RE: Non convergence error while attempting to remove element

[ryancwl95]

Hi all,
I am attempting to install a pile into a soil body using 8-node SSPbrickup elements. My first step in the analysis is to create the nodes and elements for the soil body. Then, I use the initial state analysis to run a gravity analysis. After this is completed, the initial state analysis is turned "off." I am using an Elastic Isotropic soil model for the elements. The gravity analysis was successful and gave reasonable results.

In the second analysis sequence, I tried removing the soil elements (pile zone) and created new nodes and elements (SSPbrick) to model the embedded pile. Then, I connected the nodes intersecting between the pile and soil using equaldof. After that, I ran another gravity analysis.

However, I encountered an error in the 2nd analysis sequence. I suspect that the error occurred when I removed the elements. I attempted to run the 2nd sequence without removing the soil elements in the pile zone but continued with the pile installation, and it worked. This the error:

Screenshot 2024-08-07 225851.png (18.48 KiB) Viewed 3100 times

I also tried removing 1 element and then performing pile installation and a gravity analysis (this worked). I gradually increased the number of removed elements followed by pile installation & the gravity analysis. Out of these attempts, I was only able to remove a maximum of 5 elements. I tried removing elements at different parts of the pile zone first, but I couldn't complete a successful analysis of removing all the soil elements in the pile zone, installing the pile, and then conducting the gravity analysis.

Some images of the mesh:

Picture1.png (61.61 KiB) Viewed 3100 times

Initial soil body:

Screenshot 2024-08-07 230216.png (182.43 KiB) Viewed 3100 times

Successful 2nd analysis sequence gravity analysis, removed only 5 soil elements in the pile zone, followed by pile installation and gravity analysis:

Screenshot 2024-08-07 230548.png (136.09 KiB) Viewed 3100 times

Do you have any advice regarding the observed error?

[STKO Team]

Can you share your STKO file?

[ryancwl95]

Hi,

I am using STKO to post-process my analysis, so I don't have the SPD file. Would the text/tcl file for this analysis help? Where can I send it to?

[STKO Team]

Well, debugging a plain tcl file can be hard.
However, if you have troubles (crashes or singularities) after removing elements, make sure of the following points:

• If you remove a set of elements, those will have nodes shared with other deleted elements, and nodes shared with elements still in the domain. Make sure to remove the hanging nodes (those that do not belong to any element after their removal)
• Before removing the nodes, make sure to first remove any SP/MP constraints and/or Loads!

[ryancwl95]

Well, debugging a plain tcl file can be hard.
However, if you have troubles (crashes or singularities) after removing elements, make sure of the following points:

• If you remove a set of elements, those will have nodes shared with other deleted elements, and nodes shared with elements still in the domain. Make sure to remove the hanging nodes (those that do not belong to any element after their removal)
• Before removing the nodes, make sure to first remove any SP/MP constraints and/or Loads!

Thank you for your response. The convergence error is no longer an issue since I have removed the elements followed by removing the "hanging nodes." However, I met with this error instead (shown beow). After this error was displayed on the command screen, the program immediately crashed.

Picture2.png (57.72 KiB) Viewed 3048 times

Could you explain to me this error a bit more? I am unsure what the problem is, as I have removed the hanging nodes. Your advice would be helpful.

Can I also confirm that when you say SP/MP constraints, you are referring to the fixities placed on the "hanging nodes" for the initial soil body gravity analysis?

[STKO Team]

Could you explain to me this error a bit more? I am unsure what the problem is, as I have removed the hanging nodes. Your advice would be helpful.

It says that a SP (Single-Point) constraint (fix for example) is applied on node 1, but node 1 is not in the domain anymore. You probably delete it without deleting the constraints acting on it

Can I also confirm that when you say SP/MP constraints, you are referring to the fixities placed on the "hanging nodes" for the initial soil body gravity analysis?

SP constraints = Single-Point constraints, constraints applied on single DOFs (fix for example)
MP constraints = Multi-Point constraints, relating multiple DOFs/Nodes (equalDOF for example)

[ryancwl95]

Subject: Assistance with Analysis Using 20-8 BrickUP Elements

Hi there,

Thank you for the help so far. I am currently attempting to perform the same sequence of analysis using a 20-8 BrickUP element instead of the 8-noded SSPBrickUP element for the soil materials, and a 20-brick element instead of the 8-noded SSPBrick element for the pile material

So far, I have successfully run the initial gravity analysis with just the soil body.

I then removed the soil elements with their respective nodes and SP/MP constraints, and I managed to run a dummy analysis after the removals of these soil elements (to be replaced by pile elements).

Next, I want to replace the soil elements I've removed with pile material, which I have modelled using the 20-brick elements. After generating the pile elements, I tried to run another dummy analysis but encountered an error.
Could you provide some insights on how to resolve this issue?

Screenshot 2024-11-20 001910.png (61.56 KiB) Viewed 2834 times

The second question I have is about naming the 20-8UP and 20-node brick elements. Could I confirm the correct way to number the nodes when creating these elements? Because they seemed like they are reversed (top and bottom) for 20-8UP brick element vs 20 noded brick element.

20-8UP brick element :

Screenshot 2024-11-20 002310.png (9.17 KiB) Viewed 2834 times

20 noded brick element:

Screenshot 2024-11-20 002144.png (103.77 KiB) Viewed 2834 times

Thank you!

[STKO Team]

You did the model manually or with STKO?
because it says "Negative Jacobian". It happens when the node numbering is reversed, thus the negative volume