HRSG boiler tubing 2

[Metaljon]

I am working on an HRSG SHP boiler. The existing superheater tubes are manufactured from grade A178. The original tubes had a hardness of about 150 HV. The tubes are about 10m long and this has caused a lack of flexibility in operation causing the tubes to buckle and distort. The tubes have been progressively yielding probably on each cold/ warm cycle causing the tubes to become excessively cold worked. The central bank of tubes have been removed leaving about a 0.5 m length of tube stub on each header. The bottom tube stub has a hardness of up to 300 HV whilst the top tube stub has a hardness of 500!! HV. Ideally I would normally recommend complete removal of the tube stubs and reweld new tube sections to the header but this presents a difficulty of access to the top and bottom headers as the design has very restricted access to the rear of the tubes. We are looking at PWHT options of the tubes to try and soften the steel. If PWHT is performed at 650C to recrystalise the steel prior to welding, is there likely to be any negative downside of doing such a heat treatment?

 

[metengr]

Metaljon
What you are referring to is a stress relief thermal treatment to remove the effects of cold work. ASTM A 178 is ERW Carbon and Carbon-Manganese steel tube for superheaters.

I doubt that this level of cold work in service is what is happening because you would have tube failures from low cycle fatigue. The hardness values, are they of the base material or near a butt weld in the stub tube? What is the stub tube material and that of the header? How were the hardness values obtained? Can you provide more information?

 

[Metaljon]

The tubes are attached to Grade SA106B Schedule 80 headers. The tube hardness has been tested with a Microdur MIC10 and seem to be real. The tube at the bottom has been tested at intervals of 20cm and shows hardness is very uneven and tubes appear to have un-uniformly yielded.

The tubes at the top are showing localised readings of up to 500HV which seems unbelieveable. We are retesting the tubes at the top again to see whether these hardness values are real. The abnormal hardness does not extend into the header or into the weld on the tube stub. The hardness values at the bottom of 300 HV are definitely real.

The problem with heat treatment is "strain ageing" and even if the tube is annealled at 650 deg C then the cold worked tubes could remain in the brittle condition. The only way to rectify is to normalise at a temperature from
200 - 300C above the upper critical limit of mild steel, i.e. approximately 920 deg C with subsequent cooling in still air. If the strain age hardened steel was annealed at the lower temperature (650 deg C), it may well remain in the brittle condition. This would mean that a short length about 0.5 m of the remaining tube stub would need to be heat treated at 920 deg C.

I would appreciate any advise on above.

 

[metengr]

Metaljon;
Have you actually removed tube samples for metallographic examination in the lab to confirm hardness testing in the field? Before I would undertake any course of action, I would have one or more ring samples removed from the existing stub pieces (2 to 3 rings) and evaluate the bulk hardness , chemistry and microstructure of the stub tube samples. I am still not convinced that what you have is necessarily strain age embrittlement.

I would not recommend any elevated temperature heat treatment (not a PWHT) in-situ because of concerns related to distortion and adequate support. The logistics would be a nightmare.

 

[davefitz]

You may want to address the cause of the thermal ratcheting. Probably due to liquid water entering the superheater header, either due to a mal designed spray system or water carryover from the steam drum.Likely due to spray attemporator.

 

[Metaljon]

We have remeasured hardness at top stub piece. The material is definitely work hardened as a result of ratcheting but we dont believe the readings of 500 HV are real. Very poor design with a tube span of nearly 11m with inadequate flexibility. According to design code we need at least 46mm accommodation for expansion which is not present. We have decided not to PWHT as it is too risky. I dont think the material is currently strain aged but it is poor practice to weld heavily cold worked material or anneal as it can induce strain age embrittlement. We will try and remove a section of the stub for evaluation at the next outage. Thanks for the assistance and advice.

 

[unclesyd]

What is the composition of the heating media in the boiler?

 

[Metaljon]

Environment is exhaust gas from a turbine with some auxiliary gas burner firing under certain conditions.

 

[unclesyd]

The reason I asked was that for several years we had a thermal oxidation unit with HRSG where we had a problem with some tubes being carburized if we were feeding certain process streams. We were getting carbon pickup on the superheater tubes at our normal operating temperature of 1700F at the inlet to the boiler.

 

[davefitz]

Normally, a HRSG would have an inlet gas temp of 1200 F; your 1700 F inlet temp suggests graphitization may be occurring on the outlet ( top) CS stubs.

If the unit were to undergo thermal ratcheting, the same reaction would occur at top and bottom, so they should expereince equal work hardening, so the variation in hardness top to bottom may suggest graphitization at the outlet end.

Normally on a HRSG the tubes may exceed yield in tension during a thermal ratcheting ( coldest tube) but will not exceed yield in compression due to unstable buckling .

 

[Metaljon]

We checked the tubes by replication and the structures are normal with no sign of graphitisation. The hardness was checked with a Microdur MIC10 and gave the same high readings at the outlet end at the same positions. We repolished sites again to make sure all oxide had been removed and hardness went up sligthly to 600HV. We rechecked all the tubes with a bouncing ball Equotip device (whilst this is technically not a valid method of testing for thin tube 2 to 3 mm thick it gives a relative comparison), it showed only a 10-15% difference in hardness values between the tubes showing cold work and the others. I have concluded that some of the very high hardness values recorded with the MIC10 are probably not real. The MIC10 instrument is very sensitive to external vibration and the turbine was running after service work giving a very slight background vibration. The Equotip does not suffer the same sensitivity to vibration but the probe tends to rock on the surface. Whilst I dont believe the readings with the Equotip are accurate, I believe they give a good relative indication of hardness. The tubes are definitely work hardened but only to 220 to 280 HV. Originally from new the tubes were 150HV.

 

[metengr]

Metaljon;
Now that seems more realistic. Now one other bit of advice once you determine the source of thermal ratcheting, I would recommend you perform wet fluorescent magnetic particle testing (WFMT) on lower and upper stub to header attachment welds (if accessible). You need to determine if the low cycle fatigue damage from thermal ratcheting is propagating fatigue cracks in service at any weld toe undercut.