Date of Award

Fall 2023

Project Type


Program or Major


Degree Name

Doctor of Philosophy

First Advisor

Kai Germaschewski

Second Advisor

Will Fox

Third Advisor

Kai Germaschewski


The scope of this work spans computational plasma simulation and modeling using state-of-the-art numerical and machine learning methods. The numerical work discussed involves performance and correctness modeling of two massively scalable heterogenous code bases. A gyrokinetic code used to simulate a plasma subjected to a strong magnetic guiding field and a particle-in-cell code used for simulating high energy density plasmas in a number of scenarios undergoing magnetic reconnection. Both code bases are capable of being distributed to cpu or gpu based compute nodes.

The gyrokinetic code was used to study microinstabilities and turbulence in the edge region of a magnetically confined fusion device. Plasma confinement in such a device is partially due to the configuration of a poloidal magnetic field. It has been shown in experiment that a flipping of the magnetic field orientation may help reduce edge region turbulence, thus reducing the possibility of a catastrophic disruptions. Here a gyrokinetic code was used to study the edge region and the effect of poloidal field configuration on microinstability driven anomalous transport.

The particle-in-cell numerical method of plasma physics balances a trade-off between computational cost and intrinsic noise. Inference on data produced by these simulations generally consists of binning the data to recover the particle distribution function, from which physical processes may be investigated. In addition to containing noise, the distribution function is temporally dynamic and can be non-gaussian and multi-modal, making the task of modeling it difficult.

Here, the particle-in-cell code was used to produce data which was used with normalizing flow deep learning methods to learn a smooth, tractable approximation to the noisy particle distribution function. It is demonstrated that the resulting data driven likelihood conserves relevant physics and may be extended to encapsulate the temporal evolution of the distribution function.

An alternative to particle-in-cell and gyrokinetic models are multi fluid models. The inclusion of kinetic effects into fluid models has been a long standing problem in magnetic reconnection and plasma physics. Generally the pressure tensor is reduced to a scalar which is an approximation used to aid in the modeling of large scale global systems such as the Earth's magnetosphere. This unfortunately omits important kinetic physics which have been shown to play a crucial role in collisionless regimes.

The multi-fluid 10-moment model on the other-hand retains the full symmetric pressure tensor. The 10-moment model is constructed by taking moments of the Vlasov equation up to second order, and includes the scalar density, the vector bulk-flow, and the symmetric pressure tensor for a total of 10 separate components. Use of the multi-fluid 10-moment model requires a closure which truncates the cascading system of equations. Here we look to leverage data-driven methodologies to seek a closure which may improve physical fidelity of the 10-moment multi-fluid model in collisionless regimes. Specifically we use the Sparse Identification of Nonlinear Dynamics (SINDy) method for symbolic equation discovery to seek the truncating closure from fully kinetic particle-in-cell simulation data, which inherently retains the relevant kinetic physics. We verify our method by reproducing the 10-moment model from the PIC particle data and use the method to generate a closure truncating the 10-moment model which is analyzed through the nonlinear phase of reconnection