CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CFX (https://www.cfd-online.com/Forums/cfx/)
-   -   fan characteristics (https://www.cfd-online.com/Forums/cfx/132144-fan-characteristics.html)

J.Kassai March 26, 2014 12:13

fan characteristics
 
Dear Colleagues,
I wish to calculate with CFX the characteristic curve P(Q) of an axial fan. I prepared a model of a virtual test arrangement: stationary inlet tube with a length of 3D, where D is the fan outer diameter, rotating ducted fan and stationary outlet tube with 3D length; three domains in series connected by frozen rotor interfaces. I prescribed air mass flows as inlet BC having opening with arbitrary pressure as outlet BC. I can calculate by CFX-Post areAve pressures at different cross sections of the arrangement. I arrived at basically two questions:
1.: What is the rigth position of the interfaces between rotating fan, and stationary inlet and outlet tube domains? How close should they be to the blades?
2.: Calculating the average pressures (static or total) in the different cross sections along the axis of rotation there is a pressure distribution which definitely shows the effect of the rotating fan. But where exactly should I read out these pressure values? Typically the minimum of these is at the inlet tube/fan interface and the maximum is at a certain distance from the fan/outlet tube interface. I wish to have a realistic pressure difference so as to compare with real measurements.
Thanks in advance,

singer1812 March 26, 2014 15:29

If you have a rotating fan blade, why are you prescribing a mass flow at inlet? The rotating blade should determine that via the blade design and RPM.

Total pressure should be used to characterize a fan (I know, history uses static, but you can read all the issues and problems with that, and total pressure is the correct way to do it).

In your case, if you are trying to characterize the fan itself, and not a particular place and ducting the fan is sitting in, just model the inlets and exits far from the fan to remove all effects of fan on the boundaries (for example non developed flow at fan exit).

If rather you are modeling the fan, and the measurements you are comparing to is in an actual system, then model the system. Take the pressures at the same points as in the measurement.

J.Kassai March 27, 2014 06:51

Generally, a fan characteristic curve is taken at fixed rotation speed by changing mass flow and reading out pressure difference -- if I understood well (I am not an expert of fans, unfortunately.)
My questions regard two main points. The first one is a kind of modelling question: where /how close to the rotating blades/ should I place the interfaces of rotating/stationary domains using frozen rotor method? Pressure results depend on the position of these interfaces. The second one is kind of "theoretical" question: when reading out average pressure values for the fan curve P(Q), where should I calculate these average values? I am asking this because average pressure values are a function of distance along the modelled test system according to my experience. Of course it would be the best to model a complete test arrangement with actually rotating fan (transient sliding mesh etc.) but it is not always practical. I would like to "measure" a fan characteristic by CFD possibly in the most simple way.
This latter problem proceeds toward more general question: is there a fan characteristic curve independently of the test arrangement, or does it always depend on the actual test configuration?

chaudhry_hashim July 3, 2014 10:39

I am also trying to calculate the fan performance curve. If your problem is resolved then kindly help me out how should be the length of inlet and outlet duct. I am using velocity inlet as a inlet and pressure outlet as a outlet and giving fixed rpm to the rotating region which is acting like a fan. I would be very thankful to you if you help me out in the measurement of pressure values

J.Kassai July 4, 2014 07:08

Your response is welcome; I started to think that this was such a trivial problem that almost nobody was interested in sending any answer...
Well, here comes some conclusions I arrived at (Caution! These may be wrong or not; I am not a fan expert! Conclusions are based on my investigation of technical literature and own CFD simulations.):
1./ Fan characteristic P(Q) curve does not exist in itself! It depends on how a fan is installed.
What does this mean as for CFD?
Hence it follows that your simulation model should be the same as 'reality'. 'Reality' may be a standard arrangement (see e.g. in ANSI/AMCA 210-99 "Laboratory Methods of Testing Fans...") or it may be an actual arrangement of any kind of test/installation. However, this finding is in accordance with the last statement of the previous response of singer1812.
There comes some additional problems. For example: what is the effect of a flow straightener generally applied in the arrangements of ANSI/AMCA 210-99? Should it be involved in a CFD model, too? It is still not clear for myself, too; I need further checking.
2./ I did not find any unambiguous prescription about how close to the rotating blades I should place the interfaces of rotating/stationary domains using frozen rotor method. It seems to be an area of scientific investigation and it is specific to an actual problem. In case of an axial fan modeling, now I put these interfaces as close to the blades as possible with keeping meshing still feasible.
Going back to your question:
if you have an actual test arrangement, you should use that geometry accordingly; if you have "nothing", choose an arrangement from ANSI/AMCA 210-99 and use that. (This standard is, however, very useful to read through to have a picture about fan measurements.) I think, this is a reasonable approach if you do not have lots of experience in fan modelling.

chaudhry_hashim July 4, 2014 08:39

hi
 
thanks for replying I am totally new in this field. I am using AMCA standard for testing the fan experimentally in order to get fan curve and I hope it works. But I am worrying about the CFD side. AS far as I got to know from the literature for numerical simulation we have to make just upstream and downstream extruded region across the fan but i am worrying about the boundary condition and measurements for fan curve. Right now I am using velocity inlet, fixed rotating speed to the rotating region and pressure outlet. But I couldn't succeeded yet I am confused where should I take measurement for pressure and how far from the fan

J.Kassai July 4, 2014 11:52

In ANSI/AMCA 210-99 you will find the places for measurement of pressure. If you choose to follow the ANSI standard, you can use the same positions in your CFD model. This way your CFD calculation and the results from a standard test will be comparable.


All times are GMT -4. The time now is 16:19.