Bulletin of the American Physical Society
63rd Annual Meeting of the APS Division of Plasma Physics
Volume 66, Number 13
Monday–Friday, November 8–12, 2021; Pittsburgh, PA
Session TO07: ICF: Simulations, Statistics, and SensitivitiesOn Demand
|
Hide Abstracts |
Chair: Varchas Gopalaswamy, Laboratory for Laser Energetics - Rochester Room: Rooms 315-316 |
Thursday, November 11, 2021 9:30AM - 9:42AM |
TO07.00001: Understanding Cryogenic Target Performance on OMEGA Using Statistics-Based Analysis with 2-D DRACO Simulations Duc M Cao, Varchas Gopalaswamy, Aarne Lees, Dhrumir P Patel, Riccardo Betti, Cliff A Thomas, Rahul C Shah, Wolfgang R Theobald, James P Knauer, P. B Radha, Christian Stoeckl, Sean P Regan, William Scullin, Timothy J Collins, Valeri N Goncharov In the past, great success in target performance optimization has been achieved by building a predictive capability using statistical modeling. Based on results of nearly 300 cryogenic experiments performed on the OMEGA laser, such modeling constructs scaling laws for neutron yields and areal density that are in addition to scaling predictions from LILAC 1‑D simulations [V. Gopalaswamy et al., Nature 565, 581 (2019)]. These additional scalings are functions of design parameters (target and laser beam sizes, fuel adiabat, shell thickness, convergence ratio, etc.) and hypothesized to represent multidimensional degradation effects, but they have yet to be fully explained. In this talk, we describe a new statistical model based on the same experimental data but combined with advanced 2-D simulations. This new model shows that these additional scaling terms can be understood with physics models included in DRACO 2-D [P. B. Radha et al., Phys. Plasmas 12, 032702 (2005)]. This material is based upon work supported by the Department of Energy National Nuclear Security Administration under Award Number DE-NA0003856. |
Thursday, November 11, 2021 9:42AM - 9:54AM |
TO07.00002: Multivariate Sensitivity Analysis of Radiation-Hydrodynamics Modeling of Uniaxially Driven ICF Systems Dave Chapman, Rafel Bordas, Nikita Chaturvedi, Nicholas Hawker, Martin Read, Dan Vassilev, Nathan Joiner Uncertainty quantification of radiation-hydrodynamics simulations is a fast-growing field within the ICF community. Understanding the role of interactions between model components is crucial for identifying key sensitivities and additionally serves to inform prioritization for further code development. In particular, the impact of uncertainties in the modeling of coupled phenomena such as conduction, plasma viscosity and radiation transport, each of which may vary in fidelity during different phases of target operation, are thought to be crucial for designing and diagnosing ICF systems. Recently, we undertook a preliminary assessment to modeling uncertainties of flux-limited conduction in the uniaxially driven system studied by Derentowicz et al. [1]. Our results indicated sensitivity to ionic heat flow between the fusion fuel and metallic anvil, which is quasi-nonlocal and occurs where material properties are often of limited accuracy. This submission presents a multivariate expansion of Ref. [1], covering a large options hyperspace sampled using a space-filling Latin hypercube design. The results not only confirm the importance of flux-limited conduction, but also point to uncertainties in hydrodynamic interface tracking as one of the most influential aspects of our modeling. This raises the intriguing possibility that this platform may be able to constrain predictions of advanced interfacial diffusion models. |
Thursday, November 11, 2021 9:54AM - 10:06AM |
TO07.00003: Unknown Unknowns in Inertial Confinement Fusion Peter W Hatfield Uncertainty quantification is a key part of physics; scientific researchers attempt to model both statistical and systematic uncertainties in their data as best as possible, often using a Bayesian framework. However it is well known that most statistical claims should be taken contextually; even if certain models are excluded at a very high degree of confidence, researchers are typically aware there may be systematics that were not accounted for, and thus typically will require confirmation from multiple independent sources, diagnostics, or experiments, before any novel results are truly accepted. In this work we compare two methods in the literature that seek to attempt to quantify these `unknown unknowns' - in particular attempting to produce realistic thick tails in the posterior of parameter estimation problems, that account for the possible existence of very large unknown effects. We apply these methods to the field of inertial confinement fusion to produce thick-tailed probability distributions for predictions of the outcome of future experiments. |
