Discussion Forum

Hi

if it is a rigid rotation in Solid, extract the pin domain (geometry difference) select the round boundary and use a rigid connector with user defined CoR 0,0,0 and constrained X,Y,Z position (of CoR 2nd tab) and that should do. but you are adding several equations to your system.
Another way is to add a global equation to constrain the average of the x and y position of your boundary to the CoR

there are many ways to Rome ...

--
Good luck
Ivar

Hello Ivar,

I extracted the geometry of pin (geometry difference) and selected the round boundary using rigid connecto.
My center of rotation is Xc = (0.0), and I would like to make a prescribed rotation (pi / 2) * t with time dependent = range (0,0.1,1), but my simulation did not converge.

What is wrong?

Thanks for your help!

Hi

are you in time dependent solver sequence ?

just to get consistent units write rather 2*pi*t[1/s] to clear the orange unit warning for the imposed rotation angle

for me it does converge, but I'm using arbitrary material values.

On the other hand if you want to look at the centrifugal deformations, you could rather remain in the rotational frame and apply body loads

An quick example hereafter, I might have missed a point, only applying radial forces. I leave it to you to find out why there are differences ;)

It's a v4.2a model

--
Good luck
Ivar

Hello Ivar,

The simulation ran, but did not provide the expected result. I thought then to rotate a solid 90° it would be necessary (phi = 0.5 * pi * t [1 / s]) and time=0:0.1:1s. However the simulation results, presented a very small rotation (figure attached).

How do I rotate the solid at 90°?

I'm using rigid connector with comsol v4.2a.

thanks again!

Hi

You are fooled by the automatic scaling of COMSOL. The programmes is not set up for "large deformations like that, if you search in the plot node you will find deep under the scaling tab, go to manual scaling and set a value of 1, then it should work better.

Pls notice that I'm using the structural "small angle displacements" the 0.5*solid.curlUZ, this is based on linear theory so it's not really valid or angles above some 5-6° of rotation, be aware. I did neither not turn on "large displacements" that might be usefull of your solid part (not the rotation by itself) is defoming a lot.

--
Good luck
Ivar

Hello Everyone,

I would like take this problem and surround the rotating solid with a fluid medium (water). I am currently working with the FSI module and having no luck. I keep getting errors about coverages , here is the full message.

Failed to find consistent initial values.
Segregated group 2

Attempt to evaluate real square root of negative number.
Function: sqrt
Last time step is not converged.

Can anybody provide some guidance about what I am doing wrong?

Hi

there can be many causes, but it could be a mesh not appropriate to solve a diffusion effect, which can lead to sign inversion when not expected, hence a sqrt(-1) issue. Not in the solid I expect, more often in HT or CFD related part

--
Good luck
Ivar

Ivar,

Thank you for your response. I am new to COMSOl so please pardon my ignorance. I was wondering if you give me some common solutions that may help resolve the issues I am having with my mesh. Specifically what type of mesh or meshes should I be using, how fine should the mesh be, or anything else I should change but don't know about yet.


Thanks again
Casey

Hi

the best way to learn is to exercise the model library cases, and, when possible, take a couple of 1 day courses by COMSOL, these give you really a better understanding and support for the many questions, which is important to speed up the learning curve the first weeks when starting with COMSOL

Also there are many discussions on the FORUM about good practices, try a search

--
Good luck
Ivar

Anyone who can help me with this?
It is also related to rotation.

www.comsol.com/community/forums/general/thread/24687/

Thanks!

Hello Ivar,

I had a look at your attached model, but it seems to have the same problem as my own model with rotation. The geometry becomes bigger every time step. Generation a point-plot with displacement in y-direction or total displacement, you can see values of 1,5 mdescribing the displacement. But such a value of displacement is not possible to be generated by the simulated rotation. Where is the fault in my consideration or in your model?

I am trying to model the rotation of a solid cylinder by using a boundary condition concerning one of the front surfaces, but within rotation the solid cylinder seems to deform (becomes bigger and smaller in radius). I am using the rigid connector condition as boundary condition for one front surface. All other boundaries are free because of the simulated process (open die forging). Is the rigid connector the right way to let the solid cylinder rotate?

Thanks for a hint.