Discussion Forum

Hi

COMSOL has some model reduction possibilities built in v3.5a (only partially there in current 4.0) but the doc is not very explicit. So I can only suggest that you try it out on a simple cantelever that you know analytically.

In v4.1 you have access (solver otions) to the eigenmode mass participation factors in structural (at least for the linear displaceents) to help you sort your modes

--
Good luck
Ivar

Ivar,
Thank you for your help, I haven't got anywhere with the reduction. I'm trying to develop model slowly from 2D to 3D. I can't get the freq response out of 2D. I attached the file Also I can't get the shell element to work in the 3D fluid model. Because the geometry of the the pivot bar in the middle of the two blocks will be made with a shell element. Is there a trick to use shell element and solids element's in Fluid Flow physics

Th file size is to big I don;t know how ales I can post it. I can delete all studies and most only geom and physics

file

Hi

sorry for a late reply but I do not currently have access to "my" COMSOL as I shear it with two colleagues that have been running models for 4 days and nights and have still some time to run ;)

--
Good luck
Ivar

Ivar,
Thank you, can you also look at this model when you get a chance. I can't get the shell and solid element to workout it gives me an error that shell elements aren't meshed

Ivar,
Have you had a change to look at the model, I have tried everything i could of to get eh shell to work with the solid . when I mesh everything I get an error that says " not all domains are meshed" But i have selected all domains int he system. Would you know what would be the problem? Thank you for your time

Hi

Sorry for the latre reply but I have not had access to the COMSOL licence the last days/weeks, my colleagues have saturated the workstation ;) Now they are on holiday so I have had a chance to take a look:

first of all your geometry: if you look at your mesh you will notice many elements (almost all dark graph, and the mesh quality has some low peaks) this is due to your large blocks that overlap over a few tens of microns in the lower region. Such "small features must be avoided as the mesher get lost and your solver turns for ages for nothing. First check carefully all your domains once passed the finish flag to see if they make sens, and take special care to avoid small geoemtrical details that are not crucial to the model.

Second the way you define your shell, compared to the rest of the 3D model. Shell physics solves 6 dependent variables: u,v,w the linear displacements and arx ary arz the three angles, but 3D structural and fluidics only three u,v,w (plus pressure p but that is another story) so youre three rotations are not linked and this will give you errors if you are not carefull:
1) you extend the shell element to the rim area around the fixed part and fix this extended area along u,v,w, then the rotations will be constrained by themselves
2) on the edge as you have defined tham you fix u,w,v AND set arx ary arz to "0" assuming that your large rim of polySi is "solid" and stiff compared to the rest.

Here too, I wounder if you really need to model the solid rim why not just forget about domains 2 anyhow of no interest for me, and 3 the stiff structure

But you need to define the volume under your thin flaps no, also the fluid ?. If you have a hollow cavity underneath to get the fluid flow correct through only those thin slots you would need a very fine meash thereon

Then the meshing: the shell should be meshed quite fine specially around the hinges to give correct stiffness response

Then you need to load the shell structure by the fluid pressure (surface load)

Finally you need to define correctly the fluid section fix the mesh along all fixed surfaces, drive the mesh along the flap shell surfaces, indentify the pressure variable to apply it onto the shell

a lest that's a good beginning

--
Good luck
Ivar

Ivar,
Thank you for your response I will try to go thought and figure everything out. Its just I'm beginner with Comsol, I have used ANSYS for a while. In my lab we have 24 core machine that will be better if I run Comsol. Also is there a better tutorials that can help with learning Comsol better. Like where to define DOF on the selection, and how to modify the physics to solve for specific dependent variables. One more time Thank you for your response and have a great holidays.

Hi

I would say the trick is to start simple, go through the exercices one by one, one physics at the time before you couple them and study those from the model library within COMSOL. If possible get on some of the COMSOL training courses.
For V4 get hands on the latest (patched) V4.1 (dec 2010) as the doc is better and there are very is few left in the physics (in fact I know of none, which is not the case for the earlier V4.0)

--
Good luck
Ivar

Ivar,
Thank you for all your help. you said that i need to load the shell structure by the fluid pressure (surface load), I have looked around and i only saw the body load and face load. Also how i can use the fluid pressure from the fluid structure model to apply on the shell.

Vlad

Hi

Isn't "face load" and "boundary load" or "surface load" just synonyms ?

--
Good luck
Ivar

Hi,
I got the model to work without the errors but not I'm not getting the motion of the shell which is because its not loaded with any load. you said that I need to "Finally you need to define correctly the fluid section fix the mesh along all fixed surfaces, drive the mesh along the flap shell surfaces, indentify the pressure variable to apply it onto the shell" Is there an example that I can look at. I looked in the Comsol documentation and there is no explanation on that. Thank you for your help.

Vlad,
Happy new year.

Hi
there are several examples in the model library, probably not of a "shell" but you extract the force/pressure from the fluid and apply it to a structural solid boundary just as you should do for a "shell" boundary

--
Good luck
Ivar