วีดีโอ

No worries! Happy to help out. Let me know if you have any questions!

Most welcome

Hello, I think this video can help you: th-cam.com/video/gCQzzF8Few8/w-d-xo.htmlsi=KIzys71PZdDqOteg

Not, really, no video on comparing the analytical approach. This is relatively straightfoward and I did this in this publication of mine: doi.org/10.1016/j.commatsci.2012.12.036 This paper should give you the answer you want.

@ thank you Dr Let me check it out

Hello, thanks for the question. By external libraries, I assume you mean for example user-defined material subroutines, python libraries, etc. If that is the case, then this is not straightforward. It cannot be answered easily here. I will suggest you do your own research about this.

@@MichaelOkereke Thank you for your response. I am currently developing a UMAT subroutine to predict the nonlinear behavior of materials. In part of the code, I need to solve a system of linear equations ‘Ax=B’. Initially, I used Fortran libraries such as LAPACK and PARDISO to solve the system. However, these libraries do not provide accurate solutions for the equation’s roots. In contrast, MATLAB computes these roots precisely. To address this, I established a connection between Fortran and MATLAB to calculate the equation’s roots within the Fortran code. Currently, the Fortran code runs without errors in the Visual Studio 2019 environment and produces results consistent with my previous MATLAB implementation. However, when I integrate the Fortran code into the UMAT subroutine, Abaqus cannot access MATLAB libraries during the analysis phase. To resolve this issue, I conducted extensive research and made modifications to the abq2024.bat, win86_64.env, and custom_v6.env files to define the path for MATLAB libraries in Abaqus. The MATLAB libraries required for execution are located in the following path: C:\Program Files\MATLAB\Matlab Version\extern\lib\win64\microsoft. I tried adding this path both through Windows environment variables and by defining it explicitly in the .env files of Abaqus. However, the error persists and is displayed as follows: error LNK2019: unresolved external symbol engopen referenced in function umat.R error LNK2019: unresolved external symbol mxcreatedoublematrix800 referenced in function umat.R error LNK2019: unresolved external symbol mxgetdoubles800 referenced in function umat.R error LNK2019: unresolved external symbol mxcopyreal8toptr800 referenced in function umat.R error LNK2019: unresolved external symbol engputvariable referenced in function umat.R error LNK2019: unresolved external symbol engevalstring referenced in function umat.R error LNK2019: unresolved external symbol enggetvariable referenced in function umat.R error LNK2019: unresolved external symbol mxcopyptrtoreal8800 referenced in function umat.R error LNK2019: unresolved external symbol mxdestroyarray800 referenced in function umat.R

@@MichaelOkereke Thank you for your response. I am currently developing a UMAT subroutine to predict the nonlinear behavior of materials. In part of the code, I need to solve a system of linear equations ‘Ax=B’. Initially, I used Fortran libraries such as LAPACK and PARDISO to solve the system. However, these libraries do not provide accurate solutions for the equation’s roots. In contrast, MATLAB computes these roots precisely. To address this, I established a connection between Fortran and MATLAB to calculate the equation’s roots within the Fortran code. Currently, the Fortran code runs without errors in the Visual Studio 2019 environment and produces results consistent with my previous MATLAB implementation. However, when I integrate the Fortran code into the UMAT subroutine, Abaqus cannot access MATLAB libraries during the analysis phase. To resolve this issue, I conducted extensive research and made modifications to the abq2024.bat, win86_64.env, and custom_v6.env files to define the path for MATLAB libraries in Abaqus. The MATLAB libraries required for execution are located in the following path: C:\Program Files\MATLAB\Matlab Version\extern\lib\win64\microsoft. I tried adding this path both through Windows environment variables and by defining it explicitly in the .env files of Abaqus. However, the error persists and is displayed as follows: error LNK2019: unresolved external symbol engopen referenced in function umat.R error LNK2019: unresolved external symbol mxcreatedoublematrix800 referenced in function umat.R error LNK2019: unresolved external symbol mxgetdoubles800 referenced in function umat.R error LNK2019: unresolved external symbol mxcopyreal8toptr800 referenced in function umat.R error LNK2019: unresolved external symbol engputvariable referenced in function umat.R error LNK2019: unresolved external symbol engevalstring referenced in function umat.R error LNK2019: unresolved external symbol enggetvariable referenced in function umat.R error LNK2019: unresolved external symbol mxcopyptrtoreal8800 referenced in function umat.R error LNK2019: unresolved external symbol mxdestroyarray800 referenced in function umat.R

