Welcome to the forums. Please post in English or French.

You are not logged in.

#1 2021-05-14 08:24:23

Bragadesh_Srivatsan
Member
Registered: 2021-04-23
Posts: 15

three point beam bend nodal reac vs displacement error.

Hello,

I am trying to plot force vs CMOD(crack mouth opening displacement). The simulation runs into Green and I have defined an elastoplastic material with contact properties as defined in the attached command file. I do not know where I am going wrong but the simulation curve is not complementing the experimental curve.

I have attached the zip file of the log, mess, comm, mesh, and image of curves with this post.

It will be really helpful if someone has some insights on this.

Thank you.

Regards,
Bragadesh Srivatsan.


Attachments:
3pbending.zip, Size: 390.8 KiB, Downloads: 43

Offline

#2 2021-05-14 09:03:29

mf
Member
Registered: 2019-06-18
Posts: 227

Re: three point beam bend nodal reac vs displacement error.

Hello,

I'd advise to use a 2nd order mesh, and see how it goes from there. Also consider a more refined mesh, especially in the area of the crack tip.

One further note: does the real specimen crack during the test? Or is it a VERY ductile material and we still 'only' see the blunting of the crack tip without any crack propagation? I suppose it does not crack in the attached curve?

Furthermore: take a look at the exact measurement method of CMOD (I know it as CTOD...crack tip opening displacement and I know there are many different methods) and if your 'measurement' in the simulation is comparable to the real measurement.

Mario.

Last edited by mf (2021-05-14 09:17:14)

Offline

#3 2021-05-14 09:38:08

Bragadesh_Srivatsan
Member
Registered: 2021-04-23
Posts: 15

Re: three point beam bend nodal reac vs displacement error.

Hello Mario,

Thank you for your reply.

I will try out with higher-order elements.

No, it does not crack in the attached curve. I will be trying to incorporate damage to this model afterward though.

The measurement method is to find the average displacement of the edge nodes along the longitudinal direction of the beam, as shown in the figure attached.

I also want to know if am extracting the forces from the correct nodes. I have plotted these curves from the support bearing surfaces 'left_sup' and 'right_sup' which I have fixed in all directions.

Regards,
Bragadesh Srivatsan.


Attachments:
cmod.JPG, Size: 29.21 KiB, Downloads: 42

Offline

#4 2021-05-14 09:46:40

mf
Member
Registered: 2019-06-18
Posts: 227

Re: three point beam bend nodal reac vs displacement error.

Hello,

you're welcome, I worked for years with these kind of specimens during my thesis (although a long time ago, in a galaxy far far away :-) ).

With measurement methods, I meant the method in real life: it can be done by image analysis of a video of the test, by clamps on the specimens edges, laser interferometry (also laser speckle methods), etc... all methods have different errors (also systematic ones), which might influence the recorded curve. The CTOD may also be measured by 3D-reconstruction of the cracked surface,

Mario.

Last edited by mf (2021-05-14 09:49:03)

Offline

#5 2021-05-14 09:55:58

Bragadesh_Srivatsan
Member
Registered: 2021-04-23
Posts: 15

Re: three point beam bend nodal reac vs displacement error.

Hello,

Ah sorry big_smile, I do not know the measurement method used for the CMOD data.



Regards,
Bragadesh Srivatsan

Offline

#6 2021-05-15 17:56:50

Bragadesh_Srivatsan
Member
Registered: 2021-04-23
Posts: 15

Re: three point beam bend nodal reac vs displacement error.

Hello,

I did update the model using quadratic elements but I get an error termination. Floating-point error. I get the same error when I try to use finer linear mesh.

How do I overcome this error?

I have attached the log file along with the mesh file.

Thanks and Regards,
Bragadesh Srivatsan


Attachments:
3pbend_float_er.zip, Size: 812.38 KiB, Downloads: 50

Offline

#7 2021-05-16 18:47:35

mf
Member
Registered: 2019-06-18
Posts: 227

Re: three point beam bend nodal reac vs displacement error.

Hello,

I ran your files in SM2020 (Ubuntu 20.04 LTS) and everything is alright. Sorry, I do not know about the error but it might be related to the Windows-version of SM.
Your error happens just as the contact starts, this cannot be a coincidence. Do you have enough RAM for this example (compare my .mess and yours...)? I attach my .mess and the three .resu (tables),

Mario.

Last edited by mf (2021-05-16 19:06:23)


Attachments:
3pbend.zip, Size: 25.44 KiB, Downloads: 40

Offline

#8 2021-05-16 22:04:21

mf
Member
Registered: 2019-06-18
Posts: 227

Re: three point beam bend nodal reac vs displacement error.

Hello again,

I entered the same value in 'Memory' as you did, and it also completed. So it's not a lack of memory.

Mario.


Attachments:
message, Size: 178.32 KiB, Downloads: 33

Offline

#9 2021-05-19 00:03:26

Bragadesh_Srivatsan
Member
Registered: 2021-04-23
Posts: 15

Re: three point beam bend nodal reac vs displacement error.

