I normally calculate B31 expansion stress range as an algebraic difference between two operating cases. So what should I do if the operating cases have different weight and pressure componenents?

Example:
L1 W+T1+P1 Operating below ambient conditions with liquid
L2 WNC+T2+P2 Steam Out
L3 W+P1
L4 WNC+P2
L5 L1-3
L6 L2-L4
L7 ?

L7=L1-L2? Includes variations in weight and pressure. Maybe this is right since they are cyclical. Seems very conservative though since steam out is far less than 7000 cycles.

L7=L5+L6? A scalar stress combination of two algebraic cases? Or is this a case for an absolute or SRSS stress combination?