Search NASA⌕ Search

SEARCH · Search NASA

Results for “Model Inference”

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 559 records · Page 31
All sourcesNASA NTRSDOE OSTI
DOE OSTI2026

A non-intrusive framework using acoustic signals and deep learning for boiling diagnostics in visual-limited environments

Accurate monitoring of boiling heat transfer is critical for safeguarding high-power systems operating in environments where conventional optical diagnostics are hindered by radiation fields or restricted visual accessibility. This study presents a non-intrusive framework that integrates hydroacoustic sensing with deep learning to infer near-wall boiling characteristics and enable predictive thermal assessment without visual access. In a prototypical subcooled flow-boiling facility representative of the Isotope Production Facility (IPF) at Los Alamos, hydrophones capture boiling-induced acoustic emissions that are transformed into background-removed Short-Time Fourier Transform (STFT) spectrograms. A convolutional neural network (CNN) then regresses heat flux, wall superheat, and key bubble parameters directly from these spectrograms. The CNN achieved predictive accuracy under nominal conditions and demonstrated robustness and generalization under acoustic noise for Signal-to-Noise Ratios (SNRs) down to approximately 0 dB. When integrated into an ANSYS CFX wall-boiling model, the acoustically inferred parameters reproduced boiling curve and critical heat flux (CHF) values consistent with image-based benchmarks. Furthermore, the model retained reliable performance under moderate variations in bulk temperature, flow rate, and hydrophone placement, confirming its generalizability across practical boundary conditions. These results demonstrate the feasibility of hydroacoustic-based deep learning as a viable path toward real-time, radiation-tolerant boiling diagnostics and predictive thermal safety assessment in inaccessible systems such as the IPF.

42 ENGINEERING↗
DOE OSTI2022

Proteome-scale Deployment of Protein Structure Prediction Workflows on the Summit Supercomputer

Deep learning has contributed to major advances in the prediction of protein structure from sequence, a fundamental problem in structural bioinformatics. With predictions now approaching the accuracy of crystallographic experiments, and with accelerators like GPUs and TPUs making inference using large models rapid, genome-level structure prediction becomes an obvious aim. Leadership-class computing resources can be used to perform genome-scale protein structure prediction using state-of-the-art deep learning models, providing a wealth of new data for systems biology applications. Here we describe our efforts to efficiently deploy the AlphaFold v.2 program, for full-proteome structure prediction, at scale on the Oak Ridge Leadership Computing Facility's resources, including the Summit supercomputer. We performed inference to produce the predicted structures for 40,526 protein sequences, corresponding to four prokaryotic proteomes and one plant proteome, using under 4,400 total Summit node hours, equivalent to using the majority of the supercomputer for a little over one hour. We also designed an optimized structure refinement that reduced the time for the relaxation stage of the AlphaFold pipeline by over 10X for longer sequences. We demonstrate the types of analyses that can be performed on proteome-scale collections of sequences, including a search for novel quaternary structures and implications for functional annotation.

Gao, Mu↗
DOE OSTI2019

A Novel Statistical Method for Measuring the Temperature–Density Relation in the IGM Using the b – N Hi Distribution of Absorbers in the Ly α Forest

We propose a new method for determining the thermal state of the intergalactic medium based on Voigt profile decomposition of the Lyα forest. The distribution of Doppler parameter and column density (b–N H i distribution) is sensitive to the temperature–density relation T = T 0 (ρ/ρ 0 ) γ-1 , and previous work has inferred T 0 and γ by fitting its low-b cutoff. This approach discards the majority of available data and is susceptible to systematics related to cutoff determination. We introduce an approach that exploits all information encoded in the b – N H i distribution by modeling its entire shape. We apply kernel density estimation to discrete absorption lines to generate model probability density functions, and then we use principal component decomposition to create an emulator that can be evaluated anywhere in thermal parameter space. We introduce a Bayesian likelihood based on these models enabling parameter inference via Markov Chain Monte Carlo. The method's robustness is tested by applying it to a large grid of thermal history simulations. By conducting 160 mock measurements, we establish that our approach delivers unbiased estimates and valid uncertainties for a 2D ( T 0 , γ) measurement. Furthermore, we conduct a pilot study applying this methodology to real observational data at z = 2. Using 200 absorbers, equivalent in path length to a single Lya forest spectrum, we measure $\mathrm{log}{T}_{0}={4.092}_{-0.055}^{+0.050}$ and $\gamma ={1.49}_{-0.074}^{+0.073}$ in excellent agreement with cutoff fitting determinations using the same data. Our method is far more sensitive than cutoff fitting, enabling measurements of log T 0 and γ with precision on $\mathrm{log}{T}_{0}$ (γ) nearly two (three) times higher for current data set sizes.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2021

