N+T for P91 Induction Bend

[rn01]

For the normalizing and tempering heat treatment of P91 piping after induction bending, is there one particular document or published source that can be used to establish the specific ramp up rate of heating, soak time, and ramp down rate of cooling for A335-P91 materials? Most of the current forum discussion seems to focus on the importance of maintaining 9 degrees/minute minimum cool down rate after normalizing P91 materials and not heat treating locally.

 

[metengr]

rn01;
The following heat treatment information is what I have compiled from actual procurement of Grade 91 induction bent fittings subjected to a N&T heat treatment for HRSG's we had built. The normalization heat treatment involves heating the material to the austenitizing temperature, cooling in air or quenching if thickness exceeds 3", and tempering.

Heating and cooling ramp rates should follow the Code of Construction recommendations (either B31.1 or Section I or Section VIII depending on application); for example B31.1 requires above 600 deg F, not to exceed 600 deg F/hr (I would use this as a standard ramp rate to austenitizing, and cooling down at this rate only after tempering).

1. Heat the material using the ramp rate based on the applicable Code of Construction rules, until the austenitizing temperature is reached.

2. Austenitizing temperature range should be 1900 to 1950 deg F for Grade 91 material, and holding at 1 hour per inch of thickness within this temperature range.

3. Rapidly cool the material in still air away from the furnace. The cooling rate is important to achieve 100% martensite. For material less than 3" in thickness, simply cooling in still air away from the furnace is acceptable. For thicker materials (3" or >), cooling rate is most important and liquid quenching may be required to achieve 100% transformation.

4. The material must cool to below 200 deg F before tempering.

5. Tempering shall be performed between 1375 deg F to 1420 deg F (maximum). The holding time for tempering shall be 1 hour minimum time, for material up to 2" in thickness, with 15 minutes added for each additional inch in thickness.


Follow the cool rate mentioned above down to 600 deg F, and cool in air below 600 deg F.


Most important is to monitor, monitor and monitor and record, record, record all heating and cooling cycles by the fabricator. I would also suggest some hardness testing after N&T to validate the heat treatment. Follow the requirements in ASME SA 335 for maximum hardness, and for the minimum hardness, which is not required to be reported by specification, however, I still like to confirm that over tempering did not ocurr. The hardness should be above 190 BHN.

Sources; ASME B&PV Code
ASME B31.1
The T91 Book from Vallourec Industries
EPRI

 

[rn01]

metengr:

Thank you. Another question-

For P91 induction bending, with induction temperature around 1900 F, normalizing is required with soak temperature of 1900 to 1950 F. When welding P91, only PWHT of less than 1470 F is required but welding process temperature would be over 1900F. Is the difference in required heat treatment temperatures for bends versus weld due to shorter duration of welding heat input or some other reason? I understand that a local normalizing wouldn't work due to transition zone issues but I was interested in the rationale.

 

[metengr]

rn01;
The N&T heat treatment is always preferred for post welding and forming operations on this material (pipe spools, fittings and components) because you eliminate all effects from exposure to various temperature regimes, as you indicated above, local and otherwise. The elimination of the fine grain structure in the intercritical region of the weld reduces susceptibility to Type IV cracking.

Unfortunately, one cannot always perform a N&T heat treatment when it comes to field welding. So, your only option is to accept having a girth weld to join pipe spools or welded attachment and perform a local PWHT. The key though is to be selective in locating the field welds in low stress regions or accounting for these welds during design to have them placed in low stress regions.

 

[davefitz]

In the Mannesman Vallourec booklet " the T91/P91 book", fig 6.6 is an edited version of the CCT ( or TTT) diagram . The full original version is by T. Wada " The continuous Cooling Transformation Diagram and Tempering Response of 9Cr-1Mo-V-Nb " item J-4672, from the Climax Molybdemnum Co ( 1981)

The M+V curves do not list the cooldown rates. For the curve that hugs the left boundary of ferrite + pearlite ( F+C) and yields a hardness of 406( prior to tempering), the Wada curves would show 5 C/min ( 9 F/min). This implies that to avoid the ferrite and pearlite, the item should be cooled faster than -9 F/min in the temperature range of 1475 - 1000F.

The Oak Ridge papers that discuss this alloy always make a point that they have cooled down the sample faster than -10F/min during N+T.

 

[rn01]

metengr/davefitz,

Thank you for your responses.

With regards to my P91 welding question-

For B31.1 power piping applications, there is obviously a limit as to how much spool can be normalized and tempered as a complete assembly so PWHT of a large number of P91 field welds is the only option. As stated in other threads, local normalizing of P91 piping is not an option.

At a number of power plants, it has been reported that locally or poorly heat treated P91 piping spools with bends are being cut out and the N+T process is being redone. Local temperatures that exceeded 1470 F without renormalizing is referenced as the problem. However, the local temperature issue at P91 field welds would still be present. Seems like the local temperature issue (above 1470 F)that is associated with the bending process would also be applicable to the base material adjacent to the field welded joint where renormalizing can not be performed.

Is the local transition area temperature issue at P91 bends more severe than the local temperature issue of the girth weld?

 

[metengr]

rn01;
Yes, indeed. Having an entire fitting or pipe spool that has not been properly heat treated in accordance with SA 182 or SA 335 is certainly more of a problem, and a very serious problem at that in comparison to having to deal with locally PWHT girth welds. An entire fitting or pipe spool can rupture in the longitudinal direction because of the hoop stress from pressure.

When field welding is performed on a fitting or pipe spool (with shop welds) that has already been N&T, you will have a local region on either side of the weld joint that is susceptible to Type IV cracking as a result of PWHT. This region is called the intercritical heat affected zone (ICHAZ).

Keep in mind that these local bands of ICHAZ are circumferential in orientation, so the actual service stress is only going to be 50% of the hoop stress + any bending stresses caused by thermal/mechanical loads.

The key for these girth welds in Grade 91 material is to keep the thermal/mechanical loads as low as possible. There have been studies performed in the UK that indicate a threshold stress, hoop stress + thermal mechanical stress, that would be required for increased susceptibility to Type IV cracking.

 

[rn01]

metengr,

Thank you.

Are there any maximum limits on hold times for P91 N+T?

 

[metengr]

You mean for PWHT or tempering?

 

[rn01]

Tempering

 

[metengr]

The hold time for tempering is based on 1 hour for material thickness at or below 2". In some cases others use 1 hour/inch of material thickness. If the material thickness is above 2", add 15 minutes to the 1 hour minimum for each additional inch of thickness for tempering.

 

[davefitz]

as an update, the rate of cooling required to form martensite following normalizing is a functio of the austentite grain size. The grain size is a function of the normalizing temp and time help a that temp.

The Wada results that show a need for -9 F/min is based on a 1832 F normalizing temp and an ASTM austenite grain size of 7. For a normalizing temp over 1900 F and a grain size of 10+, a 3.4 C/min (-6.5 F/min) cooling rate may be adequate.

Only way to be certain is a final plastic film replica to confirm adequate martentite.