Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Topic Modeling”

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 1,297 records · Page 72

Determination of nanovibration amplitudes using frequency-modulated semiconductor laser autodyne

The method for measuring nanovibration amplitudes using the autodyne signal of a semiconductor laser at several laser radiation wavelengths is described. The theoretical description of the frequency-modulated autodyne signal under harmonic vibrations of the reflector is presented and the relations for its spectral components are derived using the expansions into the Fourier and Bessel series. The results of numerical modelling based on the proposed method for measuring the reflector nanovibration amplitudes are presented that make use of the low-frequency spectrum of the autodyne signal from the frequency-modulated laser autodyne and the solution of the appropriate inverse problem. The experimental setup is described; the results of the measurements are presented for the nanovibration amplitudes and the autodyne signal spectra under the reflector nanovibrations. (laser applications and other topics in quantum electronics)

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Fast ion relaxation in ITER mediated by Alfvén instabilities

Abstract We address the critical issue for future burning plasmas of whether high-energy fusion products or auxiliary heating-beam ions will be confined for a sufficiently long time to compensate for thermal plasma energy losses. This issue can be mitigated by one of the most deleterious collective phenomena—the instability of low, sub-cyclotron frequency Alfvén eigenmodes (AEs), such as toroidicity-induced AEs and reversed-shear AEs in the ITER steady-state scenario. Using a revised quasi-linear (QL) theory applied to energetic particle (EP) relaxation in the presence of AEs, we find that the AE instabilities can affect both neutral beam ions and alpha particles, although the resulting fast ion transport is expected to be modest if classical particle slowing down is assumed. On the other hand, the QL theory predicts that the AE amplitudes will be enhanced by the background microturbulence, although this topic remains outside our scope due to the significant numerical effort required to evaluate these effects. We report our results for EP relaxation dynamics obtained utilizing several tools: (i) a comprehensive linear stability study of the sub-cyclotron Alfvénic spectrum as computed by ideal magnetohydrodynamic NOVA simulations for the AE eigenproblem, (ii) drift kinetic NOVA-C calculations for wave–particle interaction and AE growth/damping rates, and (iii) predictive QL modeling coupled with the global transport code TRANSP to assess the EP relaxation on the equilibrium timescale.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Innovative Strategies for Long-Term Monitoring of Complex Groundwater Plumes at DOE’s Legacy Sites (Workshop Report)

Most remaining Department of Energy (DOE) sites will require extended periods of institutional control, especially at complex groundwater sites where attenuation-based strategies have been implemented to facilitate closure. The current practice of monitoring—obtaining and analyzing contaminant concentration in groundwater samples at numerous wells—will account for a large portion of the projected life-cycle at these DOE sites unless a new approach is adopted. State-of-the-art technologies are being developed, including in situ sensors, geophysics, radiation mapping, numerical modeling and AI/ML. These technologies can optimize monitoring strategies in space and time, provide spatially extensive information at vulnerable regions and/or provide more continuous monitoring at lower cost. As part of DOE’s Office of Environmental Management (DOE-EM’s) efforts to advance long-term monitoring systems, an in-person/virtual hybrid workshop was hosted by Savannah River National Laboratory (SRNL) on January 24 and 25, 2023, in Augusta, Georgia. Because DOE-EM’s complex sites will eventually be transferred to DOE’s Office of Legacy Management (DOE-LM), representatives of DOE-LM were important participants in the workshop. The purpose of the workshop was to identify challenges and opportunities for deploying advanced technologies for long-term monitoring at DOE sites. The key questions during the workshop were: 1) the regulatory acceptance of replacing a process that traditionally has used laboratory sampling and analysis of groundwater samples, and 2) the application of this strategy to the southwestern arid sites that include many of the remaining DOE-EM and DOE-LM complex groundwater plumes. Characteristics common to most arid sites present both limitations and opportunities for advanced technologies. DOE-EM has funded a National Laboratory team from SRNL, Lawrence Berkeley National Laboratory (LBNL), and Pacific Northwest National Laboratory (PNNL) to establish the overarching framework of long-term monitoring by systematically combining advanced hardware and software technologies. This project is titled “Advanced Long-Term Environmental Monitoring Systems (ALTEMIS)” and is sponsored by the DOE-EM Technology Development Program. The multi-laboratory team is currently developing and testing innovative monitoring strategies, including the use of in situ groundwater sensors, geophysics, drone/satellite-based remote sensing, reactive transport modeling, and artificial intelligence/machine learning (AI/ML). The project’s demonstration testbed is at the Savannah River Site (SRS) F-Area Seepage Basins, where a well-characterized complex groundwater plume composed of uranium and other radionuclides is in the latter stages of remediation. The workshop included more than 70 participants, presentations, a field visit to F-Area, breakout working groups, and large group discussion. Participants developed recommendations on five topics: in situ sensors, spatially integrative tools, challenges to regulatory acceptance, AI/ML strategies, and transitioning sites to DOE-LM.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Collaborative Proposal: Improving Decadal Prediction of Arctic Climate Variability and Change Using a Regional Arctic System Model (RASM)