Particle image velocimetry analysis with simultaneous uncertainty quantification using Bayesian neural networks

Particle image velocimetry (PIV) is an effective tool in experimental fluid mechanics for extracting flow fields from images. Recently, convolutional neural networks (CNNs) have been used to perform PIV analysis with accuracy on par with classical methods. Here we extend the use of CNNs to analyze PIV data while providing simultaneous uncertainty quantification on the inferred flow field. The method we apply in this paper is a Bayesian convolutional neural network (BCNN) which learns distributions of the CNN weights through variational Bayes. In order to demonstrate the utility of BCNNs for the PIV task, we compare the performance of three distinct BCNN models with simple architectures. The first network estimates flow velocity from image interrogation regions only. Our second model learns to infer velocity from both the image interrogation regions and interrogation region cross-correlation maps. Finally, our best performing network infers velocities from interrogation region cross-correlation maps only. We find that BCNNs using interrogation region cross-correlation maps as inputs perform better than those using interrogation windows only as inputs and discuss reasons why this may be the case. Additionally, we test the best performing BCNN on a full synthetic test image pair and a real image pair from the 1st International PIV Challenge. We show that ~98% of true particle displacements from the full synthetic image pair can be captured within the BCNN's 95% confidence intervals, and that the BCNN's performance on the real image pair is quantitatively similar to that of algorithms tested in the 1st International PIV Challenge. Finally, we show that BCNNs can be generalized to be used with multi-pass PIV algorithms with a moderate loss in accuracy, which may be overcome by future work on finetuning and training schemes. So to our knowledge, this is the first use of Bayesian neural networks to perform PIV.

47 OTHER INSTRUMENTATION↗
NASA NTRS1988

How well do we know the Martian abundances of highly volatile elements

The table is given that summarizes four models based on observations and gives some reasonable inferences of each for Martian volatiles. The table columns give: (1) the volatile assumptions of the model; (2) the inferred Martian volatile concentrations relative to earth; (3) the N/36 Ar and C/36 Ar ratios; (4) the percentage of Martian degassing inferred by the model, and 5 and 6) the equivalent global column heights of liquid water and solid CaCO3 over the entire surface of Mars that would be formed from the model quantities of degassed volatiles.

Bogard, Donald↗
DOE OSTI2021

Overview of Algorithms for Using Particle Morphology in Pre-Detonation Nuclear Forensics

A major goal in pre-detonation nuclear forensics is to infer the processing conditions and/or facility type that produced radiological material. This review paper focuses on analyses of particle size, shape, texture (“morphology”) signatures that could provide information on the provenance of interdicted materials. For example, uranium ore concentrates (UOC or yellowcake) include ammonium diuranate (ADU), ammonium uranyl carbonate (AUC), sodium diuranate (SDU), magnesium diuranate (MDU), and others, each prepared using different salts to precipitate U from solution. Once precipitated, UOCs are often dried and calcined to remove adsorbed water. The products can be allowed to react further, forming uranium oxides UO3, U3O8, or UO2 powders, whose surface morphology can be indicative of precipitation and/or calcination conditions used in their production. This review paper describes statistical issues and approaches in using quantitative analyses of measurements such as particle size and shape to infer production conditions. Statistical topics include multivariate t tests (Hotelling’s T 2 ), design of experiments, and several machine learning (ML) options including decision trees, learning vector quantization neural networks, mixture discriminant analysis, and approximate Bayesian computation (ABC). ABC is emphasized as an attractive option to include the effects of model uncertainty in the selected and fitted forward model used for inferring processing conditions.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗
DOE OSTI2017

Bayesian inverse modeling of the atmospheric transport and emissions of a controlled tracer release from a nuclear power plant

