I am trying to model the effects of a subsidence event (due to mine collapse under pipeline) on a buried pipeline. I have modelled the pipeline and buried it and have data on the changed soil profile after the subsidence, but am unsure how to modify the restraints to reflect a lowering of the ground level.

Perhaps I can somehow use the different stiffnesses, K1 & K2, on the bilinear restraints Any help would be greatly appreciated.

The way to do this is to put CNODEs on your bi-linear restraints, then displace the CNODEs.

How would you set up the load cases in this situation?

That depends on the code in use. Typically support settlement would be added to the OPE case only, so its effect would be seen in the expansion case (as a single non-repeated anchor movement). However, I'm not sure a "mine collapse" was the intent of this particular "code" statement.

Would you CNODE all three bilinear restraint directions or just the vertical restraint?

A test I did gave different axial stresses in each case.

The CNODE displacements will only work in the direction of the particular restraint so using the same CNODE for the X, Y & Z restraint is unnecessary if the ground only moves in the Y direction. Only the Y restraints require the Y displacements.

If you add the CNODE to all three restraint vectors be sure to properly define the CNODE displacement set. I would guess that you are leaving the X & Z fields blank on your CNODE. A blank entry for the CNODE means that node is free to move in that direction. As such, the X & Z restraints no longer supply load to the pipe since the other ends are free.

Either specify X & Z displacement as zero or do not use the CNODE on these restraints.

One other point that may be important... CAESAR II ignores deadweight in the buried model since the pipe is continuously supported by the bottom of the trench. By dropping the bottom of the pipe, our buried assumptions are violated.