Joel E. Tohline
Alumni ProfessorDepartment of Physics & Astronomy
Louisiana State University
|
Building on our past success in simulating complex fluid flows in
Newtonian astrophysical systems, and taking advantage of the
expertise
in relativistic systems that has been brought together at LSU
(especially in collaboration with LSU Professors
Lehner,
Pullin, and
Frank),
we are developing the capability to
simulate fully relativistic fluid flows on a self-consistently evolving,
warped spacetime metric.
We will use these new simulation tools to significantly extend our
study of the stability of contact binaries and accretion flows in
semi-detached binary star systems,
to investigate various mechanisms for igniting explosions
in compact binaries,
and to better understand dynamical and secular instabilities
that arise in newly formed neutron stars.
Our relativistic simulations will incorporate AMR (adaptive mesh-refinement), arbitrarily specified killing fields, various high-order advection schemes, and in the weak-field limit will accurately represent self-gravitating Newtonian flows. By developing these computational tools within the Cactus Framework we will be able to effectively extend our simulations on individual massively parallel computers (such as SuperMike at LSU, or tungsten at NCSA) to a Grid Computing environment that ties multiple high-performance computers together via a high-bandwidth network (such as LONI within the state of Louisiana, or the National LambdaRail). The heterogeneous computing environment that permits us to routinely generate three-dimensional (3D) animation sequences to illustrate and diagnose the results of our fluid simulations is being extended to generate movies at High-Definition (HD) resolution, and to permit the real-time analysis of ongoing simulations via a web-based VRML (Virtual-Reality Markup Language) viewer or while sitting in an immersive CAVE environment. We also are exploring how digital holographic techniques may be developed into a practical tool to aid in the real-time visualization of numerical simulations.
|
| — formulated in July, 2005 |