Yukawa-Friedel-tail pair potentials for warm dense matter applications.

Accurate equations of state (EOS) and plasma transport properties are essential for numerical simulations of warm dense matter encountered in many high-energy-density situations. Molecular dynamics (MD) is a simulation method that generates EOS and transport data using an externally provided potential to dynamically evolve the particles without further reference to the electrons. To minimize computational cost, pair potentials needed in MD may be obtained from the neutral-pseudoatom (NPA) approach, a form of single-ion density functional theory (DFT), where many-ion effects are included via ion-ion correlation functionals. Standard N-ion DFT-MD provides pair potentials via the force matching technique but at much greater computational cost. Here we propose a simple analytic model for pair potentials with physically meaningful parameters based on a Yukawa form with a thermally damped Friedel tail (YFT) applicable to systems containing free electrons. The YFT model accurately fits NPA pair potentials or the nonparametric force-matched potentials from N-ion DFT-MD, showing excellent agreement for a wide range of conditions. The YFT form provides accurate extrapolations of the NPA or force-matched potentials for small and large particle separations within a physical model. Our method can be adopted to treat plasma mixtures, allowing for large-scale simulations of multispecies warm dense matter.

Multifidelity regression of sparse plasma transport data available in disparate physical regimes.

Physical data are typically generated by experiments and computations in limited parameter regimes. When datasets generated using such disparate methods are combined into one dataset, the resulting dataset is typically sparse, with dense "islands" in a potentially high-dimensional parameter space, and predictions must be interpolated among such islands. Using plasma transport data as our example, we propose a multifidelity Gaussian-process regression framework that incorporates physical data from multiple sources at multiple fidelities. The impact of the proposed framework varies from little improvement over simpler approaches to qualitatively changing the prediction with consistently increased confidence in regions lacking high-fidelity data. By varying low- and high-fidelity data sources, we demonstrate an approach for determining when multifidelity Gaussian-process regression adds value over single-fidelity regression and therefore when its additional computational costs are merited. We also examine the case in which the outputs of the low- and high-fidelity models correspond to different physical quantities, one of which may be intrinsically computationally cheaper to compute. We conclude by analyzing strategies for sampling high-fidelity data for use in this framework, and we develop a simple sampling approach for reducing regression error across large gaps in data.