Learning Forum

[boonthiam13]

Hi guys, I am keen to do the above simulation, any idea which of the solvers should I use, Fluent or CFX, appreciate if you could share your thoughts and reasons. My ultimate objective is to evaluate the performance of the fan if certain element of the blade design is changed, for both transient and steady-state analyses.nAlso, would be grateful if you could share relevant links that provide the tutorials I have done some research, there are more tutorials on using Fluent rather than on CFX.nIn regards to fan simulation using Fluent, what I understand is that a disk (air domain) is created to closely enclose the fan and the disk is surrounded by bigger air domain which resembling the ambient. Fan (solid) medium basically doesn't exist in the model but only its outer surface is captured and treated as wall. Then moving mesh (associated with rotation) is applied to the disk. I have questions as follow in regards to this method:nwhen rotational moving mesh is applied to the disk, is that only the boundaries [outer and inner (blade shape) surface] are moving, or the entire domain is moving? nIs that possible to include directly the fan (solid domain) in the model, without the enclosing disk, and then apply rotational moving mesh to the fan?nThank you in advance for your kind assistance nRgds,nBTnn

[DrAmine]

Both are fine to do this fan simulation task,nn1/the fan zone will be the rotating zone including anything in it like the resolved or the non resolved blade walls and shaft. You can however tell Fluent or CFX to not rotate some partsn2/That is the called immersed solid approach which Ansys CFX can deal with it. However this is not the most accurate way. Rely on Sliding Mesh approach to tackle this task.n

[boonthiam13]

Both are fine to do this fan simulation task,1/the "fan" zone will be the rotating zone including anything in it like the resolved or the non resolved blade walls and shaft. You can however tell Fluent or CFX to not rotate some parts2/That is the called immersed solid approach which Ansys CFX can deal with it. However this is not the most accurate way. Rely on Sliding Mesh approach to tackle this task.https://forum.ansys.com/discussion/comment/106893#Comment_106893

Dear DrAmine, appreciate your feedback. As for question 1, I am assuming 'fan' zone you mentioned is referring to air domain contained inside the disk. Meaning when we apply rotational moving mesh to 'fan' zone, air particles within the 'fan' zone rotates by themselves rather than being pushed due to the rotational motion of the fan's wall (inner surface of the disk)?n

[boonthiam13]

Both are fine to do this fan simulation task,1/the "fan" zone will be the rotating zone including anything in it like the resolved or the non resolved blade walls and shaft. You can however tell Fluent or CFX to not rotate some parts2/That is the called immersed solid approach which Ansys CFX can deal with it. However this is not the most accurate way. Rely on Sliding Mesh approach to tackle this task.https://forum.ansys.com/discussion/comment/106893#Comment_106893

Also, DrAmine, what did you mean by resolved and non-resolved blade walls and shaft?n

[DrAmine]

Resolved: walls are not zero thickness wall=> blades are as solid body.n

[DrAmine]

Yes: in the rotating zone fluid will be driven by the zone motion: in this case rotationnImmersed solid like approach in Fluent will be to rely on deforming meshes to rotate the blade geometry: I do not recommend that at all. We generally assume a volume space which is rotating and other outer domain which is kept stationary.n

[boonthiam13]

Yes: in the rotating zone fluid will be driven by the zone motion: in this case rotationImmersed solid like approach in Fluent will be to rely on deforming meshes to rotate the blade geometry: I do not recommend that at all. We generally assume a volume space which is rotating and other outer domain which is kept stationary.https://forum.ansys.com/discussion/comment/106924#Comment_106924

I agree that in fan simulation we should rotate a volume space (which is a fan in my case) while the outer domain (air medium) should be kept stationary. If you have time, could you please explain in detail how to do this, probably a video could help to explain better. nThanks and looking forward to hearing from you nn

[Karthik R]

Hello,nThere are several videos on YouTube that explain how to set-up this problem in Fluent. Please search for the following keywords in YouTube - Fluent sliding mesh fan. The first few search results should give you a good starting point.nKarthikn

[DrAmine]

Check YouTube Ansys Channels and some other videos of other Fluent's Users!n

[boonthiam13]

Hello,There are several videos on YouTube that explain how to set-up this problem in Fluent. Please search for the following keywords in YouTube - "Fluent sliding mesh fan". The first few search results should give you a good starting point.Karthikhttps://forum.ansys.com/discussion/comment/107062#Comment_107062

Thanks Kremella and DrAmine. When we apply a mesh motion to an air domain, literally we are driving the air particles in the domain to move in a defined direction. It is very common to see those available tutorial videos when people doing axial fan simulations, air domain is divided into two zones. The outer zone is stationary while for the inner zone, mesh motion (rotary motion) is applied. The inner zone of the air domain includes the blade wall. nBy default, the blade wall is set to stationary (no relative motion between blade wall and the inner zone of air domain). With these settings, since the blade wall is moving together with the air domain in the inner zone, there should be no shear (between the blade wall and the air domain in the inner zone) and consequently, no thrust force is generated?.Looking forward to hearing from you nRgds,nBTn

[Surya Deb]

Hello, nIf I understand your question correctly, then due to the no slip condition on the blade walls and due to the zero relative motion [between the blade walls and the rotating zone], the fluid surrounding the blades should have the same velocity as the blades.nThe rest is taken care of by general scalar transport type equation but with slight modification for the mesh rotation. nYou can find more information here for conservation equations for dynamic mesh. For sliding mesh, this equation will get a bit simplified as the cell volumes don't change with time.nRegards,nSD