Learning Forum

[Belliveau18]

Hi, I'm modelling a discontinued long fibre composite (DLF) in Ansys apdl for my master’s thesis. The model is based on a standard composite tensile test. Thus, virtual tensile stiffness (modulus) and tensile strength can be predicted. The goal of this model is to compare the tensile properties of DLF composites with UD fibres and woven fibres.

3D mesh is used with solid185 elements and the Puck failure criteria are used to predict failure of the elements for the UD properties. However for the Woven properties traditional Puck failure criteria cannot be used. In the ACP help manual, it’s noted that the puck failure criteria can be used for woven with the ply type set to the woven specification (which is done in engineering data, in workbench). Since I'm using APDL, is it still possible to set ply type to woven composites?

[Sean Harvey]

nAs far as I can see, the ply type set to woven is only used by ACP in the failure calculation and not by APDL. I will confirm and get back to you. nYou can find the woven specification for puck in the engineering data (which you probably already found), but I do not see any specifics in the theory. Let me see if I can get some details. That will also help us answer if this woven version can be used in APDL. I do not believe (unless you use the userfc subroutine). ACP failure and APDL failure are separate. If you needed it, and could use it in ACP Post, that is great. If you needed it in APDL because you are going to run progressive failure, then that could be a reason to need it there. With that said, can you clarify if you are using ACP or APDL?.Thank younSeannn

[Sean Harvey]

nThe developer who worked on this is not available this week, but next week I can circle back with more details. Thank you.nSeann

[Belliveau18]

nThank you for your responce!nI'm using ansys APDL.nYes exactly, I have found the woven specification in workbench engineering data. The reason for the question is that I'm currently develloping a non traditionnal FEA analysis for discontinued long fibre composites. The analysis is a progressive damage model using de MPDG method with the puck failure criteria in ansys APDL. The APDL software is used instead of ACP since I need to create random orientation of DLF chips inside a tensile specimen (3D model). I have tried to use ACP in the past for this project but was unsucessfull because of certain parameters in ACP such as plie dropoffs (not required for my specific application). I have attached a picture of my model in this comment, to help define my application. To define the orientation of the chips elements are selected (based on the chips dimensions and randome location) and a coordinate systeme is define with a randome orientations. nIf plie drop off can be turned off in ACP I could possibly switch, as the algorithm can be transfered to a Python shell in ACP to create a solid model from a surface instead of an solid model in APDL. Do you now if ACP post can predict progressive failure in the material or is this limited to ansys APDL and mechanical? And do you know if ply drop off can be disabled in ACP?nHowever if this could be applied in APDL it would save me some time to rewrite the code. Also I have already validated my model in APDL with UD chips with experimental results.nThank you again for your help!nn

[Sean Harvey]

nI was away from work for a week, and need to circle back to get more details on the implementation for woven and if that can be shared.nRegarding ply drops in ACP, have your tried checking the boxes in the solid model drop offs tab? See below.nnWith ply dropnnDisablednnThank younSeann

[Sean Harvey]

Hello Array nI wanted to circle back. I obtained some further information from a colleague. So Alfred Puck (of the Puck failure criteria) describes in his book Festigkeitsanalyse von Faser-Matrix-Laminaten that his failure theory can also be used for woven fabrics by splitting the fabric into two UD layers. Unfortunately the book is written in German with all the details and we can not translate and provide here, but that is the reference.nNow with that said, Puck requires at least 20 material properties to be used for woven, and these will be more sources of unknowns. and I might consider starting with a much simpler failure criteria. I hope this helps!nnRegards,nSea

[Belliveau18]

nThank you for the information, it was very helpfull. nAs for the ACP model, I did had not notice that you could do that. I will definitly explore that option.nFor the Failure model I have tried to use the Max stress model, however since I need to represent very complicated mode of failure this was insuficient. Meaning that the model did not match the experimental data. However, I will be exploring the ACP model more in depth since Workbench has the direct implementation of the model. nThank you again for your help.nR?jeann