOD

When analyzing field data, the uncertainty in the background continuum emission produces the majority of error in the final gamma-source analysis. The background emission typically dominates an observed spectrum in terms of counts and is highly variable spatially and temporally. The majority of the spectral shape of the background continuum is produced by combinations of cosmic rays, {sup 40}K, {sup 235}U, and {sup 220}Rn, and the continuum is similar in shape to the 15%-20% level for most field observations. However, the goal of spectroscopy analysis is to pick up subtle peaks (<%5) upon this large background. Because the continuum is falling off as energy increases, peak detection algorithms must first define the background surrounding the peak. This definition is difficult when the range of background shapes is considered. The full spectral template matching algorithms are heavily weighted to solving for the background continuum as it produces significant counts over much of the energy range. The most appropriate background mitigation technique is to take a separate background observation without the source of interest. But, it is frequently not possible to record a background observation in the exact location before (or after) a source has been detected. Thus, one uses approximate backgrounds that rely on spatially nearby locations or similar environments. Since the error in many field observations is dominated by the background, a technique that is less sensitive to the background would be quite beneficial. We report the result of an initial investigation into a novel observation scheme for gamma-emission detection in high background environments. Employing low resolution, NaI, detectors, we examine the different between the direct emission and the 'spectral-shadow' that the gamma emission produces when passed through a thin absorber. For this detection scheme to be competitive, it is required to count and analyze individual gamma-events. We describe the unique instrumental setup which we assembled to make these measurements.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Radiation asymmetry during shattered pellet and massive gas injection in DIII-D

Infrared thermography of the first wall in DIII-D is used to show the toroidal peaking of plasma radiation during mitigated disruptions with shattered pellet injection (SPI) and with massive gas injection (MGI). During MGI, the radiation peak location is shown to be due to the n = 1 magnetohydrodynamic (MHD) activity and continuously variable in toroidal phase based on externally applied error fields, consistent with previous experimental results. Furthermore, the measured toroidal peaking factor (TPF) is in agreement with the previous estimates based on radiometry, with a total TPF of 1.07 ± 0.05 when integrated over the entire duration of the disruption, and a value of 1.3 ± 0.1 during the thermal quench (TQ). For SPI, the location of the radiation peak is found to be determined predominantly by the particle source location, and thus can be varied on DIII-D by injecting from either of two toroidally separated injectors. 3D non-linear MHD simulations with the NIMROD code support this picture. The localized SPI particle source leads to higher peaking than for MGI, with a total TPF of 1.3 ± 0.1 over the full disruption and a TQ TPF of 1.9 +0.5/−0.3, a value also consistent with the NIMROD simulations. This TQ value is close to or potentially exceeding the allowable limit for the previously planned Be first wall in ITER, pointing to the importance of radiation asymmetries depending on the choice of first wall material.

Physics - Plasma physics↗

HPDR: High-Performance Portable Scientific Data Reduction Framework

The rapid growth in scientific data generation is outpacing advancements in computing systems necessary for efficient storage, transfer, and analysis, particularly in the context of exascale computing. With the deployment of first-generation exascale computing systems and next-generation experimental facilities, this gap is widening and necessitates effective data reduction techniques to manage enormous data volumes. Over the past decade, various data reduction methods, including lossless compression, error-controlled lossy compression, and data refactoring, have been developed to accelerate I/O in scientific workflows. Despite significant reductions in data volume, these methods introduce considerable computational overhead, which can become the new bottleneck in data processing. To mitigate this, GPU-accelerated data reduction algorithms have been introduced. However, challenges remain in their integration into exascale workflows, including limited portability across different GPU architectures, substantial memory transfer overhead, and reduced scalability on dense multi-GPU systems. To address these challenges, we propose HPDR, a high-performance and portable data reduction framework. HPDR is designed to enable the execution of state-of-the-art reduction algorithms across diverse processor architectures while reducing memory transfer overhead to 2.3 % of the original, resulting in up to 3.5× faster throughput compared to existing solutions. It also achieves up to 96% of the theoretical speedup in multi-GPU settings. In addition, evaluations on accelerating I/O operations at scale up to 1,024 nodes of the Frontier supercomputer demonstrate that HPDR can achieve up to 103 TB/s reduction throughput, providing up to 4× acceleration in parallel I/O performance compared to existing data reduction routines. This work highlights the potential of HPDR to significantly enhance data reduction efficiency in exascale computing environments.

