P91 Heat Treatment

[DSB123]

Dear Members,
I would like your thoughts on the following:

We have a new piping system which is in A335 P91 material. The post weld heat treatment has been carried out at 750 Deg C +/- 10 Deg C. Some of the heat traces show the temperature to be 741 Deg C. Bearing in mind the accuracy of the measuring devices (+/- 0.75%) the temperature could be as low as 734 Deg C. Is this acceptable and will the resultant metal structure be acceptable?

 

[davefitz]

As per the Metrode tech data sheets, the min permitted actual temp is 730 C.

Of equal importance are:

a) preheat to limit hydrogen embrittlement. Use of certified low hydrogen consumables and use of low hydrogen welding procedures, or else performing a "hydrogen bake" after welding and prior to cool down to 100 C

b) cool down to 100C prior to PWHT, to ensure complete transformation to martensite

c) PWHT above 730 C but below the lower critical temp to avoid overtempering. Lower critical temp is a function of consumable compostion, espesc Nickle content.

d) cannot wait more than a few days between welding and PWHT, to avoid SCC.

e) AFTER PWHT, should include add'l NDE to confirm proper heat treatment ( hardness test) and lack of H2 or SCC cracking ( dye pen , UT, mag particle wet flourescent)

 

[metengr]

To add to davefitz;
If you had already performed the PWHT - hardness testing of the weld, HAZ and base metal, and NDT must be performed to verify weld integrity.

My only question regarding PWHT is time. How long did you hold the temperature for PWHT? The holding time should be a minimum of 2 hours at temperature.

 

[DSB123]

davefitz and metengr,
Thanks for your replies. The problem is that the design of the piping system has not accounted for the fact that P91 welds are weaker ( for creep strength) than the parent material and the stress levels are high. This along with the lower PWHT value is giving me some concern over the integity of the welds for creep purposes.

 

[davefitz]

Join the crowd.

For the case of circumferential butt welds not experiencing "multiaxial stresses", there should be no issues, as the membrane stresses in that weld ( perpendicular to the plane of the weld) is only 50% of the allowable stress.

Multiaxial stress will change the picture, for example, if there is imposed a large bending moment at the weld, or a large shear stress, or transient internal stresses caused by thermal stress during startups and shutdowns. This last problem has bothered us, in the case of a 1.5" thk P91 pipe welded to a 3.2" thick steam trubine stop valve ( 1.25Cr-1Mo-.25 V). If the transiton piece does not exhibit a gradual change in wall thickness near the weld, then high internal stresses can be imposed during startups.

In the case of longitudinal welds, the membrane stresses are twice taht of circumferntial stresses, so in those cases you should be adding a "weld srength reduction factor) of about 0.6 when assigning a wall thickness, to account for the weaker HAZ.

 

[metengr]

DSB123;
Based on your information and concerns, you are going to have to keep watch on the Type IV HAZ cracking issue if you have higher piping stresses than design in the steam line.

As Davefitz pointed out, the circumferential welds do have lower service stresses. However, if you have thermal/mechanical piping stresses above design conditions, the HAZ on the base metal side of the welds will have increased susceptibility to Type IV cracking in service for sub-critically post weld heat welds. The only way out of this problem is to re-heat treat any welds, which are normally done during shop fabrication of piping components to avoid having sub-critically post weld heat treated welds in highly stressed locations.

To address your concerns, I would recommend implementing an in-service inspection program for the P91 piping system. After 5 years of operation, perform an in-service inspection of the highly stressed girth welds using surface NDT (like wet fluorescent MT) followed by surface replication to check for Type IV creep damage along the base metal side of thr HAZ. If none is found, you could increase the re-inspection interval to 7 years.

 

[DSB123]

Thanks again Davefitz and metengr,
You have the same approach/concerns that I do. The fact is the piping system has been designed by others who have not considered Weld Strength reduction factors and the operating temperature is around 565 Deg C. The bending stresses are high on the circ butts and this is what is giving me concern. Also the pipework original design remit was for two shifting but luckily the operation could well be base loaded. Taking all things into consideration and the number of recent papers on P91 weldments I am a little sceptical and as you suggest have proposed a more stringent inspection regime. The problem is that type IV cracking in P91 only shows itself at about 80% of the life of the time to failure which I think could be very short in the high stressed locations.

