[ANSYS Meshing] Ducted propeller MRF issue

cagriaydin · March 3, 2017, 07:49

So, basically, i'm working on a project that involves the usage of ducted propellers and since the tip clearance of the propeller to the duct surface is extremely small, problems occur between the interface of the rotating domain and the bottom surface of the duct.

My question is, if i include the duct into the rotating domain instead of just trying to capture the hub and blades, would the results be reliable?

Cheers,

Cagri.

Krao · November 7, 2019, 05:48

Hi Cagri,

I am trying to simulate something similar, and I am also having the same idea of including the Duct inside the rotating domain. Can you please share your experience, it would be very much valuable. I am actually using OpenFOAM to simulate the propeller with duct. But the principles should be the same.

Thanks,

K. Rao

arnie333 · November 12, 2019, 02:49

Cagri,

I don't know if it's possible to have interfaces that don't line-up on the same planar face, but I am interested to find out whether it worked or your final solution.

I had a similar problem 2 weeks ago on a pump analysis where the impeller outer diameter is almost touching the pump casing. There was not enough space to have a wall/interface between the two edges i.e. rotating domain (impeller) and the static domain (volute). The elements here were extremely small or skewed.
So, what I did was to have the rotating frame outer diameter just touch/penetrate the tip of the impeller i.e. the impeller edge 'broke through' just the domain. I think this is acceptable because I came across web images showing similar concepts.

Curious to hear how you resolved it though.

Thanks.

Krao · January 16, 2020, 04:49

Hi Arnie,

Can you please explain, how did you resolve the above issuey? It would be interesting to know, as I am stuck with this problem from long time.

Thanky ou

arnie333 · January 18, 2020, 08:22

Sometimes a picture is worth a thousand words.

I have attached a screenshot of a similar concept I found on the internet.

Note how the tip and flanks of the pump vanes break the surface of the rotating domain !

Krao · January 20, 2020, 02:23

Hi Arnie,

Thank you, for the explanation. I will see how I can implement it in OpenFOAM, for a MRF simulation involving AMI interface.

Krao · January 21, 2020, 09:53

Hi Arnie,

I have got the results, thank you. But the results are 7% lower than what I have expected to be. To study this I have varied the dimensions of MRF domain, and found out if the size of MRF domainis less than 1.3D x 0.3D the simulation results keep on decreasing. Is it same in ANSYS too?

I am curious to know, thank you.

arnie333 · January 21, 2020, 12:59

What CFD software are you using ?

I'm not sure I understand the MRF dimensions (in reference to impeller diameter) you are talking about ... can you post some screenshots ?

PS. A difference of 7% to experiments does not seem like a big amount. Rarely does one get such good results on first attempt with all the variables/assumptions with boundary conditions, operating conditions, equipment variability, etc.

Krao · January 28, 2020, 10:49

Hi Arnie,

Thank you very much for your kind reply, and I have missed my notification and sorry for the late reply. I am using OpenFOAM. The diameter of MRF region I mean. For example I have a propeller of 450 mm. If I chose the rotor region(MRF) to be around 460 mm, the simulation takes too long to converge and the results are around 20 % away from my experimental results. But if I chose any diameter of rotor region greater than 580 mm. The simulation converge fast, and the results will not change, if I increase my diameter too.

Is it same in ANSYS also? Are the results are so much dependent on MRF region size?

Waiting for your comments eagerly.

Thank you.

