Stanweld;
It seems the manufacturer of the base pipe does supply the SA-335 pipe, in straight sections , with proper material properties as per SA-335. The hot bending fabrication error occurs at the downstream foundry that uses a hot bending machine to fabricate the bends, and many of these foundries appear to be ignorant of the need to N+T the entire section of piping to correct the sections that were overtempered ( ie, zones adjacent to the local 1900 F induction heated zone).
Generally, what seems to happen is the EPC vendor obtains competitive quotes from several foundries for the bent els, and selects the lowest bidder. The bent els are then purchased , under the generic spec of SA-335, but without addending to the purchase spec a QC requirement to perform a hardness traverse of the bent piping or mandating a N+T. It often happens that the lowest cost foundry doen't have a staff metallurgist, and being unaware of the P91 material degradation during overtempereing will merely follows code stated requirements and would only add the neccesary N+T if stated in the purchase spec from the EPC vendor. As I read B31.1, the responsibility for correct fabrication procedures is the "designer's", not the foundry.
Until adequate corrections are made to the code and ASTM specs, it will be neccesary for the EPC specification, and the ultimate customer's specification, to explicitly require the N+T for hot bent P91 els and/or addend the same wording as Section I is proposing in note 4 of case BC05-1483.
Other equally common and equally damaging errors are (a) insufficient rate of cooling following N+T- it must be cooled faster than -9 F/min to avoid ferrite formation ,and (b) the max permitted thickness transition angle may not exceed 30 degrees at the OD in the vicinity of butt welds- this is explicitly defined in the codes, but lately been ignored by reputable designers, to the detrement of the end user.