RASM is a multi-disciplinary project, which brings together researchers from six state universities, one military postgraduate school, and one DoE laboratory to address the core modeling objectives of the arctic research community articulated in the Arctic System Modeling report by Roberts et al. (2010b). This report advocates the construction of a regional downscaling tool to generate probabilistic decadal projections of Greenland ice sheet retreat, evolution of arctic sea ice cover, changes in land surface vegetation, and regional processes leading to arctic amplification. Unified coupled models such as RASM are ideal for this purpose because they simulate fine-scale physics, essential for the realistic representation of intra-annual variability, in addition to processes fundamental to long term climatic shifts (Hurrell et al. 2009). By using RASM with boundary conditions from a global model, we can generate many-member ensembles essential for understanding uncertainty in regional climate projections (Hawkins and Sutton 2009). This probabilistic approach is computationally prohibitive for high-resolution global models in the foreseeable future, and also for regional models interactively nested within global simulations. Yet it is fundamental for quantifying uncertainty in decadal forecasts to make them useful for decision makers (Doherty et al. 2009). For this reason, we have targeted development of ensemble generation techniques as a core project task (Task 4.5). Environmental impact assessment specialists need high-fidelity regional ensemble projections to improve the accuracy of their work (Challinor et al. 2009; Moss et al. 2010). This is especially true of the Arctic, where economic, social and national interests are rapidly reshaping the high north in step with regional climate change. During the next decade, considerable oil and gas discoveries are expected across many parts of the marine and terrestrial Arctic (Gautier et al. 2009), the economics of the Northern Sea Route will steadily improve (Arctic Council 2009), and sovereign claims over the Arctic Ocean will increasingly be subject to international negotiations (Proelss 2009). Issues such as these have led to an expanding demand for Arctic climate projections to aid national and commercial decisions. However, detailed information from existing models is lacking. RASM will enhance the existing Arctic system modeling capabilities and align them with the scientific and societal needs outlined above. The science involved in the development of RASM will be integrated with teaching and training at a few different levels. RASM PIs will supervise at least six postdoctoral and doctoral students with additional involvement of undergraduate research assistants. RASM postdoctoral fellows and graduate trainees will benefit from being a part of a large, collaborative, and multi-disciplinary research projects employing state-of-the-art modeling and computational tools. They will also have the opportunity to contribute to the UTEP-led education and outreach effort developed specifically for RASM. The UTEP team will use RASM as a platform for education and outreach. First, we will develop products for public dissemination, such as curriculum units, lesson plans and other materials (simulations, movies, and images) for use by students and teachers in high school and university classrooms. Second, we will facilitate bringing RASM PIs, postdocs, and graduate students into the classroom, through electronic mentorship and by contributing content via online lectures and presentations. Third, by the use of problem-based learning (PBL) approaches, we will provide real-world scenarios and problems enabling students to do research and develop position papers or presentations on topics related to RASM. (Problem-based learning is a student-centered, inquiry-based approach in which students work in teams to solve challenging, open-ended problems.) Fourth, we will develop teacher-training materials that will be developed into workshops (face-to-face or online) to help teachers understand how to use the materials we supply. Overall, we will provide a broad, and long-term RASM program for education and outreach. RASM education and outreach activities will target underrepresented populations - in particular, Hispanic students and teachers. For example, the UTEP student body is more than 70 percent Hispanic and 55 percent female, and the majority of students are the first in their families to attend college. Students from underrepresented groups will be given the chance to expand their understanding of the different scientific knowledge and processes associated with RASM. We will develop an eye-catching RASM website to include information about the project, model results and educational materials, and lists of RASM publications and presentations.

