Discussion Forum

Hi Carlos

do not worry about mistakes, probably you can also write in spanish, apart that I would not manage to follow that easily, would try in all cases though ;) Anyhow, i.e. typo's arrive for all of us.

Coming back to your problem, a support of your ballast thank that behaves as a spring:
a) either you define a soft material under your thank such that the E modulus is not too high and you will have a soft spring (but you might have to turn on large displacements if its really soft); or
b)you create a spring by equations over a given "contact" area A, as a force load/pressure on the boundary, and set the boundary conditions, with a constant gravity load
If gravity is along -Z you define the constant G0=1[lbf/lb] (this is a shortcut to have COMSOL to write out the 9.81... value), then you define the subdomain force as a gravity load Fz=-G0*rho_smsld (_sm.. whatever application mode name you are using), and finally over th supporting area you define a force as Fz=-ks*w, where w is the deformation along the Z direction calculated by COMSOL, and ks is the spring constant, that you define as a Constant value before adding it on the surface load case. Note: surface loads are defined as pressures, i.e. force over area in [Pa], so you must know the total contact area and normalise your forces to get the right pressure values. T e more precise, you should consider perpendicular forces, or local normal forces, or define a calindrical coordnate wth r=sqrt(x^2+y^2) and express your foces with this.

How to dimension the first case ? well if the tank has a mass of m[kg], acting on a contact surface A[m^2] and a young modulus of your soil material of E[Pa], (let us assume nu=0.3 the poisson constant) on ground with a gravity acceleration og 1G=9.81[m/s^2], you can do the following hand calculations:

the spring stiffness in normal compression of a rod (or a large bulk structure) is :
ks[N/m]=E[Pa]*A[m^2]/(H[m]),
where H is the height of the material of section A, then you say that the total sag under gravity load acceptable is "z0" we have ks[N/m]*z0[m]=F[N]=m[kg]*9.81[m/s^2].

As probably your tank is a cylinder with a horizontal axis, your normal is not parallel to G all over the contact area, so you should take a security factor of 2 or 3 on your sag value z0.

With this you can estimate a E[Pa] Young modulus, that you can use in your simulation for the bulk material, assuming you know the surface A and you have selecteed a value H at your convinience.

Now knowing that you are form Guatemala, I imagine (have never been down there, I'm from the far north) that your earth is soft and muddy type, so the surface area might change with the penetration of the tank into the soil. This means that the surface area is not constant, and that you must consider contact elements, in "Assembly mode with contact pairs. A perfectly possibl situaton, but the explanantions become reather long, so I would suggest that you read through then the chapter related to contact behaviour and changing surfaces. You can also use plain equations with bolean operators, but this is also slightly tricky.

In any case good luck, and keep us informed how it progresses
Ivar