You are on page 1of 4

Welding and Postweld

Heat Treatment of
P91 Steels
These advanced steels are quite different from and require
significantly more attention than traditional chrome-moly alloys

BY WILLIAM F. NEWELL JR.

The P(T)91 and other creep strength-


enhanced ferritic (CSEF) steels are expe-
riencing increased use around the world. What Are CSEF Steeis?
There are various sources for base mate-
rials, welding consumables, and fabrica-
tion of components. The art is such that According to the ASME Boiler and Pressure Vessel Code, Section IX,
few welding problems are encountered. QW/QB-492 (Ref. 2), creep strength-enhanced ferritic (CSEF) steels are a
However, premature failures are being en- family of ferritic steels whose creep strength is enhanced by the creation of a
countered due to design, improper heat precise condition of microstructure, specifically martensite or bainite, which
treatment, inadequate experience, or fail- is stabilized during tempering by controlled precipitation of temper-resistant
ing to observe procedures. Heat treatment carbides, carbo-nitrides, or other stable and/or meta-stable phases.
in both original component manufacture
and completed welds during final instal-
lation appears to be the primary cause of cause of their attractive thermal, physical, postweld heat treatment (PWHT).
premature failure. Improper design and and elevated-temperature mechanical Many challenges and premature fail-
inexperienced personnel also follow as im- properties. In addition to P(T)91, other ures have risen because of the strength
portant causes. alloys in this family include Grades 911, mismatch between the superior strength
92, 23, and 24 (Table 1). All of the CSEF CSEF steel and material of lower strength.
Welding steels are designed for use in high- Even weldments involving P(T)91 and
temperature applications where maintain- P(T)22 can be challenging if adequate sec-
Provided that proper welding filler ing strength and creep resistance is re- tion thickness is not present in the lower-
metals, welding procedures, and preheat quired. Properly manufactured CSEF strength material. It is even more chal-
are implemented by personnel with ade- components and systems exhibit signifi- lenging when a larger mismatch in
quate skill, welding P(T)91 or any of the cantly more strength — up to almost dou- strength or lower critical transformation
other creep strength-enhanced ferritic ble — than that of their unenhanced coun- temperatures exist between the two steels.
steels (see boxed item) is rather straight- terparts with similar chrome and molyb- Creep strength-enhanced steels should
forward. In fact, if engineering and super- denum contents. Elevated thermal con- not be used unless needed. Where re-
vision have done an adequate job, weld- ductivity offers the designer additional quired, transitions between dissimilar ma-
ing these alloys is the easy part (Ref. 1). opportunities. terials must be made in regions of the sys-
The weld metals used to join the CSEF tem where operating temperatures and
Design steels are stronger, oftentimes stronger design stresses have descended to levels
and more resistant to tempering than the within the ability of the lower-strength
Designers choose CSEF steels, partic- base metals. Satisfactory service is ob- material to provide adequate perform-
ularly for heavy-walled components, be- tained when weldments receive a proper ance. The use of machined transition or

WILLIAM F. NEWELL IR. (wfnewell@pobox.comj is cofounder and vice president of engineering, Euroweld, Ltd., and president ofW.F.
Newell & Associates, Inc., which are both in Mooresville, N C. He is also vice chair oftheAWS DIO Committee on Piping and Tubing, and
a member of the ASME BPVIX and IIIIX, Subgroup on Strength of Weldments.

Based on a presentation made during the Chrome-Moly Steels Conference held Nov. 17 during the 2009 FABTECH International &
AWS Welding Show, Chicago, III.

