I modelled partially shrouded co-rotating disks (3 disk)using commercial software. However there's secondary flow developed at region between the disk near the partially shrouded area.

My question is 1. Are they supposed to be there? 2. If not, what went wrong?

I supposed to verify some experiment results (done by someone else). And I'm not sure which is more correct..

Thank you so much...

As soon as you don't have solid-body rotation of the flow between your rotating discs you will get secondary flows. Solid-body rotation only occurs if you don't have any kind of "disturbances" like temperature gradients (ie density gradients), leakeages, non-rotating walls etc.

I'm not sure what you mean by "partially shrouded", but it sounds as if you might have both leakages and non-rotating walls.

When you don't have solid body rotation you will get radial out-flow along the rotating walls and nice Ekman boundary layers full of cool secondary flow phenomena.

Temprature gradients can give even nicer secondary flows, depending on the rotational speed you might get completely chaotic flow or nice radial arms of cold fluid migrating outwards creating cyclons and anti-cyclons...

Cavity flows are fun. Lots of secondary flows and rotational effect leading to even more secondary flows. Many phenomena are similar to what happens in the atmosphere.

Dear Jonas,

Thanks for the quick response.

The model I'm working on is actually rotating disks inside a fixed enclosure. The enclosure itself consists of a half cylinder (hence the shroud) and a box. It's actually a simplified model of hard disk drive. In that case I do have non-rotating walls around the disks.

However in the case of full shroud (disks rotating in stationary cylinder), this problem never actually occured. Is there any explaination to this?

Also, do you know any reference or journal paper I can refer to regarding this type of flow?

Thanks again...

My guess is that when you open up the cylinder you create a big enough disturbance to get the pumping effect of the rotating discs working. The open part will allow fluid to be pumped out by the rotating discs there. With a full cylinder shroud you have a symmetric sealed case and the shroud in this case might not cause a big enough disturbance to start the secondary flows.

There has been a lot of work on this type of flows in relation to compressor-cavities in jet-engines. I don't have my references available here but a good start is to look for papers by Micael Owen, University of Bath. Here is a page from that university with a bunch of references:

http://www.bath.ac.uk/~ensmw/pubs/pubs.html

I saw several papers about rotating cavities with stationary outer casings - sounds similar to your problem.

It is correct. The secondary flow is the part of the flow (in axial direction) over the shroud gap.

Which kind of CFD package you are using, Did you verify the vortex shedding frequency? Thanks!

Dear Jen, I'm using FLUENT and I'm not sure what you mean by "verifying the vortex shedding frequency"

Thanks

That is another way to check how good is your unsteady simulation for your hard drive. Do you also concern about the vibration too? I think it is the major concern. Fluent does not have the coupled modules, I think. They are struggling to solve similar problems for the unsteady flows with correct vortex shedding frequency in Indian. Keep in touch.