Sei sulla pagina 1di 4

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/291818792

Controlling heat treatment of welded P91

Article  in  Welding Journal · June 2006

CITATIONS READS
8 502

2 authors, including:

Patric de Smet
Siemens
2 PUBLICATIONS   10 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Patric de Smet on 26 August 2019.

The user has requested enhancement of the downloaded file.


Controlling
Controlling Heat
Heat Treatment
Treatment
of
of Welded
Welded P91
P91
Hardness testing proves to be a powerful tool for checking the condition of P91

BY PATRIC DE SMET AND HANS VAN WORTEL

Modified 9Cr-1Mo, commonly re- interface, a fine-grained heat-affected trend line. Further, this figure shows that
ferred to as P91, is widely used in the zone (FGHAZ) can be distinguished. the base metal hardness does not vary
power industry because of its superior Figure 1 represents typical hardnesses much with LMP.
properties at elevated temperatures. P91 along a postweld heat-treated weld metal, Base metal hardness is found at a level
get its favorable strength and toughness HAZ, and base metal of P91. Figure 1 somewhat above 200 VHN. Only at very
properties from its microstructure. The shows that the highest hardness is found high values for LMP can a drop in hard-
desired microstructure can only be ob- in the CGHAZ, and the lowest hardness ness below, but still close to 200 VHN, be
tained using material with a well-balanced in the FGHAZ. This FGHAZ is the area found.
chemical analysis and proper heat treat- with the weakest creep properties and the This close relation between hardness
ment. Heat treatment is an especially crit- location where the so-called Type IV and PWHT parameters (expressed as
ical step during the manufacturing and cracking occurs (Ref. 1). LMP), for the weld metal, is very helpful
fabrication of P91. for checking the P91 condition after heat
P91 is a martensitic chromium-molyb- treatment.
denum steel, microalloyed with vanadium Effects of Variation in Heat If a portable hardness test is used,
and niobium, and with a controlled nitro- Treatment Parameters some care should be taken. Often the base
gen content. In the as-welded condition, during PWHT metal, weld metal, and HAZ are subjected
P91’s microstructure consists of “fresh” to hardness testing. However, consider-
martensite. This untempered martensite TNO Industrial Technology and NEM ing the probe dimension, it is not easy to
is hard and brittle. The toughness is low performed many hardness tests on P91 in test the narrow heat-affected zone. Even
and the material is prone to stress corro- different heat treatment conditions. Fig- if one could get some results out of the
sion cracking. In addition, while the creep ure 2 illustrates the effects of various heat HAZ, it is difficult to indicate if the val-
properties of fresh martensite are un- treatments on weld metal (maximum) ues come from the FGHAZ or from the
known, they can safely be assumed to be hardness. The hardness is presented as a CGHAZ. A relatively low value can be ac-
unfavorable. Therefore, tempering of the function of the Larson Miller parameter ceptable if it comes from the FGHAZ, but
formed martensite in P91 after welding is (LMP), which is used to characterize the unacceptable if it is found in the CGHAZ
necessary to obtain the required service combination of heat treatment time and — Fig. 1.
properties. Proper heat treatment will re- temperature. The LMP is valid for the tem- A fair approach is to test only the weld
sult in tempered martensite with precipi- perature of stress relief heat treatment. and base metals. Both of these zones can
tated carbides (M23C6) and vanadium/ Postweld heat treatment with an LMP be easily distinguished. Further, Brühl et
niobium-rich carbo-nitrides. The hard- between the two dotted vertical lines in al. (Ref. 1) found that the FGHAZ
ness of weld and base metal will be be- Fig. 2 leads to good material properties. hardness is approximately 20 VHN lower
tween 200 and 270 Vickers hardness num- The window of required heat treatment is than the base metal. Based on the col-
ber (VHN), and toughness will be suffi- represented by 750°C/2 h (i.e., LMP = lected data, this “fixed” difference in
cient, i.e., higher than 27 J (20 ft-lb) at 20.8) as the lower limit and 770°C/10 h hardness between FGHAZ and base
20°C (68°F). (LMP = 21.9) as the upper limit. metal is confirmed.
A relation between CGHAZ hardness
Hardness of P91 Hardness Test as a and weld or base metal hardness is less
pronounced, but it was found that the
Control of PWHT CGHAZ is roughly up to 20 VHN higher
The heat from welding will affect the
base metal. The heat-affected zone From Fig. 2, it is clear that there is a than the mean weld metal hardness.
(HAZ) starts directly adjacent to the weld direct relation between the Larson Miller Figure 2 can be used as a guide for in-
metal at the weld interface. A coarse- parameter and weighted average weld terpretation of the test results, keeping
grained heat-affected zone (CGHAZ) in metal hardness; most of the hardness val- the accuracy of portable hardness testing
the HAZ, and, farther away from the weld ues are expected to be below the average in mind. When analyzing portable test re-

