Some comments:
1) You're missing some load cases. You need to calculate W+T1+P1 by itself, and you need to subtract it from W+T1+P1+F1 as an OCC case, and then you need to add W+P1 to it also as an OCC case. This is obligatory.
2) If you then run a restraint summary extended, you'll see displacements for this node in both cases. Your pipe is too rigid to set back down onto its support after it's thermally grown upwards, even with your thermal growth applied. The proximity of the closest upstream support is the culprit, here, and you'll find that in your current model, that's likely where all your relief loads are being transmitted to.
3) For a conservative restraint load, you can also run W+P1+F1, which will be roughly the same as W+P1, with F1 applied to it. The numbers may vary based on how the various centers of gravity allow the pipe to warp.
4) You can also apply a stiffness factor for all supports, which will allow them to all sag, but possibly maintain contact with all supports. However, this can be a tedious process to input manually. It could also be done by adding structural elements to your model. It all boils down to time versus money versus expectation of realism.