Search NASA⌕ Search

SEARCH · Search NASA

Results for “random sampling”

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 1,783 records · Page 99

Composition and Interface Analysis of InGaN/GaN Multiquantum-Wells on GaN Substrates Using Atom Probe Tomography

In0.20Ga0.80N/GaN multi-quantum wells grown on [0001]-oriented GaN substrates with and without an InGaN buffer layer were characterized using three-dimensional atom probe tomography. In all samples, the upper interfaces of the QWs were slightly more diffuse than the lower interfaces. The buffer layers did not affect the roughness of the interfaces within the quantum well structure, a result attributed to planarization of the surface of the 1st GaN barrier layer which had an average root-mean-square roughness of 0.177 nm. The In and Ga distributions within the MQWs followed the expected distributions for a random alloy with no indications of In clustering.

Environmental Molecular Sciences Laboratory↗

Determining Infrared Optical Constants for Dolomite using Single Angle Reflectance and Spectroscopic Ellipsometry

The optical constants, namely the real (n) and imaginary (k) parts of the complex refractive index, are of particular interest to generate the infrared (IR) spectra of liquid and solid materials in different morphologies. To obtain n/k, however, most materials are typically not found in the monolithic forms necessary to easily measure n/k, and thus require the use of other methods such as pressing powders to form planar, specularly-reflective pellets. In this work dolomite crystals are measured using fixed-angle IR reflectance spectroscopy in both 1) a monolithic form with the crystal fixed in epoxy and polished, but data are also recorded from 2) pressed-pellet forms of powders of the same mineral using the same spectroscopic protocols. First results for the two methods are compared. It was found that the measured reflectance can vary by as much as a factor of two between the dolomite crystals and the pressed powder forms. For the two-sample preparation approaches the preliminary spectra are compared and the implications and limitations of each method for determining the optical constants of given materials are discussed. Comparison to literature data suggest that polarization effects likely account for the differing amplitude results for reflectance (and hence the k -vectors): Dolomite is known to be biaxial with significantly differing ordinary and extraordinary rays; the pressed pellet method measures an ensemble of microcrystals in randomly oriented positions, whereas the single crystal maintains just one orientation relative to the optical axis.

ellipsometry, dolomite, complex refractive index, ↗

The DESI DR1 Peculiar Velocity Survey: Fundamental Plane Catalogue

Measurements of peculiar velocities in the local Universe are a powerful tool to study the nature of dark energy at low ($z < 0.1$) redshifts. Here we present the largest single set of $z<0.1$ peculiar velocity measurements to date, obtained using the Fundamental Plane (FP) of galaxies in the first data release (DR1) of the Dark Energy Spectroscopic Instrument (DESI). We describe the photometric and spectroscopic selection criteria used to define the sample, as well as extensive quality control checks on the photometry and velocity dispersion measurements. Additionally, we perform detailed systematics checks for the many analysis parameters in our pipeline. Our DESI DR1 catalogue contains FP-based distances and peculiar velocities for $98,292$ unique early-type galaxies, increasing the total number of $z < 0.1$ FP distances ever measured by a factor of $\sim2$. We achieve a precision of $26\%$ random error in our distance measurements which is comparable to previous surveys. A series of companion DESI papers use the distances and peculiar velocities presented in this paper to measure cosmological parameters.

Ross, C. E. [Queensland U.]↗

Time Series, Stochastic Processes and Completeness of Quantum Theory

Most of physical experiments are usually described as repeated measurements of some random variables. Experimental data registered by on-line computers form time series of outcomes. The frequencies of different outcomes are compared with the probabilities provided by the algorithms of quantum theory (QT). In spite of statistical predictions of QT a claim was made that it provided the most complete description of the data and of the underlying physical phenomena. This claim could be easily rejected if some fine structures, averaged out in the standard descriptive statistical analysis, were found in time series of experimental data. To search for these structures one has to use more subtle statistical tools which were developed to study time series produced by various stochastic processes. In this talk we review some of these tools. As an example we show how the standard descriptive statistical analysis of the data is unable to reveal a fine structure in a simulated sample of AR (2) stochastic process. We emphasize once again that the violation of Bell inequalities gives no information on the completeness or the non locality of QT. The appropriate way to test the completeness of quantum theory is to search for fine structures in time series of the experimental data by means of the purity tests or by studying the autocorrelation and partial autocorrelation functions.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Influence of aqueous hexamethylenetetramine on the morphology of self-assembled SnO{sub 2} nanocrystals

