Learning Forum

[George Joseph Thomas]

I have a situation where I know only the pressure boundary conditions at the inlet and outlet (design of a nozzle). I have been hearing from my day one with CFD not to apply pressure BCs on all boundaries. Just to make sure, I calculated the flow through a simple tube with pressure inlet and outlet conditions and compared the results with the experimental results (obtaines using LDA) and the difference was around 7 % which for my application is a top-class result. So, is it okey to apply pressure BCs at all boundaries in Fluent or is it not recommended?

[Rob]

With all pressure boundaries the mass flow becomes part of the solution. Given the density is also varying we get a fairly nasty problem in a numerical sense. Ie if density alters because of local pressure, the flow velocity changes as does the mass flow: which variable is the one to mess with to get the defined pressure drop. 

What you plan is OK, but be aware of the potential for a stiff solution. Nozzles tend to be OK, but adding in shocks can complicate things. If you have LDA so know the velocity, can you use a velocity inlet then compare the pressure drops? 

[George Joseph Thomas]

Yeah, thank you for the insight, but I do not have the a physical nozzle yet - is still in the development phase. When I try pressure boundaries, I see a covergence problem with continuity. I suppose this is what you meant. Let me know if you have any other suggestions. Meanwhile, I try velocity BC for inlet and check the pressure drop between inlet and outlet - if that corresponds to my required pressure difference. That should work too, right?

Thanks again!

[Rob]

I'd need to see images to comment on the residuals, both of the flow field (contours) and residual plots. Given how stiff these solutions can be, you may need a good initial condition (not hybrid) to get the solution going. 

For fixed density systems it's common to run a velocity inlet, pressure outlet and then compare the pressure drop & mass flow with experimental results. The typical graph is a full experimental curve of pressure v flow with selected "points" from the CFD results. I did that when publishing some work on inhaler studies. 

[George Joseph Thomas]

I am in the middle of some calculations. I will get back to you with the pictures in a while. I am tring a couple of different initial conditions and BCs, and understanding the influence on the flow. Intial conditions seems to affect my convergence drastically.

Thanks for your inputs.