 Multiple
Unix and Linux platform support for either
1)
Block level parallelization using MPI or MPICH for multi-processor machines
2)
Serial execution for single CPU machines
2-D,
axi-symmetric, or 3-D single or multi-block grid topologies |
Block-to-block
interface options:
1) arbitrary, block face-to-block
face C(0) continuous connectivity
2) arbitrary, block face-to-block
face non-C(0) continuous connectivity |
Steady-state
algorithms for spatially elliptic/hyperbolic and spatially
parabolic/hyperbolic equations
1) Runge-Kutta with implicit residual smoothing
2) Diagonalized approximate factorization
3) Incomplete LU (calorically perfect flows only) |
Unsteady algorithms
1) Runge-Kutta
2) Diagonalized approximate with dual time-stepping
3) Incomplete LU with dual time-stepping (calorically perfect flows only) |
factorization
Convergence
acceleration options:
1) multigrid
2) mesh sequencing |
Gas
models
1) Single component calorically
and thermally perfect gases
2) Arbitrary multi-component
mixtures of thermally perfect gases |
Chemistry
models
1) Frozen flow
2) Arbitrary, non-equilibrium,
finite-rate chemical kinetics |
Inviscid
flux treatments
1) Jameson's central differencing
with artificial dissipation
2) MUSCL kappa scheme of
van Leer
a) Roe's flux difference split scheme with entropy fixes
b) Edwards' low dissipation flux split scheme
c) HLLC scheme of Toro |
Viscous
flux treatments for laminar or turbulent flow
1) Full Navier-Stokes
2) Thin-layer Navier-Stokes
3) Parabolized Navier-Stokes |
Mean
flow turbulence model options
1) Spalart-Allmaras
2) Wilcox k-omega (98)
3) Wilcox low Reynolds no.
k-omega (98)
4) Menter k-omega (baseline)
5) Menter k-omega (SST)
6) Abid low Reynolds no.
k-epsilon
7) Gatski-Speziale EASM k-omega
(98)
8) Gatski-Speziale EASM
k-epsilon |
Mean
flow turbulence boundary condition treatment options
1) Wilcox compressible wall
matching ( k-omega based models)
2) Solve to wall (all models)
|
Hybrid
RANS/LES options
1) Detached Eddy Simulation (two-equation based model of Strelets)
2) Hybrid RANS/LES model of Baurle and Edwards |
Turbulence/chemistry
interaction models
1) Eddy break-up model of Magnussen and Hjertager
2) Temperature fluctuations:
average turbulent reaction rate coefficient
a) Assumed gaussian probability density function (PDF)
b) Assumed beta probability density function (PDF)
3) Species fluctuations: average
species product of concentration using a
multi-variate assumed beta probability density function (PDF) |
VULCAN, named for the Roman god of fire, is a Computational Fluid Dynamics software package (available for serial and parallel computational platforms) for turbulent reacting and non-reacting flows at conditions ranging from subsonic to hypersonic speeds. The computational cost of propulsion flow analysis is reduced through the use of special turbulent wall treatments, multi-grid methods for elliptic and space marching schemes, and conditioning of the governing equations to reduce numerical stiffness. Physical modeling capabilities are improved through the inclusion of models for compressibility, Reynolds stress anisotropies, turbulent diffusivity, finite rate chemistry, and turbulence/chemistry interaction effects. VULCAN can simulate two-dimensional, three-dimensional, or axi-symmetric problems on structured multi-block grid systems. A variety of PDE-based turbulence models are available (including explicit algebraic Reynolds stress models) for use with VULCAN. Hybrid Reynolds-Averaged Simulation / Large Eddy Simulation options are also present. VULCAN offers significant geometric flexibility; boundary conditions can be imposed on any boundary or boundary subset, and the code has (C0)continuous as well as non-C(0) continuous block-to-block interface capability. VULCAN also offers thermodynamic and kinetic model flexibility. The working fluid can be simulated as a calorically perfect single component gas, as a mixture of thermally perfect gases (with or without chemical reactions).
Highly
Implicit algorithms
