Bulletin of the American Physical Society
62nd Annual Meeting of the APS Division of Plasma Physics
Volume 65, Number 11
Monday–Friday, November 9–13, 2020; Remote; Time Zone: Central Standard Time, USA
Session TI01: Invited: SimulationsLive
|
Hide Abstracts |
Chair: Nat Fisch, PPPL |
Thursday, November 12, 2020 9:30AM - 10:00AM Live |
TI01.00001: Plasma Physics in Strong-Field Regimes Invited Speaker: Yuan Shi (Rosenbluth winner) In strong electromagnetic fields, new plasma phenomena and applications emerge. In the classical regime, starting from megagauss magnetic fields, scattering of lasers becomes manifestly anisotropic. For the first time, a convenient formula for the three-wave coupling coefficient in arbitrary geometry is obtained and evaluated. By solving the fluid model to the second order, an alternative perspective of parametric instability is provided and magnetic-field effect on collective scattering is quantified. As an application, magnetic resonances are used to mediate laser pulse compression. Using magnetized plasmas, it is not only possible to achieve higher output intensity for optical lasers with more engineering flexibility, but also possible to compress UV and soft X-ray pulses that cannot be compressed using existing techniques. Taking advantage of the emerging feasibility of strong magnetic fields, a pathway to next-generation powerful lasers is identified, whose viability is supported by particle-in-cell simulations. In even stronger magnetic fields or intense laser fields, relativistic-quantum effects become important. At that point, plasma models based on quantum electrodynamics (QED) are necessary. Allowing for nontrivial background fields, a new formalism for QED plasmas is developed by computing the effective action using path integrals. The new formalism enables simple wave dispersion relations in strongly magnetized plasmas to be obtained for the first time, based on which the modified Faraday rotation and the anharmonic cyclotron absorptions near pulsars can now be quantified. Beyond the perturbative regime, real-time lattice QED is extended as a unique plasma simulation tool, especially when collective scales overlap with QED scales. Applying this tool to laser-plasma interactions, the transition from wakefield acceleration to electron-positron pair production is demonstrated for the first time when the laser fields exceed the Schwinger threshold. [Preview Abstract] |
Thursday, November 12, 2020 10:00AM - 10:30AM Live |
TI01.00002: Building Scientific Reproducibility into Plasma Research Invited Speaker: Nicholas Murphy The reproducibility crisis of modern science is the inability of scientists to reproduce roughly half of the results published in scientific journals. This crisis has affected a broad range of fields, such as psychology, chemistry, and oncology. While physicists tend to have high confidence that physics research is reproducible, no comprehensive studies have been performed to support or refute this claim. Nevertheless, the scientific, cultural, and institutional practices that contribute to the reproducibility crisis in other fields are also present in plasma science. This tutorial will describe how to implement best practices for scientific reproducibility into plasma research. The talk will begin by outlining sources of irreproducibility, such as cognitive biases, improper use of statistics, publication bias, closed access policies for data and software, and the reward system for modern academia. The talk will then describe remedies for these problems such as open access data policies; open metadata standards; open source software; training on proper use of statistics; pre-registration of research methodologies; independent methodological and statistical support; and valuing the reproducibility of research in tenure, hiring, and funding decisions. [Preview Abstract] |
Thursday, November 12, 2020 10:30AM - 11:00AM Live |
TI01.00003: Cognitive Simulation Models for Inertial Confinement Fusion: Combining Simulation and Experimental Data Invited Speaker: Kelli Humbird The design space for inertial confinement fusion (ICF) experiments is incredibly broad. Researchers rely heavily on computer simulations to traverse this enormous parameter space in search of high-performing designs. However, complex ICF multiphysics codes must still make simplifying assumptions. While highly predictive for many classes of experiment, simulations deviate from precision measurements for the most challenging implosions. For more effective design and investigation, simulations require input from past experimental data to better predict future performance. In this talk, we describe a cognitive simulation method for improving numerical models using existing empirical data. It uses deep neural network models to “calibrate” ICF simulation results through a process called “transfer learning.” The technique was originally developed for pure machine learning tasks, like image recognition, to cope with limited data. As we apply it here, it serves as a powerful, nonlinear method for calibrating ICF simulations against a wide range of experimental observables. To build our improved model, we train advanced neural networks on thousands of computer simulations before partially retraining them on sparse sets of experimental data. We use our transfer learning methods to produce elevated models that are far more accurate than simulations alone. We demonstrate improved model performance for a range of ICF experiments at both the Omega Laser Facility and the National Ignition Facility. We end by demonstrating how the methods might be used to carry out a data-driven experimental campaign to optimize performance, illustrating the key product – models that become increasingly accurate as data is acquired. This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344. LLNL-ABS-809626. [Preview Abstract] |
Thursday, November 12, 2020 11:00AM - 11:30AM Live |
