[Previous] | [Session 14] | [Next]
V.N. Gamezo, A.M. Khokhlov, E.S. Oran (Lab. for Computational Physics and Fluid Dynamics, Naval Research Laboratory)
We consider a Type Ia supernova explosion originating as a deflagration in the center of a carbon-oxygen Chandrasekhar-mass white dwarf (WD) with initial composition 0.5C+0.5O, central density 2.0 \times 109 g/cm3 and initial radius 2.1 \times 108 cm. A three-dimensional (3D) numerical model is based on reactive Euler equations of fluid dynamics coupled with an equation of state for a degenerate matter and a simplified kinetics of energy release. The energy-release model provides the correct propagation velocity for a laminar flame and takes into account carbon burning, as well as nuclear statistical quasi-equilibrium and equilibrium relaxations. The model for the turbulent burning on scales that are not resolved in the simulations is based on the assumption that burning on small scales is driven by the gravity-induced Rayleigh-Taylor (RT) instability. We performed 3D calculations for the first 1.9 seconds of explosion using an adaptively refined structured mesh. For the highest-resolution case, the minimum cell size was 2.6 \times 105 cm, and the mesh consisted of 108 computational cells by the end of the simulation. The flame, started as a sphere with the radius 3 \times 106 cm, becomes very convoluted due to the RT and Kelvin-Helmholtz instabilities on resolved scales and develops multiple buoyant plumes. As the plumes grow, the unburnt material either sinks towards the center or expands more slowly than the burnt material inside the plumes. The material burns at all distances from the center even when the larger flame plumes reach the outer layers of the star. By 1.9 seconds, some of these plumes approach the surface of the expanding WD that extends to (5-6) \times 108 cm from the center. About 50% of the material burns out releasing 1.3 \times 1051 ergs of nuclear energy. The expansion velocity at the surface reaches 1.2 \times 109 cm/s and continues to grow. A convergence study shows that at high resolutions, the results become practically independent on the computational cell size and insensitive to subgrid model parameters.
[Previous] | [Session 14] | [Next]
Bulletin of the American Astronomical Society, 34
© 2002. The American Astronomical Soceity.