Users Forum

ADONIS Category => ADONIS Users => Topic started by: myhk2 on November 19, 2025, 09:02:41 pm

Title: Model convergence
Post by: myhk2 on November 19, 2025, 09:02:41 pm

Hello,
I am trying to model a simple problem where a very large vertical stress is being applied to a steel plate placed on a thin asphalt layer overlying various pavement and subgrade layers.
I suspect the model isn't converging but am unable to identify the source(s) of the issue.

Would you be able to take a quick look at my script and point out glaring problems?

Many thanks.

Title: Re: Model convergence
Post by: Roozbeh on November 20, 2025, 06:55:38 am

Hi,

I made a simple change to your script at line 30. The Nodal Mixed Discretization (NMD) method uses an averaging technique to smooth strain across elements, but when two materials have a large stiffness contrast (steel and soil), the averaging becomes unreliable and produces distorted results. For this reason, I turned useNMD off. Please note that NMD is applicable only to T3 meshes; Q4 meshes do not require any averaging method to improve accuracy.

The revised script is attached.

-Roozbeh

Title: Re: Model convergence
Post by: myhk2 on November 20, 2025, 04:05:33 pm

Hello,
Thank you very much for your help with this.

Best wishes.