topic Re: Nonline analysis results do not agree with the experiment results. in Simulation Mechanical Forums (Read-Only)

Nonline analysis results do not agree with the experiment results.

Anonymous — Wed, 04 May 2016 08:47:31 GMT

I made examination materials and performed a structure experiment.

This purpose is to make analysis data that agree with experimental data.

This details are shown in "letter1.pdf".

I performed non-linear analysis to make reliable analysis data.

But, Analysis results do not correspond with the experiment results.

The contents of "letter2.pdf". are my parameters about contact problem and non-linear analysis.

Please tell me the necessary analysis parameters to agree with experimental data, and I would be grateful for your time.

Thank you for your time!

Re: Nonline analysis results do not agree with the experiment results.

AstroJohnPE — Wed, 04 May 2016 12:53:24 GMT

Hi,

Thanks for the detailed explanation. It is always nice to have an engineering description of the model.

Would it be possible to provide the model itself? That would make it easier to answer some questions that come to mind and try some different parameters. See "Create, Post, or Provide an Archive of your model". For example,

What type of elements are you using for the strings?
How are you handling the buckling of the strings in the analysis?
How do the calculated displacements compare to the experimental displacements?
For model B, in the linear analysis (B-2), are the cylinders really penetrating (the displacement values indicate penetration), or is it just the exaggerated displaced shape ("Results Contours > Displacement > Displaced Options") makes it look like they are penetrating? (Mathematically, there has to be some penetration in order to generate a contact force.)
What version of Simulation Mechanical are you using?

Re: Nonline analysis results do not agree with the experiment results.

Anonymous — Thu, 05 May 2016 00:41:07 GMT

Hi, Thank you very much for your time!

I want to sent the model date to you, but error showed the following message.

"The file A01LineNew.fem does not have a valid extension for an attachment and has been removed. "

-----------------------------------------------------------------------------------------------------

Q)What type of elements are you using for the strings?

A) This elements are 2D Elasticity Elements (Quadrilateral).

Thickness are as follows. String of thickness is 10mm, and Assembly of thickness is 50mm.

-----------------------------------------------------------------------------------------------------

Q)How are you handling the buckling of the strings in the analysis?

A)This buckling is elastic buckling. I performed 3D analysis before performing 2D analysis.

Type of analysis is geometric non-linear analysis with large deformation and contact problems. This gets good results.　

But, When an analysis model becomes large, the load of the computer will be large. Therefore I want to perform 2D analysis.　

-----------------------------------------------------------------------------------------------------

Q)How do the calculated displacements compare to the experimental displacements?

A)This experimental result is got in the range of elasticity. 　

The plastic deformation is not considered, but Assemblies of contact problems and Strings of elastic buckling are considered. I want to perform non-linear analysis with large deformation.

-----------------------------------------------------------------------------------------------------

Q)For model B, in the linear analysis (B-2), are the cylinders really penetrating (the displacement values indicate penetration), or is it just the exaggerated displaced shape ("Results Contours > Displacement > Displaced Options") makes it look like they are penetrating? (Mathematically, there has to be some penetration in order to generate a contact force.)

A)In case of linear analysis (B-2), The contact force between Assemblies occurs in Penetration part. On the other hand, the contact force between Assemblies do not occur in Non-Displacement part.

-----------------------------------------------------------------------------------------------------

Q)What version of Simulation Mechanical are you using?

A)This version is 2017.

Re: Nonline analysis results do not agree with the experiment results.

AstroJohnPE — Thu, 05 May 2016 12:30:41 GMT

Hi Tanaka,

The .FEM file is not the complete model; that is why the forum does not allow you to attach it. You need to create an archive of the model (click the Application button in the top left corner, then "Archive > Create") and attach the resulting .ACH file. Try making the archive with results and see if the size is small enough to attach to your reply. (If the archive with results is too large, then you can remake the archive of the model only.)

Re: Nonline analysis results do not agree with the experiment results.

Anonymous — Fri, 06 May 2016 03:41:44 GMT

Hi,

Thank you very much for telling the method of the solution to the problem to me!

