P91 welding

[akas1968]

Dear Experts,
4 inch spool of P91 weldment of 5 years age having linear crack on the weld toe. can we cut the entire weldment and do the new welding on the same p91 weld faces followed by SR?
Request to help me out.
Thanks in advance to all experts.

 

[metengr]

Not sure that you would need to cut the entire weld out. The reason I mention this is you could have fatigue crack propagation along the weld toe versus the common Type IV creep damage, which usually happens along the base metal heat affected zone. Removing the entire girth weld is certainly an option but much more work involved.

The second option would be to locally remove the linear defect at the weld toe and continue the excavation around the entire circumference following the weld toe. Once the defect is cleared, verify using surface NDT, perform weld repair, PWHT and perform final NDT.

 

[r6155]

Yes, you can.
Regards
r6155

 

[davefitz]

The weld repair might not solve the long term problem, and it may be recommended to also have the failure analyzed by a metallurgist and/or stress analyst to determine the root cause .

For example , one sort of P91 weld cracking issue now occurring at reheaters of combined cycle plants is at the junction of a thin wall P91 header link pipe, downstream of the interstage attemporator, to a much thicker F22 O-let . Rapid changes in thin wall P91 pipe , in response to alternating impacts of non-atomized spray water droplets, are occurring in the vicinity of the much thicker forged F22 o-let, which may be able to change temperature perhaps 20 times slower than the P91 pipe {rate of change of temperature due to shocks is roughly proportional to the wall thickness squared to the -2 power) . Local temperature gradients of over 250 F at this weld line , combined with a very high number of thermal cycles, can lead to a crack caused by fatigue, and it might not be creep induced. In that particular case, the practical / economical solution might be to use a forged F91 transition piece with a much smaller gradient in the rate of change of wall thickness ( less than 18 deg angle), although the boiler OEM may quote a multi million dollar fix that involves providing the proper amount of straight pipe downstream of the spray. No mention of why the proper amount of straight pipe was not provided originally.

"Nobody expects the Spanish Inquisition!"