I do not understand well the part of there being two gravel beds. Is it that the intent is to bury the lower tubulars in gravel and then start with the following layer of tubulars as in the previous row?

To be more specific I uploaded a new sketch. On the sketch you will see a top view and a front view.
In the top view you will notice 2 gravelbeds. The tubulars are imbedded in the gravel for 1000 mm.

Gtz
Johg

Forgot the sketch

• http://files.engineering.com/getfile.aspx?folder=330f6303-fe8a-4bd8-9acc-d4

My rough viewing of this, as you are planning and assuming the 450000 kg is total load, it will progressively fail unless the gravel is cemented together, as with a weak concrete.  Those two narrow rows of gravel are woefully inadequate.  The least one might need would be much wider rows.

Why not cable to together some timbers on the outsides of the bottom rows, keeping the group from spreading laterally?.  The gravel can remain as a soft bedding.

Geotechnically I would charter the round tubes as round spheres of very significant unit weights and simply do some graphical diagram of what goes on with arrows for direction of force and length of arrows as the forces.  Then lateral pressure on the gravel zone must be resisted by figuring passive resistance knowing angle of friction and unit weights there.

This is something like figuring where pool or billiard balls will go when struck by another ball.

Respect the equilibrium of forces and for a perfect geometry you have to consider say two of the tubes more to one side, the one above placed with some impact. Symmetries will eliminate the other cases. The relative position of the centers and wedging may affect the point of application of the upper load. For a rigid and stout enough bridging, the load could be being applied from above. Note also that if the bridging is to be applied quite locally an exam of the strength and deformation of the relative support is required (inner case as well, of course). Once satisfactory you enter your rolling-sliding problem. Preferably the reaction of the upper cylinder should be applied in a line passing through the center of the lower cylinder to avoid tendence to roll. A FEM model can give you the shape of the elastic reaction (or whatever material model you assign to the soil). In any case will discover inclined preassures exerted by the lower cylinder first in the gravel layer and then on the soil, so you need to ascertain that for the inclined load the pressure is within allowable limits. You have in Bowles formulas for bearing capacity under inclined loads. What I have suggested be examined through FEM can be directly on the resultant providing equilibrium to the named scheme, with less precision locally and computed general equilibrium. I also agree in that the bands look quite minimal by your design, anyway all this thing is amenable to "deterministic" evaluation of the required width of gravel once the allowables for inclined loads in the gravel and soil are established.

As a theoritical exercise it is very interesting. as a practcal matter, the stability of such stacks is difficult to model. Such stacks can and do roll. Occasionally an unfortunate worker gets caught in them. Do not use the gravel beds. Use concrete barriers or even better, driven vertical piles to contain the tubes. These stacks are dangerous and people do get hurt working on them.