Research highlights: {yields} Synthesis details self-assembly of SnO{sub 2} in aqueous hexamethylenetetramine solution. {yields} Hydrogen bonds clustering around the hexamethylenetetramine interact with the non-polar SnO{sub 2} surfaces. {yields} Self-assembly was possible with aqueous tin (II) chloride solution and not with aqueous tin (IV) chloride solution. -- Abstract: The present report details the effects of synthesis time, concentrations of hexamethylenetetramine (HMTA) and precursor tin (II) chloride solutions on the self-assembly of SnO{sub 2} nanocrystals. High-resolution electron microscopy images revealed that the structures were made of randomly attached SnO{sub 2} nanocrystals with sizes in between {approx}2 and 5 nm. X-ray photoelectron spectroscopy (XPS) showed that the Sn3d region was characterized by the spin-orbit splitting of the Sn3d{sub 5/2} ground state at {approx}487.6 eV and by the Sn3d{sub 3/2} excited state at {approx}496.1 eV, which was attributed to the Sn{sup +4} oxidation state of the SnO{sub 2} samples. We also found that the self-assembly could be achieved only with aqueous tin (II) chloride solution, and not with aqueous stannic (IV) chloride solution. A plausible growth mechanism is proposed in order to analyze the distinctive self-assembly of SnO{sub 2} nanocrystals in the presence of aqueous HMTA solution.

36 MATERIALS SCIENCE↗

Atom probe tomography characterisation of a laser diode structure grown by molecular beam epitaxy

Atom probe tomography (APT) has been used to achieve three-dimensional characterization of a III-nitride laser diode (LD) structure grown by molecular beam epitaxy (MBE). Four APT data sets have been obtained, with fields of view up to 400 nm in depth and 120 nm in diameter. These data sets contain material from the InGaN quantum well (QW) active region, as well as the surrounding p- and n-doped waveguide and cladding layers, enabling comprehensive study of the structure and composition of the LD structure. Two regions of the same sample, with different average indium contents (18% and 16%) in the QW region, were studied. The APT data are shown to provide easy access to the p-type dopant levels, and the composition of a thin AlGaN barrier layer. Next, the distribution of indium within the InGaN QW was analyzed, to assess any possible inhomogeneity of the distribution of indium (''indium clustering''). No evidence for a statistically significant deviation from a random distribution was found, indicating that these MBE-grown InGaN QWs do not require indium clusters for carrier localization. However, the APT data show steps in the QW interfaces, leading to well-width fluctuations, which may act to localize carriers. Additionally, the unexpected presence of a small amount (x = 0.005) of indium in a layer grown intentionally as GaN was revealed. Finally, the same statistical method applied to the QW was used to show that the indium distribution within a thick InGaN waveguide layer in the n-doped region did not show any deviation from randomness.

36 MATERIALS SCIENCE↗

Nanocrystallization in spark plasma sintered Fe{sub 48}Cr{sub 15}Mo{sub 14}Y{sub 2}C{sub 15}B{sub 6} bulk amorphous alloy

Spark plasma sintering (SPS) is evolving as an attractive process for the processing of multi-component Fe-based bulk amorphous alloys and their in-situ nanocomposites with controlled primary nanocrystallization. Extended Q-range small angle neutron scattering (EQ-SANS) analysis, complemented by x-ray diffraction and transmission electron microscopy, was performed to characterize nanocrystallization behavior of SPS sintered Fe-based bulk amorphous alloys. The SANS experiments show significant scattering for the samples sintered in the supercooled region indicating local structural/compositional changes associated with the profuse nucleation of nanoclusters (∼4 nm). For the samples spark plasma sintered near and above crystallization temperature (>653 °C), the SANS data show the formation of interference maximum indicating the formation and growth of (Fe,Cr){sub 23}C{sub 6} crystallites. The SANS data also indicate the evolution of bimodal crystallite distribution at higher sintering temperatures (above T{sub x1}). The growth of primary nanocrystallites results in impingement of concentration gradient fields (soft impingement effect), leading to non-random nucleation of crystallites near the primary crystallization.