Chen, Jieyang [University of Oregon]↗

Calculating tumor trajectory and dose-of-the-day using cone-beam CT projections

Purpose: Cone-beam CT (CBCT) projection images provide anatomical data in real-time over several respiratory cycles, forming a comprehensive picture of tumor movement. The authors developed and validated a method which uses these projections to determine the trajectory of and dose to highly mobile tumors during each fraction of treatment. Methods: CBCT images of a respiration phantom were acquired, the trajectory of which mimicked a lung tumor with high amplitude (up to 2.5 cm) and hysteresis. A template-matching algorithm was used to identify the location of a steel BB in each CBCT projection, and a Gaussian probability density function for the absolute BB position was calculated which best fit the observed trajectory of the BB in the imager geometry. Two modifications of the trajectory reconstruction were investigated: first, using respiratory phase information to refine the trajectory estimation (Phase), and second, using the Monte Carlo (MC) method to sample the estimated Gaussian tumor position distribution. The accuracies of the proposed methods were evaluated by comparing the known and calculated BB trajectories in phantom-simulated clinical scenarios using abdominal tumor volumes. Results: With all methods, the mean position of the BB was determined with accuracy better than 0.1 mm, and root-mean-square trajectory errors averaged 3.8% ± 1.1% of the marker amplitude. Dosimetric calculations using Phase methods were more accurate, with mean absolute error less than 0.5%, and with error less than 1% in the highest-noise trajectory. MC-based trajectories prevent the overestimation of dose, but when viewed in an absolute sense, add a small amount of dosimetric error (<0.1%). Conclusions: Marker trajectory and target dose-of-the-day were accurately calculated using CBCT projections. This technique provides a method to evaluate highly mobile tumors using ordinary CBCT data, and could facilitate better strategies to mitigate or compensate for motion during stereotactic body radiotherapy.

60 APPLIED LIFE SCIENCES↗

Efficient Training of Deep Neural Operator Networks via Randomized Sampling

Neural operators (NOs) employ deep neural networks to learn the mappings between infinitedimensional function spaces. Deep operator network (DeepONet), a popular NO architecture, has demonstrated success in the real-time prediction of complex dynamics across various scientific and engineering applications. In this work, we introduce a random sampling technique to be adopted during the training of DeepONet, aimed at improving the generalization ability of the model, while significantly reducing the computational time. The proposed approach targets the trunk network of the DeepONet model that outputs the basis functions corresponding to the spatiotemporal locations of the bounded domain on which the physical system is defined. While constructing the loss function, DeepONet training traditionally considers a uniform grid of spatiotemporal points at which all the output functions are evaluated for each iteration. This approach leads to a larger batch size, resulting in poor generalization and increased memory demands, due to the limitations of the stochastic gradient descent (SGD) optimizer. The proposed random sampling over the inputs of the trunk net mitigates these challenges, improving generalization and reducing the memory requirements during training, resulting in significant computational gains. We validate our hypothesis through three benchmark examples, demonstrating substantial reductions in training time while achieving comparable or lower overall test errors relative to the traditional training approach. Our results indicate that incorporating randomization in the trunk network inputs during training enhances the efficiency and robustness of DeepONet, offering a promising avenue for improving the framework’s performance in modeling complex physical systems.

Karumuri, Sharmila [Department of Civil & Systems ↗

Sensor-Assisted Fault Mitigation in Quantum Computation

We propose a method for assisting fault mitigation in quantum computation through the use of sensors, co-located near physical qubits. Specifically, we consider using transition edge sensors co-located on silicon substrates hosting superconducting qubits to monitor for energy injection from ionizing radiation that has been demonstrated to increase decoherence in transmon qubits. We generalize from these two physical device concepts and explore the potential advantages of co-located sensors to assist fault mitigation in quantum computation. In the simplest scheme, co-located sensors beneficially assist rejection of calculations potentially affected by environmental disturbances. Investigating the potential computational advantage further required development of an extension to the standard formulation of quantum error correction. In a specific case the standard three-qubit, bit flip quantum correction code, we show that given a 20% overall error probability per qubit, approximately 90% of repeated calculation attempts are correctable. However, when sensor-detectable errors account for 45% of overall error probability, the use of colocated sensors uniquely associated with independent qubits, boosts the fraction of correct final-state calculations to 96% at the cost of rejecting 7% of repeated calculation attempts.