Hello,

Thank you, Mario. That really did the trick. The windows version was the problem.

But I am still having the problem with Force, CMOD plot. I have a plot from Abaqus for the same model and it is complementing with the experimental results, so I do not think it may be from the CMOD measurement. I have attached the plot with this post.

I have computed the resultant reactive forces DY from the supports, and for some reason, it shows zero reaction forces when I try to compute it at the load-bearing.

I can't figure out where I am going wrong.

Regards,
Bragadesh Srivatsan


Attachments:
3pbend.zip, Size: 271.84 KiB, Downloads: 35

Offline

#10 2021-05-19 10:36:18

mf
Member
Registered: 2019-06-18
Posts: 227

Re: three point beam bend nodal reac vs displacement error.

Hello,

glad this worked out.

I looked at your results, I think your model behaves correctly (it has contact, nothing wiggles around etc.). But I do not know which nodes you looked at, but see my attached images.

Your model also has some gaps left, are these also in your geometry? If not, maybe you want to refine the mesh of the cylindrical supports a little to close these gaps (also choose 2nd order for the supports). This shouldn't have influence on the curve, but let's see. However, in terms of simulation time, contact will be earlier on the time scale if you close the gaps.

Also: RESI_GLOB_RELA could be lower, try 1e-6. This will increase calculation time, though.

There is still a little to do, but I think you are getting there. If you want to compare it with other solvers, you should be very careful and compare selected elements, mesh, material data, material model, accuracy in the calculation (residuals etc.) etc..

Mario.

Last edited by mf (2021-05-19 10:45:09)


Attachments:
screenshots.zip, Size: 1.24 MiB, Downloads: 38

Offline

#11 2021-05-19 12:40:06

sameer21101970
Member
Registered: 2019-09-06
Posts: 338

Re: three point beam bend nodal reac vs displacement error.

INTERPOL=('LOG','LIN', ),

try above.

Offline

#12 2021-05-25 22:37:04

Bragadesh_Srivatsan
Member
Registered: 2021-04-23
Posts: 15

Re: three point beam bend nodal reac vs displacement error.

Hello,

Thank you for the suggestions Sameer and Marco, I tried that out and I get the Force Curve included in the zip file. Is there any other suggestions that will improve the result ?

I was not able to run the simulation with second order elements in support due to the computational restrictions on my personal Computer.

I was trying to simulate the model which I have included in the zip file, The model takes too much iterations to converge. Is there some way to make it converge faster ?

Thanks in advance.

Bragadesh Srivatsan


Attachments:
3pbending1.zip, Size: 902.88 KiB, Downloads: 37

Offline

#13 2021-05-26 08:10:03

sameer21101970
Member
Registered: 2019-09-06
Posts: 338

Re: three point beam bend nodal reac vs displacement error.

COMPORTEMENT=_F(DEFORMATION='GROT_GDEP',

                                        GROUP_MA=('beam_full', ),

                                        RELATION='ELAS'),

change above and converge

Offline

#14 2021-06-01 10:58:07

Bragadesh_Srivatsan
Member
Registered: 2021-04-23
Posts: 15

Re: three point beam bend nodal reac vs displacement error.

Hello,

Thank you Sameer for the suggestion,

1. The model converges if the bearings are of mesh size of 6mm but does not converge when the bearing size is 5mm. I have attached the model again with this post.

2. I want to define a cohesive contact for the beam, I did read about Cohesive zone model. For the cohesive zone model we need the cohesive elements to be modeled to define the crack plane. I do not want to model extra elements but just model the cohesive zone as a contact. Is it possible in code_aster ?
As of now I am trying to model this zone in a cube. I have attached the .Comm file(In the next post) .I modeled using the validation examples. I want to know where I have gone wrong


Attachments:
beam3p_1_6.zip, Size: 505.95 KiB, Downloads: 32

Offline

#15 2021-06-01 11:01:56

Bragadesh_Srivatsan
Member
Registered: 2021-04-23
Posts: 15

Re: three point beam bend nodal reac vs displacement error.

Hello,

Files for cohesive contact model.


Attachments:
Cohesive.zip, Size: 72.07 KiB, Downloads: 30

Offline

#16 2021-06-14 14:12:47

Bragadesh_Srivatsan
Member
Registered: 2021-04-23
Posts: 15

Re: three point beam bend nodal reac vs displacement error.

Hello !

Anybody?

Offline

#17 2021-07-21 12:09:08

ywan459
Member
Registered: 2020-08-18
Posts: 4

Re: three point beam bend nodal reac vs displacement error.

What does the CMOD mean? Is the term for the openning width of between node group "right_crack" and "left_crack"? And what force are you referring here? the supports reaction force or the force measured at the top fixture?

I have download your first posted pakcage. It produces lots of outputs in two table files. I am totally confused. Sorry.

Offline