A High-Order Accurate Particle-in-Cell Method for Vlasov-Poisson Problems Over Long Time Integrations
Speaker: Andrew Myers, Lawrence Berkeley National Laboratory
Presentation
Seminar Slide Deck (PDF-6.5MB)
Abstract
The Vlasov-Poisson system of equations is used to model collisionless fluid flows in both plasma physics and cosmology. Particle methods—especially the popular Particle-in-Cell technique—are a common approach to obtaining numerical solutions to this system. However, these methods have several downsides. First, they are usually limited to 2nd-order accuracy. Second, they are known to suffer from particle noise effects that degrade the accuracy of the numerical solution over long time integrations. In this talk, I will describe a new Particle-in-Cell method does not suffer from these problems. Our method is 4th-order accurate in both space and time and uses a 6th-order phase-space remapping technique to control particle noise at late times. I will describe the implementation of our algorithm using the Chombo framework for solving partial differential equations and demonstrate its performance on a set of standard test problems.
Biography
Andrew Myers received his Ph.D. in Physics from the University of California, Berkeley, where he used radiation-magnetohydrodynamic simulations to study star formation in the interstellar medium. After that, he was a postdoctoral researcher in the Applied Numerical Algorithms Group at Lawrence Berkeley National Laboratory, where he focused on particle methods for astrophysics and plasma physics simulations. He is also interested in a variety of other topics within scientific computing, including adaptive mesh refinement algorithms, reproducibility, and data analysis and visualization.