WELDING JOURNAL
I I 16" 1.87" Thick
16" x 1.87"
. -^-

2-1/2 Hour Soak

D- Top

"^/tW" ID 13500F min


/

Btai IMmm

Fig. 1 — Temperature gradient in heavy-wall pipe PWHT without in- Fig. 2 — Temperature gradient in heavy-wall pipe PWHT with in-
creased heated band. Minimum ID temperature required to be 1350°F creased heated band. Entire ID is greater than I350°F (730°C) min-
(730°C) (Ref. 5). imum (Ref. 5).

an alloy more closely resembling the sig-


Table 1 — Nominal Compositions of CSEF Steels nificantly less strong Grade 9. Most prob-
lems are related to either not achieving
Grade Cr Mo V w Other adequate temperatures or exceeding per-
missible temperatures.
91 9 1 0.2 — Nb, N Those performing and monitoring pre-
911 9 1 0.2 1.0 Nb, N heat and PWHT activities must be trained,
92 9 0.5 0.2 1.75 B experienced, and follow formal proce-
23 2.25 0.25 0.25 1.5 B,Nb, N
dures. The CSEF steels demand that heat-
24 2.25 1 0.25 — B, N, Ti
ing-related tasks become a primary func-
tion and not be considered a secondary or
unimportant activity. Use of AWS D10.10,
"pup" pieces of similar or even interme- ness level. A similar thought process may Recommended Practices for Local Heating
diate strength alloys is popular for miti- be extended to other CSEF steels. When of Welds in Piping and Tubing, is becom-
gating such challenges. values are discovered that are slightly less ing more common as a guide to better per-
The designer, fabricator, installer, and or much less than acceptable, it must be form local PWHTs.
owner-operator all normally assume that determined that measurements are not The base metals typically respond to
the material is as it is marked or stamped being taken in carburized surface mate- tempering and at broader and lower tem-
and will perform as shown on the material rial. Sometimes only a metallurgical ex- perature ranges than the weld metal. This
test report(s). However, many instances amination of the microstructure will pro- is particularly true when P(T)91 is joined
have surfaced where "soft spots" (not de- vide enough accurate information to make with the matching E/ER90XX-B9 compo-
carburized surface material) or even en- a proper determination (Refs. 1, 3). sitions. The matching "-B9" filler metals
tire components (fittings) have been iden- do not temper as well. This is due, in part,
tified where the material was degraded and Heat Treatment to the fact that a narrow range exists be-
not P(T)91 or other CSEF steels anymore. tween the minimum temperature required
This issue was originally discovered in the By the very definition of a CSEF (see for tempering and the maximum permit-
course of checking (using hardness tests) boxed item), these steels obtain their me- ted. This range may be as narrow as 50°
to see that welds received PWHT. Now it chanical properties by exhibiting a spe- to 750F (10o-24oC). The upper limit is dic-
is routine to check new base material be- cific microstructure. Assuming that the tated by the composition, especially the
fore it is used or installed. Where this con- chemical composition is satisfactory, the nickel plus manganese (Ni+Mn) content,
dition is discovered, two options normally enhanced properties or microstructure which affects and depresses the lower crit-
exist: scrap the material or perform a nor- can only be achieved by proper heat treat- ical transformation temperature as the
malize and temper heat treatment on the ment of components and product forms, sum of their weight-percent content in-
entire component. including welds. These materials can be crease. A maximum Ni+Mn content of
Hardness is the means normally used significantly altered by improper heat 1.5 wt-% has been established in many do-
to initially evaluate material. Most base treatment to the point that they may per- mestic codes of construction. If the actual
material and components have code cri- form significantly different than intended. composition is unknown, the user is re-
teria for the maximum hardness. New cri- For example, if P(T)91 is heated close to stricted to a tempering temperature range
teria are being considered for P(T)91 and or above its lower critical transformation of 1350° to 14250F (730° to 7750C). A
92 to impose a 190 HBW minimum hard- temperature, it can actually be turned into Ni+Mn of less than 1.0 wt-% permits the

APRIL 2010
range of 1350° to 1470oF (730° to 800oC).
The actual code of construction must be
consulted for specific rules (Refs. 1, 2).