77 NANOSCIENCE AND NANOTECHNOLOGY↗

The Milky Way Tomography with SDSS. 2. Stellar Metallicity

In addition to optical photometry of unprecedented quality, the Sloan Digital Sky Survey (SDSS) is producing a massive spectroscopic database which already contains over 280,000 stellar spectra. Using effective temperature and metallicity derived from SDSS spectra for {approx}60,000 F and G type main sequence stars (0.2 < g-r < 0.6), we develop polynomial models, reminiscent of traditional methods based on the UBV photometry, for estimating these parameters from the SDSS u-g and g-r colors. These estimators reproduce SDSS spectroscopic parameters with a root-mean-square scatter of 100 K for effective temperature, and 0.2 dex for metallicity (limited by photometric errors), which are similar to random and systematic uncertainties in spectroscopic determinations. We apply this method to a photometric catalog of coadded SDSS observations and study the photometric metallicity distribution of {approx}200,000 F and G type stars observed in 300 deg{sup 2} of high Galactic latitude sky. These deeper (g < 20.5) and photometrically precise ({approx}0.01 mag) coadded data enable an accurate measurement of the unbiased metallicity distribution for a complete volume-limited sample of stars at distances between 500 pc and 8 kpc. The metallicity distribution can be exquisitely modeled using two components with a spatially varying number ratio, that correspond to disk and halo. The best-fit number ratio of the two components is consistent with that implied by the decomposition of stellar counts profiles into exponential disk and power-law halo components by Juric et al. (2008). The two components also possess the kinematics expected for disk and halo stars. The metallicity of the halo component can be modeled as a spatially invariant Gaussian distribution with a mean of [Fe/H] = -1.46 and a standard deviation of {approx}0.3 dex. The disk metallicity distribution is non-Gaussian, with a remarkably small scatter (rms {approx}0.16 dex) and the median smoothly decreasing with distance from the plane from -0.6 at 500 pc to -0.8 beyond several kpc. Similarly, we find using proper motion measurements that a non-Gaussian rotational velocity distribution of disk stars shifts by {approx}50 km/s as the distance from the plane increases from 500 pc to several kpc. Despite this similarity, the metallicity and rotational velocity distributions of disk stars are not correlated (Kendall's {tau} = 0.017 {+-} 0.018). This absence of a correlation between metallicity and kinematics for disk stars is in a conflict with the traditional decomposition in terms of thin and thick disks, which predicts a strong correlation ({tau} = ?0.30 {+-} 0.04) at {approx}1 kpc from the mid-plane. Instead, the variation of the metallicity and rotational velocity distributions can be modeled using non-Gaussian functions that retain their shapes and only shift as the distance from the mid-plane increases. We also study the metallicity distribution using a shallower (g < 19.5) but much larger sample of close to three million stars in 8500 sq. deg. of sky included in SDSS Data Release 6. The large sky coverage enables the detection of coherent substructures in the kinematics-metallicity space, such as the Monoceros stream, which rotates faster than the LSR, and has a median metallicity of [Fe/H] = ?0.95, with an rms scatter of only {approx}0.15 dex. We extrapolate our results to the performance expected from the Large Synoptic Survey Telescope (LSST) and estimate that the LSST will obtain metallicity measurements accurate to 0.2 dex or better, with proper motion measurements accurate to {approx}0.2-0.5 mas/yr, for about 200 million F/G dwarf stars within a distance limit of {approx}100 kpc (g < 23.5).

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

PowerNet: Multi-agent Deep Reinforcement Learning for Scalable Powergrid Control

This paper develops an efficient multi-agent deep reinforcement learning algorithm for cooperative controls in powergrids. Specifically, we consider the decentralized inverter-based secondary voltage control problem in distributed generators (DGs), which is first formulated as a cooperative multi-agent reinforcement learning (MARL) problem. We then propose a novel on-policy MARL algorithm, PowerNet, in which each agent (DG) learns a control policy based on (sub-)global reward but local states and encoded communication messages from its neighbors. Motivated by the fact that a local control from one agent has limited impact on agents distant from it, we exploit a novel spatial discount factor to reduce the effect from remote agents, to expedite the training process and improve scalability. Furthermore, a differentiable, learning-based communication protocol is employed to foster the collaborations among neighboring agents. In addition, to mitigate the effects of system uncertainty and random noise introduced during on-policy learning, we utilize an action smoothing factor to stabilize the policy execution. To facilitate training and evaluation, we develop PGSim, an efficient, high-fidelity powergrid simulation platform. Here, experimental results in two microgrid setups show that the developed PowerNet outperforms the conventional model-based control method, as well as several state-of-the-art MARL algorithms. The decentralized learning scheme and high sample efficiency also make it viable to large-scale power grids.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Machine learning reveals strong grid-scale dependence in the satellite N d –LWP relationship

