Propeller Simulation - Mesh control for accurate and fast results

[HHOA1307]

Hello all,

I'm trying to work out a very simple propeller simulation and I would like to get your thoughts on the scenario for verification.

First, I applied uniform mesh on the leading edges and spread the changes. The mesh resulted in ~1,3M total elements. (I couldn't apply uniform mesh to the Rotating Region because the number of elements exceeded 200 million!!!)

** FINITE ELEMENT SUMMARY FOLLOWS...
302794 Total Nodes,   277803 Fluid Nodes ,   24991 Solid Nodes
1359427 Total Elements,  1124175 Fluid Elements ,   235252 Solid Elements
0 Inlets   0 Outlets   0 Unknowns

The mesh looked pretty fine and it took my PC ~15 hours to solve 1000 iterations. The convergance plot was very clean too.
Result was 4,9m/s velocity on Y axis; which eventually generates 109gr thrust based on my calculations. (I was expecting a bit more thrust but I think the result could be somewhat feasible)

On the second scenario, I did NOT apply uniform. The mesh resulted in ~200K total elements.

** FINITE ELEMENT SUMMARY FOLLOWS...
46549 Total Nodes,   46114 Fluid Nodes ,   435 Solid Nodes
207601 Total Elements,  196433 Fluid Elements ,   11168 Solid Elements
0 Inlets   0 Outlets   0 Unknowns

Obviously, the mesh looked very bad but it took my PC ~3 hours to solve 1000 iterations. The convergance plot was full of weird fluctuations.
Result was 9,2m/s velocity on Y axis; which eventually means 381gr thrust based on my calculations. (I would think this could be feasible too )

So, here are my questions:

1. Am I on the right track at all or am I doing something totally wrong?
2. How would you apply mesh to this propeller so that the result is accurate and the solver runs in reasonable time too?
3. Is there a way to monitor the "thrust" parameter in SimCFD or do I have to calculate it externally?
4. If you could run the simulation on your side, could you please let me know how many Total Elements did your mesh created and what is the Velocity result on Y axis (Vy)
   I've uploaded the file to my 360. It's the single file in the Folder. I don't know how I can pass you the link -sorry 🙂

Many thanks in advance for your support!

Best,

Kaan

[Jon.Wilde]

Hi Kaan,

The mesh estimation is not always 100% accurate, it might be worth trying it.

A 1.3m element model for a propeller sounds pretty low. Did you refine all the leading and trailing edges also? Have you seen the guide here? I think this would be the best starting point.

Kind regards,

Jon

[HHOA1307]

Hi Jon,

Thank you very much for the quick reply!

I left the trailing edges too because of the size of the model. Anyway, if millions of elements are normal for a propeller analysis then I should give another try. 

The scenario is aligned with the guide and the hangout session -except the 2 last points:

• Run with intelligent solution control Off.  This is the software default for transient anslysis.
• The newer SST model works very well for turbomachinery.  It is suggested to consider using this model instead of K-e

On top of the previous scenario, I will apply uniform mesh to the leading edges, as well as to the RR. I will turn off the intelligent solution control, and use the newer SST model.

I will let you know of the results. Though this could take my poor laptop weeks to solve 

Best regards,

Kaan

[HHOA1307 · Accepted solution]

Ok, I applied the steps I mentioned above and : out of memmory after 5 hours in "Volume meshing - creating volume mesh".

I conclude simulation is too much resource consuming in my case. I better print the propellers in 3D and test them in real...

Best,

Kaan