Boundary conditions in the case of a pseudo static analysis

Question

Hi, I am doing pseudo static analysis in earth dam to calculate the yield acceleration for different geometry and property configurations. I have questions about the boundary conditions to calculate this value, because my ky_crit change depending if my lateral and lower Line Displacements are 'Free' or 'Fixed', I don't know which situation is more similar to the real world. 
 Thanks!

Faseel Khan · Answer

With the line displacement set to Fixed , do you specify any value or just keep it as 0?

Faseel Khan · Answer

In a pseudo-static analysis, conventional static displacement boundary conditions are considered, consisting for example, of restraining the horizontal displacements along the lateral boundaries (X-fixed) of the model and both horizontal and vertical displacements (XY Fixed) along the bottom boundary of the model. Applying a line displacement is not a pseudo-static analysis as it is similar to applying a load/displacement at one end of the model. Whereas in a pseudo-static analysis, the acceleration coefficients are applied to the weight of the mass for the whole mesh simultaneously. You can read through the following articles for more information and guidelines on the pseudo-static analysis: Fundamentals of pseudo-static analysis in PLAXIS Pseudo static acceleration

Faseel Khan · Answer

In a pseudo-static analysis, interfaces are not required at the boundaries of the model. In order to use the options compliant base and free field the manual creation of interface elements along the full model boundary in Structures mode is required. By creating an interface we create a so called node pair which allows for the necessary decoupling to be able to apply the input motion (load history) while also being able to absorb incoming waves. The lateral boundaries are usually fixed in the x-direction in 2D. This is to assume that boundary effects (stress/strain) are minimal and are not affecting the results and that the model boundaries are sufficiently far enough to cause any impact on the computed stress/strains. The fixity in the x-direction is much closer in fulfilling this assumption compared to the fixity in xy-directions and therefore a fixity in x-direction in 2D is usually used. If your model boundaries are too far already then fixing the xy-direction at the boundary will also be the same as fixing the xy-direction.