Hello Nigel,

We do not see failures in B16.9 Welding Tee's (there were quite a few before the Code reduced the allowable stress for 335, P-11 (1 1/4 Cr, 1/2 Mo) material - due to creep-fatigue interaction). We do see failures in fabricated branch connections (reinforced and non-reinforced) but we think these reflect welding workmanship more than anything else.

Something that should be considered is that the B31 Codes DO NOT SPECIFY requirements for Tee's, so the Manufacturers of Fittings (Tee's) DO NOT (are not asked to) comply with the B31 Codes. The B31 Codes refer to ANSI Standard B16.9 for standard welding fittings. It must be understood that VERY LITTLE geometry is standardized by B16.9. End-to-end (and centerline-to-end) dimensions are standardized, and the "squareness tolerances" of the fittings are standardized - nothing else, including crotch radii, thickness or shape is standardized. If you compare the shape of several Tees's that DO COMPLY with the requirements of B16.9, you will see some of them are quite "spherical" at the "cheek" area, some are barrel shaped and some are very cylindrical on all their "legs". But they all are B16.9 Tee's.

The B16.9 fittings (e.g., Tee's) "must be able to be shown" to be as strong (against internal pressure loading) as the straight pipe welded to them - i.e., the pipe will burst under pressure before the mating fitting ruptures. The wall thickness is not standardized throughout the fitting but the manufactures are required to match the wall thickness (schedule) at the weld line so there will be good fit-up (minimum mismatch). Most fitting manufacturers have traditionally made fittings that are thicker than needed to pass burst tests and they line bore them to match the mating pipe schedule.

Most of the SIF's that are (still) in the B31 Codes are those that came out of the Markl (and colleagues) experimental work that was the basis of "new" Code fatigue rules in 1955 (B31.3 and the B31 Mechanical Design Committee are actively working to update the rules). Markl's work was done at the facilities of Tube Turns Company, so the testing used that company's fittings (and all the fittings and pipe were NPS 4, standard schedule, A-53, Gr B). Since NO TWO MANUFACTURERS MAKE FITTINGS THE SAME SHAPE (except to comply with B16.9 Standards) the tests and the fatigue data (and SIF's) that resulted from the testing could, strictly speaking, be applied only to the test sponsoring company's product. It should be recognized that given the WIDE variation in the shapes of the fittings (especially Tee's and reducers) provided by the manufacturers, the Appendix "D" data may or may not be accurate. B31 Codes have always told the designer that if he/she had better data to apply (these days it would be from FEA) then the designer was free to employ this data. Having no better data, the Code SIF's (and FF's) are better than nothing. Some time after the original fatigue rules (of Appendix "D" etc.) were included in the Code, words were added (in several "phases") to adjust for crotch radii, pressure stiffening, flanged ends, weld-on and weld-in branch fittings, etc. As an aside, The manufacturers of proprietary weld-on and weld-in fittings have had to perform their own tests and determine SIF's for their products (they will provide these for you if you request them).

The flexibility factors (FF's) for Tee's were always (even before Mark's quantification) known to be other than 1.0 (when compared to straight pipe), but they were left at 1.0 for various branch connections. That is USUALLY conservative.

The B31 Codes will be changing to reflect a more contemporary understanding of how various piping components perform in service. I am confident that as the B31 Codes adopt new rules, Caesar will exactly reflect these new rules.

Just my opinion.

Best regards, John.