Lifetime Assessment of NH -Plants:

Approach for Lifetime Extension

An overview of effects that may limit a plant's lifetime and why lifetime extension is recommended.

C. J. Dekker
Continental Engineering B.V., Amsterdam, The Netherlands

Introduction to carefully assess its current state. These inspection

programs, and later the consequential maintenance and
Why lifetime extension instead of a shining new renewal programs, should be planned well in advance
plant so that they may coincide as much as possible with
scheduled plant shutdowns.
When contemplating such an overhaul to extend the

T
o put it bluntly; Bucks, Quid (U.K.) or Poen
(Dutch). plant's lifetime far beyond the original designed life-
time, it may be advantageous to also consider the inclu-
Quite a few ammonia plants were built in the 1960s sion of a further energy saving program and/or mod-
and early 1970s and these plants are now approaching ernizing the main reaction of the plant by adopting a
and/or exceeding their original design lifetime, usually preconverter. This is detailed by De Wit and Riezebos
estimated to be about 20-25 years. However, replacing (1999, 2000) and Verduijn and De Wit (2001).
them with newly designed plants has become exceed-
ingly expensive and, with the current volatile markets Overview of ammonia plants worldwide
(and politics), a gamble with respect to being able to
recoup the initial investment costs within a reasonable hi terms of bulk petrochemical production ammonia
timespan. Besides, where would one be able to build a takes pride of place internationally. The best feedstock
new petrochemical plant? Suitable sites are few in for the production of ammonia is natural gas though
number, whereas opposing pressure groups are many. other raw materials like naphtha and coal are also used.
For all these reasons and, no doubt, many more, all However, the majority (about 90%) of the internation-
directly or indirectly translatable into U.S. dollars or al production of ammonia is from natural gas or oil.
into the new European currency unit EURO, it makes Ammonia production has currently grown to an esti-
sense to reconsider closing a plant just for lifetime rea- mated level of 150 million tons/year. Since 1970 the
sons. Although, to continue operating an old plant in a ammonia production in Western Europe has remained
safe and economical way beyond its original design almost constant, the introduction of new plants was
lifetime requires a diligent inspection program in order limited, while old plants have become more produc-

AMMONIA TECHNICAL MANUAL 22 2002

live. Effects that Limit the Plant Lifetime

European ammonia plants General description

The ammonia industry in the European Community Prolonged use of piping and pressure equipment may
produces approximately 11 million tons of ammonia result in construction materials deteriorating so much
per year (1994), in about 50 plants. Figure 1 shows an that the integrity of the piping and pressure equipment
overview of the age of these plants. Generally speak- cannot be guaranteed any longer. The main effects that
ing, all recently constructed EC plants have capacities may limit the lifetime of equipment can be categorized
in excess of 1,000 tons per day (tpd). Capacities up to roughly as follows:
1,800 tpd are not uncommon, although the typical size (1) Corrosion/erosion: A chemical reaction of kon at
of a large single train plant is in the range of 1,000 tpd exposed surfaces of steel with (usually) oxygen in one
to about 1,500 tpd. It is also worth noting that recently form or another in the contained medium. This reaction
developed technology allows high operation perform- may be enhanced by physical effects, such as high-liq-
ances at low capacity levels of about 400 tpd. uid velocities.
Of the 18 ammonia plants built in Europe between (2) Creep: A process whereby the material slowly
1967 and 1973, it is estimated that more than two- and progressively deforms with time under the action
thirds are in their original state, with little or no modi- of stress, for steels usually at elevated temperatures
fications. These plants are prime candidates for lifetime (400°C). Depending on the actual amount of stress, the
extension programs coupled with energy saving pro- temperature and the material, the creep process will
grams (De Wit and Riezebos, 1998; Riezbos, 2000; result in either rupture or unacceptable deformations in
Verduijn and De Wit, 2001). the structure.
(3) Chemical attack, penetrating: to be distinguished

10
9

8
.Q

l 7

6
§

I 4
'S
l 3
3
2
1

O 4

Figure 1: Number of ammonia plants built in the European Union against year of first commission.

