I've seen "Free Thermal Analysis" referenced and have very little understanding of exactly what this means and how useful it is to perform.

Do forum members routinely use a free thermal analysis when designing a new system?

Or is it more useful in troubleshooting an existing system that is overly restrained?

Would someone describe the steps or reference where the steps are described in some detail? What I understand is that all restraints other than supports for deadweight are removed, the operating case run. Then the thermal displacements are reviewed to aid in determining strategic locations for restraints.

This goes back to the old days of 60's and 70's.
Imagine a a hot pipe requiring sevaral supports and having vertical lifts between the end points. Wherever there is a lift or drop in the line one or more supports will need to be springs. Caesar sizes these automatically.
In those old days a first run would be "deadweight only" with the springs taken as hangers. There would be no expansion. The load output would tell us what load the springs should take.
The second run would be "free thermal". The springs and surrogate hangers would be removed and only the true rests and supports would remain. There would be expansion but no weight of pipe, fluid, insulation, rigids or anything else. The displacement output would tell us how much movement the springs should accomodate.
From the load and movement and applying rules such as no more than 25% load variation we can select suitable springs from the manufacturer catalogue. If the spring moved down the preloads would be the "deadweight only" loads. If the spring moved up the preloads would be "deadweight only"+displacement[from "free thermal"]x spring rate. This latter rule is to ensure the spring load is at minimum equal to "deadweight only" for any operating or sustained load case.
The model would then have the spring data input and be re-run as a last check.
So you see "free thermal" is only used to select springs. Caesar now does this automatically.

Thank you, I can understand how the free thermal was used to size springs.

Based on another response outside the forum I understand the weighless free thermal run also gives an indication as to whether there was sufficient inherent flexibility in the piping configuration. If the free thermal failed, it would only make matters worse once restraints were added, so an alternate configuration would have to be sought.

"Deadweight only" stresses by comparison with the sustained allowable would tell if there were enough supports.
"Free thermal" stresses by comparison with the stress range allowable SA would reveal any obvious overstress requiring a configuration change although everybody knows it's better to check the loads on fixed supports for equal but opposite "telephone number" loads and also large uplift loads requiring a change.
Once the spring data is in the analysis the "deadweight only" case is clearly a sustained case and the "free thermal" is clearly a stress range case.

Nice reply deira