I am trying to create a numerical simulation of fluid flow through an axial cooling fan. I am aware that there are three models available for the representation of the moving blades, these beign 1.The Body Force Method, 2.Multiple Reference Frame (MRF)and 3.Sliding/Transient Method.

I will be running my model in FLUENT5 and have found literature with regards to the MRF and Sliding Mesh methods. However I am having difficuty in understanding how to set the simpler Body Force Method. Could anyone please instruct me?. Any suggestions would be greatly appreciated.

Thanks in advance,

Jonathan.

Jonathan,

I think that, in your question, you are confusing different techniques which have different purposes.

Basically, MRF and Sliding Mesh methods are used when the computational domain includes stationary AND rotating components simultaneously (e.g. rotor + stator, wheel + volute, etc). As far as I know the naming conventions of Fluent and other commercial CFD codes, MRF is used to calculate a steady or quasi-steady solution, whereas the SM method is used to simulate an unsteady solution, which is obviously much more CPU-consuming.

For your application, if you are mainly interested in the performances of the cooling fan, you need to simulate only 1 rotating component (the fan itself). Therefore, you do not need to make use of those two techniques (MRF,SM). Instead, a 'simple' steady simulation of 1 blade passage of the cooling fan (either in the relative or in the absolute frame of reference) should provide you with a very good solution. If you need to deal with unsteady phenomena (off-design conditions), an unsteady approach such as the phase-lagged technique would be fine.

In conclusion, do not care too much about the name of the technique that should be used. You should first wonder about the results you are looking for, and the most appropriate technique should be easy to identify (if the CFD vendor's manual is clear).

Now, let me give you my personal opinion : I have been working in a company designing fans for years, and we use to use FINE/turbo from NUMECA as a CFD code for all our flow simulations. It would have taken me less time to make the simulation of your fan with F/T than to answer your question ! No other CFD code that we tested was so quick and accurate.

This being said, Good luck.

I would be interested to hear aobut the references you have for the MRF and the sliding-mesh technique. I have not been able to find many references regarding these. As to your question I'm afraid I have no idea! Sorry.

Can I ask you one thing?

As I heard from someone, when I use a MRF, I have to give a rotational velocity to fluid zone. At that time, I didn't know exactly about the concept of MRF & SLIDING MESH. So, I gave a rotation velocity to solid zone(impeller part) but there was an rotational effect. I could see the rotational velocity component when I cut the cross-section of my system.(3D problem)

What I want to know is..."In case of MRF appication(when steady-state), do I have to give a rotational velocity to fluid zone or solid zone?"

Thank you.

Well, I am afraid that I'd need to know more about Fluent's conventions before I could answer this question. As I said, I use another commercial CFD code (FINE/Turbo), which requires to give a rotational speed to both the fluid zone and the solid zone. It makes sense to me : if you are dealing with an unshrouded fan, the shroud wall is not rotating and that needs to be told to the system (i.e. Solid zone information is needed) ; if you are dealing with multiple frames of references, you need to define the rotating speed of each frame of reference, i.e. Fluid zone information is needed.

In conclusion, I think that the answer to your question is : to BOTH

Alex

Thanks Alex.