I was able to make ACH file. But this file size is 60MB.

I was not able to sent ACH file to you.

"A01LineNew.fem" was uploaded to the web below.

https://www.datadeliver.net/receiver/file_box.do?fb=d5e39baf947b48709921fe6dfc40acd1&rc=95957ac49dfd4fd7999a6460b5487d94&lang=ja

Download method is shown below.

----------------------------------------------

1) You display the web below.

https://www.datadeliver.net/receiver/file_box.do?fb=d5e39baf947b48709921fe6dfc40acd1&rc=95957ac49dfd4fd7999a6460b5487d94&lang=ja

2) You click "A" by referring to "dateben.PNG".

3) New, you push "B".

4) You click "A01LineNew.fem" by referring to "dateben2PNG".

5)You can get "A01LineNew.fem".

----------------------------------------------

I would be really grateful if you could get.

Re: Nonline analysis results do not agree with the experiment results.

AstroJohnPE — Fri, 06 May 2016 04:09:46 GMT

Let me make sure that I understand.

The archive file (.ach) without the results is 60 MB. Is that correct?

I see from the download website that the fem file (.fem) you uploaded is 60 MB. However, the fem file is not useful to us because it does not include all of the input. We need the archive file. Can you put the archive file on the download website?

Note that you can also read the information in this thread to decrease the size of the archive file if that is necessary. "Create, Post, or Provide an Archive of your model"

Thanks.

Re: Nonline analysis results do not agree with the experiment results.

Anonymous — Fri, 06 May 2016 05:26:06 GMT

I am very sorry.

The archive file (.ach) is 40MB.

I put this archive file on the download website below.

https://www.datadeliver.net/receiver/file_box.do?fb=84428d16a0d44a0892e8dc3245d2e2cd&rc=3dc2d4d55d0647d1a578c12eecb45be4&lang=ja

As explained earlier, you can get "A01LineNew.ach".

I would be really grateful if you could get.

Re: Nonline analysis results do not agree with the experiment results.

AstroJohnPE — Fri, 06 May 2016 16:23:48 GMT

Hi Tanaka,

I made the following simplifications to the MES model.

I realize that your description of the experiment included "strings with bending stiffness", but I could not think of a good way to model the strings that would allow bending but also include the "buckling" effect. Since the "string" was only 0.3 mm thick in your model, I suspect that the bending stiffness is approximately 0, and therefore the strings mostly act in tension, insignificant in bending, and 0 compression. So I replaced the 2D elements that you were using for the strings with coupling elements. (Coupling elements are lines that only work in tension.)
I created a load curve for the nodal force. In your model, the load curve multiplier was 0 for the entire analysis, so that would result in 0 load.
I did use a coarser mesh so that the analysis would run faster.

The results are shown in the attached image. (Note that I magnified the displaced shape to make it easier to see the motion.) How does this displacement compare to the experiment?

I have also attached my archive of the model. Note that I changed the extension to .ZIP because of a problem that I am having posting archive files (.ACH).

Here are some other things to think about for my version of the model:

By changing the "Coupling Length" of the strings (under the "Element Definition"), you can create an initial slack or preload. You can try to change this depending on high tight your device is when assembled.

I think that this experiment cannot be analyzed in linear stress -- at least not very easily -- because the buckling of the strings will be difficult to simulate.

The element type in linear stress that is equivalent to the coupling elements that I used in the MES analysis are Gap elements (with the "Element Definition" type set to tension with no gap). But because of the motion if mostly perpendicular to the direction of the strings, the small displacement limitation of linear stress will not give correct results.
If you model the strings with 2D elements, you would need to remove the strings that are in compression and buckle, and run the analysis again. So it becomes a manual operation to iterate to find the correct solution.

Feel free to reply if you have any questions.

Re: Nonline analysis results do not agree with the experiment results.

Anonymous — Sat, 07 May 2016 10:16:45 GMT

Hi,

I made analysis model ago by referring to your model.

I got result that is closest to my image.

I improve analysis precision from now on.

Your advice solved a problem. Thank you.