Learning Forum

[masud407]

Dear friends

1) I want to simulate the compression test of  ABS specimen(its an honeycomb structure) in ANSYS.

2) Should i use the tensile test data of standard ABS specimen(obtained experimentally)  to model the material for plasticity or

3) Will I have to do compression test of a standard cube element of ABS and then use the data obtained to model the material for plasticity?

please reply

[ramgopisetti]

Hi Array , I guess you are close, i hope by reduce the element size ( run convergence study with smaller difference) , i hope you might be close to the experiment. Try to use the displacement of your compression (Load curve) as function of time and compare the results. Comment down your results.
Cheers,
Ram

[masud407]

Hi Ram,

Thanks a lot for your response. So you are saying the refined mesh and convergence study with a smaller difference might help to reduce the gap? I can understand the effect of meshing, but can you please describe what you meant by " convergence study with smaller difference" in a Static structural simulation? In addition. how can I express the displacement of compression as a function of time in a static structural simulation? Isn't it for transient analysis?

[ramgopisetti]

Convergence is an Tree object which does heavy mesh refinement work for you. Initiate the model with the desired mesh and invoke the convergence option in the context menu and set the allowable change say 10 or 5% and run the model, the solver will change the mesh by ad hoc and for each refinement point it will give you the graph, aslong as the graph is flat, you can interpret the results and this should be reinforced by your validation.
Compression test are Quasi Static meaning that the application of load from the end-effector is slow ( strain rates are small ), if you have done the experiement, try to get the displaement curves, it is generally the XY plot with time in X axis and Displacement in Y axis, that said, you can use substeps for each time time increment and define the load ( displacement and use the Semi implicit option in Static structural system. But if you want to look into the failure of the lattices as you posted, you need to define an EPS or failure criteria in your material model and use the Explcit code for full implcit scheam to look into the failure of the body.
Cheers,
Ram

[masud407]

.

Hello Ram,

Thanks again for your response. I have no idea how to invoke the convergence option in the context menu and set the allowable change say 10 or 5%. Can you please mention some steps to start this process?

I have the experimental data (load, time, deflection). Now I applied the corresponding deflection of 5.2sec from the experiment (attached figure). On the other hand, I am calculating the reaction force at the top surface and trying to see if the reaction force matches with the experiment one. I am wondering if it is the right process or not. I need to calculate the elastic modulus, so I am not concern about the plastic regions. Thanks again and sorry for my ignorance. Looking forward to hearing back from you.

.

[ramgopisetti]

.

Ok Lets start with some idea around convergence, you can invoke this by selelcting an criteria say vm-Stress or Strain and right click it u will see an some thing like this

By choosign it, U will see its properties as follows

ypu can set the Allowable change base on your requried criteria ( dont be aggressive to make it 0, say 5% is good) and run the system, after some set runs, the plot in above will update for each solution instnace and an table will appear with all description if it. If you are moving to an newer version of ANSYS yu can have an snapshot of all runs in soluion output option. Try this steps and update it here, if you are sucessful in this step we shall move towards the using the load curves

Cheers,

Ram

.

[masud407]

Hello Ram,

Thanks for your response and good to know more about the convergence. In my previous run (0.10% strain for 5.2 sec and used substeps), I found the reaction force around 817 N, however, it was only 26 N in the experimental data, meaning 0.1% strain occurred after 26 N load.I am not understanding if just the convergence/ mesh refinement is responsible for this large difference.

Unfortunately, when I right-click as like you, I am getting the tan for convergence. How can I add that extension tab in AnSYS 19.2? Is it in the extension manager?

[masud407]

.

Hello Ram,

I managed to manage to invoke the convergence option in the context menu and set the allowable change say 20%. However, there is still differences in the reaction forces. Can you tell me the next steps?

.

[ramgopisetti]

Hello Array , Looking at your loading conditions, Does your lattices showed any signs of Damage in the experiemnt or do they just deformed, can you commnet down,
Cheers,
Ram

[ramgopisetti]

.

Hi masud407 ,

I have did same simulation but the material properties i used was Stainless Steesl NL with Failure strain with full implicit scheam, Based on your other post your defromation on the body was 2mm at 5.2 sec

my load curve -displacement type.

my spc forces for given lattices are as follows, but the maximum force generated is around 32.9kN at the fixed side of the lattices

Can you elobrate the force positions for futher support

Cheers, Ram

.

[masud407]

Hey Ram,

I am not getting what you meant to say. Can you run the simulation (for 0.1724 mm displacement at the top and fixed the other end) for the following material properties:
density -1180 kg/m3; young's modulus 1920 Mpa, poison ratio 0.36, geometry : lattice 17.24*17.24*17.24 mm lattice
Link for my geometry; https://drive.google.com/file/d/1_lEhJWFJ15Atek5yfz7oCIc2KRRtZt-S/view?usp=sharing
I am getting around 8000 N reaction force for that 0.1724 mm of displacement, but according to the experiment, it should be on 26 N (no kN). I am wondering if there is any chance that I am doing any mistake in material properties as I am getting much higher reaction forces (meaning less deformation for a particular force).