TI01.00004: Quantum Algorithms for Efficient Classical Plasma Simulation Invited Speaker: Alexander Engel Quantum computers show promise to solve some computational problems exponentially faster than classical computers. This naturally includes the simulation of quantum many-body systems, yet quantum algorithms can also achieve speedups for problems unrelated to quantum physics. For example, a large speedup may be obtained if a computational problem can be mapped to a quantum system. We successfully apply that strategy to a specific kinetic plasma physics problem: linear Landau damping [1]. We find that a set of variables describing this system evolves unitarily, which allows us to develop and test a quantum algorithm that can simulate this evolution efficiently. Our algorithm has a computational complexity only logarithmic in the number of grid points, but also has a cost factor of 1/E, where E is the measurement error. Extensions to higher dimensions and electromagnetics appear straightforward, but incorporating nonlinearity is more challenging. We introduce techniques for mapping nonlinear dynamical systems to infinite-dimensional, linear dynamical systems. When the resulting linear system can be truncated, yielding a finite linear system, quantum algorithms can be applied to perform the associated evolution. We discuss conditions under which such an approach can accurately and efficiently reproduce outputs for nonlinear dynamical systems. Progress on applying this strategy to the Vlasov-Poisson system will be reported. [1] A. Engel, G. Smith, and S. E. Parker, Quantum algorithm for the Vlasov equation, Phys. Rev. A 100, 062315 (2019). [Preview Abstract] |
Thursday, November 12, 2020 11:30AM - 12:00PM Live |
TI01.00005: Understanding the Fusion Yield and All of Its Dependencies Using Statistical Modeling of Experimental Data Invited Speaker: Aarne Lees Statistical modeling of experimental and simulation databases has enabled the development of an accurate predictive capability for OMEGA DT layered implosions, leading to new target designs and record fusion yields, threefold higher than previously achieved.[V. Gopalaswamy et al., Nature 565, 581 (2019)] In addition to enhancements in fusion performance, a new application of statistical modeling has been devised to greatly improve our understanding of the underlying physics, various dependencies, and all degradation mechanisms affecting the fusion yield of OMEGA implosions. Since the statistical framework relates the outputs of 1-D simulations to experimental results, a judicious choice of simulation outputs can identify and quantitatively assess the different dependencies and degradation mechanisms. Each dependency is validated by comparison with trends in 3-D simulations. We find that the yield is reduced by four factors: the ratio of laser beam to target radius (a proxy for laser beam geometry mode); the variance of inferred ion temperatures (a proxy for l $=$ 1 mode from offset and mispointing); the time span over which the tritium fuel has decayed (a proxy for tritium damage and 3He buildup, subsequently included in codes as a result of this work); and the normalized pulse length [Tpulse/(R/Vimp)], related to the in-flight aspect ratio (a proxy for the growth of short wavelength modes from sources like laser imprinting). We find that the degradation from beam geometry illumination nonuniformity is greater than predicted by 3-D simulations and accounts for 30{\%} to 40{\%} reduction in yield in best-performing implosions. The degradation from short wavelength modes limits the yield at convergence higher than best performers. The degradation from DT-fill age is significant, and is mitigated by reducing fill-to-shot time to under three days. The l $=$ 1 mode is only important when Ti asymmetries exceed 10{\%}. This material is based upon work supported by the Department of Energy National Nuclear Security Administration under Award Number DE-NA0003856. [Preview Abstract] |
Thursday, November 12, 2020 12:00PM - 12:30PM Live |
TI01.00006: Structure-Preserving, Geometric, Particle-in-Cell Algorithms for Tokamaks Invited Speaker: Hong Qin Recently, structure-preserving, geometric discretization methods have been applied across a wide range of fields, including plasma physics, fluid dynamics, and astrophysics. Structure-preserving geometric algorithms preserve the geometric structures of the physical systems, such as Poincare symmetry and local energy-momentum conservation, gauge symmetry and charge conservation, and symplectic structure and phase space volume. They are especially suited for exascale computing hardware. They possess the long-term accuracy required, but unavailable using conventional algorithms, in the study of the multi-scale, complex dynamics of space and laboratory plasmas [PRL 100, 035006]. To preserve the symmetries and geometric structures of physical systems in discrete space-time lattices, the new algorithms utilize modern mathematical techniques, such as discrete manifold, interpolating differential forms, and non-canonical symplectic integrators [PoP 22, 124503]. The talk will focus on the recent development of the structure-preserving geometric Particle-in-Cell (PIC) algorithms [PoP 19, 084501; PoP 22, 112504; NF 56, 014001; NF 59, 106044], whose advantages are now apparent in a variety of important problems. For example, the long-term accuracy and fidelity of the algorithms made possible the first-ever whole-device 6D kinetic simulations of tokamak physics [arXiv:2004.08150] and enabled us to confirm numerically, over several orders of magnitude, Villani's Fields-Medal-winning theory on nonlinear Landau damping. In addition to the new generation of PIC methods, MHD simulations using the structure-preserving algorithms have now provided the strongest numerical confirmation so far of Parker's conjecture of current singularity. And structure-preserving algorithms for the Klein-Gordon-Maxwell system enabled the first real-time lattice QED simulations of laser-plasma interactions. These important developments and discoveries will be systematically reviewed as well. [Preview Abstract] |
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