The relationship between cloud droplet number concentration ( N d ) and liquid water path (LWP) is highly uncertain yet crucial for determining the impact of aerosol-cloud interactions (ACI) on Earth's radiation budget. The N d -LWP relationship is examined using a machine learning (ML) random forest model applied to five years of satellite data at grid resolutions ranging from 10° to 0.05° in 12 distinct regions. In the subtropics, the shape of the N d -LWP relationship switches from an inverted-V at 1° grid-resolution to an “M” shape at 0.1° resolution with decreased $\frac{\textrm{dln⁡LWP}}{\textrm{dln⁡}N_d}$ sensitivity. Tropical and midlatitude regions generally show a more positive sensitivity. Cloud sampling and filtering also influence this slope, wherein the exclusion of thin clouds, as commonly performed to reduce retrieval uncertainty, leads to strongly negative sensitivity across all regions. Precipitation is primarily responsible for driving the strength of the sensitivity, with strong positive slopes in raining clouds and negative and/or neutral responses found in non-raining clouds. A new method to compute radiative forcing from the ML model shows a robust Twomey radiative forcing across all regions and grid resolutions. However, LWP and cloud fraction adjustments to the radiative forcing, which are ∼50 % or smaller than the Twomey effect, decrease to negligible values with higher spatial resolution data. As Earth system models move toward higher spatial resolutions in the future, evaluating the LWP and CF adjustment contributions to the radiative forcing budget at these finer resolutions will be essential for evaluation and model development.

Aerosol-Cloud Interactions↗

Search for 2{nu}{beta}{beta} Decay of {sup 130}Te to the First Excited State of {sup 130}Xe with an Ultra-Low-Background Germanium Crystal Array

The goal of searching for zero-neutrino double-beta (0{nu}{beta}{beta}) decay is to probe an absolute neutrino mass scale suggested by the mass-splitting parameters observed by neutrino oscillation experiments. Furthermore, observation of 0{nu}{beta}{beta} decay is an explicit instance of Lepton-number non-conservation. A sensitive measurement of two-neutrino double-beta (2{nu}{beta}{beta}) decay can provide critical input to Quasiparticle Random Phase Approximation (QRPA) calculations of the nuclear matrix elements in models similar to those used to extract the absolute neutrino mass from (0{nu}{beta}{beta}) decay experiments. Tellurium-130, an even-even nucleus, can undergo 2{nu}{beta}{beta} decay to the first 0+ excited state of {sup 130}Xe producing three possible {gamma}-ray cascades as it transitions to the ground state. The Cascades detector is a high purity germanium (HPGe) crystal array consisting of two ultra-low-background copper cryostats each housing a hexagonal array of seven crystals. The project is currently being developed at Pacific Northwest National Laboratory in Richland, WA (USA), and aims to obtain very high {gamma}-ray detection efficiency while utilizing highly effective and low-background shielding. GEANT4 simulations of the detector are performed for a {sup 130}Te sample in order to determine the optimum size and geometry of the source for maximum detection efficiency and predict its sensitivity for measuring 2{nu}{beta}{beta} decay to the first 0+ excited state of {sup 130}Xe. These simulations are validated with calibration sources and presented.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Uncertainty-Aware Machine Learning for Small-Angle X-ray Scattering Analysis in Autonomous Experimentation

Small-angle X-ray scattering (SAXS) is a powerful high-throughput characterization tool for probing nanoscale structure in native sample environments, providing real-time morphological information such as nanoparticle size and shape during synthesis. However, automated SAXS data analysis for extracting meaningful structural parameters is non-trivial and remains a bottleneck in closed-loop experimentation towards autonomous materials discovery, which demands fast, reliable, and uncertainty-aware data analysis. Here, we develop a machine-learning approach for automated SAXS analysis tailored to closed-loop nanoparticle synthesis. A Random Forest (RF) regression model is trained on 100,000 synthetic SAXS curves generated from polydisperse spherical nanoparticles with realistic background contributions. Using normalized one-dimensional SAXS intensity profiles as input, the RF model directly predicts nanoparticle radius, size polydispersity, and background parameters, while the ensemble standard deviation across trees provides built-in uncertainty quantification (UQ). On synthetic data, we show that combining fit-quality metrics (R 2 , MAE) with thresholds on prediction uncertainty reliably identifies accurate parameter estimates without access to ground truth. We then apply the trained model to 365 experimental SAXS profiles of citrate-reduced gold nanoparticles synthesized using an automated droplet-flow microreactor with in situ SAXS at a synchrotron beamline, classifying the results into high- and low-confidence subsets based on UQ metrics. Finally, we integrate RF-based SAXS analysis into a simulated closed-loop optimization campaign using Gaussian process Bayesian optimization to minimize nanoparticle polydispersity, benchmarking against conventional automated Levenberg–Marquardt fitting. The RF-guided campaign exhibits substantially faster convergence and lower relative opportunity cost (∼0.07 vs ∼0.3), demonstrating that uncertainty-aware machine-learning SAXS analysis significantly enhances the efficiency and robustness of autonomous nanomaterials synthesis workflows.

