[euclid]

Hello all,
finally I managed to create a scenario analysis of blade propeller.
After getting results I return back and change dimensions of fluid domain (in my case is a block).
Running again the scenario I getting different results??
How domain affects my case?
My propeller in real works into a tunnel which is included into domain.
pls check attached image.
Should I avoid the domain? Which should be the right dimensions of domain?
Thank you

[ghorrocks]

What you have done is something which should be checked for any CFD simulation - the sensitivity of the results to the proximity to the boundary conditions. When moving the boundaries affects your results it shows your boundaries are close and are effecting the results. If you are trying to model the object in a large space (far field) then your model will be wrong.

If you are trying to model a far field situation then you have to make the domain larger and larger until the variation is smaller than a tolerance you can accept.

If you are trying to model the device in a wind tunnel then I would model the actual wind tunnel structure. Make sure that the wind tunnel people do not do "corrections" for wind tunnel blockage factor. If you model the wind tunnel directly then your results should be comparable to the wind tunnel with no blockage factor corrections.

[euclid]

Thanks for the reply ghorrocks,
My propeller works into a tunnel under sea water.
The goal is to calculate thrust.
My domain should be the cylindrical tunnel to achieve accurate results or an outer bigger block (which includes cylindrical tunnel)?
What I did until now……
-Import the geometry of the propeller, and defined the rotational domain and the stationary domain (the block showed previously). The propeller was then suppressed in order to model only the fluid domains.
-The mesh was left as default at a first stage. Later on I will make a more detailed mesh for more accurate results.
-Set up: I used transient mode analysis ( from other posts I believe I have to make a steady state analysis)
Boundary -> Inlet with initial velocity U=V=W= 0m/s (as I want to calculate the static thrust)
All the other parts of the block were defined as openings.
The propeller surfaces were defined as a rotating non slip wall.
I defined fluid-fluid interfaces between the stationary default domain and rotating propeller domain.
The frame change option is set on both fluid-fluid interfaces to frozen rotor.
Attached is an image of real part.

[ghorrocks]

The most accurate way is to simulation everything. The skill in a CFD practitioner is to know what can be ignored and simplified without unduly compromising accuracy.

So the most accurate method would be to model the rotor in the duct and boat stern. Modelling the rotor in the duct only (without the surrounding fluid) is less accurate as you do not capture the flows around the duct. But is it accurate enough? You will have to do sensitivity analyses to determine that.