Ceiling Fan Simulation

[Anonymous]

I am using Autodesk Simulation CFD to simulate the air delivery of a 3 blades ceiling fan.

I need to measure air velocity 1.5m below the fan.

However, the air velocity measured is only half of the testing data.

Attached are the setting for rotating region and solving:

I did use uniform mesh at rotating region.

There is no inlet or outlet assigned, closed room is assumed.

Results:

The geometry for this simulation had been run on other software which produce results that were quite close to the teasting data.

Hopes someone can give me advice on the settings and how i can get a more accurate results in Autodesk Simulation CFD.

Thank you.

Best Regards,

Mathew

[Jon.Wilde]

Hi,

I have a couple of comments for you that I think will help.

1. Check out this webinar on rotating regions
2. Your mesh looks pretty coarse - I think that after watching that video you will choose to refine it anyway

Hope that helps,

Jon

[Anonymous]

Hi Jon,

Thank you for the reply and link.

May i know when to use either "degrees per time step" or "number of blades"?

Will "degrees per time step" gives better accuracy?

Regards,

Mathew

[Jon.Wilde · Accepted solution]

Did you watch the webinar? Or see this article? I think it should give you a comprehensive answer 🙂

[Anonymous]

Dear Jon,

Good news!

I managed to get results close to testing data.

I'm using workstation with Xeon 2.0Ghz, 6 cores processor and 64GB Ram for this simulation.

My model has 1.4 million mesh, run until 5000 iterations and the simulation took 44 hrs to complete. Is this normal?

Thanks for your help.

Best Regards,

Mathew

[Jon.Wilde]

Hi Mathew,

Fantastic! That is great to hear.

It is possible it could take so long, CFD would be using 4 cores (we used 2^n cores) and at 2GHz I would expect longer run times than I see myself. Saying that, 1.4m elements is a pretty small model. 5000 iterations is quite a lot though, this could take time. 

Could you run for less and still get a good result?

Thanks,

Jon

P.S - I just ran a motion model to 500,000 iterations 😄

[Anonymous]

Dear Jon,

I think the simulation can stop at 4000 iterations.

I set to 5000 because i'm not sure when the values will start to converge.

BTW, 500k iterations is unimaginable!

Best Regards,

Mathew

[Jon.Wilde]

Yes, this looks very repetitive from 3000-4000 and can likely be stopped. Hopefully the results then are the same and that saves you a bit of time?

Yup, half a million iterations is crazy and it was only run one 😄

[me20b0167ZE8Q]

Hi, can I know what were the initial and boundary conditions u used, as I am trying to do the same but I am stuck?

Thanks