PWHT of T23 waterwalls- required or not?

[davefitz]

There have been several recent papers that indicate it is neccesary to PWHT T23 membrane waterwalls , but I understand the current claims by boiler mfrs is that it is not neccesary.

As per the below link, there will occur fatigue cracking in the weld HAZ as the material ages at service temperature due to a loss of ductility, unless the panel originally undergoes a PWHT.

<

http://www.nanocon.cz/data/metal2007/sbornik/Lists/Papers/116.pdf>

So, if the panel is supplied without PWHT and eventually cracks, do we then get to hear from the mfr that it is caused by "operator error" or " incorrect feedwater treatment", or is the clam valid that PWHT is requried?

 

[metengr]

When T23 material was introduced as a Code Case material and ultimately accepted for use, this material was described as a bainitic/martensitic low alloy steel that because of the low carbon was weldable and would not require PWHT.

Now, I have heard of several cases where new installation of T23 waterwall panels in supercritical boilers has resulted in cracks at buckstay attachment welds after hydrotesting. I could not obtain clear information as to the cause of the cracking but I suspected it was due to hydrogen pick-up from field welding and unrelated to shop welding.

I reviewed the paper and like other papers arguments for an against PWHT can be made. I did not understand the statement regarding no preheat below;

This paper deals with influence of PWHT of T23, T24 and 15 128 welds on their
mechanical properties. The T23 and T24 steel were designed for welding membrane water
walls without preheating and without PWHT.


If this is true this is totally incorrect would indeed pose a problem. A preheat is required for this material, period.

I do have installed T23 tubing that is loose and butt welded as reheater jumer tubes. These have performed very well without PWHT.

 

[davefitz]

Thnaks for the input, MetEng.

There are other papers that indicate one other source of cracking is "hot cracking" during membrane wall shop construction, due to too-fast weld speed plus excess martensite in the supplied membrane plate material. But the eye-opener is the concept of HAZ material aging at service temperature leading to precipitates and a continuously decreasing ductility with time.

In the case of a radiant furnace membrane wall, the membrane "fins" run at a significantly higher metal temperature than does the pressure containing tubes. While the material may be specified for the expected worst tube metal temperature at desing pressure, the length of the fins are usually limited by the effects of thermal stress due to fin tip temp vs membrane wall composite average metal temp, and due to corrosion / oxidation limits. In any event the "service temperature" of the fin HAZ can be significantly higher than that of the tube design metal temp.

Fatigue cracking due to loss of ductility due to service aging mgiht escape first year's warranty obligation, but may be a really big headache 1 day after the warranty runs out. I would expect the T23 is mostly used on modern supercritical units only in the upper , vertical tubed section of the furnace ( where it is easy to repair) , but if T23 were used in the lower spiral wound section of the furnace ( where repair is a B&&ch), we're talking about 50% availability.

 

[stanweld]

davefitz,
We have only installed T23 tubes in the superheater sections (no membrane) of conventional and supercritical boilers. So far no cracking problems. We have found that without preheat, hardness can readily exceed 350 Hv. With 300 F minimum preheat and interpass temp < 600 F, hardness values consistently less than 320 Hv were readily obtained. As a result, we specify this minimum preheating/interpass temperature in our welding procedures. With 200 F preheat, hardness values were less than 350 Hv ranging from 300 Hv to 345 Hv.

 

[davefitz]

stanweld:

Yes, I am sure that loose tubes do not have the alternating thermal stress issues of a membrane wall in supercritical applications. So, a loss of ductility in the weld HAZ does not lead to a fatigue crack in loose tube applications, except perhaps near a tube -to -header weld.

If it is true that skipping PWHT causes MX precipitates at the HAZ with in-service aging and a resulting loss of ductility, one could address most of that by PWHT the panels in a shop furnace, but the field weld of panels would be a tough nut to crack. Maybe the longitudinal panel field weld could use an incolnel filler? just guessing.

 

[metengr]

Inconel filler will not prevent secondary hardening of the T23 tube base material heat affected zone, IF secondary hardening is indeed proven to be the case. VM and Sumitomo (original developers) have done extensive research on this material, and none of this has been reported. I do question how the T23 material was manufactured because there were problems with how the Code Case was orignally developed regarding critical alloying elements.

What does bother me is a reported case of a new boiler installation where reports I have heard from my peers indicated much higher than normal hardness (45 HRc) of failed attachment welds (not membrane) on new T23 WW panels (for a new ultra supercritical boiler install in CO).

I have not seen the met results. If this hardness is indeed the case, this is a metallurgical notch and will be susprtible to various failure modes - mechanical fatigue, hydrogen embrittlement, SCC.

 

[davefitz]

Meteng;

I do not know why V+M or Sumitomo have not reported this result- but when I proposed use of T23 for waterwalls w/o PWHT back in 1995 , IHI had warned at that time that their mock-ups and testing had shown that PWHT was neccesary for membrane walls.

Similar fatigue failures on waterwall mock-ups were reported in reference (7) of the above link, by Karlson , Rasmussen and Melanie Montgomery in 2000.

 

[pvy]

Is there any pratical experience with a local pwht of a T23 membrane wall ?

 

[Guest102023]

My understanding is that for non-PWHT T23, secondary hardening in the CG-HAZ can lead to reheat (a.k.a. stress relief) cracking. The problem is worse in the upper end of the allowable operating temperature range (I will dig up a reference for that).

Dealing with CSEF alloys is all about carbide management

 

[davefitz]

brimstoner,

I've just read confimration of the stress relief cracking- some recnet papers at the Chicago P91 conference last month. Masuyama and others discuss minimizing this effect by control of Ti + N in the weld consumable-

I'm not yet convinced the iinstability of T23 and T24 alloy is yet fully appreciated.The increase in hardness and loss of ductility in the HAZ after only 10 hrs service suggests more long term testing on membrane wallpanels may be needed.

 

[Guest102023]

The more I learn about T23 the less I like it, because of its instability (especially in thick sections). This is on top of inherently poor creep ductility. Based only on literature and not first-hand experience, it appears to me that T24 is less prone to some T23 issues.

A big problem has been going from development by expert, well-equipped scientists in controlled environments to carbon steel fabricators welding this stuff on dirt floors without benefit of engineering knowledge of the intricacies and pitfalls, under EPC pressure to minimize costs. On top of that there are the metallurgical issues that needed a decade to be understood and fully appreciated. But I wouldn't be the first to express these opinions here.

'Support Your Local Welding Engineer' IIRC was the title of a 60s film.