Most welcome

Yes, correct. It is ABAQUS style. Material reference frames are 123 when they align with global reference frame XYZ. I have set the model up so that this is the case in which case 1=X, 2=Y and 3=Z. If you orient different so that the fibre axis (currently Z) lines up with the X-axis, then you would have the fibre modulus as E11 as traditionally reported in literature. In the case here, my fibre axis is Z-axis so E33 would be equal to E11 and the two transverse directions i.e. Y- and X- axis becomes the 2- and 3-material axis leading to E22 and E33 as reported in literature.

I really like your good comments TonyABbeyFETraining. I agree with all of them. I know the message about modes are incomplete. I just put something together quickly for illustration and it could be better. I did not make any effort to justify the predictions with analytical values (especially for simple pillar models). If that was done, then it could be a first step to justify/verify the solution before attempting a complex system as a thin-walled cylinder (in 3D). Thanks for all the insights you gave. I will look at them more. As regards triggering necking, you are right. I really like that video but it does not seem to draw a lot of audience to it since the points made in it is a bit nuanced. I am happy that there another nerd in my community like me in these things. Thank you and of course I will check out your channel.

Hello, This is a good question. I think it can be done. You do a similar thing but just make sure you specify it as a surface (i.e. in terms of x and y values for example z = f(x,y)). The analytical fields can specify expression in terms of x and y. I have not tried it before but I think this is possible. Maybe in future I will explore this.

Thanks

Thanks

Yes... you use RVE to generate properties for your complex geometry problem.

@@MichaelOkerekeSir, I further need a clarity on it. For example my structure is a honeycomb panel consisting of many cells. Direct honeycomb structure could be employed for RVE analysis? Or RVE properties from a simple rectangular domain should be utilized as input for bulk structures analysis? One more question is how the varying weight percentages of nanoparticles is defined in the matrix? Thank you for your time and consideration to this.

What does this mean?

Meaning?

Meaning?

2:16

You need to read the odd file with a python script and see stress in select elements.

Use the approach I use in my model using a referral point note and kinematically link it to the edge. Apply your load on that node and extract your history output from that reference point node.

Thanks

No idea of how to do so... sorry!

You could be right... I need to check carefully... the principle remains though.

Sorry can't help

No ideas about this

Sorry no idea of how to do this

Thanks

Glad you thought so!

I will imagine the localised approach since you are targeting the strains at specific locations on the beam.

@MichaelOkereke thank you for your reply. But how could I choose section points, bottom and top of the shell e.g., and select them in history output so that i can plot their strain values. I'm thinking something like two strain gages fixed back to back in a given location of a panel

I think you can, but I'm not sure, definitely, as I do not use it.

Hello, this video from my channel will try to give you sufficient answers to this question in more details for you: th-cam.com/video/YeCBDkjEmDU/w-d-xo.html

It's on the base of the TH-cam short where you watched it.

It is better you contact me via email by joining my CM Videos Insider Group through this link: cmvig.cmvideos.org where you get a welcome email from me, and from there just get in touch.

What do you mean by May? Am not sure!

@@MichaelOkereke which software?

Which part? Please explain the question.

Please go to the about us page on my TH-cam channel page. You will see the address.

I suggest you look at Google scholar on papers about this. I am not able to help now.

You're welcome

Thanks for the query. The fibre I used (E-glass) does not show anisotropic properties. If you want to use an anisotropic property for say carbon fibre, then by all means go ahead. You have to slect the right model for anisotropic properties modelling from ABAQUS material library.

i do not believe the excel file was part of this. I have videos on the channel on pbcs and stress-strian which you can extract thhe files there.