Analyze failed , returned:-3 when adding springs.

[Zouatine]

Dear Everyone,

I have been working on a four-story model that was performing well until I added some springs at the base to incorporate soil-structure interaction. I defined zero-length elements with six components of stiffness and damping, using the following properties:

Stiffness:

Kx = 2.38E+08
Ky = 2.46E+08
Kz = 2.80E+08
Kxx = 2.18E+10
Kyy = 5.49E+10
Kzz = 5.52E+10
Damping:

Cx = 9.53E+06
Cy = 9.53E+06
Cz = 2.83E+07
Cxx = 8.44E+08
Cyy = 2.50E+09
Czz = 1.43E+09

To ensure the soil remains elastic, I used a zero-length element and applied Steel01 material with a high yield strength (Fy). However, after making these changes, I encountered convergence issues when running the model, as shown in the attached image:

OpenSees_Failed.PNG (166.74 KiB) Viewed 2626 times

I tried even to fix the Z direction thing that will solve the issue but even with that the issue persists. Here is my current model. I would appreciate it if you could take a look and provide any insights or suggestions.

Four_Story+SSI.zip

(375.77 KiB) Downloaded 145 times

Thank you.

[Zouatine]

Dear STKO,

Thank you for your supports in my previous posts,

I have updated the opensees solver in my PC and the problem with my gravity analysis is solved , but i still have the issue with THA , i enconter an issue around 28% like the following :

Failure 28% .JPG (294.38 KiB) Viewed 2580 times

for the sake of clarificaiton i wanted to also show which GMs is running on which processor :

Processors.JPG (217.63 KiB) Viewed 2580 times

here is the model that i am running i just added another GMs ( 8 in total) .

Four_Story+SSI_8GMs.rar

(384.19 KiB) Downloaded 176 times

However, what i find weird is that i run the model speratly (by hand) ,so i took the first 3 GMs (Acc_X0,Acc_Y0,Acc_Z0)and then run them speratly (no parallel processing) , and do the same for all (Acc_X1,Acc_Y1Acc_Z1) .......(Acc_X7,Acc_Y7,Acc_Z7) , i have not obeserved any convergence issue . the model i used for this case is the following ,

Single Set.zip

(378.26 KiB) Downloaded 137 times

thank you again for your support , and waiting for your kind response !

[STKO Team]

When you use OpenSeesMP for parametric analysis you should not use ParallelRCM numberer and MUMPS solver. They are meant to be used in Partitioned analysis