Some considerations, not necessarily recommendations-
-Can you control the expansion through use of guides and stops
-Does material and number of cycles permit the use of stress range factor greater than 1 as per latest B31.3, if this is indeed your code.
-Are the design conditions realistic
-Are you using the liberal stress range
-If so, does sustained stress play much of a role, eg what happens when you vary the 'sustained stress SIF multiplier'
-Have you treated the branch as an anchor with applied displacements? If so, modelling more of the header may reduce calculated stresses.
-Make sure you consider the full temperature range from Max to Min temperatures, and multiple operating scenarios if significant
-Is there a corrosion allowance and if so have you considered 'all cases corroded'
-Have you considered the WRC 329 SIF
-Can you use a thicker wall, or a welding tee or a weldolet
-Discuss all this with your lead
-I find you can almost always get a line rerouted if need be
regards
Martin