T23 Boiler Tube - Curious Fracture 7

[Guest102023]

T23 Reheater Tube Fracture in Weld -- Anyone seen this fracture morphology before?


Several T23 reheater tubes failed in a similar manner after less than 6 months and ~250 cycles from commissioning of the HRSG. The circ. fracture originates in the weld, at the front (where you would expect bending force to be greater) then moves into the base metal as it propagates around to the back side. Meanwhile, a ~4" long tear happened on the front side. It seems like there is more than one thing wrong here.

Unfortunately I did not see the header side of the story (another case of repair first, ask questions later), so it is hard to confirm the lack of penetration that me and the repair contractor suspected.

Particulars: 2" T23, welded to P22 header with manual TIG, ER9018-B3, PWHT. Unknown tube origin, but they are significantly older than the plant. All zones appear to have normal microstructure (no ferrite at least). The correct filler was used. Weld hardness is 285HV1, HAZ up to 335HV1, base ~192HV1. I have found both tube surfaces decarburized. Scattered corrosion pits up to 3 mils deep on outer surface. A few very short secondary cracks and sometimes oxide along prior austenite GBs.

Any thoughts or suggestions? The hardness seems on the high side for a PWHT joint. I thought about reheat cracking, but did not think that can affect the weld also(?)

I am a newbie here so I hope this gets posted correctly ...

 

[Guest102023]

I googled and quickly found the same paper, even though I already had it in a giant directory I have created while researching this stuff. Thanks again!

Interesting but useful that this material is so insensitive to preheat; it allows one to make a better conjecture about the PWHT. With conventional steels it can be more difficult to know whether high HAZ hardness was caused by insufficient preheat or PWHT.

 

[davefitz]

The T23 CCT curve indicates that it is feasible to develope ferrite if it is cooled too slowly. Perhaps it may become neccesary to fabricate thick walled headers with short tube stubs of a non-bainitic alloy, as the cool down rate for a bainitic alloy close to a thick header might not achieve the fast rates required to avoid ferrite.The T23 would then be field welded to the short ferritic tube stub.

 

[Guest102023]

davefitz,

I would hope that under shop conditions the entire header would be preheated using electric resistance elements with adequate thermocouple control. For field repairs things get more difficult of course.

In this case I am sure that PWHT, or lack thereof, is the culprit. The HAZ hardness level in this tube is typical of as-welded construction when it should have been ~30-40 points lower. Also, T23 as-welded hardness does not seem to be particularly sensitive to preheat level. Some ferrite probably is not uncommon in T23 base metal, but I did not see it anywhere in this case.

 

[unclesyd]

In one of the references I post a link to mentioned the problems of welding T23 to P22 in a header arrangement such as yours. I haven't had a chance to chase down the origin.

if you ever need to PWHT something like your headeer you might want to discuss the problem with these people.

http://www.hotwork.com/html/hpi_vesselpostweldheat.htm

 

[Guest102023]

unclesyd,

I would appreciate that ...

Different materials, widely different thickness and thermal mass, high PWHT temperature, 45 foot long tubes ... I am curious about the PWHT procedure that is used by OEMs.

 

[unclesyd]

You can work out all your variables so you can cover with one set of heating conditions. Your highest required temperature probably want be a problem with the other materials and a little extended time want hurt anything.

These people can heat some mighty large boilers and vessels with excellant control. Remember that you can overcome a lower temperature by extending the time.

In the majority of materials about 95% of your PWHT is accomplished in the first 15 minutes or less.

 

[Guest102023]

Thanks for the info, they have a location not too far away. The local heat treaters are not up to snuff in my experience with them.

BTW, what is the optimum PWHT temperature for T23? And does it depend on thickness at all?

 

[wyogoob]

Could you or did you get hardness readings close to the HAZ where there was no stretching? Just curious.

Looks like improper PWHT problems to me. It can be very tough to PWHT the header and not overheat the tube just past the socket weld, especially if PWHT is performed on in-boiler repairs. Many times the PWHT techs will move the TIs down the tube to a "cooler" spot. This avoids over-temps showing up on the chart, but the tube directly above the socket weld is cooked and/or cooled down too fast.

 

[Guest102023]

thanks wyogoob,

HAZ hardness just next to the fusion line was 330-335 HV, as I have since learned a little high for welds that were PWHT. The hardness is in fact similar to the as-welded condition. But metallurgy notwithstanding, the immediate root cause appears to have been water hammer. Failure
analysis gets tricky when more than one thing goes wrong!

But I agree TC placement is crucial, and I have witnessed 'temperature management' by PWHT techs who are mainly concerned with obtaining a perfect-looking chart.

