P91 PWHT

By worgz01 Date 04-04-2008 01:19 Edited 04-04-2008 01:22

I have a dilemma in relation to PWHT temperature. We have just conducted PWHT (Stress Relief) on a 660mm x 24 mm Hot Reheat Pipe via local circumferential bands during our night shift operation on site. The stress was set up with four type K thermocouples positioned equally around the weld line. As the temperature reach 1310 deg, one of the elements failed and temperature at one TC location dropped to 1210 deg. The remaining elements and TC continued to raise and set at soak temperature of 1330 deg in accordance with the qualified WPS. Soak commenced at this temperature and continued for 80 minutes before the problem was noticed by the technician. The technician after finding the problem increased the soak set to 1365 degs, which brought the "working" elements up but no effect to the non "working" element (of course). He then decided to increase the soak time by another 2 hrs. All the time the one element that is not "working" has a TC which is recording a temperature of 1200 ~ 1210 degs constant. The total sooak time has now been nearily 3.5 hrs

The question which I need to ask is what is the likely effect of having one temperature gradient up to 120 deg of recorded temperature and 100 deg below the standard requirements as minimum? Aussie Standard (1310 ~ 1400 deg)

By GRoberts (***)

Date 04-04-2008 04:14

The temperature gradiant can leave residual stresses in the pipe. The microstructure of the low temperature area will not be tempered properly. However, by re-heat treating the joint, both of these problems will resolve themselves with the 2nd go-round. I don't have any information handy, so I couldn't review your temperatures. Perhaps someone else can confirm their adequacy.

By js55

Date 04-04-2008 13:38

That sounds kinda strange. Are you sure it was the element and not the thermo couple? The reason I say that is that given extended soak, thermal conductivity would have a tendency to homogenize the temps throughout the pipe. Or at least you should have seen some increase under the 'failed' element due to conductivity from the other elements.
The other thing is, 1330 is too low for P91. You should be cooking it in the upper range anyhow. You're tensiles will be good at low PWHT temps but your toughness will not improve as much as it should generally.
When you qualed your procedure did you run Charpy's with that PWHT temp?

By GRoberts (***)

Date 04-04-2008 19:04

based on some heat treating we have been doing recently, I would not be surpised in a steady state soak, that a non-working heat treat pad (element) would drop that much. It isn't too hard for the pipe to suck the heat out the end. It would depend on the width of the heated band and insulated band though. I've typically found that when a TC fails, it give erratic readings, but haven't found "wrong" readings that I know of when the temperature wasn't fluctuating all over the place. In either case, a re-heat treat is necessary, and some hardness checking would probalby be in order after the 2nd heat treat as well.

By worgz01 Date 04-06-2008 00:20

The WPS qual was conducted to a PWHT temperature of 1390 deg. Charpy V were conducted at HAZ and Weld with results of all 3 tests in the HAZ 148 J and between 39~62 of the 3 tests in the weld. Aussie requirement is 18 J @ 0 deg min. Aussie standard allows the temperature to to be with in a given range with out requalification. the 1330 is still within that range however yes our target is still 1390 deg.

This was the target temperature the PWHT was heading for before the element failure. Yes it was an element failure, tail connection burn out. The chart record has a mid slope decrease to to the 1210 deg on the failed element, then it flat lined, with no erratic movements that would be expected with a TC problem.

I am concerned that another PWHT, even at the correct temperature, will reduce yield strenght, tensile strength and notch toughness, as I have seen is some documents particularily if soaktime (accummulated) exceeds 6 hrs

Thanks Ken

By js55

Date 04-07-2008 14:12

In my opinion, and experience with far longer cooks than 6 hours, you will not see a reduction in yields and tensiles for a 6 hour, or more, cook. And you will most likely see an improvement in notch toughness, especially if you measure it in lateral expansion.
If I were youI'd take Greg's advice and cook it again.
You don't even really have a choice.

By Fredspoppy (**)

Date 04-07-2008 16:40

js55,
I don't know if your "cook it again" advice is sound. If the WPS gives a temperature range and time, you can't arbitrarily do more PWHT (ASME) to correct a mistake in the 1st go around. Section IX requires at least 80% of the production PWHT time to be used during qualification, therefore to justify 6 hours of PWHT, qualification would have to have been for 4.8 hours, considering the comment that the WPS was done with impacts. I don't know what your Code says regarding time at temperature, but I would guess it is addressed.

By js55

Date 04-07-2008 17:11

FP,
The 80% requirement is a supplementary essential variable for impact regimes. You won't have a code impact regime for Grade 91 since it is intended for high temp applications.
If 80% is required on the WPS, I'd revise the WPS. I have never encountered a metallurgical or mechanical issue with going from a 2 hr/3hr cook to a 6 hr or more cook on Grade 91.
I've cooked 3 1/2" Grade 91, which when considered as a minimum of 1 hr per inch of thickness would be 3 1/2 hours. 2 1/2 hrs will be of little consequence.
We commonly had engineering specifications that required 4 hours minimum.
In fact, the Mannesmann recommendation for PWHT regimes for Grade 91 limits ramping to ~150deg per hr. (if memory serves). With no up fast to 600 or 800, which is allowed by code. Which means that if you're cookin at ~1400 it should take you over 8 hrs to get to temp. Now granted you aren't at soak temp but the stuff will be reacting (slowly precipitating V/Cb carbonitrides) below soak temp.
The only other alternative is to cut out the welds. Nobody wants to do that.

By js55

Date 04-07-2008 17:17

One other thing, you do impacts on Grade 91 for information. That does not constitute and impact regime. An impact regime kicking in the 80% requirement is a code impact regime, not a self imposed one.
The impact concern for Grade 91 has to do with loading material handling practices and plant start up issues.

By Fredspoppy (**)

Date 04-07-2008 18:18

I would be very careful on giving advice regarding P91 materials. I don't know if you are aware of historical problems with this material, but just Google "P91 Metal" and you get some horror stories such as the one at http://www.eng-tips.com/viewthread.cfm?qid=172967&page=1. PWHT issues seem to run a consistent thread in these articles.

By js55

Date 04-07-2008 18:47

FP,
Your concern is duly noted. And certainly care with Grade 91 is warranted. But rest assured that a variance of a couple of hours in a PWHT (as additional) is well within standard operating parameters for Grade 91.
Yes, heat treament does seem to be a recurring theme in some of the incidences. I have been aware of perhaps a few. In the particular case you posted it was not PWHT that was at issue however. It was the possibility of pipe manufacturing heat treat. And the mystery traceability as well.