Simulation of an Actuator Disk

dusan · January 1, 2019, 8:08pm

I’m trying to simulate a propeller disk, as in the momentum theory.

I’m a bit unclear about how to set the boundary layer of my disk to represent the energy imparted from the propeller to the air, as it will break the conservation of laws as represented by NS equation.

project link: https://www.simscale.com/projects/dusan/propeller_disk_simulation/

Also my geometry surface splitting operation failed for some reason.

dusan · December 26, 2018, 2:41am

I was unable to split the geometry of the stl cylinder created in blender.
I solved this by creating a cylinder in freecad and exporting as step.

jousefm · December 26, 2018, 5:55am

Hi @dusan!

You can have a look at public projects for propellers and you can see that only the propellers themselves are refined with inflated boundary layer cells and not the cylinder which is only used for the rotating zone (MRF). If you cannot find the example let me know and I will help you out!

Best,

Jousef

dusan · December 26, 2018, 5:25pm

Hi Jousef,

I don’t think you got the gist of what I’m trying to get. There is no actual blade or propeller. In momentum theory, the propeller is modelled as a simple disk imparting energy as a pressure increase on the flow. As this energy is externally applied, it breaks the conservation of momentum of air crossing the disk.

Regards,
Dusan

jousefm · December 26, 2018, 5:46pm

Hi Dusan,

looked up the actuator disk case and now know what it is about, quite interesting! Will have to skim the documents and papers I can find and see if there needs to be a special BC implemented for this case and if not I can help you to validate the results from Mikkelsen (2003) - guess that’s what you are trying to do, right?

Best,

Jousef

dusan · December 26, 2018, 6:09pm

Hi Jousef,

Glad you are looking into this, hope you will find a solution.

I found an interesting document specifically addressing this problem in openFoam: http://www.tfd.chalmers.se/~hani/kurser/OS_CFD_2010/erikSvenning/erikSvenningReport.pdf

Maybe you’ll have a look and decide how to proceed best.

If the mentioned approach will not work, I propose to do this in several steps:

As a very rough first estimate, a model without swirl and the force across the disk constant
Implementation of variable force dependant on radial position
Implementation of swirl

Thank you for your support.

Regards,
Dusan

dusan · December 30, 2018, 6:00pm

Any update on this?

jousefm · December 30, 2018, 8:17pm

Hi Dusan,

have not given it a try yet, did you? Can you tell me the measurements of your actuator disk? I will then start a first run and we can see if my simulation fits your expectations.

Best,

Jousef

dusan · December 31, 2018, 2:12pm

Hi Jousef,

Thank you for helping me. It is much appreciated.

I have the following experimental data:
propeller diameter: 1.8m, constant rotational speed: 2450rpm

airspeed(m/s) thrust(N)
11.6 m/s 2625N
5.14 m/s 3314N
1.78 m/s 3536N
0 m/s 3559N

If you need any more details, just let me know.
Happy New Year!

Regards,
Dusan

jousefm · December 31, 2018, 4:51pm

Hi Dusan,

do you also have data for the surrounding air speed because I think the airspeed you stated are for the actuator disk, right?

Best,

Jousef

jousefm · January 1, 2019, 12:19am

Hi Dusan,

here is the first sim which is still running though. Also initialized with small values. Can you then try adapting it for your purposes? Project: Actuator Disk - Simulation

Cheers,

Jousef

dusan · January 1, 2019, 7:35pm

The speeds are for the unperturbed incoming air.

jousefm · January 1, 2019, 7:35pm

Here a simulation with 1 \frac{m}{s} freestream velocity and 5 \frac{m}{s} from the fan. You can play around with it and see if you can validate the results from the paper. Also I have just taken an estimate depth (y-coordinate) for the fan and have not adapted every numerical setting like in the document - actually one time I had an error and simply set back the settings for the velocity.

Would be happy to see some validation results from you

Cheers,

Jousef

dusan · January 1, 2019, 7:37pm

Hi Jousef,

This is great! Thank you for doing this! This made my day(year!)

Regards,
Dusan

dusan · January 1, 2019, 7:46pm

Hi Jousef,

I have trouble understanding what you did. As far as I understand the Part 1 is the domain, but how did you add the Cylinder 1? Also I cannot find where are the settings for the cylinder, the boundary conditions are for the domain only.

Thanks,
Dusan

jousefm · January 1, 2019, 8:05pm

Hi Dusan.

First about the cylinder settings and the “missing” boundary conditions. Momentum sources can be used to simulate fans, ventilators and other similar devices. Whilst normally, the user would have to provide a detailed geometry and a fine mesh of these components, this feature allows one to define a volume which generates momentum, without any further complication.

\underline{\textbf{1. Create the box with OnShape}}

I ignored the recommendations from the paper in the beginning and just created a bigger domain by intuition. Feel free to follow the paper and see if the size has an influence.

\underline{\textbf{2. Upload the geometry via Onshape plugin on SimScale}}

After the message “successfully imported” appears you can close the tab.

Import%20from%20Onshape

\underline{\textbf{3. Create the cylinder}}

Adding%20a%20geometry%20primitive

\underline{\textbf{4. Add the Momentum Source}}

All the best,

Jousef