Orrell, John L.↗

Coriolis force compensation and laser beam delivery for 100-m baseline atom interferometry

The Coriolis force is a significant source of systematic phase errors and dephasing in atom interferometry and is often compensated by counter-rotating the interferometry laser beam against Earth's rotation. We present a novel method for performing Coriolis force compensation for long-baseline atom interferometry, which mitigates atom-beam misalignment due to beam rotation, an effect which is magnified by the long lever arm of the baseline length. The method involves adjustment of the angle of the interferometer beam prior to a magnifying telescope, enabling the beam to pivot around a tunable position along the interferometer baseline. By tuning the initial atom kinematics and adjusting the angle with which the interferometer beam pivots about this point, we can ensure that the atoms align with the center of the beam during the atom optics laser pulses. This approach will be used in the MAGIS-100 atom interferometer and could also be applied to other long-baseline atom interferometers. An additional challenge associated with long baseline interferometry is that since long-baseline atom interferometers are often located outside of typical laboratory environments, facilities constraints may require lasers to be housed in a climate-controlled room a significant distance away from the main experiment. Nonlinear effects in optical fibers restrict the use of fiber-based transport of the high-power interferometry beam from the laser room to the experiment. We present the design of and prototype data from a laser transport system for MAGIS-100 that maintains robustness against alignment drifts despite the absence of a long fiber.

Glick, Jonah↗

The MP{sub N} method: a new angular discretization method based on piecewise polynomial interfaces fluxes

In transport calculations, it is well known how S{sub N} method is extremely inefficient in problems where the particle physics is dominated by streaming. The ray-effect eventually produced by the insufficient angular discretization, appears to be extremely persistent with respect to the refinement of the angular quadrature. The MP{sub N} method, that relies on continuous angular representation, offers a robust remedy to such an issue. MP{sub N} is based on the decomposition of the unit sphere into solid angles and on a piecewise continuous definition of interface fluxes, which are expanded in polynomials in each solid angle. This allows propagating more than one angular degree of freedom simultaneously while maintaining unaltered the block-diagonal pattern of the displacement plus removal operator. The method is therefore well suited for the flux resolution by means of a conventional sweep algorithm. Furthermore, unlike the S{sub N} method, MP{sub N} does not rely on discrete directions and, thus, on an angular quadrature formula, but rather constructs a set of linear equations solving for the angular moments of the flux for all discrete solid angles within the sweep. MP{sub N} shows an error convergence rate higher than S{sub N} at the expense of an increased size of the coefficient matrices, so of the computational cost. Although MP{sub N} is not free from ray-effect, the latter is effectively mitigated and less persistent with respect to the increase of the angular refinement order. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Discretising the velocity distribution for directional dark matter experiments

Dark matter (DM) direct detection experiments which are directionally-sensitive may be the only method of probing the full velocity distribution function (VDF) of the Galactic DM halo. We present an angular basis for the DM VDF which can be used to parametrise the distribution in order to mitigate astrophysical uncertainties in future directional experiments and extract information about the DM halo. This basis consists of discretising the VDF in a series of angular bins, with the VDF being only a function of the DM speed v within each bin. In contrast to other methods, such as spherical harmonic expansions, the use of this basis allows us to guarantee that the resulting VDF is everywhere positive and therefore physical. We present a recipe for calculating the event rates corresponding to the discrete VDF for an arbitrary number of angular bins N and investigate the discretisation error which is introduced in this way. For smooth, Standard Halo Model-like distribution functions, only N=3 angular bins are required to achieve an accuracy of around 10–30% in the number of events in each bin. Shortly after confirmation of the DM origin of the signal with around 50 events, this accuracy should be sufficient to allow the discretised velocity distribution to be employed reliably. For more extreme VDFs (such as streams), the discretisation error is typically much larger, but can be improved with increasing N. This method paves the way towards an astrophysics-independent analysis framework for the directional detection of dark matter.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Overview of Results from the National Spherical Torus Experiment (NSTX)

