is automatically converted and output in the inertial frame
(ground reference frame).
An optional section of the file allows forcing functions to be defined. These are enclosed in the
FORCING tag. The forcing type is enclosed within the FORCE tag and expressed in the file
as:
1<FORCE TYPE="[NAME]"> 2 ... 3</FORCE>
The force type can be any one of the following.
This force type allows the user to apply an absorption layer (essentially a porous region)
anywhere in the domain. The user may also specify a velocity profile to be imposed at the
start of this layer, and in the event of a time-dependent simulation, this profile can be
modulated with a time-dependent function. These velocity functions and the function defining
the region in which to apply the absorption layer are expressed in the CONDITIONS section,
however the name of these functions are defined here by the COEFF tag for the layer, the
REFFLOW tag for the velocity profile, and the REFFLOWTIME for the time-dependent
function.
1<FORCE TYPE="Absorption"> 2 <COEFF> [FUNCTION NAME] <COEFF/> 3 <REFFLOW> [FUNCTION NAME] <REFFLOW/> 4 <REFFLOWTIME> [FUNCTION NAME] <REFFLOWTIME/> 5 <BOUNDARYREGIONS> 1,4 <BOUNDARYREGIONS/> 6</FORCE>
If a list of BOUNDARYREGIONS is specified, the distance to these regions is available as additional
variable r in the definition of the COEFF function:
1<FUNCTION NAME="AbsorptionCoefficient"> 2 <E VAR="p" EVARS="r" VALUE="-5000 * exp(-0.5 * (3*r / 0.4)^2)" /> 3 <E VAR="u" EVARS="r" VALUE="-5000 * exp(-0.5 * (3*r / 0.4)^2)" /> 4 <E VAR="v" EVARS="r" VALUE="-5000 * exp(-0.5 * (3*r / 0.4)^2)" /> 5</FUNCTION>
This force type specifies the name of a body forcing function expressed in the CONDITIONS
section.
1<FORCE TYPE="Body"> 2 <BODYFORCE> [FUNCTION NAME] <BODYFORCE/> 3</FORCE>
This force type allows the user to apply synthetic turbulence generation in the flow field. The Synthetic Eddy Method is implemented. The approach developed here is based on a source term formulation. This formulation allows the user to apply synthetic turbulence generation in any specific area of the domain, not only in the boundary condition as most methodologies do. So that, after defining a synthetic eddy region (box of eddies), the user can randomly release eddies inside this box which are going to be convected downstream and will produce turbulence depending on the flow conditions. Each eddy that leaves the synthetic eddy region is reintroduced in the inlet plane of the box, so this mechanism re-energise the system, roughly speaking.
Below it is shown how to define the Synthetic Eddy Method for a fully three-dimensional
Navier-Stokes simulation. Note that this definition is under the FORCING tag. Firstly, in the
TYPE entry, we define the force type as IncNSSyntheticTurbulence for the incompressible
solver and CFSSyntheticTurbulence for the compressible solver. In the BoxOfEddies tag,
under the FORCE tag, the center plane of the synthetic eddy region is defined. The coordinates
of its center are given by x0, y0, z0 and lengths of its sides are lyref and lzref
in the y- and z-directions, respectively. Note that the length in the x-direction is
defined in the characteristic length scale function (see below), so that l00 defines
the value of lx. In the Sigma tag, we define the standard deviation (sigma) of the
Gaussian function with zero mean, which is used to compute the stochastic signal.
After that, the bulk velocity (Ub) of the flow must be provided in the BulkVelocity
tag.
1<FORCE TYPE="IncNSSyntheticTurbulence"> 2 <BoxOfEddies> x0 y0 z0 lyref lzref </BoxOfEddies> 3 <Sigma> sigma </Sigma> 4 <BulkVelocity> Ub </BulkVelocity> 5 <ReynoldsStresses> [ReynoldsStresses FUNCTION NAME] </ReynoldsStresses> 6 <CharLengthScales> [LenScales FUNCTION NAME] </CharLengthScales> 7</FORCE>
In order to define the Reynolds stresses (ReynoldsStresses tag) and the characteristic length
scales (CharLengthScales tag) of the eddies, the name of the functions which define them
must be given. These functions must be placed under the CONDITIONS tag. Both functions are
provided below. It is worthy mentioning that it is possible to define space-dependent functions
for each Reynolds stress. In other words, the user can, for instance, provide the analytical
solution of the Reynolds stresses close to the wall (boundary). This information is essential to
calculate the velocity fluctuations.