Thursday, November 11, 2021 10:06AM - 10:18AM |
TO07.00004: Estimating propagation of uncertainty in the instrumental function into reconstructions of neutron sources in ICF implosions at NIF Kevin Lamb, Noah W Birge, Christopher Danly, Matthew Freeman, Verena Geppert-Kleinrath, Carl H Wilde, Laurent Divol, David N Fittinghoff, Arthur E Pak, Alex B Zylstra, Petr L Volegov Neutron imaging is an important component of the nuclear diagnostics suite at the National Ignition Facility (NIF). The Nuclear Imaging System (NIS) provides a means to measure neutron sources in inertially confined fusion (ICF) implosions. Two- and three-dimensional neutron source reconstructions of an ICF implosion are key indicators of its overall performance. |
Thursday, November 11, 2021 10:18AM - 10:30AM |
TO07.00005: Data-driven models for improved preshot predictions of ICF experiments at the NIF Kelli D Humbird, Jay Salmonson, Luc Peterson, Bogdan Kustowski, Brian K Spears Standard computer simulations for indirect drive inertial confinement fusion (ICF), without platform-specific corrections, often show discrepancy with experiments. In this talk, we present a machine learning based method for training models that correct for this discrepancy, and are thus more predictive of National Ignition Facility (NIF) ICF experiments than simulations alone. |
Thursday, November 11, 2021 10:30AM - 10:42AM |
TO07.00006: Applying Machine Learning Techniques on ICF Performance using Experimental Data Michael Pokornik, Andrew Maris, Shahab Khan, Luc Peterson, Kelli D Humbird At the National Ignition Facility (NIF), the largest and most energetic laser in the world, cutting edge research is undertaken towards the goal of sustained nuclear fusion reactions. This research could lead to a remarkable energy source with an energy density far greater than nuclear fission. However, achieving a sustained burning plasma is difficult due to several possible implosion degradation mechanisms and limited laser energy. Recent breakthrough experiments are a testament that we have not fully optimized the experimental design in ICF experiments. Simulations, the traditional way of designing experiments, is limited in capturing all the physics involved in actual experiments, making optimization a challenging task. A machine learning tool can provide new insight into poorly understood physics and lead targeted exploration of the parameter space and guide future experimental design and simulation. |
Thursday, November 11, 2021 10:42AM - 10:54AM |
TO07.00007: Post-shot Simulations with Uncertainty: Fast, Approximate Inference Using NIF Data Jim A Gaffney, Kelli D Humbird, Michael K Kruse, Eugene Kur, Bogdan Kustowski, Ryan C Nora, Luc Peterson, Brian K Spears Post-shot simulations, where radiation hydrodynamics are tuned to match experimental data after it is collected, are foundational to the interpretation of inertial confinement fusion (ICF) and high energy density physics (HEDP) experiments. The post-shot process provides the best picture of drive parameters, stagnated plasma conditions and degradation mechanisms for ongoing experiments at the NIF and form the basis for our understanding of capsule and hohlraum designs. The standard approach, however, involves tuning a single simulation to match a small set of observables and so cannot provide uncertainty information, or allow the possibility of multiple competing explanations of the observed ICF performance. |
Thursday, November 11, 2021 10:54AM - 11:06AM |
TO07.00008: Determining the yield amplification of Bigfoot shots with Bayesian methods and Stochastic Collocation Michael K Kruse, Gemma J Anderson, Jim A Gaffney, Ryan C Nora, Kevin L Baker, Kelli D Humbird, Luc Peterson, Brian K Spears Several NIF campaigns have recently demonstrated the ability to consistently reach neutron yields of 1016 and above. With such yields we are approaching the burning plasma regime where the energy deposition of the 3.5 MeV alpha particle into the hotspot and cold fuel is significant such that it raises the temperature of the plasma and the DT reactivity. If enough alpha particles deposit their kinetic energy into the DT fuel a sustained burn-wave is formed in the ice layer eventually leading to fusion ignition. The yield