Problems fan simulation
 

Hello all,
I have a strange problem doing a simulation of a simple centrifugal fan. I have done the simulation before in Ansys13 but the problem occurs now when I try to continue in Ansys14.

One of the boundary conditions is the mass flow at the inlet, in this case 0,42kg/s (the other one being rotation speed) , and usually I get this figure back in the result from CFX, correctly expressed as Volume flow, with the rest of the figures, pressure, efficiency, power etc.

But now it does not work!! The flow in the result is wrong (??!) and all the other figures as well of course... I have tried several time with a set-up used before. It seems to occur randomly, three days ago I run two simulations giving correct flow but yesterday I could not repeat them.

Please, some advices, what have I done wrong??

The setup is very simple, the wheel looks like this, it is only wheel, no volute.https://docs.google.com/file/d/0B9kB...it?usp=sharinghttps://docs.google.com/file/d/0B9kB...it?usp=sharing
https://docs.google.com/file/d/0B9kB...it?usp=sharing
I just go from Bladegen to TurboGrid then to CFX. I do not have much experience of CFD so I do it right as described in the tutorials. And earlier I got some results with very good concordance with the real tests. (Especially tests with plenum fans, i.e. fan without volute)
https://docs.google.com/file/d/0B9kB...it?usp=sharing
I have tried to save some of the project files on my Google Drive (bladegen, turbogrid CFX-files, almost 400Mb..).
https://docs.google.com/file/d/0B9kB...it?usp=sharing
https://docs.google.com/file/d/0B9kB...it?usp=sharing
https://docs.google.com/file/d/0B9kB...it?usp=sharing
If anyone has the possibility to take a look and give some feedback I would be very grateful. I do not recommend copying the design shown because it is not very good, I have an idea of a better one which I want to simulate of course…Will post it later if I get this to work.
Best Regards
Rackham

What is wrong?

Your mesh is far too coarse. The solver will be producing random numbers on a mesh that coarse, and this could lead to all sorts of problems. You need a finer mesh than that to get anything close to accurate.

Hello Glenn,
Thanks for your time. I got it working again, got some "floating point error.." if I start with a very high resolution. Seems better to start with a bit more coarse grid and the refine it for the second run (do not know if it is the correct way..). The same answer with different resolutions is it a sign of "grid independent" results?

The old fan design showed to be better than I thought actually...

I was just wondering where do you measure your pressure if you do not have a volute around your centrifugal fan.\?????

At inlet and outlet .... Is volute necessary to record the pressure? http://www.pump-zone.com/topics/pump...unction-volute

I would prefer a volute because its a centrifugal fan. Secondly, a customer will never purchase a fan when it is not tested as per AMCA standards. And as per the standard, if you are testing a centrifugal fan, the fan is tested in a test rig with a volute...

Hello sainath,

You are absolutely right that 90% of all centrifugal fans are delivered with a volute. The plenum fans (or plugin fans) that I mentioned earlier are fans without volute, used in ventilation applications, furnaces etc.. In those cases the performance test is often done in a "custom" way not according to any standard. The fan wheel can be placed in a "mock-up" of the customers equipment, and real flow and pressure are mesured in this particular set-up.

Since I have very limited experience of CFD I have only simulated wheels so far, but I will try to model a complete test rig (ISO compliant) later on.

I have checked the accuracy of the simulations against plenum fan tests (i.e. without volute), it is quiet good.

My purpose now is only to compare different wheel designs in order to find optimum blade angles, inlet and outlet diameter, blade shape etc.. Essentialy I use old design rules from Bruno Ecks book. The results are sometimes interesting but it is difficult to draw any conclusions. I have used it in one real case so far. The fan was actually accepted by the customer without any on-site test... the average power consumption was low enough.

Will get back ont this.

I have also designed plenum fans especially for ventilation of A.C. Generators.

Performance of fans differ when they are with a volute and without a volute. In cases where your domain is huge and your approximations on the system side pressure drop is approximate, the pressure rise made by your fan is very critical and selection of wheel diameter.

There will be a difference in pressure rise made by your fan when your fan is put inside the customers setup. Moreover there will be a lot of errors on measuring the flow velocities phyically.

Hence it is always advised to test your fans in a standard test rig and generate your performance curves.

This is just a suggestion.

Thanks for info. :)

Yes you are correct. I agree.