Thermal Gradients
Recently, it has become apparent that
due to the thermal conductivity, especially
P(T)91 and 92, the heated and soak bands
need to be greater than those used (Ret 4)
on the traditional chrome-molybdenum
steels and as outlined in AWS D10.10. This
is especially true on heavy sections where
the thermal gradient between the outside
and inside of a component may be such that
the inside may not be seeing the required
PWHT temperatures. On a 2-in. (50-mm)
or thicker section, thermal gradients of 70°
to 200oF (21° to 930C) are not unusual.
Therefore, verifying procedures with
mock-ups or extensive monitoring of the
weldment during PWHT becomes criti-
cal. Just because the outside or surface
where the heaters are placed reaches tem-
perature, it cannot be assumed that the Fig. 3—Arrangement of preheating electrical resistance pads on 30- x 1-in. (760- x25-mm)
inside or surface opposite the heaters has pipe (Ref. 6).
reached temperature. Normally on heavy
or complex sections, the heated band must
be increased and the maximum possible high as 2.4 wt-%. Weld metal with higher conducted properly, but would exceed the
PWHT temperature used. This is illus- nickel content is used to satisfy some off- lower critical transformation temperature
trated in Fig. 1. The outside reached and shore toughness requirements. Such val- of the adjacent high Ni+Mn weld metal,
maintained the 1400oF (760oC) target ues will result in a lower critical transfor- thus potentially degrading its properties.
temperature, but the inside varied from mation temperature as low as 13350F It is being proposed that during manu-
1340° to 1360oF (727° to 7380C). The min- (7240F), which would be less than that re- facture of components (pipe, fittings, etc.),
imum permitted temperature was 1350oF quired by code or that necessary to tem- material specifications will limit the nickel
(730oC). When an extended heated and per "-B9" weld metal. The problem arises plus manganese to 1.0 wt-% maximum to
soak band is to be utilized, the required when the installer or repair organization avoid this issue. Similar consideration, ex-
temperatures can be achieved — Fig. 2. makes a weld near or repair of this weld cept a 1.5 wt-% maximum, is being given
metal and uses domestic (ASME) code to code rule revisions. In the interim, many
Soft Spots criteria for PWHT. The PWHT could be domestic owner-users are specifying a 1.0

Where soft spots have been found in ^^^^ Temp»f


CSEF components, most are not even near
a weld. Their cause was initially a mystery;
however, recent work by EPRI in a collab- 900
orative program with industry (Ref. 6) sug-
gests such hot areas that resulted in soft 1000
spots may have initiated during heat treat-
ment, even preheating. Figures 3 and 4 il- 1100

[i fVmn
lustrate one arrangement for preheating 1200
and the resulting effect. These findings fur-
ther substantiate the need for additional 1300
monitoring thermocouples, not just the
minimum number recommended in AWS 1400
D10.10 or other sources.
1500
Weld Metal Composition —
New Issue, Fabrication vs.
Installation 1700

A new issue concerning Ni + Mn con-


tent of weld metal actually used during
shop fabrication vs. the PWHT tempera-
tures required for installation or repair
has been identified. In certain parts of the
world, fabrication has been conducted Fig. 4 — Inside of the pipe from Fig. 2 showing heat required to obtain 400oF (204°C)
with weld metal with Ni+Mn contents as preheat in the P91 weldment region (Ref 6).

