Coding Faliure Criterion

[ashu28]

Hello everyone,
I am simulating machining of polycarbonate and I need to code the failure model. I would want to know how can I do that ? What subroutine do I have to write ? Is it appended in UMAT ? Some sample coded failure models would be really helpful.

Regards,
Ashu

[Jorgen]

Machining, hmm, interesting. What type of machining do you have in mind? Polycarbonate is a notch sensitive material. That is, the material can be quite brittle if notched, and ductile is smooth. Are you interested in smooth or notched components?

You might want to check out the work by Gearing and Anand on failure of PC. (http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6VJS-49Y3WT9-7&_user=10&_coverDate=02%2F29%2F2004&_rdoc=1&_fmt=&_orig=search&_sort=d&view=c&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=7586844def64d9e54b63bd3b99d789a7).

You can implement your failure condition directly into your ABAQUS UMAT or VUMAT. Note, however, that only ABAQUS/Explicit currently supports element deletion based on a user-defined failure condition.

- Jorgen

[ashu28]

Thanks for the reply Jorgen. Let me explain my work a bit more.

Please look at the following picture:

https://netfiles.uiuc.edu/dikshit2/shared/Research%20Idea.bmp?uniq=-psvzpt

So I would be having a workpiece (CNT reinforced PC) fixed at bottom and left edge and then a tool will travel to the left starting from bottom at a particular depth of cut. So initially the workpiece doesnt have any notch in it. Thus I would need a ductile failure criterion (am I wrong ?) to fail elements which then I would delete.


Do you think that I need the brittle failure criterion once the tool advances and fails some of the elements (because now there is a notch or a cleavage in the workpiece) ?

I am unable to find a ductile failure criterion. Also how would I code 2 failure criterion ?

/Ashu

[Jorgen]

Looks like a cool project. I agree that initially the failure is likely ductile. I am not sure what the failure mode will be after initiation. As you mentioned, you can always code in two failure conditions. If the failure is brittle then the brittle strength will be lower than the ductile strength, and vice versa.

There is no problem to implement multiple failure conditions, since only one will be active at time.

-Jorgen

[ashu28]

Hello Jorgen,
I went through the literature and found one brittle (notched) and one ductile criterion. As you know I wrote VUMAT for PC. I have somewhere close to 80k nodes in my material (CNT reinforced PC). In order to speed things up and converge in practical time, I simulated a lot many stress strain plots at different strain rates and fitted the *Elastic, *Plastic options to those curves. This way I got elastic plastic option calibrated for PC at different strain rates upto large strains. Thus I dont use my VUMAT anymore. Everything works through the input file only.

The question I have now is where should I code this failure criterion ? Which file ? I am really not sure where should I write these equations. I would really appreciate your comments and also can you give me a couple of example codes of some failure criterion if you possess any ?

Regards,
Ashu

[ashu28]

Hello everyone,
I am trying to run a machining simulation and am using *elastic, *plastic option with tabular values. Also I am using *SHEAR FAILURE, ELEMENT DELETION=YES with a value of 0.6. I want to study the behavior of workpiece when the tool moves in with a specified depth of cut. I have defined a fine mesh at the top near the tool. I define surface to surface explicit interaction between the tool and all the nodes the tool might come in contact with. The workpiece is fixed at left and bottom edge and tool moves in from right. Please have a look at the following picture.
http://netfiles.uiuc.edu/dikshit2/shared/Simulation/Input.bmp

There are a couple of issues I am having in this.
1. When I view my result ( http://netfiles.uiuc.edu/dikshit2/shared/Simulation/result.bmp ), I see that some of the element have lost stiffness and still attached to the workpiece ( they seem to be flying all over the place). I suppose they are failed elements, so I go to Create Display Group-->Result Value-->(change Field Output to PEEQ) --> Change the type to |--> and enter 0.6 in the Min value tab and after this, press Remove tab at the bottom. I observe that some of those elements that were flyin are removed from the viewport but still I see some which have PEEQ ( equivalent plastic strain) < 0.6, having undergone a very heavy deformation ( http://netfiles.uiuc.edu/dikshit2/shared/Simulation/plasticonly.bmp).

2. I defined adaptive remeshing for elements near the top region of the workpiece but I dont see any changes in the mesh taking place with time.

The result of all this is that the simulation stops and I get the following error in the .log file:

***ERROR: Zero or negative mass in element number 21221 of instance PART-1-1
***ERROR: Excessive distortion of element number 21221 of instance PART-1-1
Step Summary Diagnostics for Adaptive Meshing: Step 1
Note, only adaptive mesh increments in which at least one node is moved
are considered in calculating these quantities.
***ERROR: There is only one excessively distorted element

I am not sure why am I getting this error. What does zero or negative mass imply and why is this excessive distortion error popping up. Is it for the elements that have crossed the failure criterion or some other. How can I get rid of this.


I would highly appreciate if someone can help me out in here. I have uploaded my .inp file as well ( http://netfiles.uiuc.edu/dikshit2/shared/Simulation/mac7nodecontact.inp )

All these files along with .ODB can also be downloaded from ( http://netfiles.uiuc.edu/dikshit2/shared/Simulation/MachiningSimulation.zip )

Eagerly waiting for some advice from ABAQUS users.

Ashu

[Jorgen]

You are asking great questions. The first thing I would try is to use a newer version of ABAQUS (v6.7 just came out). I believe they added better output capabilities for simulations with element deletion.

- Jorgen

[ashu28]

Hello Jorgen,
Thank you for the reply but unfortunately I dont have ABAQUS 6.7 in the university. I have only 6.6 but right now I am running simulations on 6.5. I will try to run them on 6.6 and update you on the results.

Btw, I observed that the elements undergoing excessive distortion are the ones which have not yet crossed my failure criterion(PPEQ less than 1.0 when failure criterion was *SHEAR FAILURE, ELEMENT DELETION=YES at 1.0). I am not sure as to why are they getting so much deformed. This is what is messing up my simulations. Have you ever carried out simulations of PC wherein you allowed strains to go as high as 1.0 ? I am sure it should be ideally fine, but something is preventing my simulations to run properly.

/Ashu

[Jorgen]

I have done many simulations of thermoplastics deformed to large strains. Typically these simulations work fine, but any time when you have true strains of the order of 1.0 or more then you run the risk of experiencing convergence problems due to element distortions. Are the elements that are cauing your problem well formed?

- Jorgen

[ashu28]

Thanks Jorgen for the reply.
To start with I have a simple brittle failure criterion:



Pressure=p1, temperature=t1
Pressure=p2, temperature=t2
Pressur=p3, temperature=t3

Is there a way to express it in input file. I would be glad if I can get away somehow using some of Abaqus's inbuilt command to express this condition in input file, as I dont wanna code this in UMAT.

Please let me know if it is possible.

Regards,
Ashu

[Jorgen]

I typically use various failure conditions that I have built-in into my UMAT/VUMAT subroutines so I am not quite sure about this, but as far as I recall you cannot directly specify a failure condition in terms of the pressure.

- Jorgen