CFD Flow Analysis of Rotating Object - What Does the Data Tell Us?

[Wetmelon]

Hey guys, I've got a CFD question about the image linked below.

Basically, I've got the thing designed and I ran a simple flow simulation on it, but I'm curious about the information it's telling me. I can see that the flows are definitely going down the rifling, but I'm not sure what the velocities are truly telling me.

Hopefully this thing will begin rotating due to the force of air in the channels. Once it gets to speed, will I get nearly laminar flow through the channels? Or will I get a lot of turbulence over the channels?

 

[minger]

What are you looking to get from it? Also, it appears that this problem is periodic. You can get much more accurate results by looking at a "slice" (i.e. you can use the same number of grid points in a much smaller area).

As for rotation, well for that you need a net torque. Could you perhaps do a conservation of angular momentum over the entire domain?

 

[Wetmelon]

The end result I'm looking for is the object's stability through air (at about 100-150m/s) along its longitudinal axis (i.e. no yaw or pitch deviations). I assume this means an adequately high rotational speed

I'm just using FloXpress in SolidWorks, so I'm not sure about how much capability I have, but if you give me an idea of how to do a slice I can look into it.

As for the rotation, how do I get the force of the air on the ridges?

 

[minger]

You could look at the static pressures on the faces. If your solver can get an element normal then you can figure out the torque generated about the axis.

To slice it, you could do it in any solid modeler. Then just take your cut edges and make them periodic.

 

[Wetmelon]

I managed to get the proper Flow Simulation software going; the good stuff. As I have it (The full pellet), I'm generating about 1.525e^-5 Nm of torque, giving me an angular acceleration of about 200,000RPM/s. This might be adequate, but it's hard to tell. I'll have to get some actual fired pellets to see if they work :)

I'm still not sure what you mean by "slicing" Do you want me to slice just one channel section?

 

[minger]

Yes, your pellet (as you're calling it) is symmetrical. So, if you "slice" your pellet around the grooves such taht you can "repeat" each section n times to get a full pellet, then you have periodicity.

That let's you use the same amount of mesh in a much smaller area giving you a much more accurate results. There is very rarely a need to model an entire domain of a periodic body.