The mission of NSTX is the demonstration of the physics basis required to extrapolate to the next steps for the spherical torus (ST), such as a plasma facing component test facility (NHTX) or an ST based component test facility (ST-CTF), and to support ITER. Key issues for the ST are transport, and steady state high β operation. To better understand electron transport, a new high-k scattering diagnostic was used extensively to investigate electron gyro-scale fluctuations with varying electron temperature gradient scale-length. Results from n = 3 braking studies are consistent with the flow shear dependence of ion transport. New results from electron Bernstein wave emission measurements from plasmas with lithium wall coating applied indicate transmission efficiencies near 70% in H-mode as a result of reduced collisionality. Improved coupling of High Harmonic Fast-Waves has been achieved by reducing the edge density relative to the critical density for surface wave coupling. In order to achieve high bootstrap current fraction, future ST designs envision running at very high elongation. Plasmas have been maintained on NSTX at very low internal inductance l i ~0.4 with strong shaping (κ ~ 2.7, δ ~ 0.8) with β N approaching the with-wall beta limit for several energy confinement times. By operating at lower collisionality in this regime, NSTX has achieved record non-inductive current drive fraction f NI ~71%. Instabilities driven by super-Alfv´enic ions will be an important issue for all burning plasmas, including ITER. Fast ions from NBI on NSTX are super-Alfv´enic. Linear TAE thresholds and appreciable fast-ion loss during multi-mode bursts are measured and these results are compared to theory. The impact of n > 1 error fields on stability is a important result for ITER. RWM/RFA feedback combined with n=3 error field control was used on NSTX to maintain plasma rotation with β above the no-wall limit. Other highlights are: results of lithium coating experiments, momentum confinement studies, scrape-off layer width scaling, demonstration of divertor heat load mitigation in strongly shaped plasmas, and coupling of CHI plasmas to OH ramp-up. These results advance the ST towards next step fusion energy devices such as NHTX and ST-CTF

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

The IEC 61850 Sampled Measured Values Protocol: Analysis, Threat Identification, and Feasibility of Using NN Forecasters to Detect Spoofed Packets

The operation of the smart grid is anticipated to rely profoundly on distributed microprocessor-based control. Therefore, interoperability standards are needed to address the heterogeneous nature of the smart grid data. Since the IEC 61850 emerged as a wide-spread interoperability standard widely accepted by the industry, the Sampled Measured Values method has been used to communicate digitized voltage and current measurements. Realizing that current and voltage measurements (i.e., feedback measurements) are necessary for reliable and secure noperation of the power grid, firstly, this manuscript provides a detailed analysis of the Sampled Measured Values protocol emphasizing its advantages, then, it identifies vulnerabilities in this protocol and explains the cyber threats associated to these vulnerabilities. Secondly, current efforts to mitigate these vulnerabilities are outlined and the feasibility of using neural network forecasters to detect spoofed sampled values is investigated. It was shown that although such forecasters have high spoofed data detection accuracy, they are prone to the accumulation of forecasting error. Accordingly, this paper also proposes an algorithm to detect the accumulation of the forecasting error based on lightweight statistical indicators. The effectiveness of the proposed methods is experimentally verified in a laboratory-scale smart grid testbed.

97 MATHEMATICS AND COMPUTING↗

Overview of results from the National Spherical Torus Experiment (NSTX)

The mission of the National Spherical Torus Experiment (NSTX) is the demonstration of the physics basis required to extrapolate to the next steps for the spherical torus (ST), such as a plasma facing component test facility (NHTX) or an ST based component test facility (ST-CTF), and to support ITER. Key issues for the ST are transport, and steady state high beta operation. To better understand electron transport, a new high-k scattering diagnostic was used extensively to investigate electron gyro-scale fluctuations with varying electron temperature gradient scale length. Results from n = 3 braking studies are consistent with the flow shear dependence of ion transport. New results from electron Bernstein wave emission measurements from plasmas with lithium wall coating applied indicate transmission efficiencies near 70% in H-mode as a result of reduced collisionality. Improved coupling of high harmonic fast-waves has been achieved by reducing the edge density relative to the critical density for surface wave coupling. In order to achieve high bootstrap current fraction, future ST designs envision running at very high elongation. Plasmas have been maintained on NSTX at very low internal inductance l(i) similar to 0.4 with strong shaping (kappa similar to 2.7, delta similar to 0.8) with beta(N) approaching the with-wall beta-limit for several energy confinement times. By operating at lower collisionality in this regime, NSTX has achieved record non-inductive current drive fraction f(NI) similar to 71%. Instabilities driven by super-Alfvenic ions will be an important issue for all burning plasmas, including ITER. Fast ions from NBI on NSTX are super-Alfvenic. Linear toroidal Alfven eigenmode thresholds and appreciable fast ion loss during multi-mode bursts are measured and these results are compared with theory. The impact of n > 1 error fields on stability is an important result for ITER. Resistive wall mode/resonant field amplification feedback combined with n = 3 error field control was used on NSTX to maintain plasma rotation with beta above the no-wall limit. Other highlights are results of lithium coating experiments, momentum confinement studies, scrape-off layer width scaling, demonstration of divertor heat load mitigation in strongly shaped plasmas and coupling of coaxial helicity injection plasmas to ohmic heating ramp-up. These results advance the ST towards next step fusion energy devices such as NHTX and ST-CTF.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