In your simulation, you are getting also in the kN range for steel which makes sense to me and matches what I am getting. But my problem is I am having a much lower load (26 N) for the 0.1724 mm displacement in the experiment.

[masud407]

My structure was printed using the stereolithography technique. Can it be a reason for having a different result?

[ramgopisetti]

Hi Array , I have run the simulation for 5.6 sec with the top loading of =-0.17294 as you specified in early post.
You specified some material properties, like basic elastic properties which are not enoough to capture full scale deformation(elastic and plastic), so you are definatly making error in the material properties, i hope you correct them by specifying some kind of plasticity and EPS as well.
Compression test are Quasi static with some set strain rates, these meas time is to be accounted and this is supported by implicit scheam.
Ansys or Lsdyna doesnot know what process you have followed, it only makes difference in results as there is change in the material, You need to apply the material properties based on the test.
I will try to reproduce with your material properties an show you the difference.
Cheers,
Ram.

[masud407]

.

Hello Ram,

Thanks a lot for your response. I have later come to know that the layer-by-layer 3D printing process actually induces anisotropic properties. Can you please tell me how I should insert the anisotropic behavior in ANSYS. It asks for six coefficients (in the anisotropic tab) to replicate anisotropic nature. I have attached the total material properties supplied by the manufacturers.

I search for it in the ANSYS manual but I haven't found much information. In addition, how can I know what type of plasticity behaviour I should select as there are a couple of options (Bilinear Isotropic, Bilinear Kinematic, multilinear isotropic etc.)

.

[masud407]

.

Hello Ram,

Thanks a lot for your response. I actually performed the compression up to 50% strain. But there was no damage in the structure for 0.1724 mm load. In the experiment, it required around 26 N (not kN) force to cause 0.1724 mm of displacement. However, I am getting a reaction force of 820 N if I tried to apply 0.1724 mm of displacement in the simulation. In your simulation, you are also getting a very high force (in kN) at Y direction although your material properties are different.

My material properties are 1920 MPa (young's modulus), 0.36 (poisson's ration), and 1180 kg/m3 (density). I am not considering the plastic deformation here. Can you check how much reaction force you are getting at the top or bottom surface for a lattice (17.24*17.24*17.24 mm) if you apply displacement of .1724 mm using the same material properties?

Here is the link for the geometry:

I am suspecting that my materials properties for the simulation are not the same as the experiment one. Can it be possible as I am getting less deformation in the simulation than the experiment for a particular force?

.

[masud407]

.

Hello Ram,

Thanks a lot for your response. I actually performed the compression up to 50% strain. But there was no damage in the structure for 0.1724 mm load. In the experiment, it required around 26 N (not kN) force to cause 0.1724 mm of displacement. However, I am getting a reaction force of 820 N if I tried to apply 0.1724 mm of displacement in the simulation. In your simulation, you are also getting a very high force (in kN) at Y direction although your material properties are different.

My material properties are 1920 MPa (young's modulus), 0.36 (poisson's ration), and 1180 kg/m3 (density). I am not considering the plastic deformation here. Can you check how much reaction force you are getting at the top or bottom surface for a lattice (17.24*17.24*17.24 mm) if you apply displacement of .1724 mm using the same material properties?

Here is the link for the geometry:

I am suspecting that my materials properties for the simulation are not the same as the experiment one. Can it be possible as I am getting less deformatio

.

[masud407]

Hello Ram,

Sorry for my last 2 replies as they are the copy of one of my comments. I don't know why those copies were re-posted here.

[ramgopisetti]

.

Hi masud407 , I will try to take your values and use no plasticity terms on your model and update you and for your reference the following is the matrix with terms u can hse in your anisotropic material table.

.

[ramgopisetti]

Couldnot able to download your model, try puttign an direct link
Cheers,
Ram

[masud407]

Hello Ram,

Sorry, I think you have access now.

[masud407]

Hello Ram,
Here is the necessary anisotropic properties. I applied those properties,but I am still getting a very large reaction force. Can you please check?

EXZ = 2,076.9 MPa, EY =1,677.1 MPa, VXZ = 0.28, VXY = 0.4, GXY = 613.7 MPa
V=poison ratio

[Mansoureh]

.

I hope you answer me and can see my message.

Dear masud407 may I ask some questions about your modeling to help me ?

Dear ramgopisetti, you wrote in previous comments:

"You specified some material properties, like basic elastic properties which are not enough to capture full-scale deformation(elastic and plastic), so you are definitely making errors in the material properties, i hope you correct them by specifying some kind of plasticity and EPS as well."

what do you mean about EPS?

and if we just have true elastic-plastic data as input data to Ansys, can we achieve failure in our model? and another question is, I want to apply a compression test like Masud407. I want to do it in explicit dynamic. my strain rate is 1.8 mm/min in 5 minutes.

how can I apply this in the "analysis setting" section? should I increase the number of steps?

thanks if you guide me.

.