Probability distribution functions (PDFs) of model inputs that affect the transport and dispersion of a trace gas released from a coastal California nuclear power plant are quantified using ensemble simulations, machine-learning algorithms, and Bayesian inversion. The PDFs are constrained by observations of tracer concentrations and account for uncertainty in meteorology, transport, diffusion, and emissions. Meteorological uncertainty is calculated using an ensemble of simulations of the Weather Research and Forecasting (WRF) model that samples five categories of model inputs (initialization time, boundary layer physics, land surface model, nudging options, and reanalysis data). The WRF output is used to drive tens of thousands of FLEXPART dispersion simulations that sample a uniform distribution of six emissions inputs. Machine-learning algorithms are trained on the ensemble data and used to quantify the sources of ensemble variability and to infer, via inverse modeling, the values of the 11 model inputs most consistent with tracer measurements. We find a substantial ensemble spread in tracer concentrations (factors of 10 to 10 3 ), most of which is due to changing emissions inputs (about 80%), though the cumulative effects of meteorological variations are not negligible. The performance of the inverse method is verified using synthetic observations generated from arbitrarily selected simulations. When applied to measurements from a controlled tracer release experiment, the inverse method satisfactorily determines the location, start time, duration and amount. In a 2 km × 2 km area of possible locations, the actual location is determined to within 200m. The start time is determined to within 5min out of 2h, and the duration to within 50min out of 4h. Over a range of release amounts of 10 to 1000kg, the estimated amount exceeds the actual amount of 146kg by only 32kg. The inversion also estimates probabilities of different WRF configurations. In conclusion, to best match the tracer observations, the highest-probability cases in WRF are associated with using a late initialization time and specific reanalysis data products.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2017

Numerically Modeling the First Peak of the Type IIb SN 2016gkg

Many Type IIb supernovae (SNe) show a prominent additional early peak in their light curves, which is generally thought to be due to the shock cooling of extended hydrogen-rich material surrounding the helium core of the exploding star. The recent SN 2016gkg was a nearby Type IIb SN discovered shortly after explosion, which makes it an excellent candidate for studying this first peak. We numerically explode a large grid of extended envelope models and compare these to SN 2016gkg to investigate what constraints can be derived from its light curve. This includes exploring density profiles for both a convective envelope and an optically thick steady-state wind, the latter of which has not typically been considered for Type IIb SNe models. We find that roughly ∼0.02 M{sub ⊙} of extended material with a radius of ≈180--260 R{sub ⊙} reproduces the photometric light curve data, consistent with pre-explosion imaging. These values are independent of the assumed density profile of this material, although a convective profile provides a somewhat better fit. We infer from our modeling that the explosion must have occurred within ≈2–3 hr of the first observed data point, demonstrating that this event was caught very close to the moment of explosion. Nevertheless, our best-fitting 1D models overpredict the earliest velocity measurements, which suggests that the hydrogen-rich material is not distributed in a spherically symmetric manner. We compare this to the asymmetries that have also been seen in the SN IIb remnant Cas A, and we discuss the implications of this for Type IIb SN progenitors and explosion models.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

Dictionary Learning with Accumulator Neurons

The Locally Competitive Algorithm (LCA) uses local competition between non-spiking leaky integrator neurons to infer sparse representations, allowing for potentially real-time execution on massively parallel neuromorphic architectures such as Intel's Loihi processor. Here, we focus on the problem of inferring sparse representations from streaming video using dictionaries of spatiotemporal features optimized in an unsupervised manner for sparse reconstruction. Non-spiking LCA has previously been used to achieve unsupervised learning of spatiotemporal dictionaries composed of convolutional kernels from raw, unlabeled video. We demonstrate how unsupervised dictionary learning with spiking LCA (\hbox{S-LCA}) can be efficiently implemented using accumulator neurons, which combine a conventional leaky-integrate-and-fire (\hbox{LIF}) spike generator with an additional state variable that is used to minimize the difference between the integrated input and the spiking output. We demonstrate dictionary learning across a wide range of dynamical regimes, from graded to intermittent spiking, for inferring sparse representations of both static images drawn from the CIFAR database as well as video frames captured from a DVS camera. On a classification task that requires identification of the suite from a deck of cards being rapidly flipped through as viewed by a DVS camera, we find essentially no degradation in performance as the LCA model used to infer sparse spatiotemporal representations migrates from graded to spiking. We conclude that accumulator neurons are likely to provide a powerful enabling component of future neuromorphic hardware for implementing online unsupervised learning of spatiotemporal dictionaries optimized for sparse reconstruction of streaming video from event based DVS cameras.

