[glik]

I would like to use two differents buoyancy values (in x axle), depending on the height (y axle).

I tried something like that :
if(y<9[m],g1,g2)*if(y>0[m],1,0),
but of course it doesn't work.

Does someone have an idea plz ?

[ghorrocks]

This sounds like an XY problem to me..... (http://xyproblem.info/)

Can you explain what you are trying to do? Also what do you mean by "change buoyancy"? I have no idea what the variables g1 and g2 mean.

[glik]

Hello ghorrocks, thanks for helping.

In fact I would like to simulate a tank which is submitted by an earthquake, but in a static problem, and using Housner (impulsive and convective modes-accelerations).

So I have 1 acceleration from the bottom of the tank to the three quarters to the height (for example), and a second acceleration from the three quarters to the top of the tank.
But my problem is, these accelerations are not in the verticale axle.

Is it clearer ?

[ghorrocks]

Isn't the image you show the forces applied on the tank by the fluid, not the acceleration on the fluid? So this is not a varying acceleration, but a varying force - and a force applied to the tank.

How can you model a transient earthquake using a static model? You can do modal analysis in FEA, but CFX does not support this.

[glik]

No it shows the acceleration of the earthquake.
Accelerations g(impulsive) and g(convective) are my input datas.

But maybe I can do something with the force, by calculating it from the acceleration.

And I just want to submit the fluid to these accelerations, and show the maximal pressure under the wall.
So it's not necessary for me to model a transient earthquake.

[ghorrocks]

Why not do the obvious thing and use moving mesh in a transient simulation to move the tank following the earthquake displacement versus time? Then the fluid will do its motion and generate the appropriate pressures on the tank walls.

This is the standard way of doing this sort of simulation.