mathias
Forum Replies Created
-
Posts
-
Yes, that is correct! Have a look at our upcomming spring school – it could be interesting. I can also make contact to our turbulence expert developer. In any of the two cases, please contact me directly.
Best
Mathias J. (Krause)Dear Arash81068,
the two line asign the kind of bc. The value is set be “sLattice.defineRho”. You can define that within the time loop for changing boundary pressure values as it is done for the velocity in that very example.
Best
MathiasDear Ryan,
Thanks for your interest.
1/ In Navier-Strokes euations, pressure is defined up to a constant. At the outlet we fix this constant. So it is a reference pressure. Taking the computed pressure distripution you can add any constant pressure you want. So a negative value is fine and still physical.
2/ The paper refers to a different example. There we fixed the mass flow by fixing the velocity. If you want do that in the mentioned example you may set velocity boundary conditions at the two outflows and a pressure condition at the inflow.Best
MathiasThere is no new inflow velocity. The viscosity is chenaged simply. Best Mathias
There is no restriction. The problem must be caused be something else! Best Mathias
You can combine it easily but it was not tested yet. For 2D it should work all right. But maybe using a standard LES modell is a better choice since LES was studied a lot in combination with LBM.
Best
MathiasDear John,
basically you set the pressure! You may leave it to the values 1 and chance the charRho in..
std::vector<T> rho0;
rho0.push_back( 1 ); //first lattice
rho0.push_back( 1 ); //second lattice -> e.g. change it to 2
PsiEqualsRho<T,T> interactionPotential;
ShanChenForcedGenerator3D<T,DESCRIPTOR> coupling( G,rho0,interactionPotential );Best
MathiasWhich boundary conditions did you use? Bounce back would be my first try! Then, please check if the initial conditions are fine. In particular check, that they cause no big jumps in the density. Best Mathias
CharRho is a value in Si units which is linked to rho=1 in LB simulations. Details can be found in the implemation of our unitConverter as well as in a general form in the book of Krüger on LBM. Best Mathias
Please, read our documentation and have a look at our example. We cannot give a complete introducion in this forum. For it, you would need to come to us in person, e.g. to our next spring school! Best Mathias
In OpenLB in the example folder! You may consider coming to our spring school next March! Best Mathias
Dear afnene,
Please have a look at our example “phaseSeparation2d/3d”.
Best
MathiasIn the new release we will provide an example. I recomment that you come to our next spring school in March to get more details! Best Mathias
You can set any value you want at the pressure outlet. I also said, you may come to Karlsruhe before! Best Mathias
Dear Vaibhav,
With LBM, we usually store one or several distribution functions f at each node. From that we compute physical quanteties like the density, pressure, velocity, temerature, stress, … as moments of f. There is theoretical evidence that you get macroskopic equations like Navier-Stokes, convection-diffision, even Maxwell up to a discretization error! Have a look in the standard literature or (and ;-)) come to our next spring school which takes place in Karlsruhe.
Best
Mathias