Lateral drifting of whole FLAC2D model during liquefaction analysis

1

Hello

Greetings to Itasca support team!

I simulate soil liquefaction of level ground using FLAC2D & P2Psand model. Attached boundary conditions were imposed on lateral boundaries while the acceleration was applied along the bottom boundary. I find the whole model to drift laterally to a large extent. I request your help in addressing this anomaly.

Thank you

2

Hello @dineshn

This may be an artifact of baseline correction - Dynamic Modeling Considerations — Itasca Software 9.6 documentation
I’m not sure but you may want to check that out.

3

Hi

Thank you for the response.

Baseline correction is applied to ensure zero displacement once the shaking ceases. However, I am observing the drifting of model during the shaking. Would you like to check my datafile?

4

Hi @dineshn

I am involved with projects where I apply Norsand and PM4Silt cmodels to liquefaction analyses in FLAC2D. I at times observed some lateral drift, but it is largely minimised now by my my current workflow ensuring baseline correction and filtering to the applied ground motions. Although i must note I employ compliant base theory in my models where stresses (via velocities) are applied at the base rather than rigid base theory with accelerations applied at the base.

I can have a look at you datafile(s) and share some observations based on my experience.

2 Likes
5

You may want to check again your baseline corrected motions to make sure you have a zero displacement.

1 Like
6

This does sound like a boundary condition issue:

  1. check bottom boundaries - I assume that if you are applying an acceleration directly, you have a rigid boundary where you want to fix both x and y.

  2. check your motion - first trim (if you are trimming) then baseline correct. Check that your baseline correction does what it’s supposed to do. Sometimes filtering alone is enough, sometimes you need to do some order of a polynomial-based correction. Chekc acceleration, velocity, displacement time histories.

  3. Ideally, actually, and if this one of your first experiences in dynamic modeling with liquefaction, before 1 and 2 make sure you have tested /validated smaller models with some sort of idealized motion (e.g. a ramping up and down sinusoidal one)

2 Likes