Note-4 for table D300 of ASME B31.3 reads, “Where flanges are attached to one or both ends, the values of k and i in the Table shall be corrected by the factors c1, which can be read directly from Chart B, entering with the computed h.” Now, chart B on previous page implies the upper limit for correction factor c1 & flexibility characteristic h as 1 (because the values greater than 1 are not plotted).

For 3” sch40 pipe/pipe bend with bend radius 5D, Caesar-II’s ‘Bend SIF scratchpad’ utility gives following flexibility factors ( In this case h = 1.202, greater than1)
Blank-No flange = 1.373
1-Single Flange = 1.416
2-Double Flange = 1.460

Shouldn’t the flexibility factor (when h>1) decrease from 0-flange to 2-Double flange?

That's a subtle point - that Chart B, giving c1 for various h's, stops at c1=1 and h=1. Good catch.

What you expect makes sense. The added stiffeners (the flanges) make the elbow stiffer, that is, the flexibility factor (k) should decrease as the elbow gets stiffer.

We apply the equation which has no limit on h or c (as IMPLIED by the chart).

I ran your bend through the FEATools processor FEBend to see what FEA would say. While I agree that, technically, the program should limit c1 to a maximum of 1.0, there seems to be little impact here: