I can see your point of view. If the re-pad is so wide that it covers the area of high stress around the nozzle (opening) then stress beyond the re-pad would be typically membrane stress in the shell/head.

If I remember from WRC-368, T here is re-pad thickness plus vessel thickness and R is mean vessel radius. So, it has to be a very wide pad to see this case. We probably do not need to compute the stress at the edge of the re-pad in this case.

WRC-368 ignores the re-pad if it is less wide than 1.65* SQRT (RT).

Another point, if re-pad is so wide, it may not behave as one unit with shell. Are you adding re-pad thickness to the shell thickness?

Can you share your dimensions?