1<FUNCTION NAME="ReynoldsStresses"> 2 <E VAR="r00" VALUE="1e-3" /> 3 <E VAR="r10" VALUE="10*y+y^2+5*y^3" /> 4 <E VAR="r20" VALUE="0.0" /> 5 <E VAR="r11" VALUE="1e-3" /> 6 <E VAR="r21" VALUE="0.0" /> 7 <E VAR="r22" VALUE="1e-3" /> 8 </FUNCTION>
Also, in the Synthetic Eddy Method implemented here, an isotropic or anisotropic turbulence can be described depending on the values provided in the characteristic length scale function. For an isotropic turbulence, all the values must be the same.
1<FUNCTION NAME="LenScales"> 2 <E VAR="l00" VALUE="1.0" /> 3 <E VAR="l10" VALUE="0.085" /> 4 <E VAR="l20" VALUE="0.125" /> 5 <E VAR="l01" VALUE="0.4" /> 6 <E VAR="l11" VALUE="0.085" /> 7 <E VAR="l21" VALUE="0.125" /> 8 <E VAR="l02" VALUE="0.4" /> 9 <E VAR="l12" VALUE="0.170" /> 10 <E VAR="l22" VALUE="0.25" /> 11</FUNCTION>
Note that the synthetic turbulence generator is only supported for fully three-dimensional simulations.
This force type allows the solution of incompressilbe Navier-Stokes in moving frame of
reference. The moving frame is attached the to body and can have translational, rotational or
both motions. Although the Navier-Stokes equations are solved in a moving reference frame,
our formulation is based on the absolute velocity and pressure (in inertial frame). However,
note that these absolute velocities and any other vector quantities are expressed using the
coordinate basis of the moving frame. Further, note that if you are using the FilterAeroForces,
the force vector is automatically converted and output in the inertial frame
(ground reference frame).
To use this formulation the user need to specify the force type inside the FORCING tag as
follwos:
1<FORCE TYPE="MovingReferenceFrame"> 2 <FRAMEVELOCITY> [MRF FUNCTION NAME] </FRAMEVELOCITY> 3 <PIVOTPOINT> x0, y0, z0 </PIVOTPOINT> 4</FORCE>
Here we are required to provide one function for this force type which defines the linear
velocity and the angular velocity of reference frame or both. In the case of rotating frame, i.e.
when we are prescribing the angular velocity of reference frame, we can provide a coordinate
of PIVOTPOINT, around which the frame is rotating. If no pivot point provided, the origin of
coordinates in the moving reference frame will be used as the pivot point. Note that the
frame velocities (both linear and angular velocities) must be defined in the inertial
stationary frame of reference,i.e. ground reference frame (and expressed using the basis
of inertial stationary frame), however, the Poivot point is in the moving reference
frame.
Examples of linear and angular velocity funcitons together with their usage in the Forcing is shown below:
1<CONDITIONS> 2 3<FUNCTION NAME="VelMRF"> 4 <E VAR="u" VALUE="2*sin(PI*t)" /> 5 <E VAR="v" VALUE="0.1" /> 6 <E VAR="w" VALUE="0" /> 7 <E VAR="Omega_x" VALUE="0" /> 8 <E VAR="Omega_y" VALUE="0" /> 9 <E VAR="Omega_z" VALUE="0.3*cos(2*PI*t)" /> 10</FUNCTION> 11 12</CONDITIONS> 13 14<FORCING> 15 16 <FORCE TYPE="MovingReferenceFrame"> 17 <FRAMEVELOCITY> VelMRF </FRAMEVELOCITY> 18 <PIVOTPOINT> 0.2, 0.0, 0.0 </PIVOTPOINT> 19 </FORCE> 20 21</FORCING>
The moving frame function defines the velocity of the body frame observed in the inertial reference frame
This means that these functions (such as the VelMRF in the above example) are defined and
expressed in the stationary inertial frame (ground frame).
Here, u0 = (u, v, w) is the translational velocity, Ω = (Omega_x, Omega_y, Omega_z) is the angular velocity. x0 = (0.2,0.0,0.0) is the rotation pivot and it is fixed in the body frame. Translational motion is allowed for all dimensions while rotational motion is currently restricted to z (omega_z) for 2D, 3DH1D and full 3D simulations.