This is another example of where having an appropriately trained engineer overseeing a task would be better - someone with a knowledge not just of welding but heat transfer, distortion, etc. It is near impossible to ensure correct heat treatment by reviewing a generic procedure while sitting in my office. One of the unfortunate trends with EPCs is cutting back on direct surveillance of fabricators ... maybe a good idea for a whole new thread!

 

[Guest102023]

thanks wyogoob,

HAZ hardness just next to the fusion line was 330-335 HV, as I have since learned a little high for welds that were PWHT. The hardness is in fact similar to the as-welded condition. But metallurgy notwithstanding, the immediate root cause appears to have been water hammer. Failure
analysis gets tricky when more than one thing goes wrong!

But I agree TC placement is crucial, and I have witnessed 'temperature management' by PWHT techs who are mainly concerned with obtaining a perfect-looking chart.

This is another example of where having an appropriately trained engineer overseeing a task would be better - someone with a knowledge not just of welding but heat transfer, distortion, etc. It is near impossible to ensure correct heat treatment by reviewing a generic procedure while sitting in my office. One of the unfortunate trends with EPCs is cutting back on direct surveillance of fabricators ... maybe a good idea for a whole new thread!

 

[Guest102023]

oops

 

[Guest102023]

Postscript - can someone recommend a company who can come in and perform temperature measurement of these headers and tubes? Thanks in advance.

 

[metengr]

For what, PWHT or in-service temperature measurement?

 

[Guest102023]

In-service temperature measurement.

 

[metengr]

This can be done in-house with purchasing TC's, TC wire and a chart recorder for much less the price of having an outside consulting group performing this activity. This is not that difficult to do.

 

[Guest102023]

Then I could probably do that for them - any guidance available as to the equipment and techniques? I would think that a laptop + software have replaced strip chart recorders. Plus I would not know how to interpret the data, I am not an operations guy.

Anyway, their staff is bare-bones and I'm sure they can afford to hire the experts.

 

[metengr]

Ok, if you need help I would first go back to the HRSG original equipment manufacturer (OEM) and inquire on obtaining in-situ temperature measurements of the header and terminal tube metal during start-up, steadystate and shutdown conditions.

 

[davefitz]

I would review some past issues of "combined cycle journal" to obtain rough recommendations as to where to place the T/C's. See this link for this useful, free magazine <

http://www.combinedcyclejournal.com/archives.html>

. Some of the articles on field testing of pressure parts by Bob Anderson may be a place to start.

Another source would be the EPRI workgroup on HRSG's .

If you only wish to focus on the above described failures, I would suggest the followign field T/C's in addition tothe standard operating T/C's:

a)10 inlet tube stubs , as close to the header as feasible
b) 2 on header top CL, 2 on hdr bottom CL
c)1 at reheater spray water pipe as close to spray attemporator nozzle as feasible
d) one on each header drain pipe, 1 ft below header connection

In the old days, field T/C's installed in thick walled tubes or headers would be simplydrilled and peened , to avoid welding. But this might not be feasible with thin wall reheater tubes. There may be a tack weld pad type available for these thin tubes, but you need to do some homework on weld procedures on P91 tubing.

 

[Guest102023]

Thanks Davefitz, I have all the free CCJ stuff and will search through there. This is a little niche that I think I could fill if I could gain enough expertise in the interpretation of the data. We have a business in town that can supply the TCs and the recording equipment, and do the attachment.

These are T23 tubes, so I am investigating the implications of attachment by welding. I think B31 may have something to say about it.

A question about b) - how close to the tubes should the header TCs be placed?

Thanks very much!

 

[davefitz]

The header t/c's have several purposes:
a) top to bottom DT- layout of water on the bottom can cause the bottom to operate much colder than the top during water induction events ( from the drains or from overspray). A top to bottom DT of opposite sign occurs during initial synch of the STG on units with a dry reheater ( ie no steam to reheater steam bypass system)

b) tube stub to header DT- header top T/C's minus temp of coldest or hottest stub T/C provides some indication of shear stress imposed at stub to hdr weld. The stubs can change temp perhaps 25- 50 times faster than the header ( varies by the 2nd power of wall thickness, (0.15" /0.75")^2 = 1/25 . A fast scan of stub temps is required to indicate how many cycles are occuring due to alternating wetting- dryout during overspary events. The header T/C does not need to be near the stub weld.

c) spray water t/c- will show the severe subcooling at this location at teh initiation of spray use- very subcooled spray water does not evaporate easily and leads to carryover of liquid to the downstream tubes and header.

d) drain line t/c near hdr- will show water induction from blowdwon tank or superheater drain manifold tie-in.