Stuck? 
support@mecway.com
support@mecway.com
Mecway
FEA
Non-linear Contact Modal Analysis with CCX
Hi All,
I'm long time lurker in this forum. I've come across what I consider an interesting - possibly new, if proven valid - feature when it comes to modal analysis with CCX.
To give a litte context: I work primarily with metal sheet, so I make a lot of use of shell elements, and recently my models are only getting bigger, so I decided to use CCX as solver, altough I understand it doesn't truly use shell formulation.
The issue is that when creating the contacts for the parts of my model, I've realised that a I can only use *TIE formulation with true solid elements. That is a big issue, since I need a linear contact model to run my modal analysis - as is an essentially linear analysis only - and I want to remain using expanded shells.
Reading the documentation, I discovered that there is another tied formulation under the *CONTACT PAIR that it's possible to establish between solid and shell that results in true linear contact formulation. So I tried to use that along with a modal analysis. To no surprise, it didn't work... The solver doesn't establish the contact and the parts behave as if independent from each other.
However, if the model is subjected to a pre-stress modal analysis, suddenly, it works! I suspect that after the static step, the stiffness provided by the contact is included in the pre-stressed stiffness matrix and, so, the model behaves as if the contacts were created during the modal analysis. So, the work around is to introduce a dummy load to trigger the pre-stress modal analysis, bearing in mind that this load may affect the true behaviour. So, as low as possible load should be imposed.
I've attached two models, both of them with the same or close to the same simple BC: one with a part modeled with solids and the other with shells, creating the custom card of the aforementioned tied contact between then; the other, both parts are modeled with solids and a bonded/tie contact is used. The screenshots show that there is a good agreement across the models. To get that result, I had to fiddle with the constants defined under the contact pair, to simulate just a stiff of contact as the bonded-tied one.
To finish up, I know that this may not be the most useful tool for most users, however the fact that it apparently works really intregues me. So, I'd like to read you guys' take on it. See if there's some theoretical back-up or other tests that may increase the confidence in using this method,
DeÂngelis.
Notes:
- The title may be misleading, because the contact between the beam and plate can't be non-linear in a modal analysis, as I've metioned before. It's just that CCX, to my understanding, interprets any contact pair as non-linear from the get go and is one of the reasons why it doesn't work directly with modal analysis;
- To see that the modal analysis fail without pre-stress, just supress the dummy load.
I'm long time lurker in this forum. I've come across what I consider an interesting - possibly new, if proven valid - feature when it comes to modal analysis with CCX.
To give a litte context: I work primarily with metal sheet, so I make a lot of use of shell elements, and recently my models are only getting bigger, so I decided to use CCX as solver, altough I understand it doesn't truly use shell formulation.
The issue is that when creating the contacts for the parts of my model, I've realised that a I can only use *TIE formulation with true solid elements. That is a big issue, since I need a linear contact model to run my modal analysis - as is an essentially linear analysis only - and I want to remain using expanded shells.
Reading the documentation, I discovered that there is another tied formulation under the *CONTACT PAIR that it's possible to establish between solid and shell that results in true linear contact formulation. So I tried to use that along with a modal analysis. To no surprise, it didn't work... The solver doesn't establish the contact and the parts behave as if independent from each other.
However, if the model is subjected to a pre-stress modal analysis, suddenly, it works! I suspect that after the static step, the stiffness provided by the contact is included in the pre-stressed stiffness matrix and, so, the model behaves as if the contacts were created during the modal analysis. So, the work around is to introduce a dummy load to trigger the pre-stress modal analysis, bearing in mind that this load may affect the true behaviour. So, as low as possible load should be imposed.
I've attached two models, both of them with the same or close to the same simple BC: one with a part modeled with solids and the other with shells, creating the custom card of the aforementioned tied contact between then; the other, both parts are modeled with solids and a bonded/tie contact is used. The screenshots show that there is a good agreement across the models. To get that result, I had to fiddle with the constants defined under the contact pair, to simulate just a stiff of contact as the bonded-tied one.
To finish up, I know that this may not be the most useful tool for most users, however the fact that it apparently works really intregues me. So, I'd like to read you guys' take on it. See if there's some theoretical back-up or other tests that may increase the confidence in using this method,
DeÂngelis.
Notes:
- The title may be misleading, because the contact between the beam and plate can't be non-linear in a modal analysis, as I've metioned before. It's just that CCX, to my understanding, interprets any contact pair as non-linear from the get go and is one of the reasons why it doesn't work directly with modal analysis;
- To see that the modal analysis fail without pre-stress, just supress the dummy load.
Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Comments
I've found some improvements:
You can make the contact using the Elastic option in bonded contact instead of writing the cards by hand. The only special change needed is to add the dummy load.
You can make it more pure by disabling the dummy load with CCX -> Modify keyword and omit the *CLOAD block(s). Then Mecway still generates the *STATIC step but doesn't put the load in it.
I'll put it on my list to test this more thoroughly and if it's good, make Mecway generate an empty *STATIC step whenever you have Elastic bonded contact with modal analysis so it works transparently.
Thanks for taking the time to analyse this feature and providing improvements!
Certainly, it's way more convenient to use MW bonded conctact definition than it's to create the cards by hand. It will make my life much easier.
It's very interesting that the load can be erased completly, e.g. that it's just necessary to run the *STATIC step, prior. That is a solid argument in favor of the validity of the method, in my opinion. I was worried that the load had to, somehow, "activate" the contact. But that just makes sense, because the contacts that we are imposing are linear in nature, so they behave independently.
I'll try and use this in bigger, more complex models. I'll post comments along the way, if I see anything note worthy. And I'll wait to see the results of the test you plan to make. If valid, this feature will help a lot!