CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Phoenics (https://www.cfd-online.com/Forums/phoenics/)
-   -   Soft boundary conditions (https://www.cfd-online.com/Forums/phoenics/51351-soft-boundary-conditions.html)

Roustam February 19, 2002 03:19

Soft boundary conditions
 
Can anybody clarify the following way off setting of outlet boundary conditions? We usually fix pressure P1 in the OUTLET by setting Value and Coefficient. How to set derivative of velocity along the outlet direction is zero in the OUTLET, or so called "soft boundary conditions" via Virtual reality tools of PHOENICS-3.3 instead of setting pressure? E.g. d(U1)/dx=0 in the OUTLET area, which is located normal to the x coordinate lines.

Richard February 19, 2002 03:37

Re: Soft boundary conditions
 
You can set the X velocity, y velocity and Z velocity as IN-CELL in the outlet attribute.

Roustam February 21, 2002 10:10

Re: Soft boundary conditions
 
Dear Richard, The way of setting in-cell for u1,v1,w1 in the outlet attributes doesn't work, as it is only used if flow should enter. In sells where flow is out of the domain, the settings of external velocity value including in-sell will be ignored. Probably, I don't understand some detail?

David Glynn February 25, 2002 11:02

Re: Soft boundary conditions
 
Dear Roustam,

You are correct - "in-cell" specifies that when the OUTLET actually has inflow (and an OUTLET is really of course a fixed-pressure condition) then the value of the variable to be convected in is the value already in the cell. (This is specified as "SAME" in the Q1).

To achieve a zero-gradient condition, I have always used Ground when I have wanted to do this. You specify the outlet as "FIXFLU, GRND" and compute the flow rate and velocity at the exit from the values at the penultimate cell, in Ground. You do of course need to have a recompilable version of the code to do this.

I would need to think whether there is any way of doing this using the VR facilities.

By the way, are you aware that you can introduce a measure of "softness" into the boundary condition by reducing the coefficient? The coefficient is the mass flow rate through the boundary divided by the pressure difference, i.e. pressure in cell adjacent to boundary minus external pressure (also divided by the area). So you can use the coefficient to specify whether there is some "softness" in this pressure.

I don't know if this would be any help in your application. It comes in useful sometimes, e.g. for problems with buoyancy and with a vertical open boundary.

Hope this is helpful,

Best regards,

David Glynn

Flowsolve

Roustam February 26, 2002 09:09

Re: Soft boundary conditions
 
Dear David Glynn,

Thank you. I have recompilable version of PHOENICS-3.3 and Compaq Fortran 6. But I have never managed to write my own ground and recompile it. Unless the subroutine for zero gradient velocity outlet boundary conditions you mentioned is already built-in so I could just type (fixflu, grndN) in user-defined object attributes via VR-reality editor. I got a two phase gravitational separation problem, so zero gradient boundary conditions would be helpful in my application Roustam


All times are GMT -4. The time now is 22:42.