artificial intelligence↗
DOE OSTI2023

Machine Learning Based System for Automated Estimation of Crystal Unit Cell Parameters from X-ray Diffraction Powder Patterns (MLCell) v0.1

This software uses machine learning models to perform inference on X-ray diffraction powder patterns, output from a diffractometer, to estimate all 6 unit cell parameters of crystalline lattices. Novel aspects of this software originated from the incorporation of concepts from crystallography to an ML-based approach; The resulting method achieves new performance capabilities without significant loss of accuracy and with less computationally intensive models.

Oklejas, Vanessa↗
DOE OSTI2025

Quantum optimal control of superconducting qubits based on machine-learning characterization

Implementing fast and high-fidelity quantum operations using open-loop quantum optimal control relies on having an accurate model of the quantum dynamics. Any deviations between this model and the complete dynamics of the device, such as the presence of spurious modes or pulse distortions, can degrade the performance of optimal controls in practice. Here, we propose an experimentally simple approach to realize optimal quantum controls tailored to the device parameters and environment while specifically characterizing this quantum system. Concretely, we use physics-inspired machine learning to infer an accurate model of the dynamics from experimentally available data and then optimize our experimental controls on this trained model. We show the power and feasibility of this approach by optimizing arbitrary single-qubit operations in detailed numerical simulations of a superconducting transmon qubit. Furthermore, we demonstrate that this framework produces an accurate description of the device dynamics under arbitrary controls, together with the precise pulses achieving arbitrary single-qubit gates with a high fidelity of ∼99.99%.

Artificial neural networks↗
DOE OSTI2023

Quantifying the Variation in the Number of Donors in Quantum Dots Created Using Atomic Precision Advanced Manufacturing

Atomic-precision advanced manufacturing enables unique silicon quantum electronics built on quantum dots fabricated from small numbers of phosphorus dopants. The number of dopant atoms comprising a dot plays a central role in determining the behavior of charge and spin confined to the dots and thus overall device performance. Here in this work, we use both theoretical and experimental techniques to explore the combined impact of lithographic variation and stochastic kinetics on the number of P incorporations in quantum dots made using these techniques and how this variation changes as a function of the size of the dot. Using a kinetic model of PH3 dissociation augmented with novel reaction barriers, we demonstrate that for a 2 × 3 silicon dimer window the probability that no donor incorporates goes to zero, allowing for certainty in the placement of at least one donor. However, this still comes with some uncertainty in the precise number of incorporated donors (either one or two), and this variability may still impact certain applications. We also examine the impact of the size of the initial lithographic window, finding that the incorporation fraction saturates to δ-layer-like coverage as the circumference-to-area ratio decreases. We predict that this incorporation fraction depends strongly on the dosage of the precursor and that the standard deviation of the number of incorporations scales as ~√n, as would be expected for a sequence of largely independent incorporation events. Finally, we characterize an array of 36 experimentally prepared multidonor 3 × 3 nm lithographic windows with scanning tunneling microscopy, measuring the fidelity of the lithography to the desired array and the final location of PH x fragments within these lithographic windows. We use our kinetic model to examine the expected variability due to the observed lithographic error, predicting a negligible impact on incorporation statistics. We find good agreement between our model and the inferred incorporation locations in these windows from scanning tunneling microscope measurements.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

Implementation of a practical Markov chain Monte Carlo sampling algorithm in PyBioNetFit

