T91 and P91 HRSG piping weld life - when to inspect?? 2

[cschelha]

I am interested in any information that involves the lifetime of HRSG welds in high energy piping or how often these welds should be inspected. The piping uses P91 and T91 steel and in the HP sections of the HRSG temps of over 1000 F and pressures of over 900 psi. More specifically I am interested in the life of the welds and when the welds should periodically be inspected. I have been searching and have not came across any good information, although I have learned a lot about the subject.

If you can help in any way, it would be greatly appreciated.

Thanks

 

[metengr]

I would first conduct an evaluation of your Grade 91 pipe hanger system (hot and cold walk downs every two years, maximum). The actual inspection of the steam line itself depends on when the line when into service. From what I have researched and in discussion with others in the Power Industry, I would perform an inspection every 5 years.

The inspection would consist of selecting terminal points, and several short or long radius elbows (concentrating at tangent points) and fittings. Locate the girth welds, locally strip insulation, grit blast and perform a wet fluorescent MT and obtain surface replicas. The weak link with Grade 91 material will be Type IV cracking along the base metal heat affected zones of girth welds. The reason I mentioned surface replicas is to evaluate the fine grained region of the base metal heat affected zone because this will be the location that will develop creep cavitation damage in service. Keep in mind that if your piping system has been adequately designed and fabricated, you should have plenty of time before Type IV creep damage would manifest itself. This is the reason for using a 5 year re-inspection interval. If hanger support problems are found, you can adjust the pipe re-inspection interval.

Regarding T91 tube material, I would conduct visual inspections of the T91 tube assemblies during your scheduled outages. I would probably remove several T91 tube samples for metallurgical evaluation after the first 10 years of operation (with 7 year intervals) UNLESS you have sudden failures of tubing.

 

[davefitz]

From the failures we have had in the last 3 years, nearly all large part P91 weld failures have occured adjacent to significant thickness transitions, such as from a forged F22 saddle branch to P91 pipe ( 1.5" thk to 0.5" thk) and low alloy turbine stop valve to P91 pipe ( 3" thk to 1.5" thk).

The very fast thermal transients associated with startups of combined cycle HRSG's plus lack of use of smooth ,gradual thickness transitions has lead to very short fatigue lives of these transitions. NOne of the weld failures were due to the weld procedure per se, but due to
(a) cyclic fatigue aggravated by
(b) severe disproportion of parts wall thicknesses due to the 2:1 ratio of material strengths between P91 and F22/low alloy plus
(c) failure to provide a gradual thickness transition ( ie transition piece at a thickness slope of less than 0.33:1 at the OD thickness transition) ( see EU PED fatigue analysis tables for weld geometry severity factors , less then a 0.33:1 ratio is a severe fatigue risk when coupled with a greater than 2:1 thickness ratio)plus
(d) weld styles specified by inexperienced engineers not familiar with the ASME code, P91 metallurgy, fatigue analyses, or stress analyses

I would recommend yanking out the plant( main steam piping and boiler) P+ID dwgs and find ALL P91 to P22 or low alloy grade transitions, then critically review the detailed mechanical dwgs of those transitions.

If a P91 transition piece was not provided , with a thickness transition slope of less than 0.33:1, then you should attach a thermocouple to the 2 different pieces and archive all temp vs time data from one cold startup to full load thru to the next scheduled shutdown . Process this data as a temperature differential ( Temp of thick P22 part minus temp of thin P91 part), and if these temp differentials EVER exceeds +-180 F, then consider either conducting a detailed fatigue analysis ( finite element) or schedule routine NDT to detect fatigue cracks.

Finally , if yo find that the desing is not fatigue friendly but the vendor is adamanant it meets current ASME code, get involved with ASME and change the code to finally recognize fatigue damage- the free pass given to mfrs and designers for the last 100 yrs needs to be yanked.

 

[stanweld]

I can attest to the commentary by davefitz. Do examine as he has stated, especially if you have frequent shut down/stertups. Also examine dissimilar stainless steel to P91 welds at instrument connections (venturi flow meters & thermowells. If you have auxilliary gas/oil burners in the HRSG, check for overheating of tubes in the hot reheat and superheat sections. T91 tube bursts have occurred in these locations in as few as 5000 hrs. Finally, check heat treatment history of induction bends, many have received improper heat treatments after bending. In cases where process steam is generated instead of power generation steam and the pipe is fabricated to ASME B31.3, the required post bend heat treatment is stress relief at the PWHT time and temperature only. Advise checking for all of these cases within the first year of operation and annually thereafter.