FEATURES OF SOLVER –MP NS
fluidyn–
MPÔ
NS is
a general purpose Computational Fluid Dynamics software package to
simulate fluid flow in and around complex geometrical
configurations. It can simulate transient/steady
incompressible/compressible flows on 3D unstructured meshes with
heat transfer, radiation, turbulence, chemical reactions,
multi-phases with free surfaces, droplets, etc.
a)
Analyzes incompressible flows and compressible at subsonic,
transonic and supersonic speeds (with the shock-capturing).
b)
Analyzes
steady and transient flows.
c)
Analyzes
viscous or inviscid flows.
d)
Finite
volume scheme based on unstructured/multi-block mesh.
e)
Supports
the following mesh features:
-
unstructured/multi-block
mesh
-
hexahedral,
tetrahedral, prismatic and other commonly used general polyhedral
mesh
-
non-conformal
mesh blocks
-
moving/deforming
meshes
-
sliding
meshes for moving/rotating components
1. Core
Hydrodynamic Solver
NS has three solvers:
|
Module
|
Description
|
|
TVD
|
Density-based
fully coupled explicit/implicit time-marching method for
unstructured mesh.
|
|
MB
|
Pressure-based
semi-implicit time-marching method for multi-block mesh
|
|
NT
|
Pressure-based
fully-implicit segregated method for unstructured mesh.
|
All solvers use finite volume methods based on Cartesian velocity
components. While TVD and NT solvers use collocated method, MB uses
cell-vertex scheme. The NT code, though written for 3D flows,
automatically determines if the problem is 2D and if so, solves only
the relevant velocity components.
The TVD code uses the coupled method to solve the governing
equations. MB uses a semi-coupled strategy while NT uses a
segregated method.
All the solvers support flow through porous media.
1.1
Convection Schemes
The
convection schemes available in different solvers are as follows:
|
TVD
|
MB
|
NT
|
|
Van
Leer flux-vector splitting
|
Weighted
partial donor cell scheme
|
Central
difference
|
|
Roe
flux-difference splitting
|
Quasi-second
order upwind scheme
|
Gamma
differencing (2nd order)
|
|
AUSM
|
|
Flux
limiter (TVD-k)
schemes (3rd order):
-
van
Leer smooth Limiter
-
van
Albada Limiter
-
MinMod
Limiter
-
Super-Bee
Limiter
-
Linear-kappa
scheme
-
SMART
scheme
-
UMIST
Limiter
-
van
Leer MUSCL Limiter
-
ISNaS
Limiter
|
|
HLLC
|
|
|
|
Preconditioned
Riemann solvers for incompressible/low-speed flows
|
|
|
1.2
Time-Differencing Schemes
|
TVD
|
MB
|
NT
|
|
Multi-stage
Runge-Kutta explicit
|
Euler
implicit (1st order)
|
Euler
implicit (1st order)
|
|
Jacobi
implicit
|
|
3
time-level scheme (2nd order)
|
1.3 Pressure
Computation
|
TVD
|
MB
|
NT
|
|
Weiss-Smith
preconditioning
|
SIMPLE
|
SIMPLE
|
|
Jacobi
implicit
|
|
SIMPLEC
|
|
|
|
PISO
|
Non-orthogonal terms are accounted while solving for pressure
1.4 Linear
Equation Solvers
The linear equation solvers used to solve the implicit difference
equations are:
|
TVD
|
MB
|
NT
|
|
Point-Jacobi
|
Conjugate
Residual
|
SIP
(for structured mesh)
|
|
|
|
ICCG
for symmetric matrices
|
|
|
|
CGSTAB/BiCGSTAB
|
|
|
|
GMRES
|
1.5
Thermodynamic Models
The thermodynamic models offered are:
|
|
TVD
|
MB
|
NT
|
|
Incompressible
|
density
variation with temperature
|
density
variation with temperature
|
density
variation with temperature
|
|
Compressible
|
Perfect
gas
|
Perfect
gas
|
Perfect
gas
|
|
Ideal
gas
|
Ideal
gas
|
Ideal
gas
|
|
Ideal
mixture
|
Ideal
mixture
|
Ideal
mixture
|
|
Equilibrium
air
|
|
|
|
JWL
(high-explosives)
|
|
|
|
polynomial
|
|
|
|
Mie-Gruneisan
|
|
|
|
Two-phase
expansion EOS for vapor-liquid equilibria
|
|
|
1.6 Boundary
Conditions
All solvers support the following boundary conditions:
-
Prescribed
inflow
-
Prescribed
outflow
-
Specified
pressure
-
Impermeable
wall and baffle surfaces
-
Cyclic
(periodic) boundaries
-
Symmetry
plane
-
Specified
stagnation condition
-
Free
stream
-
Transmit
condition
-
Characteristic
based Riemann conditions
1.7 Gravity
Models
Following gravity models are available in all solvers:
-
Full
gravity Model
-
Buoyancy
Model
-
Boussinesq
Model
2 Additional
Modeling Features
2.1
Turbulence
The
following turbulence models are available in NS:
-
Standard
k-e
with compressibility corretcions
-
Re-normalization group (RNG) k-e
-
Chen-Kim
k-e
-
Mixing
length models, Baldwin-Lomax, Cebeci-Smith
-
Large eddy simulation using the Smagorinsky’s Sub-Grid Scale model (SGS)
Non-equilibrium
law-of-wall condition is used at solid walls.
