But I think that is still physically incorrect, but it might be (much) better, it is as if it is downstream as if you had infinite many of them stacked and you look at the last one

3 hours 'a long time', damn life is good

I mean, a single blade/channel and all stages, like https://cdn.discordapp.com/attachments/726377664213942323/1445117325023121530/image.png?ex=692f2de3&is=692ddc63&hm=46f252b1aaff27bdebcff277c4c4925f2f19d9bb8468abb8ced982ed8e040850&

Keep discussing. I'll eagerly await the result 🤣. When two experts disagree, there is always something to learn 😅

Ow I do not disagree

it's a common practice, not my idea 🙂

This is the correct approach

Industry standard approach

or well "correct" I should state that its a simplification which gives accurate results while keeping simulation times reasonable

thats always the preferred answer

excactly

So, have I understood correctly? It is sufficient to analyze only one stage for the blade shape?

yes, if you apply the correct boundary conditions

no sorry, you should do one Blade through all the stages

because for only one stage with 1 stator and 1 rotor for accurate results you would have to know not only the inlet, maybe atmospheric conditions, but also the following stages pressure distribution, which you probably dont have

Ah ok. Thank you 👍 👍 👍

Before I continue working on it today, I have one more question. The entire construction is very small. The outer diameter of the blades will be approximately 60 mm. How large should the distance be between: 1. Rotor blades and stator blades 2. Stator blades and rotor blades

(2. means to the next stage)

So I don't need a perfect formula, just a value that you would take based on your experience.

Try half a chord or 1/3 a chord

Thank you very much for all your tips. I will construct a complete prototype and upload it here as a .step file soon 👍

So for an incompressible study. I have a zone where there i some kind of separation. K-omega-SST dont converge on my GCI, so i found an article that uses to k-epsilon to model the same thing as i do ( a bend pipe). and wanted to see if i could get a better GCI study here. My nusselt number and friction_factor are highly unstable (see picture). I haven't used k-epsilon that much before. My y+=0.17. Are there anything obvious i have missed?? https://cdn.discordapp.com/attachments/726377664213942323/1445415606370173120/image.png?ex=693043af&is=692ef22f&hm=3f78ed3040af1552a127096f8843595a8af8126661e6a47b18609a4d702cf1d4&

AFAIK, k-epsilon isn't too happy with low-re meshes?

Yes ive heard that as well. Wolfdynamic notes also mention this. I am just very confused that several studies are using ths approach with good results?

There's a low-re correction somewhere in the settings but I don't have experience with that. Never actually managed to run k-epsilon successfully

the ``lowReCorrection`` in the ``epsilonWallFunction``? https://gitlab.com/openfoam/core/openfoam/-/blob/master/src/TurbulenceModels/turbulenceModels/derivedFvPatchFields/wallFunctions/epsilonWallFunctions/epsilonWallFunction/epsilonWallFunctionFvPatchScalarField.H

Just found this the very moment!

There is also a kLowRe function?

and nut

Yes. But i just realised. I adjsuted all my wall to have different wall functions for my k-omega-SSt case. I used cp -r 0.oirg to replace it in my relevant case (many by now) and ive just learned that it doesnt repalce the file within the directioy you copying if they already exist... So that could also explain my problem with convergence in GCI... - But thanks!