Bayesian optimization↗

Anomalous size effect on yield strength enabled by compositional heterogeneity in high-entropy alloy nanoparticles

High-entropy alloys (HEAs), although often presumed to be random solid solutions, have recently been shown to display nanometer-scale variations in the arrangements of their multiple chemical elements. Here, we study the effects of this compositional heterogeneity in HEAs on their mechanical properties using in situ compression testing in the transmission electron microscope (TEM), combined with molecular dynamics simulations. We report an anomalous size effect on the yield strength in HEAs, arising from such compositional heterogeneity. By progressively reducing the sample size, HEAs initially display the classical “smaller-is-stronger” phenomenon, similar to pure metals and conventional alloys. However, as the sample size is decreased below a critical characteristic length (~180 nm), influenced by the size-scale of compositional heterogeneity, a transition from homogeneous deformation to a heterogeneous distribution of planar slip is observed, coupled with an anomalous “smaller-is-weaker” size effect. Atomic-scale computational modeling shows these observations arise due to compositional fluctuations over a few nanometers. These results demonstrate the efficacy of influencing mechanical properties in HEAs through control of local compositional variations at the nanoscale.

36 MATERIALS SCIENCE↗

Metal-ion-controlled growth and nanoindentation response of 3D, bicontinuous Cu–Fe thin films

Bicontinuous, nanocomposite thin film morphologies depend largely on the deposition conditions applied during physical vapor deposition. With the introduction of high power impulse magnetron sputtering (HiPIMS), the range of potential morphologies achieved during deposition has been increased. In this work, we compare the deposition outcomes between traditional direct-current magnetron sputtering (DCMS) and HiPIMS for a thin film co deposit of Cu and Fe. Modular control of the columnarity, porosity and roughness was achieved by varying the Cu and Fe metal ion currents during deposition. The directionality of the nanostructured phase-separated morphology was also controlled as the ion current increased. At zero ion current for both Cu and Fe sputtered species during DCMS, the film exhibited lateral concentration modulations of Cu and Fe. The directionality of the Cu- and Fe-rich phases shifted to vertical concentration modulations at low ion currents of I Fe = 1A and I Cu = 0:1A and to lateral concentration modulations at relatively moderate ion currents of I Fe = 5A and I Cu = 2A: At high ion currents of I Fe = 18A and I Cu = 2A; a more randomized phase domain structure was observed on the nanoscale. This structural shift is rationalized using an interdiffusion model. Here, the role of different kinds of phase separated morphologies, achieved during DCMS deposition, on the mechanical properties has also been studied. Results indicated increase in hardness, indentation modulus and flow strength values with the increase in indentation strain rates. Bicontinuous Cu-Fe nanocomposites are found to be stronger than multilayer Cu-Fe samples.

36 MATERIALS SCIENCE↗

THE ARAUCARIA PROJECT: THE DISTANCE TO THE SCULPTOR GALAXY NGC 247 FROM NEAR-INFRARED PHOTOMETRY OF CEPHEID VARIABLES

We have obtained deep near-infrared images in J and K filters of four fields in the Sculptor Group spiral galaxy NGC 247 with the ESO VLT and Infrared Spectrometer and Array Camera. For a sample of 10 Cepheids in these fields, previously discovered by GarcIa-Varela et al. from optical wide-field images, we have determined mean J and K magnitudes and have constructed the period-luminosity (PL) relations in these bands. Using the near-infrared PL relations together with those in the optical V and I bands, we have determined a true distance modulus for NGC 247 of 27.64 mag, with a random uncertainty of {+-}2% and a systematic uncertainty of {approx}4% which is dominated by the effect of unresolved stars on the Cepheid photometry. The mean reddening affecting the NGC 247 Cepheids of E(B - V) = 0.18 {+-} 0.02 mag is mostly produced in the host galaxy itself and is significantly higher than what was found in the previous optical Cepheid studies in NGC 247 of our own group, and Madore et al., leading to a 7% decrease in the previous optical Cepheid distance. As in other studies of our project, the distance modulus of NGC 247 we report is tied to an assumed Large Magellanic Cloud distance modulus of 18.50. Comparison with other distance measurements to NGC 247 shows that the present IR-based Cepheid distance is the most accurate among these determinations. With a distance of 3.4 Mpc, NGC 247 is about 1.5 Mpc more distant than NGC 55 and NGC 300, two other Sculptor Group spirals analyzed before with the same technique by our group.

