Can I simulate centrifuge with no inlet or outlet?

I want to spin a closed system with no flow in or out. Will this simulate?

How do I set the RPM rate of the system and how many times it spins?

Tried that. No difference.

Tried that. No difference. I am now seeing that I can simulate p with no rotation monitoring without errors. I was looking at pressure of the volume, but changing that to pressure on the faces, it works. As soon as I turn on rotation monitoring, I get the same error message and the rotation is a flat line.

When I run Vol, it produces a colored image, but there are a lot of voids in the volume. It seems to remove a lot of faces and edges. Is this normal?

Moving Reference Frame does not move anything

If you are still getting the same warning "The system cannot find the path specified" then the solver is not running, so you are not seeing calculated results.

You will also have to set the following File->Preferences:
Application->MPI = C:\Program Files\Microsoft MPI\Bin\mpiexec

The rotation monitor is for geometric rotation, not for a Moving Reference Frame (MRF). For more background on MRF see "Moving Reference Frame for Computational Fluid Dynamics."

To view the mesh you have to set the view properties according to the tip "View a Mesh."

Progress, but sim failing

The mesh looks good and no errors now, but will not rotate during sim run.

When I try to run a Result like U or p, I get no rotation. If I try to monitor the rotation, all of the scalar and vector displays are greyed-out so I cannot select them. If I try to run a sim with only rotation monitored, it runs, but is a flat line, no rotation.

In Physics:
I am setting the rotation around Y-axis to 3000 degrees, 0 3000 0 by selecting axis and the only volume
This is 52.36 radians/sec and the face farthest from the Y-axis is about .09m, so
U in Water is set to 4.7124 0 0 (not initial, which is 0 0 0)

This should give me 500 RPM with .09m furthest extent of the structure from Y-axis.

Must the .09m measurement be exact for it to work? I don't seem to have any method in Geometry to make precise measurements.

When I run p, I get this error:

volume [www.6627.5] RANS Flow: Failed
Unexpected process exit.
The system cannot find the path specified.

The bar at the bottom is all blue, but nothing shown on my structure.


MS MPI version

Assuming you are on Windows, the problem could be an old install of MS MPI. You can download the latest version from:

and select/install:

High accuracy not the answer

I tried medium and high accuracy. With high accuracy, the mesh is good over the length of the test cylinder.

However, when I try high accuracy on my actual structure, which is a pipe loop, the long straight tube part of the loop has almost no triangles on the middle of the tube. Actually worse than low or medium accuracy. High accuracy makes the mesh good on the rest of the structure, so that is an improvement, but where it loops back on itself to form a closed loop with the long tube, the mesh is missing in the middle. I don't understand how I can fix just this cylinder face part of the structure.

Use the Accuracy Tool, Custom option

For explicit element sizing control you need to set Accuracy Tool->Custom->Max Element Size. For more background see the tip "Accuracy Tool."

Mesh falling apart on simple cylinder

I have a long simple decomposed cylinder that is the only volume. Two circles connected by one line. End of the cylinder is at the intersection of all 3 axis. I fill it with water like the rotating paddle example and I set all of the params the same, except for the angular velocity and the Phase-U-Fixed value, which matches the angular velocity. I swing it around the Y-axis. No Cyclic added.

When I mesh the volume by doing Scalar-A, the middle of the cylinder has very little meshing. When I try to add force arrows, it crashes the app.

Shouldn't the cylinder have uniform mesh over the entire length? How can a simple cylinder have issues? What am I doing wrong?

Accuracy Tool and mesh metrics

To better control the mesh grading you need to use the Accuracy Tool and examine mesh metrics.

Specifically you need assign the Accuracy Tool with a maximum size on the wall of your cylinder to get a uniform surface mesh. It will save a lot of time if you first focus on the surface mesh rather than force a simulation run (and therefore a volume mesh) by requesting a solver specific flow field.

Cyclic settings??

I need to set the cyclic for 360 degrees and the same face for the "from" and the "to" (from, to). The tool gives me an error because it only takes one face or does not present a "done" to click on when I try to set the same face twice.


Cyclic condition only for axisymmetric geometry

If you are simulating the entire 360 then you don't need to use a cyclic condition.

Face already in front error

Okay, I'm trying to simulate and I get 3 Triangulator errors "face already in front" and one Tetulator error "face already in front"

What does it mean?

I have an enclosed volume rotating on the Y-axis with one side on the axis.

I was not able to assign the angular velocity to a wall. It only gave me the volume to select.

Review mesh and solver failure FAQ

For your meshing problem review the FAQ "How do I fix mesh and solver failures?"

For your assignment problem try shading the faces and then select them. If you have only one face that is also the volume shell then you need to split up the faces.

Moving Reference Frame for a Centrifuge

Yes you can simulate a rotating fluid with a Moving Reference Frame (MRF) as described in the tutorial "Rotating Paddle using Moving Reference Frame."

Note you do not need to use a multi-block structured mesh as in the tutorial, which will likely make your case easier to model.