Does B31.3 require stress intensification factors at reducers like B31.1?

Can someone please explain why this should even be considered? I can only think of two reasons reducers would cause a localized increase in stresses. There are the welds, which are already included in Code allowables. Then there is the discontinuity of section properties, but this is accounted for by computer modeling.

There are two approaches to handling the SIF / Flexibility of a reducer. These approaches are:

B31.3 Approach: This code states that ...
- The flexibility factor is 1.0
- The SIF is 1.0

B31.1 Approach: This code states that ...
- States that the Flexibility Factor is 1.0
- States that the SIF is:
2.0 max or 0.5 + .01*alpha* SQRT(D2/t2)
where: D1 and t1 are the diameter and thickness of the large end.
D2 and t2 are the diameter and thickness of the small end.
Alpha is the reducer cone angle in degrees.

Notes: alpha can not exceed 60 deg
The larger of D1/t1 and D2/t2 can’t exceed 100


All of the codes supported by CAESAR II subscribe to one of these approaches. The codes that state that the reducer SIF is 1.0 don't require you to apply an SIF. However, many people always assume an SIF of 2.0 for reducers, which is conservative.