Example 3.4.d How to model a laminated composite using a Composite Layup in Abaqus

[@이재호-n4f]

Thank you for the detailed analysis procedure. This is really helpful to understand your book, I bough it, in detail. I have one question about mesh quality in multi-layered plates. As shown at 12:31, The horizontal length seems a lot longer than the vertical height. In general, mesh quality such as aspect ratio is very important to make the analysis result reliable. But the thing is if I refine mesh dimensions to 1:1:1 (width, height, depth), there are too too many meshes which makes the analysis time too long. In short, is it OK to ignore aspect ratio in the thin-multi-layered wide plate?

[@EverBarbero]

I love this question. Note that the shell is discretized using "continuum shell elements" (CSE), not standard "continuum elements" (CE). If CE were used, the aspect ratio limitation would require a very fine mesh, no more that 10 times the thickness of each layer, but CSE circumvent that limitation by using a constrained 2D internal kinematics, namely linear deformation through the thickness of each lamina (constant intralaminar shear) and zero transverse strain epsilon_zz). Once you stack them, they behave as ZigZag elements. Essentially, CSE "look like" 3D solid elements but behave as First Order Shear Deformation (FSDT) shell elements in each lamina. See E.J. Barbero. 3-D finite element for laminated composites with 2-d kinematic constraints. Computers and Structures, 45(2):263-271, 1992.

[@shravanbhadoria4104]

Thanks Prof. Barbero. Why is that sometimes the results differ for continuum shell element and hexahedral element for the same laminate problem (1 order of magnitude difference in displacement)?

[@polopolimero8376]

¿Existen programas libres para este tipo de cálculos?

[@Omar.a.b.b]

Thank you for this video. I did the same procedure but in the end, I got an error! "Error in job Job-1: 2800 elements have missing property definitions. The elements have been identified in element set ErrElemMissingSection" Could you please help me to solve this error because I also got the same error when I tried to model a scarf joint from a paper.

[@ranulfocarneiro5251]

Hello Dr. Ever, thank you for the video. What does mean "element relative thickness"? Why did you put "1" for all plies? What's the difference between the ply thickness and element relative thickness?

[@civilalgeria8087]

Many thanks, excelent work

[@EverBarbero]

Thank you too!

[@amrabdullah1995]

thank you professor barbero for the great video. i have a question please. i am strengthening a pipe by 4 FRP layers. 2 in the transverse direction 90 degrees and 2 in the longitduinal direction 0 degrees with respect to pipe longitudinal axis. now should i use a lamina of 4 plies with different fiber angle orientation as you did here ? or should i use a separate ply for each one then tie them together using tie constraint in interaction step? which is more realistic. taking into consideration that each FRP layer is unidirectional. thanks alot

[@EverBarbero]

Method 3-4(d) is the most complex, only to be used if you already know that interlaminar stresses s13 and/or s23 are large. If your pipe is subject to internal pressure, the load will be in-plane (as a membrane) and unlikely to generate s13, s23. Even if you have bending, a pipe is a think walled structure, still works as a membrane. I can see using 3-4(d) if you have impact loads, or if the laminate is very thick, or if you need to transition to a 3D solid region, then 3-4(d) is the way. Otherwise, 3-4-a, b, or c, are much simpler and usually give good results.

[@amrabdullah1995]

@@EverBarbero Thank you professor for the great help.

[@sirajudheench8182]

Example 3.3 is available in youtube?

[@CP-zi3eg]

Prof. Barbero thank you for you very instructive videos. I would like to ask if there is other way to input layered section point than to do this by hand. I find it is very risky, in case of many layers or very complex geometry, to forget some number or input wrong one.

[@EverBarbero]

Have you seen the videos for examples 3.4.a, 3.4.b, and 3.4.c. They provide easier ways to define the laminate stacking sequence (LSS). Another way is to use a python script but that in itself is very hard to learn. I don't have videos for scripting. Also, look at the reply to the next question kzbin.info/door/OnbXkWC_LRHtXK1Rze1EQA

[@CP-zi3eg]

I elaborate a bit more my question. In Edit Field Output Request, there are some options for layered section points. One of these is "Specify" where the sequence of integration points is input by hand. I wonder if this task by hands is necessary given that , during the composite layup process, the number of integration points has been already specified (it is set 3 as default ). At the exit of Composite Layup, I would expect that the program has all information to extract all the numbers I input in the "Specify" field above. What am I missing ?

[@sadiqamin6425]

I have copied the same procedure, but job aborted, it says Error in job Job-1: 2800 elements have missing property definitions. The elements have been identified in element set ErrElemMissingSection.

[@EverBarbero]

I don't know. You must have done something different than the video or in a different version of Abaqus.

[@stevenongszekuan9932]

Can I ask how do you pivot one side of the model? where the load will only applied on the opposite side. Thanks

[@EverBarbero]

Engineering constants are "lamina properties" like moduli E1, E2, etc. and strength F1t, F1c, etc. These you find by testing the manufactured "lamina." If you do not have the lamina made but you are trying to design it, meaning to choose a fiber, a matrix, a fiber volume fraction, etc., then you need to use en.cadec-online.com/

[@EverBarbero]

Boundary conditions to restrain the model are shown at time 13:26 in the video. Is that what you want?

[@turanturan2831]

Prof. Barbero, Thanks you for these videos. I just was wondering why we need partition here? we can just go with the composite layup, right?

[@EverBarbero]

Please compare Ex.3.4.c. and Ex.3.4.d. they are different.

[@binbeo5855]

Could you please answer why you use SWEPT MESH instead of STRUCTURED MESH. And why the element type you use is Conventinum shell? Thank you and Best regards.

[@EverBarbero]

In part (d) I use Continuum because I want to demonstrate all options and by the way, the continuum option is the most accurate and most time-consuming. As for meshing, it is a "try and see" situation; you use whatever option that gives you the results you want.

[@binbeo5855]

@@EverBarbero Thanks a lot

[@anguspotter5667]

Could you do this with a thermal load on one side instead please?

[@EverBarbero]

You can do it for a steel plate, find the exact solution somewhere, and play with Abaqus until you get it, that’s the way I learned a lot of things.

[@sayedghalleby2522]

Can you please model machining on composit materials 3D

[@EverBarbero]

Sorry, I do not know "machining".

[@sayedghalleby2522]

Can you explain how to model damage evolution in VUMAT

[@caiomenezes4147]

if there are 8 layers of 1mm and the applied force is 100N/mm the calculation should not be 12.5Mpa and not 10Mpa