P91 Low Hardness + Good Microstructure 2

[qaqceng]

Hi guys,
I'd like to get some experience-based feedback on a problem we've run into. We're building heat transfer modules for a HRSG project that has P91 headers and fin tubes. We've checked hardness after PWHT on the base materials and welds, but have come up with many low readings (some as low as 160-170 BHN). As is common, we performed replications on these same areas to check microstructure, and the results have confirmed a uniform tempered martensite consistent with "good" P91. We've sent the micrographs to another lab for second opinion, but from what I've seen in the past it looks comparable to good material; no abnormalities that I can identify.

We've tried many tricks in the book to recheck the hardness (remove more decarb, recalibrate on block of P91, etc.), and results still come back low. Anyone run into this before? I tend to think the microstructure is the most important factor, meaning material is good, but haven't found explanation for the hardness testing. We double checked heat treat setup, tc locations, charts, etc. and it all checks out.

Any advice would be appreciated.

Thanks!

If it helps, the portable tester being used are MIC10/MIC20's with UCI method. Results reported in Brinell. And yes, we did PMI and confirmed it is P91 material

 

[metengr]

Yes, we have seen this before with Grade 91. It would appear that some residual ferrite is present that is not observable by optical microscopy. The residual ferrite would have happened from orignal heat treatment due to inadequate quenching or lower hardness from carbide coarsening due to over tempering time.

Regardless, there is no minimum hardness requirement (only maximum) unless required by contract or engineering specification. The reason for no minimum hardness specified in Code or by material specification is that there is no consensus on a real minimum hardness value that results in inferior creep strength. With low hardness all you can do is to monitor this material after it is placed into servce by periodic inspections using surface NDE at high stress locations.

 

[pradipgoswami]

Hi qaqceng,
Few Questions: Did you check the results(hardness and replications) in other locations/areas of the assembly? If so are they similar to what's reported here or are different?
As narrated by you,assuming that portable hardness testing method is perfectly calibrated& accurate. Is there a way to send a small sample(BM and weld) to laboratory and Re-check hardness?

How sure you're that the microstructure is Tempered Martensite not Bainite. I've cases where finer bainitic could be mixed up with over tempered martensite.

The following article has some thoughts.
:

https://www.twi-global.com/technica...nite-martensite-bainite-pearlite-and-ferrite/

Also the earlier discussion has generated similar thoughts as related to your query.

http://www.eng-tips.com/viewthread.cfm?qid=217600

I would say the practical considerations for hardness in Tempered P-91 alloys/welds are, Max-260-270 BHN, Min-190 BHN.

As per the industry experiences hardness of 160-170 BHN may be the precursor to premature failure, in the temperature regime of operations for P-91 alloys.
Thanks.

Pradip Goswami,P.Eng.IWE
Welding & Metallurgical Specialist
ca.linkedin.com/pub/pradip-goswami/5/985/299

 

[qaqceng]

Thanks for the advice guys.

@Pradip, there were other areas tested on the assembly, and similar results. But, there arent any trends. There are even cases where one side of the weld is good and the other is not.

We’ve considered removing a tube for lab testing, but are hoping we can resolve nondestructively.

qaqceng

 

[metengr]

qaqceng,
Unless your client has a requirement that specifically states a minimum hardness, your objective would be to evaluate the source of the lower hardness. I think a tube sample is in order, if anything, to evaluate microstructure and verify chemical analysis and bulk hardness.

Again, lower hardness (below 190 BHN) does not mean failure in service. Other factors need to be considered, similar to a fitness for service or you reheat treat the material.

You might find this paper of value to your efforts

https://www.researchgate.net/profil...ess-for-Service-of-Degraded-Grade-91-Pipe.pdf

 

[weldstan]

While I have never seen hardness below 167 BHN with correct microstructure, I have seen many instances of acceptable microstructures with hardness in the 170s. Did the microstructure consist of coarse martensite with coarse precipitates? Did you have this type of low hardness in the weld metal?

 

Do not rule out mill decarburization, especially when using hardness testers that make very small indents. P-91 is an alloy that demands you know the pedigree and entire processing history, because defective mechanical properties cannot be guaranteed to be detectable, even with in situ metallography and hardness testing. Not all CSEF products had a good start in life.

"Everyone is entitled to their own opinions, but they are not entitled to their own facts."

 

p.s., I see that you did look out for decarb.

"Everyone is entitled to their own opinions, but they are not entitled to their own facts."

 

[qaqceng]

Thanks again everyone for your opinions and advice.

@weldstan,
I would call it coarse martensite with coarse precipitates. The weld metal was also tested, with good hardness of greater than 200BHN at all locations. It's another point that oddly doesn't make much sense to me. ~300 tube welds, ~100 BM locations with low hardness, Zero locations with low weld hardness. If it were subjected to overheating or other processing errors (no signs that's the case, but it'd have to be if the HT values were true), shouldn't a tube with 160 BHN hardness have welds lower than 200 BHN.... I don't get it. Which is why I've reached out to the industry experts like yourselves for past experience. Appreciate the help.

 

[metengr]

qaqceng;
Did you review the mil test reports to evaluate the original hardness? Also, the N:Al ratio of at least 4:1 or more is most important with regards to resistance to tempering, during fabrication.

 

[weldstan]

Your issue more than likely stems from the thermal history provided by the tube manufacturer and it could have occured during either normalizing or tempering, or both. And I have seen some truly whackey thermal cycles from tube manufacturers. It is noted that there is emprical and lab tesing evidence to indicate a reduced creep life of such coarse structures in P91.