54 ENVIRONMENTAL SCIENCES↗

Path sampling of recurrent neural networks by incorporating known physics

Recurrent neural networks have seen widespread use in modeling dynamical systems in varied domains such as weather prediction, text prediction and several others. Often one wishes to supplement the experimentally observed dynamics with prior knowledge or intuition about the system. While the recurrent nature of these networks allows them to model arbitrarily long memories in the time series used in training, it makes it harder to impose prior knowledge or intuition through generic constraints. In this work, we present a path sampling approach based on principle of Maximum Caliber that allows us to include generic thermodynamic or kinetic constraints into recurrent neural networks. We show the method here for a widely used type of recurrent neural network known as long short-term memory network in the context of supplementing time series collected from different application domains. These include classical Molecular Dynamics of a protein and Monte Carlo simulations of an open quantum system continuously losing photons to the environment and displaying Rabi oscillations. Our method can be easily generalized to other generative artificial intelligence models and to generic time series in different areas of physical and social sciences, where one wishes to supplement limited data with intuition or theory based corrections.

59 BASIC BIOLOGICAL SCIENCES↗

Stochastic journeys of cell progenies through compartments and the role of self-renewal, symmetric and asymmetric division

Abstract Division and differentiation events by which cell populations with specific functions are generated often take place as part of a developmental programme, which can be represented by a sequence of compartments. A compartment is the set of cells with common characteristics; sharing, for instance, a spatial location or a phenotype. Differentiation events are transitions from one compartment to the next. Cells may also die or divide. We consider three different types of division events: (i) where both daughter cells inherit the mother’s phenotype (self-renewal), (ii) where only one of the daughters changes phenotype (asymmetric division), and (iii) where both daughters change phenotype (symmetric division). The self-renewal probability in each compartment determines whether the progeny of a single cell, moving through the sequence of compartments, is finite or grows without bound. We analyse the progeny stochastic dynamics with probability generating functions. In the case of self-renewal, by following one of the daughters after any division event, we may construct lifelines containing only one cell at any time. We analyse the number of divisions along such lines, and the compartment where lines terminate with a death event. Analysis and numerical simulations are applied to a five-compartment model of the gradual differentiation of hematopoietic stem cells and to a model of thymocyte development: from pre-double positive to single positive (SP) cells with a bifurcation to either SP4 or SP8 in the last compartment of the sequence.

Science & Technology - Other Topics↗

Effects of confinement and electron transport on magnetic switching in single Co nanoparticles

This work reports the first study of current-driven magnetization noise in a single, nanometerscale, ferromagnetic (Co) particle, attached to normal metal leads by high-resistance tunneling junctions. As the tunnel current increases at low temperature, the magnetic switching field decreases, its probability distribution widens, while the temperature of the environment remains nearly constant. These observations demonstrate nonequilibrium magnetization noise. A classical model of the noise is provided, where the spin-orbit interaction plays a central role in driving magnetic tunneling transitions.

MAGNETIC DEVICES↗

Molecular mechanism of activation of human musk receptors OR5AN1 and OR1A1 by ( R )-muscone and diverse other musk-smelling compounds

Understanding olfaction at the molecular level is challenging due to the lack of crystallographic models of odorant receptors (ORs). To better understand the molecular mechanism of OR activation, we focused on chiral ( R )-muscone and other musk-smelling odorants due to their great importance and widespread use in perfumery and traditional medicine, as well as environmental concerns associated with bioaccumulation of musks with estrogenic/antiestrogenic properties. We experimentally and computationally examined the activation of human receptors OR5AN1 and OR1A1, recently identified as specifically responding to musk compounds. OR5AN1 responds at nanomolar concentrations to musk ketone and robustly to macrocyclic sulfoxides and fluorine-substituted macrocyclic ketones; OR1A1 responds only to nitromusks. Structural models of OR5AN1 and OR1A1 based on quantum mechanics/molecular mechanics (QM/MM) hybrid methods were validated through direct comparisons with activation profiles from site-directed mutagenesis experiments and analysis of binding energies for 35 musk-related odorants. The experimentally found chiral selectivity of OR5AN1 to ( R )- over ( S )-muscone was also computationally confirmed for muscone and fluorinated ( R )-muscone analogs. Structural models show that OR5AN1, highly responsive to nitromusks over macrocyclic musks, stabilizes odorants by hydrogen bonding to Tyr260 of transmembrane α-helix 6 and hydrophobic interactions with surrounding aromatic residues Phe105, Phe194, and Phe207. As a result, the binding of OR1A1 to nitromusks is stabilized by hydrogen bonding to Tyr258 along with hydrophobic interactions with surrounding aromatic residues Tyr251 and Phe206. Hydrophobic/nonpolar and hydrogen bonding interactions contribute, respectively, 77% and 13% to the odorant binding affinities, as shown by an atom-based quantitative structure–activity relationship model.