amplification due to alpha-particle energy deposition serves as a metric to judge the progress along the path to fusion ignition. We study the yield amplification for a series of Bigfoot shots in which pairs of hydrodynamically similar shots were performed with DT or THD fuel fractions that produce or inhibit alpha particle production, respectively. Two recent developments have enabled us to provide an estimate of the mean and the variance of the yield-amplification. The Bayesian-Super-Postshot (BSPS) analysis infers the posterior-predictive set of 9 input parameters to the radiation-hydrodynamics code HYDRA that are consistent with several experimentally observed quantities for each Bigfoot shot. The Stochastic Collocation (SC) technique then exploits properties of orthogonal polynomials with respect to the BSPS input probability distribution to form a "small" set of simulation nodes in carefully placed locations that provide a reliable estimate of the mean and variance of an experimental observable. The yield amplification mean and variance determined by the BSPS and SC is found to be in good agreement with one-dimensional models. Lastly, this methodology is not limited to yield amplification and can be extended to other observables. |
Thursday, November 11, 2021 11:06AM - 11:18AM |
TO07.00009: Experimental Reproducibility in the Alpha-Heating Regime Luc Peterson, Jim A Gaffney, Michael K Kruse, Brian K Spears As inertial confinement fusion experiments begin to self-heat, the question of experimental reproducibility becomes more acute: alpha-particle bootstrapping can amplify small design deviations to produce large changes in observed nuclear yield. Quantifying the expected level of yield variation from known sources of uncertainty is important for assessing the degree to which design choices are intentionally self-heating. Such analysis could also help inform whether experimental results that differ from design calculations are surprising or to be expected. To that end, we use radiation hydrodynamic simulations to explore the sensitivity of various National Ignition Facility indirect drive experiments to known levels of target, laser, and physics uncertainties. First, we confirm the intuition that designs experience greater variability (and hence lower reproducibility) as they approach the self-heating regime. However, we also find that the expected level of alpha heating variation depends not only on the design but also on the path by which that design approaches ignition. In other words, modifications to the same baseline design with identical expectation values of alpha heating can have different variances. Since different design variants show different levels of reproducibility, it may be possible to intentionally seek out self-heating designs that are explicitly more reproducible. |
Thursday, November 11, 2021 11:18AM - 11:30AM |
TO07.00010: Approximate modelling of non-local electron transport Tony Bell Electron transport in laser-produced plasmas is non-local because the mean free path of energy-carrying electrons is comparable with hydrodynamic scale-lengths. Ideally, the Vlasov-Fokker-Planck (VFP) equation should be solved in tandem with hydrodynamic simulation. However, solution of the full VFP equation in two or three spatial dimensions with a full dynamic range of densities and temperatures is too slow. The inclusion of magnetic field presents an additional challenge. Approximate VFP solvers have proved effective but there is room for improvement. We propose a new approximate method that has the potential to be fast, robust and accurate. |
Thursday, November 11, 2021 11:30AM - 11:42AM |
TO07.00011: Preheat and ablation broadening due to nonthermal electron energy transport for Z>1. Wallace M Manheimer Over the past few years, the NRL theory group has presented and published a variety of theoretical works on the effect of nonlocal electron energy transport on fuel preheat and ablation layer broadening in direct drive laser fusion targets. The goal is not an exact model arrived at by say a Monte Carlo simulation, but rather a model accurate enough and simple enough that it can be utilized at each time step of a fluid simulation. These works involved both Krook models and a combination of Krook and Fokker Planck (FP) models. The latter works used the Krook model to determine the nonlocal energy flux at its maximum point (within a mfp or so their formation) and an approximate FP model to determine their deposition. A characteristic of the FP model is that it predicts orders of magnitude less fuel preheat than Krook models; after many mfp’s the Krook and FP models behave very differently. Earlier publications treated planar and spherical cases for Z=1. This work extends the theory to include Z > 1 ablators. Unlike the diffusion model, which simply has a diffusion coefficient going as 1/Z, a more accurate FP model does not have such a simple coupling of Z to single spatial equation, but has the Z only in the velocity terms, and not in the spatial terms. However, the solution of this more complex equation does show a decreasing fuel preheat with an increase in Z. |
