In case of having plates at the external model boundary (without any other cluster surrounding these plates at at least one side), it might happen that the external waterloads applied on the plates are wrong.
A cluster phreatic level is set in an inactive cluster in between two plates. Note that on the exterior sides of the plates there is no cluster.
The results of the plastic calculation after the phase definition are correct, and the water loads are applied as it was expected.
If however the following calculation phase is set as Safety Analysis or as Dynamic Analysis, it might happen that the external waterloads are applied wrong, as shown in the illustration below: In some cases they can be duplicated on both sides (right wall) or completely removed (left wall).
In order to avoid this problem, there is a workaround available that consists of the use of dry clusters where these “unexpected” loads might appear.
First, introduce clusters around these walls to replace the “empty domains” surrounding the plates:
Then, in your phases, set these clusters to dry.
Now, the following safety or dynamic phases will use the correct and expected external water loads: