Good day
My query concerns the selection of an appropriate weight parameter "w" when modelling a plate element in Plaxis 2D. Does Plaxis automatically account for buoyancy for plate elements that are below the water table? In other words, for a submerged plate element, must the specified weight parameter "w" be based on the equivalent buoyant unit weight of the plate material, or the unsubmerged (i.e. "dry in air") unit weight?
Many thanks
Dear Sam,
PLAXIS takes buoyancy effects implicitly into account in the finite element calculation.
For the unit weight of the plate, if you consider it as non-porous then you can use the dry unit weight.
When it comes to using the plate element, the unit weight of the soil should be subtracted for the unit weight of the plate, as plates are superimposed on a continuum. If the plate is submerged, then the unit weight of the soil to take into account is the saturated weight.
Dear Stefanos,
I did a simple check with Plaxis 2D and I am pretty sure that Plaxis 2D does not account for buoyancy for plate elements.
I put a horizontal plate element on top of a soil layer and activated it in a phase 1. Water level is defined way above. Whatever number I type in as w (plate weight in kN/m/m) it equals exactly the reaction stresses in the soil elements below the plate. Whether the plate is submerged or not the weight seems not to change at all.
So, to account for buoyancy it is actually necessary to input the buoyant plate weight. Please correct me if I am wrong.
Dear Nataly and Sam,
You are correct as to that PLAXIS does not take into account the buoyancy effects for plate material.
This means that the calculation kernel does not adjust the unit weight of the plate based on the elevation the phreatic level is.
Therefore, for taking into account the fact that a plate is submerged, the user should manually adjust the unit weight to consider the buoyancy that exists in reality.
The above is added extra to any adjustments needed for the case of thick plates which overlap with the soil, as mentioned correctly before.
Sorry for the confusion caused regarding this. Indeed, the best solution is always to create a simple model to understand how the program performs the analysis.
Thank you, Nataly, for spotting this inconsistency!