tempering parameter for PWHT of P22- hardness vs ductility

[davefitz]

We have multiple cases of craze cracking near the butt welds downstream of attemporator spray nozzles at our HRSG's. The hardness of the zone near the P22 butt weld is consistently harder than the parent metal, and I suspect this is related to the lower ductility near the weld zone. The lower ductility would lead to increased vulnerability to craze cracking.

Other recent published correlations for P91 "tempering parameter" indicates that for P91 there is a LMP larson miller paremeneter that correlates PHWT Temp, time, and hardness , and I suspct this same parameter can predict ductility for P91 and other materials , including P22.

Kindly advise if there is a known tempering parameter for P22 that correlates PWHT temp, time and hardenss. In lieu of this parameter I expect we will simply double the PWHT min dwell time over the code-required minimum time in order to improve the ductility for those butt welds immediately downstream of a spray.



"Whom the gods would destroy, they first make mad "

 

[davefitz]

one published recommendation by Cedric Chauvy for P22 is : LPM=T*(20 +Log(t)) , range of acceptble LMP = 19.61E3(min) - 20.72 E3( max), for example a thin wall can be PWHT at 710 C for 1 hr ( min holding time ), while a thick wall ( 250 mm) shell may be held at 690C for 33 hrs ( max) . Some correlation with low temp charpy impact energy is provided, but the particular issue of craze cracking at the spray outlet my be notch toughness in the temperature range 300-565 C, unless the cracking is occurring during outages at RT due to residual stresses caused by prior thermal downshock at elevated temperatures.

"Whom the gods would destroy, they first make mad "

 

[davefitz]

It looks like it is time to forward this to a real metallurgist. The relationship of LMP to UTS and YS may be parallel to the hardness readings, but not to charpy impact energy at RT. In fact, a longer PWHT dwell time ( larger LMP) , while softening the alloy, does not impart better ductility but lowers the charpy impact energy at RT- this is the wrong direction insofar as reducing the tendecy for cracking in a thermal shock environment. Perhaps the zone of cracking, identical to the zone of PWHT, is due to loss of mechanical properties due to PWHT. TBD

"Whom the gods would destroy, they first make mad "

 

[weldstan]

There are ample examples of cracked joints immediately downstream of the attemperator sprays in P22 to P91 welds. Have you performed thermography to determine temperature and temperature deltas about the circumference. P22 is subject to temper embrittlement as well.

 

[metengr]

Kindly advise if there is a known tempering parameter for P22 that correlates PWHT temp, time and hardenss. In lieu of this parameter I expect we will simply double the PWHT min dwell time over the code-required minimum time in order to improve the ductility for those butt welds immediately downstream of a spray.

Yes, there is a tempering parameter for Grade 22 steel. It is well defined. As mentioned by stanweld, P22 can become thermally embrittled with loss of ductility. However, before I would venture down this path or become involved with paralysis by analysis, you should have this confirmed by proper met analysis - sample removal. It sounds to me this could be thermal fatigue.

 

[davefitz]

Meteng,
Yes,it is thermal fatigue, but in several cases the cracking only develops near the butt weld zone ( or reheat zone) that was PWHT'd ( +/- 4" from weld line), and hardness readings show that zone uniformly has a higher hardness than parent metal further away.

We will send the damaged pipe sections to SI and request a determination of impact energy for both parent metal and the reheated zone and an analysis of carbides in each, perhaps to determine if there is a preferred PWHT procedure that would better retain ductility so as to avoid the fatigue cracking. The major thru wall cracks are circumferential , while a large number of randomly oriented craze craks are at the ID.

Thermographs of typical spray stations ( reheat dump to ACC, interstage reheater )show tremendous thermal gradients ( axial and circumferntial) downtream of the spray nozzles, implying bending moments in the piping that would be consistent with a circumferential failure.

The Chauvy paper includes a plot of toughness vs LMP ( used as a tempering parameter, calculated for PWHT hold times), showing a sharp drop off in toughness at an LMP=19500 for P11, and I suspect the same is true for P22. If this loss of RT ductility is the cause of the major failure cracks, then the residual sresses are formed at operating temperatures ( due to excessive thernmal gradients )and the cracks are occurring during outages.

"Whom the gods would destroy, they first make mad "

 

[weldstan]

Was the weld locally heat treated or heat treated in a furnace? If local there will be a considerable range in the tempering temperature(s) experienced.