I guess that, to create each hole, you first create a cylinder and then use ASBA to subtract from the plate. Then repeat with for/end.

My alternative: create first all cylinders with for/end, then select ALL and finally use ASBA to obtain the holes. No loop on ASBA.

Dr. Benasciutti,

Thanks for your response. I really appreciate that you understood my question and provided me a response worth trying out. I am still facing issues with ASBA command described below. It seems that I have too many entities in my model to operate with ASBA.

Example: Old Script

/PREP7

BLC4,-22,-22,44,47 #Make area (plate)

*DO, J, 1, 67

*DO, I, 1, 67

CYL4, -19.9 + 0.6*(I-1) , -19.9 + 0.6*(J-1), .2 #Make area (holes)

ASBA, 1,2 (subtract as we go along)

NUMCMP, AREA (re-numbering)

*ENDDO

*ENDDO

In the new method, which you suggested, I executed

/PREP7

BLC4,-22,-22,44,47

*DO, J, 1, 67

*DO, I, 1, 67

CYL4, -19.9 + 0.6*(I-1) , -19.9 + 0.6*(J-1), .2

*ENDDO

*ENDDO

After creating all the areas, I went to Component Manager, selected all the areas for all the holes and made a component named "hole".

After that, I executed,

ASBA 1, hole

Here 1 refers to the area (plate) and the hole refers to the component containing all areas.

I get an error that the new entity created is the same as the old entity and the process is terminated.

Start simple: try a for/end loop with just 1 hole, then 2 etc, to see what's going on

Thanks for your response. I tried debugging this problem in multiple ways.

When I use loops, the initial process is fast but as the holes increase, it just slows down eventually stops responding. For about 150 holes, it is quite fast but after that it becomes slow.

There is no difference in the characteristics of holes if they are first few holes our to 3600 holes (e.g.) or they are just a few holes. The way they operate is the same.

One potential solution I feel is using the GUI. When I used GUI, I was able to select around 300 holes each time and ASBA command successfully executed. If I chose holes more than those, it threw errors.

I can potentially use GUI, but I am not aware of any way to automate the selection in GUI using the command line. It seems that the GUI is disconnected with the command line operation of APDL.

I am also trying to get in touch with ANSYS but apparently, they are not interested in working out a solution for problems in their tools. I might have to move to another tool.

Thanks for your help and suggestions!

Final suggestion: for a simple selection of holes in APDL, before creating the holes you can set the numbering, for example, at 10000, so all the new entities (holes) will have numbers 10001, 10002, etc

Hello Dr. Pedrazzi,

Thanks for your comments. I tried the suggestions you gave. Apparently, there is some issue with the ASBA command because it doesn't work after almost 800 holes are created in a plate. I tried the method suggested by you and ANSYS stops responding after 3rd command with 250 holes in each command. I contacted ANSYS support in this regard, they reported it as a bug to their developers but it led to no resolution as the developers have stopped working on the geometry aspect of APDL. I am just worried that automating such a process in any other tool which they have bought or in any other software such as ABAQUS might not be feasible.

Also, by GUI, I meant SpaceClaim/Workbench and not the APDL GUI. I should have been more clear.

Thanks for taking out time and writing your suggestions.

Regards

Mohit

Dr. Pedrazzi and Dr. Benasciutti,

I should have responded earlier but I got caught up. Finally, here is what I did to solve the problem. There is indeed a way to semi-automate in ANSYS Workbench. It is not as great as writing a script in APDL, but it works and helps to reduce the manual work to a fraction of the total work. There is an option in ANSYS Workbench to make some variables (with dimensions) as parameters and change the parameters from outside the environment (both modeling and mechanical). This way, I was able to introduce the number of holes, the distance between them, their radii all at once as design parameters. I was also able to apply boundary conditions based on co-ordinates, so even if the holes were changing places, there wasn't a need to pick anything from the GUI manually, which saved a lot of time.

I was sad that APDL had a bug which slowed the process down but the Workbench could solve this problem with a lot of head hitting. For the outputs, I simply wrote an APDL script in Workbench which stored the outputs in a notepad file, which I post-processed using a Python script.

Dr. Pedrazzi,

I also appreciate your suggestion to try different modeling techniques but currently as a researcher, I work on developing computationally efficient methods to solve such problems whose speeds are several orders of magnitude faster than FEM. They often involve complex form of mathematics such as Variational Asymptotic Methods. Once the development of a particular feature or problem is over, I use 3-D FEA to validate my work. If I can not model exactly using 3-D FEA, the results would not be that useful for me. That is why I was focused on solving this problem using 3-D FEA only.