PATRIC DE SMET (pdsmet@nem.nl) is welding and materials engineer for NEM b.v., The Netherlands, and HANS VAN WORTEL
(hans.vanwortel@tno.nl) is senior project manager at TNO Industrial Technology, The Netherlands.

42 JUNE 2006
CGHAZ

Base Metal

FGHAZ

Fig. 1 — Typical hardness across P91 weld metal, HAZ, and base Fig. 2 — Trend of hardness as function of Larson Miller parameter
metal in PWHT condition. (LMP).

Fig. 3 — Charpy-V impact toughness as function of Larson Miller pa- Fig.4 — Effect of PWHT above lower critical temperature A1.
rameter (LMP).

sults, keep in mind that the accuracy is weld metal with varying LMP for weld ment at a too low LMP will result in a too
lower than that of laboratory test results. metal of submerged arc welds (SAW) is high weld metal hardness, and can easily
The effect of inadequate handling of the illustrated by Fig. 3. be found from hardness test results. Heat
test equipment, poor surface preparation For SAW, it was found that the tough- treatment beyond 770°C/10 h (i.e., too
at test location, or testing on a decarbur- ness in the as-welded condition is around high LMP) results in low hardness. As
ized or heavily deformed surface layer can 7 J. Only after a PWHT above LMP of long as the heat treatment temperature is
also lead to inaccurate results. 20.8 can a toughness higher than 27 J be below the lower critical temperature A11,
guaranteed. This is in line with the values the hardness will stay around 200 VHN,
Toughness of filler metal suppliers. Most of the sup- even at a very long PWHT duration.
pliers of P91 filler metals give data for Heat treatment above the lower criti-
In Europe, a Charpy-V impact tough- weld metal with a heat treatment around cal temperature A 1 will ruin the mater-
ness of 27 J (20 ft-lb) is generally consid- 760°C (1400°F)/2 h (LMP = 21). ial’s properties (Ref. 4). Heat treatment
ered as sufficiently ductile. Directive 97/23 Finally, it is interesting to note that above A1 results in (partial) transforma-
of the European Parliament (Ref. 2), re- NEM and TNO found that in the PWHT tion of martensite into austenite.
ferred to as PED (pressure equipment di- condition weld metals with acceptable The formed austenite will be trans-
rective) adopted this 27 J at 20°C (or at toughness showed hardnesses lower than formed into fresh, hard martensite at
lowest operating temperature) as a safe 300 VHN. rather rapid cooling rates and into soft
limit (Ref. 2). Some codes, however, re- ferrite at slow cooling rates. Both situa-
quire even higher values of 41 J (30 ft-lb) Intercritical PWHT
(Refs. 3, 5).
Regarding toughness of a welded joint, In some unfortunate cases, compo- 1. A1 is the lower transformation tempera-
the weld metal is considered most critical. nents may be postweld heat-treated out ture. On heating, ferrite is partly trans-
Results of Charpy-V impact testing of of the required range. Postweld heat treat- formed into austenite (a Æ a +g).