March 3, 2017, 07:49	Ducted propeller MRF issue	#1
cagriaydin New Member Çağrı Join Date: Feb 2015 Posts: 11 Rep Power: 11	So, basically, i'm working on a project that involves the usage of ducted propellers and since the tip clearance of the propeller to the duct surface is extremely small, problems occur between the interface of the rotating domain and the bottom surface of the duct. My question is, if i include the duct into the rotating domain instead of just trying to capture the hub and blades, would the results be reliable? Cheers, Cagri.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[ICEM] ICEM propeller issue	metmet	ANSYS Meshing & Geometry	1	July 1, 2016 21:52
meshing ducted propeller	amateurCFD	ANSYS	1	April 1, 2016 14:06
Possibly serious MRF implementation issue	Ali Blues	OpenFOAM Bugs	1	December 16, 2015 07:04
simulation of water flow though a ducted propeller	spacewatcer	FLUENT	0	April 22, 2009 10:52
Ducted propeller, Boundary conditions	Johannes	FLUENT	0	May 7, 2008 10:21

November 7, 2019, 05:48		#2
Krao Senior Member Kmeti Rao Join Date: May 2019 Posts: 145 Rep Power: 8	Hi Cagri, I am trying to simulate something similar, and I am also having the same idea of including the Duct inside the rotating domain. Can you please share your experience, it would be very much valuable. I am actually using OpenFOAM to simulate the propeller with duct. But the principles should be the same. Thanks, K. Rao

November 12, 2019, 02:49		#3
arnie333 New Member Arnie Join Date: Mar 2017 Posts: 27 Rep Power: 9	Cagri, I don't know if it's possible to have interfaces that don't line-up on the same planar face, but I am interested to find out whether it worked or your final solution. I had a similar problem 2 weeks ago on a pump analysis where the impeller outer diameter is almost touching the pump casing. There was not enough space to have a wall/interface between the two edges i.e. rotating domain (impeller) and the static domain (volute). The elements here were extremely small or skewed. So, what I did was to have the rotating frame outer diameter just touch/penetrate the tip of the impeller i.e. the impeller edge 'broke through' just the domain. I think this is acceptable because I came across web images showing similar concepts. Curious to hear how you resolved it though. Thanks.

January 16, 2020, 04:49		#4
Krao Senior Member Kmeti Rao Join Date: May 2019 Posts: 145 Rep Power: 8	Hi Arnie, Can you please explain, how did you resolve the above issuey? It would be interesting to know, as I am stuck with this problem from long time. Thanky ou

January 20, 2020, 02:23		#6
Krao Senior Member Kmeti Rao Join Date: May 2019 Posts: 145 Rep Power: 8	Hi Arnie, Thank you, for the explanation. I will see how I can implement it in OpenFOAM, for a MRF simulation involving AMI interface.

January 21, 2020, 09:53		#7
Krao Senior Member Kmeti Rao Join Date: May 2019 Posts: 145 Rep Power: 8	Hi Arnie, I have got the results, thank you. But the results are 7% lower than what I have expected to be. To study this I have varied the dimensions of MRF domain, and found out if the size of MRF domainis less than 1.3D x 0.3D the simulation results keep on decreasing. Is it same in ANSYS too? I am curious to know, thank you.

January 21, 2020, 12:59		#8
arnie333 New Member Arnie Join Date: Mar 2017 Posts: 27 Rep Power: 9	What CFD software are you using ? I'm not sure I understand the MRF dimensions (in reference to impeller diameter) you are talking about ... can you post some screenshots ? PS. A difference of 7% to experiments does not seem like a big amount. Rarely does one get such good results on first attempt with all the variables/assumptions with boundary conditions, operating conditions, equipment variability, etc.

January 28, 2020, 10:49		#9
Krao Senior Member Kmeti Rao Join Date: May 2019 Posts: 145 Rep Power: 8	Hi Arnie, Thank you very much for your kind reply, and I have missed my notification and sorry for the late reply. I am using OpenFOAM. The diameter of MRF region I mean. For example I have a propeller of 450 mm. If I chose the rotor region(MRF) to be around 460 mm, the simulation takes too long to converge and the results are around 20 % away from my experimental results. But if I chose any diameter of rotor region greater than 580 mm. The simulation converge fast, and the results will not change, if I increase my diameter too. Is it same in ANSYS also? Are the results are so much dependent on MRF region size? Waiting for your comments eagerly. Thank you.