Abstract Summary Bayesian inference in biological modeling commonly relies on Markov chain Monte Carlo (MCMC) sampling of a multidimensional and non-Gaussian posterior distribution that is not analytically tractable. Here, we present the implementation of a practical MCMC method in the open-source software package PyBioNetFit (PyBNF), which is designed to support parameterization of mathematical models for biological systems. The new MCMC method, am, incorporates an adaptive move proposal distribution. For warm starts, sampling can be initiated at a specified location in parameter space and with a multivariate Gaussian proposal distribution defined initially by a specified covariance matrix. Multiple chains can be generated in parallel using a computer cluster. We demonstrate that am can be used to successfully solve real-world Bayesian inference problems, including forecasting of new Coronavirus Disease 2019 case detection with Bayesian quantification of forecast uncertainty. Availability and implementation PyBNF version 1.1.9, the first stable release with am, is available at PyPI and can be installed using the pip package-management system on platforms that have a working installation of Python 3. PyBNF relies on libRoadRunner and BioNetGen for simulations (e.g. numerical integration of ordinary differential equations defined in SBML or BNGL files) and Dask.Distributed for task scheduling on Linux computer clusters. The Python source code can be freely downloaded/cloned from GitHub and used and modified under terms of the BSD-3 license (https://github.com/lanl/pybnf). Online documentation covering installation/usage is available (https://pybnf.readthedocs.io/en/latest/). A tutorial video is available on YouTube (https://www.youtube.com/watch?v=2aRqpqFOiS4&t=63s). Supplementary information Supplementary data are available at Bioinformatics online.

59 BASIC BIOLOGICAL SCIENCES↗
NASA NTRS1998

Final Report

This grant was made to fund two accepted proposals to observe with the ASCA satellite: (1) Debunking the myth of two-temperature coronae of active stars, and (2) Dynamic coronae of dMe stars. We obtained the requested observations and have now completed two papers that will be published in the Astrophysical Journal. The RS CVn binary star UX Ari was observed for 14 hours with all four detectors onboard the ASCA satellite. The X-ray emission was at a constant, quiescent level during the first 12 hours, after which time a powerful flare with a peak luminosity of 1.4 x 10(exp 32) ergs/s started. The flare was observed until shortly after its peak. We present a spectral and temporal analysis of the UX Ari observations and analyze the data with a two-ribbon flare model including estimates for cooling losses. A time-dependent reconstruction of the emission measure (EM) distribution shows that two separate plasma components evolve during the flare (one being identified with the quiescent EM). Most of the flare EM reaches temperatures between 50 MK and 100 MK or more. Magnetic confinement requires the loop arcade to be geometrically large, with length scales on the order of one stellar radius. The electron densities inferred from the model decrease from initial values around 10(exp 12)/cc early in the flare to about 10(exp 11)/cc at the flare peak. The best-fit models require surface magnetic field strengths of a few hundred G, compatible with the maximum photospheric fields expected from equipartition. The flare parameters imply a (conductive and radiative) cooling loss time of less than one hour at flare peak. The elemental abundances increase significantly during the flare rise, with the abundances of the low-FIP elements Fe, Mg, Si, and Ni typically increasing to higher levels than the high-FIP elements such as S or Ne. The Fe abundance increases from (17 +/- 4)% of the solar photospheric value during quiescence up to (89 +/- 18)% at flare peak. A fractionation process that occurs during the chromospheric evaporation phase may selectively enrich low-FIP elements as in the solar corona; alternatively, the chromospheric evaporation may itself bring metal-rich plasma into the metal-poor corona.

Linsky, Jeffrey L.↗
NASA NTRS1999

A TRMM Rainfall Estimation Method Applicable to Land Areas

Methods developed to estimate rain rate on a footprint scale over land with the satellite-borne multispectral dual-polarization Special Sensor Microwave Imager (SSM/1) radiometer have met with limited success. Variability of surface emissivity on land and beam filling are commonly cited as the weaknesses of these methods. On the contrary, we contend a more significant reason for this lack of success is that the information content of spectral and polarization measurements of the SSM/I is limited. because of significant redundancy. As a result, the complex nature and vertical distribution C, of frozen and melting ice particles of different densities, sizes, and shapes cannot resolved satisfactorily. Extinction in the microwave region due to these complex particles can mask the extinction due to rain drops. Because of these reasons, theoretical models that attempt to retrieve rain rate do not succeed on a footprint scale. To illustrate the weakness of these models, as an example we can consider the brightness temperature measurement made by the radiometer in the 85 GHz channel (T85). Models indicate that T85 should be inversely related to the rain rate, because of scattering. However, rain rate derived from 15-minute rain gauges on land indicate that this is not true in a majority of footprints. This is also supported by the ship-borne radar observations of rain in the Tropical Oceans and Global Atmosphere Coupled Ocean-Atmosphere Response Experiment (TOGA-COARE) region over the ocean. Based on these observations. we infer that theoretical models that attempt to retrieve rain rate do not succeed on a footprint scale. We do not follow the above path of rain retrieval on a footprint scale. Instead, we depend on the limited ability of the microwave radiometer to detect the presence of rain. This capability is useful to determine the rain area in a mesoscale region. We find in a given rain event that this rain area is closely related to the mesoscale-average rain rate. Based on this observation, in this study we have developed a method to estimate the mesoscale-average rain rate over land utilizing microwave radiometer data. Because of the high degree of geographic and seasonal variability in the nature and intensity of rain, this method requires some tuning with 15-minute rain gauge data on land. After tuning the method, it can be applied to an independent set of rain events that are close in time and space. We find that the mesoscale rain rates retrieved over the period of a month on land with this method shows a correlation of about 0.85 with respect to the surface rain-gauge observations. This mesoscale-average rain rate estimation method can be useful to extend the spatial and temporal coverage of the rainfall data provided by the Precipitation Radar on board the Tropical Rainfall Measuring Mission (TRMM) satellite.

