# Complex eigenvalue post non linear analysis- analysis steps

## Recommended Posts

Hello,
I'm doing a simulation with Optistruct on a brake system and I want to perform a non linear quasi static analysis with frictional contact first and a complex eigenvalue analysis accounting for the frictional interaction then (i.e. asymmetric terms on the stiffness matrix).
I also read all the relative posts on this forum, but i'm still having problems. Do you have any suggestions or procedure to set the two analyses?
I have another question: in my model there are: a force that pushes one pad against the disk and a rotation of 5 degrees that puts  the contact between disk and pads in sliding; i have these two step at the same time, but is possible to put in contact only with the force first and to rotate in a successive step?
Thank You.

##### Share on other sites

Hello,

as I explained in the last post, I have to do a complex eigenvalue on a brake system.

I tried to exctract the stiffness matrix, but I didn't succeed with the superelement approach because Hypermesh does not allow to use it in a non linear quasi static analysis.

I understood that this method is not the way to reach a result, so I would like to know if is possible to obtain the stiffness matrix asimmetric terms (relative to the frictional contacts) in another way.

I saw this tutorial http://www.altairhyperworks.com/hwhelp/Altair/hw14.0/help/hwsolvers/hwsolvers.htm?rd-2110_complex_eigenvalue.htm, but in my case I haven't a mesh corrispondence among the surfaces because i'm studying a general case. Is possible to use this approach in any way?

If this approach isn't correct, will be possible to extract the stiffness matrix using Hypermath??

THANK YOU!

##### Share on other sites

Hello,

I didn't have any answer so I suppose that Hypermesh does not allow this operation.

Am I wrong??

##### Share on other sites

Did you check for STATSUB PRELOAD.Is it case of Brake sequal?

Attached deck for same.

brake_squeal_example.fem

##### Share on other sites

Thanks Raul.

Now it seems to work!

##### Share on other sites

Thanks for the update.

##### Share on other sites

Hello,

I tried to launch the attached model, but I received this error:

*** ERROR # 4754 ***
STATSUB(PRELOAD) of subcase 2 points to a nonlinear subcase and
frictional CGAP/CGAPG element (or frictional N2S contact) is present,
which is not supported.

Does exist any solution for this?

Because I'm interested in the contact approach used in this sample deck!

Thanks very much

##### Share on other sites

Which version of Optistruct you are using?

##### Share on other sites

I'm using 14.0 with the student license

##### Share on other sites

I tried in full version HW solvers patch 14.0.210

Will check and update you shortly.

##### Share on other sites

Hello Raul.

Thank You very much!

##### Share on other sites

Hi,

Sorry for the late reply.As you know that we are improving & adding more features in Optistruct latest version.Shared model run fine with upcoming version of Hyperworks solver 14.0.220 patch which is yet to officially release.(14.0.210 patch do not runs through)

For Student edition we dont release any patch so it cannot be possible with student edition 14.0 at your end.If you are running similar cases then i would request you to share your deck & i would run and update you.

##### Share on other sites

Hello Rahul.

I can use the full version on the PC of my referent professor at University.

Thank very much for the support!!

##### Share on other sites

Hello Rahul.

I'm setting up my model with the same approach of the example you attached here (brake_squeal_example.fem).

I'm having troubles to create CGAP elements, especially to extend these elements to the entire surface where contact is present.

Otherwise do you have any suggestions?

Thank you!

##### Share on other sites

Do you have node to node connectivity or matching pattern in that region.

##### Share on other sites

I have the same mesh on the two surfaces, so i think is a node to node connectivity.

When I try to create CGAP l Can choose only from one node to another, instead I have to connect all the nodes of a surface to the nodes of the other surface.

Hello Rahul,

##### Share on other sites

Hello,

do there exist any way to create CGAP elements automatically for the entire surface?

Or I have to do it one by one?

##### Share on other sites

If you have matching mesh then 1d page linear 1d is good option to create CGAP element. let me know if you need any help on same.

##### Share on other sites

Hello Rahul,

I tried to remake CGAP elements on the model that you sent to me, but I didn't succeed.

I varied all the commands in linear 1d like alignment,element config and density, but HM always gives back the same error!!

In the image attached you can see it; the problem is that the mesh's elements orders are the same!!!

Do you have any suggestion for solving this??

##### Share on other sites

Just checked same at my end.It seems linear 1d works fine with 2d elements.I tried with 2d approach by extracting faces & it worked.Will prepare video & share to you here asap.

Its all about knowing & handling Hm meshing tool efficiently.

##### Share on other sites

As far as i know Hm i would have opted attached option to create gap elements.Create sets of that region elements & make use of faces for creating gap elements.

##### Share on other sites

Hello Rahul,

thanks for the help.

I have an update that could be useful for you: I also succeeded to create CGAP elements with the spotweld function that allows to create 1d elements from nodes to nodes.

Now I'm having another problem: the results of the modal complex eigenvalue analysis are bad.

If I use fixed slope approach (fixed stiffness K_t) for the frictional contact I obtain complex modes with too high dampings and too low frequencies; instead if I use elastic slip distance approach I obtain good dampings and frequencies, but all the complex modes are stable (real parts of eigenvalues are negative) even if i'm in squeal classical conditions.

I tried to change lots of parameters like friction coefficients, number of roots desired (Both on EIGRL and EIGC cards to see if could be an effect of the modal method's truncation),etc. but I didn't obtain any good results.

Do you have any suggestions on this situation??

Thanks very much

## Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

×   Pasted as rich text.   Paste as plain text instead

Only 75 emoji are allowed.