 

[unclesyd]

I don't believe I would reduce the inspection interval after five years, if anything I would lower the interval. This is especially true if any indication of anything happening is detected at the first inspection.


Are you talking about a large number of welds?

 

[DSB123]

unclesyd,
The line is around 100m long so there are a number of welds (approx every 6m). There are a number where the axial stress levels are around the same magnitude or higher than the hoop stress level.

 

[metengr]

DSB123;
Based on your information, I would give serious consideration in having the pipe configuration re-analyzed, if you have not done so already. From expierence, I have seen various models and people that perform this activity with some variation in results depending on level of expertise and assumptions used in the piping analysis.

Perhaps minor modifications to the existing piping support system can be made to reduce the axial stresses to an acceptable level.

 

[StressGuy]

DSB123,

I'm curious, what piping code was this system designed by? I would guess B31.1 since I've typically heard of P91 being used for power plant applications.

When you say the piping design did not account for the weld metal being weaker than the parent metal, I'm curious to know how that compares with the allowable stresses from the piping code. Are the code stresses too high becuase they are based only on the parent metal properties?

Edward L. Klein
Pipe Stress Engineer
Houston, Texas

"All the world is a Spring"

All opinions expressed here are my own and not my company's.

 

[DSB123]

StressGuy,
The system has been designed by two Contractors with B31.1 used for one section and BS806 used for the other. BS806 does recognise that P91 has inherently weaker welds than the parent metal and applies a weld strength reduction factor above 535 Deg C. However the Code case giving this is 6 years old and well out of date considering the latest papers/information on Weld strength reduction factors that should be applied. The problem is that the weld strength reduction factors have not been applied for the section designed to BS806.

 

[davefitz]

I understand that B31.3 has explicit weld creep strength reduction factors- according to reviews of new handbooks that explain the refinery piping code. You may want to review the new recommendations of B31.3.

 

[DSB123]

davefitz,
Can you provide a reference/paragraph because as far as I know B31.3 (the edition I have) does not strictly address creep but merely says it has to be accounted for.

 

[davefitz]

It is not in the latest published edition online. Instead, it is in the 2004 edition, not yet released ( due 12/04). See reviews of ASME traning courses and new handbook by Charles Becht on the new release of B31.3

Dr Becht is on the B31.3 committee, so you can contact him via e-mail thru the ASME B31.3 webpage.

 

[StressGuy]

As it happens, I was just reviewing that section of Dr. Becht's 2nd Edition of his Process Piping book. I convinced my boss to order a copy for the company so we could get up to speed on what sound to be some pretty heavy changes to B31.3 for 2004.

I see that they are adding a change to the sustained stress calculation. Instead of the allowable being Sh, it will be Sh x W, where W is the "Weld Joint Strength Reduction Factor". The book indicates that W will be 1 up to 510C and start dropping from there.

Looks like this addition to the code is going to address exactly the issue that DSB123 is concerned with. I would assume that B31.1 is going to incorporate this as well, particularly since P91 is more commonly being seen in Power applications.



The times they are a changing.

Edward L. Klein
Pipe Stress Engineer
Houston, Texas

"All the world is a Spring"

All opinions expressed here are my own and not my company's.

 

[DSB123]

StressGuy,
I went to ASME Site and almost ordered the book but opted out at the end with the postage being $70 (nearly the price of the book). Do not think it's worth the high postage costs. I will see if I can get a copy through the local library first.
It's about time weld strength reduction was identified in the Code.

 

[davefitz]

According to the new B31.3 handbook by Becht, the code will encourage designers to use weld creep strength reduction factors that are based on lab tests of sample welds, using same parent material, thickness, filler material , welding process, preheat, PWHT, QC, etc as the commercial piece. If that is too much work , then the user can use a one-size-fits all factor W, as follows ( eyeballed from curve):
900 F, W=1.0
950 F, W=1.0
1000 F, W=0.97
1050 F,W=0.91
1100 F, W=0.86

These values seem too low for P91, based on my review of other works. These universal factors are claimed to be an average of the W for the specific alloys 304 H, 316H, 800H,2-1/4 1Mo, 9Cr-1Mo ( V=?, Ni=?)

 

[davefitz]

err, "these factors are too HIGH for P91"

 

[DSB123]

davefitz,
Totally agree with you. The values seem to be similar to those given in BS806 and these are from 1998 Enquiry case which are high to my mind in view of the latest reports.