The code does not tell specifically what has to be considered to calculate and check the redistributed sustained stress due to non activeness of some supports in operating condition, an issue popularly known as "hot sustained stress". This stress can be checked accurately in the computerized analysis by using the concept of non-linear supports. But there is always a question on the allowable. Hot-sustained stress has got the characteristic of both primary and secondary stresses ( that is what I feel). Primary because it is a dead weight redistribution effect. Secondary because it is related to temperature and will be back to the original value when the pipe sits back after the thermal cycle. There is a school of thought that if the pipe sags under the wt.and falls back it is a case of self limitation (perhaps on this line of thought the French Petrochemical code ,I presume CODE T1 qualifies this as a secondary stress). Such a case is possible only if the lift off is small and stresses are in the elastic regime. A higher lift off and the resulting sag should be a case of plastic deformation and hence cannot make the stress redistribution justify it as self-limiting.
2004 edition of B31.3 addresses this issue indirectly in the Appendix P by consideration of "operating stress check" .In my a opinion the code committee should come out with more specific information on this question. Maybe a sustained stress check with an allowable of 1.33Sh (to keep these stresses within yield point and not allow a limit load type of failure) can be considered.This again is my personal opinion only. 1.33Sh is to keep the computed stresses within yield point.

Regards