59 BASIC BIOLOGICAL SCIENCES↗

Surface-screening mechanisms in ferroelectric thin films and their effect on polarization dynamics and domain structures

For over 70 years, ferroelectric materials have been one of the central research topics for condensed matter physics and material science, an interest driven both by fundamental science and applications. However, ferroelectric surfaces, the key component of ferroelectric films and nanostructures, still present a significant theoretical and even conceptual challenge. Indeed, stability of ferroelectric phase per se necessitates screening of polarization charge. At surfaces, this can lead to coupling between ferroelectric and semiconducting properties of material, or with surface (electro) chemistry, going well beyond classical models applicable for ferroelectric interfaces. In this paper, we summarize recent studies of surface-screening phenomena in ferroelectrics. We provide a brief overview of the historical understanding of the physics of ferroelectric surfaces, and existing theoretical models that both introduce screening mechanisms and explore the relationship between screening and relevant aspects of ferroelectric functionalities starting from phase stability itself. Given that the majority of ferroelectrics exist in multiple-domain states, we focus on local studies of screening phenomena using scanning probe microscopy techniques. We discuss recent studies of static and dynamic phenomena on ferroelectric surfaces, as well as phenomena observed under lateral transport, light, chemical, and pressure stimuli. Finally, we also note that the need for ionic screening renders polarization switching a coupled physical–electrochemical process and discuss the non-trivial phenomena such as chaotic behavior during domain switching that stem from this.

36 MATERIALS SCIENCE↗

Slow thermalization and subdiffusion in $U$(1) conserving Floquet random circuits

Random quantum circuits are paradigmatic models of minimally structured and analytically tractable chaotic dynamics. We study a family of Floquet unitary circuits with Haar random U(1) charge conserving dynamics; the minimal such model has nearest-neighbor gates acting on spin-1/2 qubits, and a single layer of even/odd gates repeated periodically in time. We find that this minimal model is not robustly thermalizing at numerically accessible system sizes, and displays slow subdiffusive dynamics for long times. We map out the thermalization dynamics in a broader parameter space of charge conserving circuits, and understand the origin of the slow dynamics in terms of proximate localized and integrable regimes in parameter space. In contrast, we find that small extensions to the minimal model are sufficient to achieve robust thermalization; these include (i) increasing the interaction range to three-site gates, (ii) increasing the local Hilbert space dimension by appending an additional unconstrained qubit to the conserved charge on each site, or (iii) using a larger Floquet period comprised of two independent layers of gates. Furthermore, our results should inform future studies of charge conserving circuits, which are relevant for a wide range of topical theoretical questions.

1-dimensional spin chains↗

Measurements of He-Ar, He-N 2 , and He-air thermal creep slip and accommodation coefficients at high temperatures

Understanding gas flows in capillaries has many applications in modeling the transport of gases in nano-structured, porous, or fractured media. A network of capillaries can often approximate these media, and also information on gas-surface interactions obtained from capillary experiments can be used for modeling flows in these media. Experimental data on flows of different mixtures of He-Ar, He-N 2 , and He-air in the slip regime and in stainless steel capillaries at high temperatures were obtained by using a two-bulb apparatus. An accurate expression for the thermal creep slip coefficients with the Lennard-Jones potential parameters and diffuse-specular reflection gas-surface interaction conditions were then used to obtain the accommodation coefficients for the different gases. The experimental data are best described with values of accommodation coefficients in the range of 0.1–0.3 for He and 0.5–1.0 for Ar, N 2 , and air. Further, the use of values in this range is suggested for modeling gaseous flows in capillaries and nano-structured, porous, or fractured media if other direct measured values for a particular medium are unavailable.