Prabhakara, C.↗
NASA NTRS2014

Turbulent Concentration of MM-Size Particles in the Protoplanetary Nebula: Scaled-Dependent Multiplier Functions

The initial accretion of primitive bodies (asteroids and TNOs) from freely-floating nebula particles remains problematic. Here we focus on the asteroids where constituent particle (read "chondrule") sizes are observationally known; similar arguments will hold for TNOs, but the constituent particles in those regions will be smaller, or will be fluffy aggregates, and are unobserved. Traditional growth-bysticking models encounter a formidable "meter-size barrier" [1] (or even a mm-cm-size barrier [2]) in turbulent nebulae, while nonturbulent nebulae form large asteroids too quickly to explain long spreads in formation times, or the dearth of melted asteroids [3]. Even if growth by sticking could somehow breach the meter size barrier, other obstacles are encountered through the 1-10km size range [4]. Another clue regarding planetesimal formation is an apparent 100km diameter peak in the pre-depletion, pre-erosion mass distribution of asteroids [5]; scenarios leading directly from independent nebula particulates to this size, which avoid the problematic m-km size range, could be called "leapfrog" scenarios [6-8]. The leapfrog scenario we have studied in detail involves formation of dense clumps of aerodynamically selected, typically mm-size particles in turbulence, which can under certain conditions shrink inexorably on 100-1000 orbit timescales and form 10-100km diameter sandpile planetesimals. The typical sizes of planetesimals and the rate of their formation [7,8] are determined by a statistical model with properties inferred from large numerical simulations of turbulence [9]. Nebula turbulence can be described by its Reynolds number Re = L/eta sup(4/3), where L = ETA alpha sup (1/2) the largest eddy scale, H is the nebula gas vertical scale height, and α the nebula turbulent viscosity parameter, and η is the Kolmogorov or smallest scale in turbulence (typically about 1km), with eddy turnover time tη. In the nebula, Re is far larger than any numerical simulation can handle, so some physical model is needed to extend the results of numerical simulations to nebula conditions.

protoplanetary nebula↗
DOE OSTI2017

Deuterium charge exchange recombination spectroscopy from the top of the pedestal to the scrape off layer in H-mode plasmas

Recent completion of the thirty two channel main-ion (deuterium) charge exchange recombination spectroscopy (CER) diagnostic on DIII-D enables detailed comparisons between impurity and main-ion temperature, density, and toroidal rotation. Sixteen sightlines cover the core of the plasma and another sixteen are densely packed towards the edge, providing high resolution measurements of the pedestal and steep gradient edge region of H-mode plasmas. The complexities of the D α spectrum require fitting with a comprehensive model, as well as using iterative collisional radiative modeling to determine the underlying thermal deuterium ion properties. Large differences in the structure and magnitude of impurity (C 6+ ) and main-ion (D + ) toroidal rotation profiles are seen in the H-mode pedestal. Additionally the D + temperature can be half the value of the C 6+ temperature at the separatrix and shows more of a pedestal structure. Typically only the impurity properties are measured and the main-ion properties are either assumed to be the same, or inferred using neoclassical models, which require validation in the steep gradient region. Furthermore, these measured differences have implications for transport model validation, intrinsic rotation studies, pedestal stability, and the boundary conditions for scrape off layer and plasma material interactions studies.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