RE: Use of InitialstateAnalysis, unexpected results due to the use of InitalStateAnalysis

[ryancwl95]

Hi all,
I am attempting to install a pile into a soil body using 8-node SSPbrickup elements and there after applying a horizontal load at the top of the pile. I am using an Elastic Isotropic soil model for the elements and have set the permeability of all u-p elements to 1E-5m/s. My first step in the analysis is to create the nodes and elements for the soil body then followed by a gravity analysis (analyze 5 5e6) with large time steps to obtain the initial stresses in the soil. The second step in the analysis will be to remove the soil elements in the pile zone and replace them with pile elements (using 8-node SSPbrick elements); this is followed by another gravity analysis (analyze 5 5e6). The first analysis and second analysis were carried out with InitialStateAnalysis on and then InitialStateAnalysis off after the second analysis. I had also ensured that material: InitialStateAnalysisWrapper was used in the elements when they were generated.

For the first analysis step there was some downward displacement due to gravity and hydrostatic pressure generated were reasonable:

Screenshot 2024-08-15 165600.png (91.84 KiB) Viewed 2466 times

For the second analysis, there was further downward displacement of the soil, perhaps due to the additional weight of the pile. There were also minor upward displacement in the pile. This analysis ends at 2.5e7s. Displacement countors were also stable at the end of this analysis steps. After the 2nd analysis, the model should be stable with no further displacements due to gravity and InitialStateAnalysis was switch to "off " :

Screenshot 2024-08-15 170147.png (162.79 KiB) Viewed 2466 times

Please remember the following information:

The third step in the analysis involved applying a horizontal load to the pile head. A pile load of H0 = 1000 kN was applied to the top of the pile head over a period of 10 seconds, with 1 second (dt) representing 0.1 of load factor 1.
Below is the Uz displacement of the model, there is still some downward displacement being generated at the first second.

Picture3.png (1.7 MiB) Viewed 2466 times

At a time period of 10 seconds, the uz displacement in this analysis roughly doubles. Below is the screenshot showing uz displacement at 10 seconds for the analysis.

Picture4.png (1.59 MiB) Viewed 2466 times

The load H0 seems to have been applied, as there is some displacement in the x-direction at the pile top, so that seems okay for now. However, I'm wondering why there is still some downward displacement generated in the third analysis step after turning off the initial state analysis. Also, the displacements at the end of the second analysis were stable. Shouldn't the downward displacement Uz be very close to 0 (at third analysis), since displacements are reset to 0 once it is switched off (at the second analysis)?

Could you share your thoughts on this and the use of initalstateanalysis? I am using STKO as a post processor only and thus would not be able to share the STKO model file. but some advice and help would be appreciated!

[STKO Team]

It's hard to tell, however there is something I'm not sure about:
In your first stage you just apply gravity and then remove elements.
Then you add other solid elements for the pile. At this point there is a displacement jump between the soil and the pile solid elements. How did you handle that? is there some zeroLength to create an interaction between them? (because zerolength elements account for initial displacement so that initial gap will actually be accounted for)