42 ENGINEERING↗

Hyperselective carbon membranes for precise high-temperature H 2 and CO 2 separation

More than 90% of the world’s hydrogen (H 2 ) is produced from fossil fuel sources, which requires energy-intensive separation and purification to produce high-purity H 2 fuel and to capture the carbon dioxide (CO 2 ) by-product. While membranes can decarbonize H 2 /CO 2 separation, their moderate H 2 /CO 2 selectivity requires secondary H 2 purification by pressure swing adsorption. Here, we report hyperselective carbon molecular sieve hollow fiber membranes showing H 2 /CO 2 selectivity exceeding 7000 under mixture permeation at 150°C, which is almost 30 times higher than the most selective nonmetallic membrane reported in the literature. The membrane is able to maintain an ultrahigh H 2 /CO 2 selectivity over 1400 under mixture permeation at 400°C. Pore structure characterization suggests that highly refined ultramicropores are responsible for effectively discriminating the closely sized H 2 and CO 2 molecules in the hyperselective carbon molecular sieve membrane. Modeling shows that the unprecedented H 2 /CO 2 selectivity will potentially allow one-step enrichment of fuel-grade H 2 from shifted syngas for decarbonized H 2 production.

Science & Technology - Other Topics↗

Support for the 2024 American Conference on Theoretical Chemistry (ACTC) (Final Report)

Funds are requested in support of the 2024 American Conference on Theoretical Chemistry. Funds are being requested from the Department of Energy in support of conference registration fees for graduate student and post-doctoral researcher registration fees. This conference will be held in North Carolina, and between 200 and 250 participants are expected. This is the major North American meeting of theoretical chemists, showcasing diverse developments in all aspects of modern theory, including methods development and applications. Applications span all aspects of chemistry, biochemistry, and materials science, including important applications to energy science. The conference will include four days of seminars and poster sessions. As well as serving as a meeting ground for sharing scientific developments and discoveries, the meeting also serves as a locus for mentoring and career development involving younger scientists. The meeting will be chaired by Prof. David Beratan of Duke University, and he will be assisted by other faculty from the region: Profs. Weitao Yang (Duke), Patrick Charbonneau (Duke), Yosuke Kanai (University of North Carolina - Chapel Hill), Zhiyue Lu (University of North Carolina - Chapel Hill), and Elena Jakubikova (North Carolina State University). This group will form a conference commitee that will define thematic topics for the conference, will invite speakers, and will assist with running the conference.

14 SOLAR ENERGY↗

Plasma Interaction with Electron-Emitting Surfaces

Electron emission from surfaces occurs in many plasma systems. Several types including secondary, thermionic and photon-induced emissions are intense under certain conditions. Understanding the effects of emission on the sheaths that govern plasma-surface interaction is important. This dissertation predicts some emitting sheath phenomena that were not reported in past studies. For example, most previous theoretical models assumed that an emitting sheath potential is always negative and that ions always accelerate into the wall. We show when the emission is intense that the sheath potential can become positive, fundamentally changing how the plasma and wall interact. In this inverse sheath state, ions are repelled, suggesting for instance that (a) no presheath exists in the plasma interior, (b) emitting walls could be used in applications to stop sputtering. Another topic considered is the transit of emitted electrons across the plasma to other surfaces, which is possible in low collisionality plasma systems. When transit occurs, the flux balance is a complex global problem where the sheaths at opposite surfaces are coupled through their exchange of emitted electrons. We also show that secondary emission can trigger a variety of sheath instability phenomena that change the state of the plasma-wall system or cause oscillations preventing steady state. Lastly, we analyze a mechanism where emitted electrons return to the same surface and knock out secondaries, which return and knock out more secondaries, etc., feedback amplifying the emission intensity. The four phenomena will be analyzed theoretically and verified with particle-in-cell simulations: (a) inverse sheath, (b) sheath coupling via transiting electrons, (c) sheath instabilities, (d) returning electron amplification. Consequences of these processes on the sheath potentials, wall heating, loss rate of charge, and cross field transport (near-wall conductivity) are discussed throughout. Possible implications are suggested for fusion machines, plasma propulsion engines, probes, dusty plasmas, RF discharges, and surfaces in space.