99 GENERAL AND MISCELLANEOUS↗

Galaxy Cluster Contribution to the Diffuse Extragalactic Ultraviolet Background

The diffuse ultraviolet background radiation has been mapped over most of the sky with 2′ resolution using data from the Galaxy Evolution Explorer survey. We utilize this map to study the correlation between the UV background and clusters of galaxies discovered via the Sunyaev–Zeldovich effect in the Planck survey. We use only high Galactic latitude (|b|>60{sup ∘}) galaxy clusters to avoid contamination by Galactic foregrounds, and we only analyze clusters with a measured redshift. This leaves us with a sample of 142 clusters over the redshift range of 0.02 ≤ z ≤ 0.72, which we further subdivide into four redshift bins. In analyzing our stacked samples binned by redshift, we find evidence for a central excess of UV background light compared to local backgrounds for clusters with z < 0.3. We then stacked these z < 0.3 clusters to find a statistically significant excess of 12 ± 2.3 photon cm{sup −2} s{sup −1} sr{sup −1} Å{sup −1} over the median of ∼380 photon cm{sup −2} s{sup −1} sr{sup −1} Å{sup −1} measured around random blank fields. We measure the stacked radial profile of these clusters, and find that the excess UV radiation decays to the level of the background at a radius of ∼1 Mpc, roughly consistent with the maximum radial extent of the clusters. Analysis of possible physical processes contributing to the excess UV brightness indicates that non-thermal emission from relativistic electrons in the intracluster medium and faint, unresolved UV emission from cluster member galaxies and intracluster light are likely the dominant contributors.

79 ASTRONOMY AND ASTROPHYSICS↗

Towards Fast and Accurate Predictions of Radio Frequency Power Deposition and Current Profile via Data-driven Modeling

Three machine learning techniques (multilayer perceptron, random forest, and Gaussian process) provide fast surrogate models for lower hybrid current drive (LHCD) simulations. A single GENRAY/CQL3D simulation without radial diffusion of fast electrons requires several minutes of wall-clock time to complete, which is acceptable for many purposes, but too slow for integrated modeling and real-time control applications. The machine learning models use a database of 16,000+ GENRAY/CQL3D simulations for training, validation, and testing. Latin hypercube sampling methods ensure that the database covers the range of 9 input parameters ($n_{e0}$, $T_{e0}$, $I_p$, $B_t$, $R_0$, $n_{||}$, $Z_{eff}$, $V_{loop}$, $P_{LHCD}$) with sufficient density in all regions of parameter space. The surrogate models reduce the inference time from minutes to ~ms with high accuracy across the input parameter space.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Reconstruction of Six-Dimensional Phase Space

A phase space is a mathematical representation of all possible physical states of a system. Particle beams at Fermilab exist within a six-dimensional (6D) phase space defined by three positional components, (x, y, z) and three momentum components, (px, py, pz). To reconstruct this space implies taking measurement data from detectors and mapping out particle behavior using computational methods. The beam detectors, however, are only able to detect spatial distribution among the events of the beam, therefore being limited to positional data. Also, due to the vast number of events in a particle beam, it is extremely difficult to analyze and differentiate every single one’s behavior. However, with Machine Learning (ML), which can distinguish between patterns and map out particle behavior more efficiently. We first used the particle beam software, G4beamline, to simulate a 10,000-event muon beam, adjusting parameters such as initial momentum magnitude (p¬0) and virtual detector position. Using ten virtual detectors, we analyzed p0 values such that minimum 9,990 events were analyzed by every detector. We then input the data from these beam simulations to a C++ program, that randomly selects 100 events, and creates a 2D histogram based on spatial distribution, detector position, and event intensity. This process is repeated 100 times to create 100 histograms per p0 value. These images were then input to a modified ResNet18 Convolutional Neural Network (CNN) for training, and to predict p0 from some unseen set of histograms. The model was accurate when trained on momentum increments of 5 MeV/c and provided with denser training samples around highly variable test values. These results displayed machine learning being able to accurately predict p0 from being trained on different particle behaviors.

Shirlee, Jermain [Fermilab]↗