Thursday, November 11, 2021 11:42AM - 11:54AM |
TO07.00012: Species Separation in Polystyrene Shock Release Evidenced by Molecular Dynamics Simulations and Laser Drive Experiments Shuai Zhang, Suxing Hu, Dayne E Fratanduono, J. Ryan Rygg, Michelle C Marshall, Amy E Lazicki, Alex Shvydky, Daniel J Haberberger, Valeri N Goncharov, Thomas R Boehly, Gilbert Collins Materials shock release into vacuum generally happens in the targets of high-energy-density (HED) and inertial confinement fusion (ICF) experiments but has been challenging to study experimentally, theoretically, or computationally. Regular simulations based on single-fluid hydrodynamics may be questionable because of the ignorance of changes in materials chemistry, such as species separation as revealed in recent molecular-dynamics (MD) calculations of polystyrene (CH) upon release of a strong shock.[1] We performed extensive studies of the effect of shock strength and hydrogen isotope on CH shock release by combining MD simulations and laser drive experiments. Our simulations showed that species separation occured for strong shocks and more preferentially for CH with lighter hydrogen isotopes, consistent with experimental observations. Our simulations also considered preheated CH and showed enhanced velocities of the released low-density plasmas, which agreed with findings by hydrodynamics simulations. These results highlight the significant role of species separation in the release of compound materials and are significant for the design and analysis of HED and ICF experiments. [1] S. Zhang and S. X. Hu, Phys. Rev. Lett 125, 105001 (2020) |
Thursday, November 11, 2021 11:54AM - 12:06PM |
TO07.00013: Scaling of hot electron generation mechanisms from two-plasmon decay in ICF plasmas Edoardo Rovere, Arnaud Colaitis, Alexis Casner, Russell K Follett, John P Palastro In Inertial Confinement Fusion (ICF) scenarios, particularly in Shock Ignition (SI), Hot Electron (HE) generation is a key aspect which can have different consequences, both positive, such as shock pressure increase, and detrimental, such as loss of ablation pressure and fuel preheat. |
Thursday, November 11, 2021 12:06PM - 12:18PM |
TO07.00014: Analysis of Hot-Electron Preheat of High-Performing OMEGA Cryogenic Implosions Dhrumir P Patel, Riccardo S Betti, Alison R Christopherson, James P Knauer, Christian Stoeckl, Varchas Gopalaswamy, Michael J Rosenberg Fast electrons generated in the coronal plasma by laser–plasma instabilities can preheat cryogenic DT fuel, thereby raising its adiabat and lowering final compression. One-dimensional LILAC simulations of high-performing OMEGA cryogenic implosions show the conditions at quarter critical to be above the threshold for two-plasmon–decay (TPD) instability. Significant amount of w/2 signal, characteristic of TPD decay, is also measured. We use methods developed from dedicated preheat measurement experiments to infer fast-electron preheat and quantify resulting performance degradation for some high-performing OMEGA cryogenic implosions. We also show that replacing CD ablator with silicate plastic CHSi can alleviate hot-electron preheat. This material is based upon work supported by the Department of Energy National Nuclear Security Administration under Award Number DE-NA0003856. A. R. Christopherson et al., “Direct Measurements of DT Fuel Preheat from Hot Electrons in Direct-Drive Inertial Confinement Fusion,” to be published in Physical Review Letters. |
Thursday, November 11, 2021 12:18PM - 12:30PM Not Participating |
TO07.00015: Towards Proton Tomography of Laser-Plasma Interactions Ben T Spiers, Ramy Aboushelbaya, Qingsong Feng, Marko W Mayr, Iustin Ouatu, Robert W Paddock, Robin Timmis, Robin Wang, Peter A Norreys Proton radiography, similarly to many diagnostics, probes a plasma in a line-integrated fashion. |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700