Fix `BoundaryModelDummyParticles()` for FSI

[LasNikas]

• [x] Incorrect placement of boundary wall for multiple layers (solved in #42)
• [x] continuity_equation!() dispatched with SummationDensity is needed for the solid-fluid interaction as well. ( #59 )
• [x] Differentiate movement function for moving boundaries (#101)
• [ ] provide solid particle acceleration for AdamiPressureExtrapolation (#262)
• [x] Interaction between different particle spacing (#59 )
• [x] update solid container in update3!() (#59 )
• [x] hydrodynamic mass for solid particles (#59 )

[efaulhaber]

We could just implement setups with analytical density profiles and run the simulation without relaxing part.

Keep in mind that with the higher density at the bottom of the tank, the particles are not just closer together in y-direction, but also in x-direction. In these examples, we can't really see that, but I assume that this is the reason for non-uniform particle distributions. It is therefore not easy to determine the correct positions analytically.

[svchb]

@LasNikas This seems to be a small issue left open. Can you finish this?

[LasNikas]

see #262 Note, that the simulation falling_spheres_2d.jl needs about twice the number of time steps with #262 @svchb @efaulhaber