Example 9.1 Transverse damage crack density UGENS & Python script in Abaqus for laminated composites

[@Pishgul]

Prof, Barbero, Would you please let me know whether failure criteria discuss details of the kind of failure in composite materials? For example, does hashin gives us the type of failure?

[@EverBarbero]

In www.amazon.com/exec/obidos/ASIN/1138196800/booksoncompos-20 page 232 after eq. (7.5) "... maximum strength FC gives information about the mode of failure...", same for maximum strain FC, then for the "interacting FC" in section 7.1.4, you get either fiber or matrix failure, but you don't know if matrix failure is due to shear or traction or both, same deal for Hashin FC. There are other considerations in chapter 7, such as "7.14. nonlinear behavior," see after eq. (7.14), "7.2.1. in situ strength", "7.3.2. truncated maximum strength" also tell you what mode of failure is active (fiber tensile, fiber compression, or fiber shear). I think TMS is best because if "failure means ultimate, that is, the laminate breaks" then, matrix failure is not an issue "only when the fiber breaks the laminate breaks" unless the LLS is not well designed (also discussed in Ch. 7 and elsewhere in examples that textbook).

[@Pishgul]

@@EverBarbero You are great, man

[@Pishgul]

Dear Prof. Barbero, In order to apply any failure criterion for composite materials, what should be considered? Could we use any criterion for our analysis? I want to compare my experimental results (type of damage) with the results given by failure criterion in abaqus. I am not sure which one I should apply.

[@EverBarbero]

Failure criterion and damage models are not the same. I think in chapter 7 of Intro to Comp Mater Design ISBN 978-1-138-19680-3 has a good explanation of the limitations of Failure Criteria, and Chapters 8 and 9 of FEA of Comp Mater Using Abaqus ISBN 978-1-4665-1661-8 the same thing for damage models, but anyway, failure and damage criteria are very specific for type of materials, type of loading, fatigue or not, temperature or not. It is not an easy question to answer. It would be nice to have "guidelines for failure criteria to use" and "guidelines for damage models to use," but they don't exist yet.

[@Pishgul]

@@EverBarbero Professor, would you please let me know which case is longitudinal shear and which one is transverse shear in your book page 229? I think F5 is longitudinal and F4 is transverse, right? What about F6?

[@refquesog8104]

Thank you very much for this video. i tried to test the example 11.1 with the thermomechanical fatique damage. i am getting the error: This error may be due to a mismatch in the Abaqus user subroutine arguments. These arguments sometimes change from release to release, so user subroutines used with a previous release of Abaqus may need to be adjusted. I think the Subroutine.for needs adjustments, because i am using Abaqus 2022. Any hints how to solve it?

[@Pishgul]

Professor sorry for lots of questions. You mentioned that "one has to be careful with the solution strategy " due to the material nonlinearity. Therefore, in the increment section you defined corresponding values. How should we do that? How do you know which increment size and time and etc we should zpply?

[@EverBarbero]

trial and error. experience.

[@Pishgul]

Professor, did you apply UGEN subroutine here? I do not understand the part with 30 parameters exactly. GIc,GIIc,deltaT,E1,E2,G12,nu12,nu23,alpha1,alpha2,thetak,lamina thickness,lambdak,D2,D6, and 3*N state variables. These are 15+9 = 24 parameters

[@EverBarbero]

Yes, Ex_9.1 is done with UGENS. The DDM plugin requires 3+9N parameters and uses 3N state variables, where N is 1/2 the number of laminas. The laminate is assumed to be symmetric, so we enter only 1/2 of the LSS. The first 3 parameters are: GIc, critical ERR mode I GIIC, critical ERR mode II Dummy (for future use) Then, 9 parameters per lamina (7 material properties, plus the angle and the thickness). E1 E2 G12 Nu12 Nu23 Cte1 Cte2 thetak tk Total: 3+9N parameters (material and geometrical). Internally, the DDM plugin uses 3 state variables per lamina (crack density, transverse damage, shear damage) for a total of 3N. In Abaqus, state variables are denoted by SDV. In the results for this example, SDV1 is the crack density for the bottom lamina, SDV4 for the center lamina, and SDV7 for the top lamina. The 2 damages are not true state variables, only crack density is a state variable, but we keep D2 and D6 as state variables to simplify the computations, internally.

[@Pishgul]

@@EverBarbero Professor thanks for explanation. Therefore, 0.254 is assigned for GIc and 100000 is assigned for GIIc, right? In he mechanical constant sections

[@Pishgul]

Since you did not mention in this video about adding subroutine in file I think. I don't understand the .obj file since it would not be opened in notepad by the way I think I confused

[@EverBarbero]

yes, correct.

[@EverBarbero]

The code is here: barbero.cadec-online.com/feacm-abaqus/Examples/Chapter_9/Ex_9.1/Abaqus2020/ abaqusddm-std.obj is the plugin that the script puts in the Job. It is compiled for Abaqus 2020. You need visual studio and Intel Fortran installed properly for Abaqus to link the .obj file to Abaqus itself. It is done automatically when you click Submit.

[@Pishgul]

Professor barbero, in my created .inp file there is a number 3 before the section shell thickness 2.736. I think it is related to integration points in that section. You did not have it in yours by the way. Did you delete it? Also, From which part of the code abaqus understand that there is 4 sdvs? I have 15 sdvs in my result

[@EverBarbero]

When I did the video maybe I left the default Simpson points at 5 so nothing would be written after 2.736. If you change the Simpson points to 3 then it is written in the input file AFTER 2.736. It is better to use 3 Simpson points because one is at the top of the lamina where the stress is either highest or lowest, another at the bottom where the stress is lowest or highest, and a point in the middle that is no good for anything. If you put 5, then 3 points are a waste. The distribution of stress is linear anyway. There are 9 SDVS, Abaqus knows because in the input file it says VARIABLES=9

[@abdelraoufboukar8323]

Prof, Barbero why when I apply it for +45/-45 lamiante it does not work

[@EverBarbero]

DDM requires the laminate to be symmetric, like (+/-45)_s. Also, a +/-45 or (+/-45)_s is a bad choice for tensile load, please read about the 10% rule in my textbook Into to Com Mater Design.

[@abdelraoufboukar8323]

Hello Prof Barbero I used the DDM in two symetrique laminate the first (0/90)3s and it's work perfect but for the (+/-45)3s lamiante it does not work.

[@civilalgeria8087]

شكرا جزيلا أستاذ. بارك الله فيك و جزاك عنا كل خير

[@EverBarbero]

you are welcome!

[@Pishgul]

Hello again. Would you please create a video also about first ply and last ply failure please?

[@engrshairkhan]

Thanks a lot for your time and efforts.

[@EverBarbero]

It's my pleasure

[@Pishgul]

In my created .inp file there is number 3 after thickness of 2.736 which I think is about the integration point in section part. You do not have this in video. You deleted it or...? The job-1.inp has nodes and element numbers in my file but it does not have in yours. Again you deleted them and they do not have any effect? Also, in your book you did not mention about rotation boundary condition I think

[@EverBarbero]

2.736, 3 is correct My file has nodes and elements. See the video at 9:04 (9 minutes 4 seconds)

[@Pishgul]

Great video

[@EverBarbero]

Thanks!