Number of times you can PWHT P91 1

[CWIMET]

Does anyone have any experience or preference on the number of PWHT cycles that may be performed on Grade 91? We have a CC HRSG that has had multiple repairs (multiple PWHT's) performed on a P91 stop valve (repair of Small Bore by-pass line, replacement of same, removal and plugging of by-pass with manufacturer's blessing, and lastly repair of a cracked valve seat). I'm feeling the cracked valve seat is indicating we should be looking into replacement at this point.

I understand that chemistry, stress, PWHT temp and process are all significant factors when dealing with P91, but assuming everything has been done properly to this point and hardness ranges are still acceptable, is there an industry standard or general consensus that would limit the time at PWHT temperature? Is there any NDE method recommended or would a boat sample sent out give me any further insight? My refinery experience has me stuck on three repairs and replace, and we've gone that far already.

I can find no references in Code or technical literature and could not find anything in the forum. As a first time user, not sure if this thread would be better located here or somewhere else? Thanks for any assistance that may be provided.

 

[metengr]

There are no maximum number of cycles for PWHT. If surface NDT, hardness results and surface replication have been performed and indicate typical results and base metal HAZ microstructure, there should be no concerns.
I would not remove a boat sample because subsequent weld repair and PWHT will be required.
If you really want to determine the maximum number of times PWHT can be performed where properties are degraded, you would need to demonstrate this via multiple PWHT cycles and destructive examination to determine mechanical properties and microstuctural evaluation of a Grade 91 weld coupon.

 

[stanweld]

The primary negative effects of multiple PWHT of P91 alloy steel valves are derived by the actual temperatures experienced and the localized stresses induced on the valve due to differential heating rates and the resulting differential expansion, especially when local PWHT is performed. Differential heating and cooling in service is also a major contributor to cracking in bypass valves - thermal fatigue. It may well be time to replace the valve and reconsider its design.

 

[Guest102023]

The total hold time (and not the number of cycles, unless limited by other specification) is limited to the time qualified on the PQR (+25%).

Apart from the issue of Ni+Mn content and its effect on lower transformation temperature, some have reported that P91 welds can be overcooked at the factory or during initial field construction, with little or no allowance for possible further PWHT in future. Properties can only decline with repeat PWHT, but the question is how much and after how much total PWHT. Also, hold temperature in the field is less controllable/consistent than we would like.

 

[Guest102023]

EPRI is developing a button creep test for field samples, but I do not know how suitable that is for weld heat-affected zones. Perhaps someone else can fill in that information here.

 

[davefitz]

The lab-based metallurgists claim the parent metal can be PWHT'd many times- assuming the zone under the induction heating coils is not overtempered. Not everyone in the field follows AWS D10.10 guidelines for installing monitoring thermocouples directly under the PWHT coils- tests have shown that in some cases the zone under the induction heating coils can be overtempered during preheat and during PWHT. So, the correct statement is that many PWHT cycles are permitted, if correctly executed.

see online presentations by superheatfgh and also some recent papers by Bill Newell.

 

[Mr168]

Brimstoner: what is your basis for the soak time being limited to PQR +25% for what is presumably a B31.1 connection?

 

[JakeR82]

My company that works with team has found that PWHT up to about 3 times until the hardness starts to fall off in that area. At that point, you have to re-normalize the material. We mainly deal with making piping parts (desuperheaters) for power plants, which are rolled and welded pieces of 91 material.

 

[stanweld]

JakeR82,
Assuming that parts are long seam welded, the completed part should be normalized and tempered after welding to avoid Type 4 Creep failure in the HAZ with the permitted design allowable stress.