CFX vs. Fluent – significantly different results with the same settings, bc
 

Dear all,
I have simulated a ducted fan config (see figure) with CFX AND Fluent with the same boundary conditions - pt Inlet and p Outlet and the same RPM. I have used the default turbulence settings etc. in both cases, hence they differ slighlty to each other. However, when I am calculating the overall force (thrust) in z direction I get significantly different results for both solvers. I am aware of the importance in setting the reference pressure in Fluent correctly, since it is sensitive to the results (see second equation). I dont get the advantage of relating the pressure to a reference pressure (why is it reducing round off errors). As far as I know, CFX is estimating the force by the accumulation of pressure and viscous force on each surface but I couldnt find any mathematical definition in user guides. In Fluent I get -212N and in CFX -68N which means the force in cfx is only about a third as high as the one calculated in fluent. Even when I include the reference pressure in CFX as done in fluent, the result is completely different (changing from -68N to +266.25N). I dont think the difference can be explained by the differences in turbulence settings (while both are using SST) etc. Do you know where the difference comes from??? I am really grateful for every help you can provide!!!
https://forum.ansys.com/wp-content/u...9-mceclip0.pnghttps://forum.ansys.com/wp-content/u...0-mceclip1.pnghttps://forum.ansys.com/wp-content/u...8-mceclip2.png

It can depend on lot of things.
I have performed many calculations using Fluent and CFX. What I noticed is that using transient simulations helped a lot. That brought both together.
Regarding the pressures, both CFX and fluent have a reference pressure. The density depends on this pressure, as well as on the inlet and outlet pressure. So everything should be fine here.
I prefer CFX for rotating equipment. It is meant for that. Fluent is much less convenient for transient calculations. Regarding the forces, you can look them up in the output file. They are given by default in the summary at the end of the run. No need to determine them in Post.

Thanks a lot for your response!!!


It doesn't seem to be intuitive, cause when I turn on "include reference pressure to true" in CFX the result is completely different. When I switch it off it is similar to the fluent result. Although the reference pressure inclusion is always on in fluent. That seems contradictory to me...I would favor CFX in my case but it is not possible to use the ffowcs williams model, which is based on the transient results of fluent. So it is a data format issue, I guess.