Stopping UDS
 

I have defined two fluent volumes in gambit and wish fluid to pass between them as there is no wall seperating them. I also want to create a A UDS that only exists in one of the volumes. How can I create an insulting zone for the UDS at the interface of the two fluid volumes? I know that you can set the values and source terms within each of the volumes, but I can't work out how to implement this for my problem.

Re: Stopping UDS
 

For the first part of your problem you have to go back to gambit, and connect the faces that are the interface of your two volumes. You probably will have to remesh the volumes as well. If you want to distinguish between your two volumes you just have to define two fluid regions - still with gambit. Concerning the second part of your problem, I cannot help you. I do not have all data about your problem, but it doesn't seem very logic to solve a transport equation for only one volume if both volumes are connected.

Re: Stopping UDS
 

Thankyou for your reply, I only have the one problem which is the concern of how to stop the uds propergating from one fluid region into the other. I have created two fluid regions and named them. I do not want the UDS to be solve for a transport equation. I simply need an additional variable that will sovle poissions equation in one fluid region.

Re: Stopping UDS
 

Well, theoretically if you put a zero diffusion coefficient and zero advective term in the second fluid zone for your UDS, you should be able to stop the UDS penetrating to the second zone. The accuracy of this is open to debate. The other idea might be using large source term technique to freeze the UDS in the second fluid zone to the boundary value between the fluid zone. Therefore, you'd have no gradient in the second fluid zone and no penetration by diffusion. This may get slightly complicated, though.

Ugur

Re: Stopping UDS
 

These are both excellent suggestions and I will get around to investigating them soon. I am a little concerned about setting coefficients to zero as this screams "floating point" or "nan" errors. Setting large source terms and assuming that onlyt small amounts of the UDS will 'leaks' through is something that I had considered but I didn't really see it as an elegant solution. However, when needs must. Thanks for your thoughts.