CIS & MINDS Seminar - Joshua Burby

<p>Recorded Seminar:</p><p><a href="https://wse.zoom.us/rec/share/fF4uuo6e3HiSViKWmH7TgvI6opMtEccoEnMFctlp5r... Zoom Meeting:</p><br><a href="https://wse.zoom.us/j/93822965644?pwd=dDNHYVZGY096QU9Dem45STBsQWQ2dz09">... W. Burby, PhD</b><br><p>Staff Scientist</p><p>Los Alamos National Laboratory</p><br><p><b>“A novel interacting particle system: plasma fragments”</b></p><p><b> </b><b>Abstract:</b><span> </span><span> </span><span>Themain statistical model for the dynamics of a rarefied plasma is by now wellestablished. The single-plasma-particle probability density obeys a nonlineartransport equation, attributed to Vlasov and Landau, whose validity hinges onthe assumption of weak correlations among plasma particles. On the other hand,dense plasmas, which can be ``strongly-coupled," present serioustheoretical and computational challenges that persist to this day. I willdescribe a modeling approach for strongly-coupled plasmas that interpolatesbetween the well-established Vlasov-Landau formalism and the full Coulombmany-body formalism for a single-species, fully-ionized plasma. The set of Nplasma particles is partitioned into M ``fragments," each comprising aK-tuple of particles. Then correlations between fragments are assumed to beweak. In this manner, correlations among particles in a given fragment aretreated non-perturbatively, implying a more accurate resolution of correlationsas the fragment length K increases. When correlations between fragments arecompletely ignored, the single-fragment probability density obeys a Hamiltoniantransport equation that generalizes the Vlasov-Poisson system. When the firstperturbative effects of fragment correlations are accounted for, the mean-fieldequations are corrected by a collision integral that produces entropy whileconserving energy. When K=1 the collision integral reproduces the well-knownresult due to Landau. But more generally the collision integral cannot beevaluated in closed form due to the non-integrability of the Coulomb K-bodyproblem. I will argue that data-driven methods offer a compelling practicalpath toward incorporating the effects of fragment collisions into fragmentkinetic simulations.</span></p><br><p> </p><p><b>Biography:</b>  JoshuaW. Burby has been a Staff Scientist at Los Alamos National Laboratory since2019. After earning his degree in plasma physics from Princeton University in2015, his postdoctoral research was funded by a series of distinguishedfellowships at the Courant Institute of Mathematical Sciences, the MathematicalSciences Research Institute, and Los Alamos. His research covers mathematicalplasma physics, dynamical systems, and structure-preserving machine learning.</p><p> </p>