WELDING JOURNAL 43
tions are unwanted, but formation of fer- D. Heat treat-
rite is considered to be the worst situation ment above weld,
because creep strength drops to that for but below base
grade P22. metal A1 , followed
The lower critical temperature, A1, is by slow cooling.
a function of nickel (Ni) and manganese Rapid cooling
(Mn) (Ref. 6). For P91 base metal, A1 is refers to a speed
found at around 800°–830°C. Weld metal just high enough to
contains higher percentage Ni and Mn prevent ferrite for-
and, therefore, has a lower A1. The lower mation out of
critical transformation temperature can austenite.
be estimated using the ORNL data-based Figure 4 illus-
relation A 1 = 848–42(Ni[%]+Mn[%]) trates situation A,
[°C]. For Ni+Mn of 1.5%, the A1 is found PWHT above weld
at around 785°C (1445°F), which is close and base metal A1,
to the upper limit of PWHT. followed by rapid
Because of the poor properties of heat cooling. Because of
treatment in the intercritical area (above formation of fresh
A1, below A3 [for A3 refer to footnote 2]), martensite upon Fig. 5 — Effect of PWHT above weld metal, but below base metal A1,
it is very important to prevent such a cooling, a higher followed by rapid cooling.
PWHT. hardness is found
Depending on heat treatment temper- for the base and
ature and cooling rate, basically four ex- weld metals, as well.
treme situations are possible. Figure 4 also il-
A. Heat treatment above weld and base lustrates situation B, i.e., PWHT above As has been shown, the base metal hard-
metal A1, followed by rapid cooling weld and base metal A1, but followed by ness is not expected to change during
B. Heat treatment above weld and base slow cooling. Formed austenite is now manufacturing.
metal A1, followed by slow cooling transformed into soft ferrite, resulting in Further, it is clear that in order to ob-
C. Heat treatment above weld, but low hardness for weld and base metal. tain good service properties and tough-
below base metal A 1, followed by rapid Experiments also showed that in the ness, the window for PWHT is relatively
cooling case where the weld metal is heat-treated narrow. A good practice is to perform a
above A 1 but the base metal is below it furnace survey to ensure PWHT within
(situation D), low hardness values are the required temperature window.
found in the range indicated in Fig. 4. Al- Finally, it is emphasized that not only
though no ferrite is formed in the base heat treatment temperature, but both
metal, the hardness drops due to a soft an- PWHT time and temperature (LMP) de-
nealing effect, i.e., coarsening of carbides. termine the effect of the PWHT cycle.
Figure 5 represents situation C, i.e., for Postweld heat treatment within the
PWHT above weld metal but below base temperature range of 750° to 770°C (1380°
metal A1, followed by rapid cooling. The to 1420°F) and an LMP of around 21 will
weld metal now contains some fresh lead to good high-temperature properties
martensite besides the tempered marten- and toughness for a safe hydrostatic test
site. The base metal is softened due to and to put the installation into service. ◆
coarsening of carbides. This situation may
be somewhat difficult to detect by hard-
ness testing, since the hardness values are References
comparable to the acceptable range for
weld metal and just below the required 1. Brühl, F., et al. 1990. Behavior of the
base metal hardness. Knowing the initial 9% chromium steel P91 and its weldments
base metal hardness before PWHT will in short and long term tests. Proceedings
help to distinguish this situation. of the ASME/IEEE Power Generation Con-
ference. Boston, Mass. pp. 1–10.
2. Directive 97/23/EC of the European
Final Materials Condition Parliament and of the council. May 29,
Hardness is directly related to PWHT 1997, p. 23.
time and temperature characterized as 3. VdTÜV 511-2, p. 2.
LMP. This makes hardness testing a pow- 4. Henry, J. F. 2005. Growing experi-
erful tool for checking the material’s con- ence with P91/T91 forcing essential code
dition. Testing only at the end of the line changes. Combined Cycle Journal,
— after final PWHT — is a good prac- 1Q/2005.
tice. However, it is better to start with a 5. Dutch Rules for Pressure Vessels,
hardness test on incoming material. This Sheet M0110, p. 7.
helps to interpret the results after PWHT. 6. Newel, W. 2002. Guideline for weld-
ing P(T)91 materials, EPRI, June, p. 9.

2. A3 is the upper transformation tempera-


ture. On heating, ferrite is transformed into
austenite (a+g Æ g ).
Circle No. 5 on Reader Info-Card

44 JUNE 2006

View publication stats

Potrebbero piacerti anche