Back to original post of threeouts, in my opinion the load cases should be written:
(OPE) L1=W+T1+P1
(OPE) L2=WNC+T2
(SUS) L3=W+P1
(SUS) L4=WNC
(EXP) L5=L1-L3 [=T1-70F=130F; disp range from instal. to max metal temp]
(EXP) L6=L2-L4 [=T2-70F=-90F; disp range from instal. to min metal temp]
(EXP) L7=L5-L6 [=T1-T2=220F; full disp range max to min metal temp]
L5, L6, L7 are Algebraic combinations.

Another method:
(OPE) L1=W+T1+P1
(OPE) L2=W+T2+P1 [to use only for cases 6, 8 below]
(SUS) L3=W+P1
(SUS) L4=WNC
(EXP) L5=L1-L3 [=T1-70F=130F; disp range from instal. to max metal temp]
(EXP) L6=L2-L3 [=T2-70F=-90F; disp range from instal. to min metal temp]
(EXP) L7=L5-L6 [=T1-T2=220F; full disp range max to min metal temp]
(EXP) *L8=L1-L2 [*optional case for full disp range max to min metal temp]
L5, L6, L7, L8 are Algebraic combinations.
This method can be used for T2 at above or below zero deg C.

Both methods above (linear/or non-linear analysis) will provide results almost exactly the same in example 1 of B31.3 appendix-s.