Finally, note that when using MovingReferenceFrame force type, for any open part of the
computational domain that the user specifies the velocity, such as inlet and free stream
boundary conditions, the USERDEFINEDTYPE="MovingFrameDomainVel" tag can be used for
velocity components. For example if boundary ID=2 is the inlet with Uinfx and Uinfy the
values of inlet velocities defined as parameters, the boundary condition for this boundary
becomes:
1<REGION REF="2"> 2 <D VAR="u" USERDEFINEDTYPE="MovingFrameDomainVel" VALUE="Uinfx" /> 3 <D VAR="v" USERDEFINEDTYPE="MovingFrameDomainVel" VALUE="Uinfy" /> 4 <N VAR="p" USERDEFINEDTYPE="H" VALUE="0" /> 5</REGION>
for the wall boundary conditions on the surface of the body, we need to use MovingFrameWall
tag as shown below:
1<REGION REF="0"> 2 <D VAR="u" USERDEFINEDTYPE="MovingFrameWall" VALUE="Uinfx" /> 3 <D VAR="v" USERDEFINEDTYPE="MovingFrameWall" VALUE="Uinfy" /> 4 <N VAR="p" USERDEFINEDTYPE="H" VALUE="0" /> 5</REGION>
The outlet and pressure boundary conditions are the same as before. In practical, It is not
necessary to set the MovingFrameDomainVel or MovingFrameWall tag for all velocity
components. In the circumstance when the analytical solution of a velocity component in the
body frame is available, the value can still be prescribed as a normal Dirichlet-type boundary
condition.
In addition to the prescribed motion, the frame velocity can also be determined from the fluid-body interaction. One example is as follows:
1<CONDITIONS> 2 <FUNCTION NAME="VelMRF"> 3 <E VAR="u" VALUE="0.0" /> 4 <E VAR="v" VALUE="0" /> 5 <E VAR="Omega_z" VALUE="cos(t)"/> 6 </FUNCTION> 7 8 <FUNCTION NAME="InitDisp"> 9 <E VAR="x" VALUE="0" /> 10 <E VAR="y" VALUE="0" /> 11 <E VAR="Theta_z" VALUE="0"/> 12 </FUNCTION> 13 14 <FUNCTION NAME="ExtForce"> 15 <E VAR="fx" VALUE="-1" /> 16 <E VAR="fy" VALUE="0" /> 17 <E VAR="Mz" VALUE="0"/> 18 </FUNCTION> 19</CONDITIONS> 20 21<FORCING> 22 <FORCE TYPE="MovingReferenceFrame"> 23 <FRAMEVELOCITY> VelMRF </FRAMEVELOCITY> 24 <EXTERNALFORCE> ExtForce </EXTERNALFORCE> 25 <INITIALDISPLACEMENT> InitDisp </INITIALDISPLACEMENT> 26 <MASS> 1.0, 0., 0, 0,1,0, 0,0,1 </MASS> 27 <MOTIONPRESCRIBED> 0, 1 ,1 </MOTIONPRESCRIBED> 28 <PIVOTPOINT> 0. , 0,0 </PIVOTPOINT> 29 </FORCE> 30</FORCING>
The meanings of the parameters are:
| Option name | Description | Default |
MOTIONPRESCRIBED | A vector that denotes a
certain degree of freedom
is prescribed
by the | 1 |
EXTERNALFORCE | External force exerted on the body, (fx, fy, fz, Mx, My, Mz). | 0 |
MASS | Mass matrix of the body system. Its size is 3 × 3 in 2D and 4 × 4 in 3D. | 0 |
DAMPING | Damper matrix of the body system. Its size is 3 × 3 in 2D and 4 × 4 in 3D. | 0 |
RIGIDITY | Stiffness matrix of the body system. Its size is 3 × 3 in 2D and 4 × 4 in 3D. | 0 |
INITIALDISPLACEMENT | A function to set the initial displacement of the body, (x, y, z, Theta_x, Theta_y, Theta_z). | 0 |
TRAVELINGWAVESPEED | A traveling wave motion can be set to simulation problem such as a traveling wave propagating along an infinite geometry. | 0 |
OutputFile | Output file name. | SessionFileName.mrf |
OutputFrequency | Output frequency of the body’s motion. | 1 |
This force type allows a forcing function to be applied directly within the code, thus it has no associated function.
1<FORCE TYPE="Programmatic"> 2</FORCE>
This force type allows the user to specify the magnitude of a white noise force. Optional arguments can also be used to define the frequency in time steps to recompute the noise (default is never) and the number of time steps to apply the noise (default is the entire simulation).
1<FORCE TYPE="Noise"> 2 <WHITENOISE> [VALUE] <WHITENOISE/> 3 <!-- Optional arguments --> 4 <UPDATEFREQ> [VALUE] <UPDATEFREQ/> 5 <NSTEPS> [VALUE] <NSTEPS/> 6</FORCE>