At the risk of sounding snarky, I would suggest that the use of CAESAR here would be better put at redesign and replacement of the line rather than the acceptability of a line that has already failed.

But as an academic exercise...
What are your displacements based on?
Is it sitting there, operating, hot, and they measured those displacements?
Did they shut off the line and then measure it?

The "right" model would be the one that would look like what would happen if you turned off the pipe right now and let it cool - with all its ugly effects and calculated forces, and then ran the model to happen to "arrive" to the present displacements if you turned the pipe on.

Edit to add:
A number of supports that have run off would look something like a +Y with gap, but for the shoes that might act as directional stops - only one for each linear run of pipe will actually function as such - and that'll be a substantial force in the sustained case - and it might have already permanently moved entire sections of lines, causing more problems and more slips.

So I doubt this will be a clear-cut answer since you have to perform forensics on what happened once the first shoe failed by slipping off, and this could have happened over more than one cycle.