Fusion↗

ARM Radar Science Recommendations from 2017 6th ARM Radar Workshop

The U.S. Department of Energy (DOE) Atmospheric Radiation Measurement (ARM) Climate Research Facility’s radar organization was formed to provide an integrated approach to meet the scientific objectives of ARM. The radar organization consists of two groups: Radar Science and Radar Engineering and Operations. The two groups work closely for planning, coordination, and communications to produce high-quality observations and products from the ARM radar network, which consists of complex radars deployed in very challenging environments. In addition, scanning cloud radars are relatively new in the community, with very limited products that are mature enough to be applied for modeling studies. The Radar Science Group and Radar Engineering and Operations Group meet regularly along with researchers and users from the community in a workshop. The overarching objective of the workshop is to maximize the usage of ARM radar data for improving our understanding of processes and, ultimately improving representation of clouds and precipitation in models. The 6th DOE ARM Radar Engineering and Radar Science Workshop was held October 24-26, 2017, at Stony Brook University. Approximately 40 participants were present; a detailed list of participants is listed in Section 5. This report documents recommendations made by members of the science community who attended the workshop. Their recommendations appear italicized in the report. ARM management will consider these recommendations in their future planning and prioritization activities. The topics addressed during the workshop are contained in the agenda (Section 4). Significant action items, recommendations, and comments that came out of the meeting are contained in Section 3. The content in Section 3 presents a path through important discussions that were pursued during the Workshop.

Southern Great Plains, North Slope of Alaska, East↗

TRACG benchmark study: 2015 Fermi thermal-hydraulic instability event

On March 19, 2015, the Boiling Water Reactor (BWR) Fermi 2 (referred to hereafter as Fermi), located in Michigan, U.S., experienced a scram initiated by the Oscillation Power Range Monitor (OPRM) system. The plant had been operating in Single Loop Operation (SLO) for approximately 10 minutes before the scram occurred. During this time, the plant operating conditions reached a region within the power/flow (P/F) map susceptible to coupled neutronic/thermal-hydraulic (T-H) instabilities. Limited code benchmarking is available in literature for T-H Instability (THI) transients occurring in plants operating at SLO conditions due to scarce plant data for this type of operational condition. This study aims to model and benchmark the THI event at SLO conditions using TRACG, the GE Hitachi (GEH) proprietary best-estimate transient system code. The TRACG model was benchmarked by comparing the results with measured plant data, such as the total reactor power and core flow, from the event. Additionally, to evaluate the inherent stability of the plant at different P/F points, Decay Ratio (DR) calculations were performed. The capability of TRACG for calculating DRs is well documented, and it has also been approved by the U.S. Nuclear Regulatory Commission (NRC) as part of ESBWR TRACG stability licensing topical reports. The results of TRACG DRs were compared to the GEH best-estimate frequency domain code ODYSY, which is the current Fermi licensing basis code for reload analysis DR calculations.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Electromagnetic scattering problems -Numerical issues and new experimental approaches of validation

Electromagnetic scattering problems, thus the question how radiated energy spreads when impinging on an object, are an essential part of wave propagation. Though the Maxwell’s differential equations as starting point, are actually quite simple,the integral formulation of an object’s boundary conditions, respectively the solution for unknown induced currents can only be solved numerically in most cases.As a timely topic of practical importance the scattering of rotating wind turbines is discussed, the numerical description of which is still based on rigorous approximations with yet unspecified accuracy. In this context the issue of validating numerical solutions is addressed, both with reference simulations but in particular with the experimental approach of scaled measurements. For the latter the idea of an incremental validation is proposed allowing a step by step validation of required new mathematical models in scattering theory.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Modeling laser-driven ion acceleration with deep learning

Developments in machine learning promise to ameliorate some of the challenges of modeling complex physical systems through neural-network-based surrogate models. High-intensity, short-pulse lasers can be used to accelerate ions to mega-electronvolt energies, but to model such interactions requires computationally expensive techniques such as particle-in-cell simulations. Multilayer neural networks allow one to take a relatively sparse ensemble of simulations and generate a surrogate model that can be used to rapidly search the parameter space of interest. In this study, we created an ensemble of over 1,000 simulations modeling laser-driven ion acceleration and developed a surrogate to study the resulting parameter space. A neural-network-based approach allows for rapid feature discovery not possible for traditional parameter scans given the computational cost. A notable observation made during this study was the dependence of ion energy on the pre-plasma gradient length scale. While this methodology harbors great promise for ion acceleration, it has ready application to all topics in which large-scale parameter scans are restricted by significant computational cost or relatively large, but sparse, domains.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