WELDING JOURNAL
wt-% maximum Ni+Mn content for weld it would be impossible to cover all situa- cial attention, or the low bidder got the
metal used on their components and dur- tions. Safety is the primary basis for our job. Sometimes problems arise from just
ing installation (Refs. 1-3). codes and standards. Attempting to dis- plain ignorance of the rules and technical
place blame for a problem or issue be- attributes required to successfully fabri-
ASME and AWS cause "it's not in the code" is technically cate and install CSEF steels.
indefensible. Contractually or legally it
Due to the increased attention CSEF may be true, but from an engineering Summary
steels require, the AWS D10 Committee standpoint, such a defense is unaccept-
removed P(T)91 materials from its exist- able. The problems and challenges sur- Base material development and code
ing guideline publication on welding rounding the CSEF steels cannot be re- acceptance have preceded effort and re-
chromium-molybdenum piping and tubing solved with code rules alone. There is no search in the areas of weldment proper-
(D10.8) and are preparing a new document substitute for sound engineering judg- ties and welding consumables for the
(D10.21, which is still pending) for ad- ment, mock-ups, and experience. This also CSEF steels. Although the base metals
vanced chrome-molys and CSEF steels. extends to material manufacturing and can offer superior properties, these bene-
Revision of D10.10 is also being consid- shop and project supervision for all as- fits can only be realized when the proper-
ered to address thermal gradient issues pects of the work. The best specification ties are maintained and not degraded dur-
being found between existing recommen- or procedure is relatively worthless unless ing manufacture, fabrication, or installa-
dations and CSEF issues. These documents it is followed. tion. From a welder's standpoint, the abil-
are pending consensus with efforts under- These facts should be obvious, but ity to weld the CSEF steels is rather
way to revise the appropriate sections and given the frequency and nature of contin- straightforward. For the CSEF steels,
rules of the ASME code. Efforts at both ued problems, the CSEF steels are not proper preheat, PWHT, and monitoring
ASME and AWS are an evolving process being given the respect and attention re- are not optional, they are mandatory.
and will continue to be so as more is learned quired. The CSEF steels are not like tra- Lessons learned with P(T)91 weld-
about these materials (Refs. 1^1). ditional chrome-molybdenum steels that ments have truly demonstrated that these
are very forgiving. Too often, the root advanced chromium-molybdenum (Cr-
cause of observed technical problems with Mo) steels are quite different and require
Supervision and CSEF steels has its genesis from a com- significantly more attention than the
Experience mercial issue, such as lack of training (no P(T)22 and more traditional chromium-
budget), inexperienced engineering or su- molybdenum alloys. Greater attention to
Codes are rarely prescriptive and only pervision (no budget), overall budget con- weld metal selection, preheat, and demon-
provide what is required, not how to do straints, unrealistic schedule, ignorance strated postweld heat treatment schedules
it. This is intentional and appropriate — of the fact that CSEF steels require spe- are some of the reasons that the CSEF al-
loys must be treated differently.
Competent, experienced engineering
and supervision are required to successfully
obtain the mechanical properties that can

WEIDHUGGER be realized with the CSEF steels. All rules,


technical requirements, and procedures
must be followed, verbatim. Overlooking
CIVER QAS DISTRIBUTION SYSTEMS
these facts will more than likely result in
problems and/or premature failure.4

References

1. Newell, W F 2009. Welding and


PWHT of P91 steels, presented at the
Welding the Chrome-Moly Steels Confer-
ence, November 17, FABTECH Interna-
Trailing tional & AWS Welding Show, Chicago, 111.
Shield Kit 2. ASME Boiler and Pressure Vessel
Code, Sections I, II, and IX, and B31.1,
New York, N.Y.: American Society of Me-
chanical Engineers.
3. Henry, J. F 2009. Cracking problems
with grade 9land other creep strength en-
hanced ferritic (csef) steels, presented at
the Weld Cracking VII: The Heat-
Affected Zone Conference, November 16,

-I FABTECH International & AWS Weld-


ing Show, Chicago, 111.
S249' 4. ANSI/AWS D10.10, Recommended
Practices for Local Heating of Welds in Pip-
ing and Tubing. 2009. Miami, Fla.: Ameri-
win can Welding Society.
5. Superheat FGH, Evans, Ga.
-LLU
<877) WtUWCR H 3447 Ftr|480l940«3»
6. Coleman, K. 2010. EPRI targets
IIWlMIIK— CSEF steel life. Energy-Tech, February.

For info go to www.aws.org/ad-index

APRIL 2010

You might also like