Table E-1, B31.8 is defining SIF for butt weld joint by following cases:

Case 1:
T> = 6.02
mismatch_max =< 1.59mm
mismatch_avg/T =< 0.13
--> sif = 1.0

Case 2:
T> = 6.02
mismatch_max =< 3.18mm
mismatch_avg/T = any value
--> sif = 1.9 max or sif = 0.9+2.7*mismatch_avg/T

My questions:
a. Which the case is the common use in sif calculation?
b. How to know mismatch_avg value?
c. I see that case 2 is more conservative. However, if using sif = 1.9 max, stress at weld joint will be very high. I will prefer to use the equation 0.9+2.7*mismatch_avg/T. Can I use mismatch_max = mismatch_avg = 3.18mm in case 2 for sif calculation?

Still looking for advice ... Any comments?

a.
I would argue that most butt welds are for identically specified spool pieces, and therefore 99% of butt welds are treated as case 1.

Case 2 is an exception to the rule where you may have two different piping specs used for a given service, but where the transition can be made up within the weld. I would hazard a guess that most cases you require a transition piece. This transition piece would fall under case 1.

b.
You should know when you have mismatching piping specifications, thicknesses, etc. Average should be at spec.

c.
In the event you are stuck with a high SIF with stresses, you should opt for a transition piece.