A213 T91 starting microstructure

[replica]

Hi all,

I want to know regarding the stress data that shown in API 530 Figure E11. for A213 T91. As I was aware that for T91 there are at least 3 different starting as received/as supplied microstructure which the same chemical composition and processes (Normalised and Tempered). As we all know that creep properties in T91 is very sensitive to microstructure hence the different starting microstructure will give different creep properties. My question is

1. What starting microstructure is used to produce the data in API 530 Figure E11 for T91/P91.
2. Is not normal to ask the sample of microstructure and to specify cooling rate during normalising when purchasing P91 rather than only hardness and N+T heat treatment in order to obtain correct fully tempered martensite structure without ferrite.

Herewith attached 3 different starting microstructure for T91 that had met the requirement for hardness and N+T heat treatment.

Comments are welcome.

 

[davefitz]

Regarding question 2, it is recommended to obtain the temp vs time plot of the N+T process, and require that all parts in the furnace be monitored to ensure that the rate of cooling is faster than -10F/min through the temperature range 1900- 1250 F. Refer to the TTT plots by ORNL ( via Climax Molybdenum corp ) and Mannesman Vallourec on the need to avoid the region that would erroneously form excess ferrite during cooling.

The initial microstructure will also be a function of several other variables, including the actual normalization temperature and time held prior to cooling ( effects the grain size and thus creep strength) , the thickness of the part (effects the cooling rate vs location within the thick part) , and the method of stacking the parts in the furnace ( the bottom part is shielded from heating and cooling compared to the top part) . Probably the most important variable is whether or not foundry has employed a metallurgist familiar with P91. Most small mom + pop foundrys that provide the most competitive quotes would not know a metallurgist from a chiropractor.

You are correct; simply measuring the hardness is not sufficient to ensure correct metallurgy, but it often provides a clue and can be indicative of excess ferrite. It is also possible ( and common) that a correctly forged P91 part can later be damaged during later fabrication process, including overtempering .

"Nobody expects the Spanish Inquisition!"

 

[metengr]

There is a new ASME B&PV Code Case for Grade 91 material that is currently out for letter ballot. This will be the use of a Class 2 Grade 91 material with tighter controls on residual elements, like Ar, Sn and Sb. It is believed that these residual elements can result in the onset of creep cavitation damage early in the life of the Grade 91 material with little margin for damage tolerance. In other words, once the on set of creep cavitation occurs, failure is rapid.

In this new Code case for Class 2 material, a cooling rate of greater than 9 deg F/min is required. Last, because of the restrictive chemistry for the new class of Grade 91 material, the stress line for the original Grade 91 material could be lowered or limits placed on the maximum service temperature.

 

[metengr]

1. What starting microstructure is used to produce the data in API 530 Figure E11 for T91/P91.

If you review the tube specification SA 213 Gr T91 is supplied in the normalized and tempered condition to produce a 100% tempered martensite microstructure with Cr and V(CN)precipitates.

 

[replica]

metergr
Thank you very much for your comments. So ....most of data available for creep is based in 100% tempered martensite with carbides and nitrides precipitates
as starting microstructure. Is there any work on creep for the other starting microstructures ? This is very important because otherwise we are using the materials with unknown creep properties for critical high temperature and pressure applications.

Regarding the newv ASME B&PV code case as you mentioned ...is it available? Do you have more info about it?

 

[metengr]

replica
The ASME Code Case has not yet been approved by BPV I standards committee. It will be letter balloted within the next several weeks so I would rather not make it available at this time. There are no other starting microstructures for Grade 91 IF the material has been supplied to an international material specification, like ASME or ASTM.

 

[replica]

davefitz
Thank you very much for your comment,
It is recommended to obtain the temp vs time plot of the N+T process, and require that all parts in the furnace be monitored to ensure that the rate of cooling is faster than -10F/min through the temperature range 1900- 1250 F.
..I will try to include this criteria in my next purchase.
metengr
Thank you..but the manufacturer also claims that the products complies with ASTM A213 T91 or ASME SA213 T91, even though they have different microstructure. Can we reject the products based on microstructure?

Our criteria for retirement for T91 under high temperature and pressure is when the hardness is below 173 Hv. The procedure is ...when the hardness has reached below 173 Hv, we do replica to look at the creep based on VGB TW 507e 2009. Do you think 173 Hv is a fair value? This value is based on minimum tensile strength for A213 T91.

 

[replica]

Sorry VGB TW 507e 2005

 

[metengr]

The Grade 91 microstructure is very complex and difficult to evaluate for other phases. Some of the microstructures I have seen had ferrite that was only confirmed by
X-Ray diffraction. If the hardness, tensile properties and chemical composition are within the material specification requirements I would accept as is.

 

[weldstan]

Replica,
Is the hardness after long term exposure or a test protocol after several years of service to determine either overheating or initial inappropriate heat treatment?

 

[weldstan]

For purchases of new T91 tube, we and many other Owners and their Engineers have specified a minimum hardness of 190 HBN. Material with lower hardness will be accepted based on micro structural analysis. After reviewing hundreds of micros of lower hardness P/T91 tubulars, a number of them had aberrant microstructures between 180 an 190 BHN. All such items with hardness < 173 BHN had unacceptable microstructures.

All of the microstructures on material with hardness of 190 BHN and below 300 BHN had typical tempered martenitic matrix with fine precipitate microstructures.

