I create a custom report for displacements summary, then bring it into Excel.
Then subtract node X displacements from node Y.

Considering directionality, I assign compression, tension, torsion, and calculate lateral displacement and bending.

Then I check every OPE load case to ensure the expansion joint's permissible individual axis movements haven't been exceeded, and then combine them to ensure their combined displacements don't exceed permissible.

It's not a bad idea to have an intermediate step of finding the load cases with largest individual components in the 4 directions and combine those into a fictitious scenario.

Doing so is more conservative than looking at each case individually.

Will it always be conservative for every piping routing conceived?

No.

There are an infinite number of pipe configurations imaginable that will result in higher deflections between T(install) and T1 that isn't accounted for.

For "reasonable" routings? Likely.