Neutrino Transport Simulations of Core-Collapse Supernovae

Dnr:

SNIC 2017/1-259

Type:

SNAC Medium

Principal Investigator:

Evan Patrick O'Connor

Affiliation:

Stockholms universitet

Start Date:

2017-07-01

End Date:

2018-01-01

Primary Classification:

10305: Astronomi, astrofysik och kosmologi

Webpage:

http://ttt.astro.su.se/~eoco

Allocation

Abstract

Core-collapse supernova (CCSN) explosions mark the death of a massive star and are one of the brightest phenomena in the modern universe. These explosions are a cornerstone of astronomy and astrophysics. The energy released plays a critical role in the formation and evolution of galaxies. The explosion itself is responsible for dispersing newly formed elements, including the building blocks of life, throughout the galaxy. CCSNe are also the birth site of neutron stars and black holes. The explosion is powered by the gravitational binding energy released when the core of a massive star collapses. Collapse continues until the densities reach values comparable to the densities found in an atomic nucleus. At this point the nuclear force between the neutrons and protons can support the matter and the collapse is halted. A shock forms, which will eventually become the supernova shock that propagates throughout the entire star and gives rise to the bright optically display. Despite over 50 years of theoretical effort, the exact mechanism and details of how the implosion is converted to an explosion are only now starting to be uncovered. What has become very clear is that the problem is a complex, multiphysics problem that can only be addressed with multidimensional, neutrino-radiation-hydrodynamic simulations. Over the past several years my colleagues and I have developed an efficient infrastructure for performing simulations of CCSNe with the FLASH hydrodynamics toolkit (Fryxell et al. 2000; Couch 2013; Couch & O'Connor 2014; O'Connor & Couch 2015). We are currently at a position where we can easily perform a large number of 2D (axisymmetry) simulations in order to systematically explore CCSNe. Furthermore, we are beginning to extend our simulations to 3D. Such simulations, especially with neutrino transport, are quite expensive (10s of millions of CPU hours per 3D simulation) and accurate approximations need to be developed to reduce this cost. With this medium size allocation on SNIC we will work towards systematically studying CCSNe in 2D. We will explore many CCSN progenitor models in order to assess how the initial conditions impact the development of the explosion. We will also systematically explore new, novel, and uncertain areas of nuclear and neutrino physics in order to make concrete statements on the sensitivity of the problem to the underlying physics. On modern processors, each 2D simulation progresses at roughly 150ms/day on 50 cores. For a typical total run time of 1 second, this results in a total CPU usage of 8 000 hours per simulation. With our request of 300 000 hours/month, this will allow us to run 30-40 2D simulations per month. For our 2D simulations we will run 3-5 simulations with different random perturbations in order to ensure our results are not stochastic in nature, 2D simulations are particularly prone to this issue. Furthermore, with this allocation we will start to explore approximations for 3D simulations. This allocation will also serve to prepare our team for submitting a large-size allocation proposal to SNIC in the future.