Dear,
If the inputs are correct, your combined model is more accurate than separate ones.
When you are considering only the six DOF displacements & SIFs at branch point for seperate flare analysis, the analysis ignores the effect of six DOF force-moment generated from this analysis on the main header analysis. If you go to-and-fro between these separate analyses with displacements from main header & force-moment from branches, these errors will reduce.
regards,
sam
_________________________
_