 

[replica]

metengr

Thank you very much ..agree with you that ferrite can only be confirmed by XRD rather than solely looking at the microstructure but slow cooling rate can produce ferrite and sometime is visible under optical microscope. .


weldstan

The hardness is after a long term exposure under creep loading. Sometime the starting hardness of T91 (as recieved new T91 material) can be as low as 190 Hv...hence rejecting the tube when reaching 173 Hv after exposure to high temperature and pressure is not a good choice...that is what I think, so that is the reason why we also confirm by in-situ metallography (replicaation) to ensure that the formation of creep voids and coarsening of carbides are within the range before the tube is retired.

 

[weldstan]

My experience with long term temperature exposure is limited to approximately 45,000 hours and short term exposures to over heating. In all cases where over heating did not occur, the micros with hardness less than 173 BHN were due to inappropriate heat treatment during manufacture or during fabrication and not due to service.

 

[metengr]

Replica;
I would echo stanweld's statement. The current issue with Grade 91 is the behavior of this material which can be unpredictable from the onset of creep cavitation to final failure. This is what I described as damage tolerance of materials. There are several factors that go into this lack of damage tolerance, original heat treatment and residual alloying elements. EPRI has done significant work in this area and I have been fortunate to have been exposed to this information first hand.

This material does not necessarily exhibit long term softening if a proper microstructure has been achieved. We have seen starting microstructures look acceptable and after 80,000 hours with no appreciable softening, sudden Type IV creep cavitation damage developed surface and subsurface in a weld region. Most important is that creep ductility was reduced to less than 20% with no appreciable change in strength (or hardness).

I would be more focused with using PAUT at butt welds to monitor the onset of microcracking from Type IV creep damage.

 

[replica]

weldstan
Thank you very much for sharing your experience. Are you saying that hardness will not change during service if the correct starting microstructure, hardness, operating temperature and pressure are used? As far as I know prolong heating even below the creep threshold temperature will reduce the hardness of F/T/P91 to some extent (appreciable or not) eventhough the correct N and T microstructure are used. Are the coarsening of carbides, recovery of the subgrains in tempered martensite, formation of lave phase only affect the creep properties and not the hardness? If this is true then it may explain the reason why I sometime encountered creep failure eventhough the hardness did not change much. Is the mechanism for creep in P91 NOT well researched?

metengr

I agree with you that the failure due to creep can happen even before the formation of microvoids visible in situ metallography. Agree with you that some other factors involved in creeping of P91. Type IV crack happened at HAZ as you mentioned probably due to incorrect microstructure at HAZ or wrong PWHT. Can you explain more on lack of creep tolerance (original heat treatment and residual alloying element)? Any link that you can share?


For your information we have many new inspection stub (nipple) made of SA182 F91 (fitting) that was welded to the header attached to final superheater outlet manifold. After welding the average hardness of the the base metal is 193Hv to 202Hv. The HAZ is 218Hv to 230 Hv and the weld metal is 238 Hv to 249 Hv. We are going to do PWHT soon. I know that the PWHT should be done as soon as possible to avoid SCC or delayed cracking. My worry is that the base metal hardness will be reduced due to PWHT and fall below the 190Hv. Is the current hardness acceptable?

When doing in situ metallography on the weldments before PWHT we observed intergranular cracks at HAZ at two separate weldments. Can this type of crack be removed by grinding?

 

[replica]

The low hardness of base metal due to decarburization is unlikely because we measured the hardness at the same location for replica and we can see clearly the coarse tempered martensite structure indicating that the decarburised layer if presence was completely removed. The base metal microstructure is almost similar to weld metal. We are using a Kräutkrämer, Microdur MIC 10 equipment and calibrated on the known standard block (with +- 5 Hv variation) but of course not on P91 standard block. Is F/T/P91 standard calibration block available commercially?

 

[davefitz]

See thread794-331821 for a summary of the NEM correlation of P91 hardness vs PWHT time and temperature using a LMP Larsen-miller parameter.

"Nobody expects the Spanish Inquisition!"

 

[metengr]

replica;
The 190 BHN lower end hardness is not a lower bound requirement for Grade 91. This is why lower bound hardness will not be a requirement because there is little consensus around what the value should be. If the hardness values end up at 185 BHN after PWHT does this mean the material will not last in service, no. Other variables like stress and temperature are as equally important as the lower bound hardness in service.

When doing in situ metallography on the weldments before PWHT we observed intergranular cracks at HAZ at two separate weldments. Can this type of crack be removed by grinding?

Yes. These are most likely cold or toe cracks from hydrogen. I have seen them before and they tend to reside at the toe of the weld in the coarse grained region of the BM HAZ.

 

[replica]

Davefitz
Thank you ...I did not realize the P/T/F 91 had been widely discussed in this forum before.
Metengr
Thank you very much for your comments..

 

[weldstan]

I have read of some reports of minor (a few points using lab equipment) decreases in hardness after long term exposure - near 100,000 hours. I have seen micros of grade 91, tempered at a minimum temperature of 1450 F, exhibiting hardness of 178 Hv10 with coarse martenitic matrix and coarse precipitates. This plate material was manufactured in 1995 and was used for successful welding procedure qualification. This 1.5" thick material also exhibited centerline delta ferrite. Finally, I also saw some remnant 14" diameter by 1.5" wall pipe with appropriate microstructure and 230 Hv hardness on the OD to a depth of about .65" then transferring to a ferritic structure with hardness < 160 Hv on the ID. This pipe was manufactured about 2005 and I have no idea of the manufacturer.