Generalized Hardening Soil in Local Seismic Response with Soil Structure Interaction ( PLAXIS 3D)

Dear All,

I hope you are all fine.

I am experiencing some problems during local seismic response analysis with soil-structure interaction in Plaxis 3D (coupled approach). In particular, all materials were modeled using Hardening Soil Small Strain with condition Undrained A. The problem is that the seismic signal undergoes heavy deamplification as it passes through the ground. The result is that minimal acceleration (I would say illogical accelerations) arrives at the base of the structure.

For this reason I did some research and found this document "">communities.bentley.com/.../site-response-analysis-and-liquefaction-evaluation". In this document rather than using the classic HS ss, is used the Generalized Hardening Soil. In particular, the GHS allows you to select different ingredients from the classic HS ss. In this paper it is selected that stiffness is a function of the INITIAL stress state and not the CURRENT stress state. If I replicate the same choice in my analyses I have no deamplification issues outlined and the problem seems to be solved with a very nice profile of PGA. However, this solution seems to be illogical because the stiffness MUST be a function of the current stress state and not just the initial stress state.

So, here are my questions:
1) First of all, I would like to know if there is any manual or notes about the "User Defined" material model called "Generalized Hardening Soil". The only information I have found is in the link I wrote about ( see above)

2) Why in undrained conditions  the stiffness should be a function of the initial stress state only and not of the current one if we use GHS?

3) A further doubt is this. I'm working with Plaxis 3D with materials in undrained conditions (Undrained A), do the boundary conditions of type "Free Field" work well? If not, how can I simulate correct boundary in my analysis?

Thanks to everyone who will help me out.