2.2
Non-Matching Mesh
The TVD and NT solvers can handle non-matching or non-conformal mesh
interfaces. This is useful for sliding meshes, for fine grid
embedding, and also for component-by-component meshing.
2.3 Moving
Mesh
All the solvers can handle moving meshes. User has to code (in
Fortran) the velocities of grid node movement.
2.4 Reactive
Flows
The
code can handle chemically reactive flows with reaction rates being
computed from one of the following 4 models:
·
Arrhenius
model
·
Eddy-dissipation
concept of Magnussen
·
Minimum
of the above two
·
Combined
Time Scale Model
2.5 Thermal
Radiation
The NT module has the P1-aproximation model for modeling thermal
radiation transport through participating media.
2.6
Free-Surface Flows
The MB and NT modules have the Volume of Fluid (VOF) for simulating
interpenetrating two-phase flows. Continuum Surface Force (CSF)
model is used for surface tension forces. Three additional schemes
are offered for advecting VOF:
·
Inter-Gamma
Differencing
·
High
Resolution Interface Capturing (HRIC)
·
Compressive
Interface Capturing Scheme for Arbitrary Meshes (CICSAM)
2.7 Dispersed
Flow
The MB and NT modules have the Lagrangian particle tracking method
for simulating dispersed flows. The salient features of this method
are:
o
Lagrangian
particle tracking method
o
Monte-Carlo
sampling for particle injection/turbulent dispersion
o
Polydisperse
particle distributions:
-
c2
distribution
-
Rosin-Rammler
-
Upper
Limit Log Normal (ULLN)
o
Selection
of models for particle momentum exchange with carrier phase
o
Droplet
evaporation with heating
o
Turbulent
dispersion
o
Droplet
breakup/collision models
2.8
Multiphase Flows
The NT module has the Eulerian multiphase flow model for simulating
multiple fluid-streams in different physical states coupled to each
other. It uses a fully-coupled solution of multi-phase flow
equations along with a single combined pressure field. The
interfacial area transport equation is solved to compute local
interface areas needed for evaluation of phase exchange terms. It
also has automatic regime mapping and activation of phenomenological
models for phase exchange terms.
2.9 Magneto
Hydrodynamics (MHD)
The NT module can solve MHD flows with constant or varying magnetic
field. It has both the electric potential method and magnetic
induction method. It can also simulate, jointly with the FE solver,
a jump in electric potential/magnetic fields.
2.10 Porous
Media
All
the modules can take into account the volume porosity. The NT module
can model flows and pollutant dispersion described by Darcy law
using direction dependent permeabilities. MP as a whole can model
heat transfer in porous media using local thermal non-equilibrium
model.
User Coding
User
coding is available for:
·
Thermodynamic
equation of state
·
Boundary
conditions
·
Initial
conditions
·
Viscosity,
conductivity, and other fluid transport properties
·
Momentum,
enthalpy, turbulence, chemical and other source terms
·
Turbulence
model
·
Convective
scheme
·
Diffusion
scheme
·
Chemical
reaction rate
·
Radiation
properties
·
Moving
grid
·
Solution
of linear equations
·
Time
step for transient calculations
·
Coefficient
of drag in Lagrangian particle tracking
·
Dispersed
phase injection
·
Dispersed
phase particle size
·
Output
and post-processing
For user programming, all the geometrical and solver
data can be accessed through common blocks and subroutines.
(Note:
Fortran compiler required for user compilation.)
|