AMMONIA TECHNICAL MANUAL 23 2002

in the following subcategories: rosive if contaminated with dirt. Depending on the
(a) Hydrogen attack: Hydrogen penetrates the steel water content and their operating temperature, both
and decarburizes it by forming methane. The methane ammonia and aqueous ammonia may result in corro-
ultimately causes fissures, cracks and blisters in the sion.
metal. The occurrence and extent of corrosion can be deter-
(b) Hydrogen embrittlement: In certain low-alloy mined with ultrasonic thickness measurements and one
metal structures the penetrated hydrogen may cause a may ascertain the remaining "corrosion" lifetime from
severe decrease in ductility which may result ultimate- the actual measured thicknesses. Many of these meas-
ly in brittle fracture. urements can be done while the plant is in operation as
(c) Nitriding: Nascent nitrogen from dissociated temperatures of up to 250°C (= 485 °F) don't pose prob-
nitrogenous material (such as ammonia) penetrates the lems for ultrasonic thickness measuring devices,
metal and forms nitrides which lowers the impact although insulation does, and that has to be locally
strength considerably. removed. Obviously, pipe lines and equipment operat-
(4) Aging processes: Precipitation of constituents ing at higher temperatures have to wait for a shut-
from supersaturated solid solutions which were down. Other forms of corrosion like pitting can be
obtained by heat-treating steel. This change in phase detected by visual and/or endoscopie inspections.
(precipitation) alters the mechanical properties of the Note that the simultaneous action of corrosion and
alloy and may occur at room temperature (natural age- erosion, thereby boosting their mutual effect to unex-
ing) or at an increased rate at above room temperatures pected heavy levels, is called "corrosion-erosion" and
(artificial aging). The presence of nitrogen in the alloy usually only occurs with fast-streaming liquids.
by, for example, nitriding processes, may greatly Elbows, reducers and T-pieces are especially at risk.
increase the precipitation rate. Steel suffers from aging
by becoming harder, less ductile and losing much of its Creep
impact strength.
These potential problems will be dealt with in a more Almost instantaneous deformation takes place when
elaborate way, together with methods on how to identi- a metal is loaded. This deformation is either elastic or
fy pressure equipment with a (potential) particular elastic/plastic and is independent of time. After this ini-
material problem and an assessment method for deter- tial deformation, a continuous or semicontinuous
mining the remaining lifetime of affected pressure deformation may be observed with time if the load is
equipment. maintained at a constant temperature. Depending on
the magnitude of the stress in the metal, this ongoing
Corrosion/ erosion deformation may eventually lead to rupture or come to
a virtual halt in the case of very low stresses.
Corrosion occurs in situations where the process This process is called "creep" and it occurs in com-
medium reacts with the (construction) material at its monly used metals and steels at relatively high temper-
surface and thereby gradually destroys it. atures, at approximately 350°C for unalloyed steels and
Process media usually found in ammonia plants are: above 400°C for low-alloyed steels. An exception is
• Synthesis gas (mixture of N2 and H2 with some lead where creep also occurs at low (room) tempera-
NH3 and CH4) tures.
• Steam and steam condensate Identifying pressure equipment potentially suffering
• Ammonia from creep is relatively easy. If at the design tempera-
• Cooling water ture the (hot) yield stress of the construction material is
• Aqueous ammonia smaller than the average creep, tensile stress to produce
Under controlled conditions, synthesis gas, steam rapture after 100,000 h (or longer for prolonged use),
(+steam condensate) as well as (treated) cooling water then that equipment is operating in the "creep" range.
are noncorrosive, although the latter might become cor- To estimate the remaining "creep" lifetime, one

AMMONIA TECHNICAL MANUAL 24 2002

needs to know the operating history of the pressure in hardness and an increasing ductility.
equipment, that is, the elapsed times for all pressure- Internal decarburization is the more serious problem,
temperature combinations at which it has operated. hydrogen penetrating the material and reacting with
When enough creep data for the considered material is carbon to form methane. Accumulation of the
available (and that should not be a problem for con- entrapped methane in voids, give rise to locally high
struction materials available 25 years ago), then, by stresses which ultimately will fissure and crack the
means of the "Larson-Miller" method or the steel. The multitude of initially just microscopic fis-
"Manson-Haferd" method, it can be calculated or pre- sures causes the steel to lose much of its strength and
dicted how much creep lifetime still remains. ductility.
Later an example is given as to how the creep life- Judiciously following the so-called "Nelson curves"
time can be calculated with the help of the linear dam- as presented in, for example, API Recommended
age rule, sometimes also called the life fraction rule or Practice 941, will prevent decarburization of steel in
Miller's rule. However, in that example we won't make hydrogen service. We will assume that this was also
use of the Larson-Miller method but we will exploit done when all the existing plants were newly built.
the rectilinear relation between log(acreep) and log(rup- However, one should be aware of cumulated experi-
ture time). ence and progressing metallurgical insights having
changed since then, for example, Mo-steel is no longer
Nitriding considered a "hydrogen resistant" material. Hence, one
should be vigilant when dealing with equipment and/or
Nitriding is a process at elevated temperatures piping made from such types of material.
whereby nitrogen atoms migrate or diffuse from a suit-
able, nitrogen-containing process medium into low- Embrittlement
alloy steels. A very brittle scale with a thickness of
about 2-3 mm will form at the surface of the steel and Heat-treatment may produce high-yield strength val-
this scale has a high nitrogen content in the form of var- ues together, of course, with the corresponding
ious nitrides. microstructures. Such microstructures can be suscepti-
Process media in the plant which may cause this ble to hydrogen stress, cracking or embrittlement.
"nitriding" phenomenon are the synthesis make-up gas Especially creep-resistant 2V4Cr-lMo steels are sus-
and the reactor effluent which contains nitrogen and ceptible to embrittlement at temperatures from 375°C
ammonia. Experience has shown that 400°C is the min- to 580°C. The main characteristic of this temper
imum temperature for nitriding to start. embrittlement is an increase of the impact energy-tran-
To prevent nitriding, low-alloy steels shall be sition temperature and this embrittlement phenomenon
cladded, lined and/or butter-welded with, for example, is closely associated with the presence of the elements
Inconel or stainless steel. Internal inspection and hard- Si, Mn, P, Sb, As and Sn in the steel.
ness measuring may reveal whether nitriding occurred Nowadays, materials for new equipment are tested to
or not. determine the likelihood of hydrogen embrittlement.
This test is called "step cooling test" (American
Hydrogen Petroleum Institute, 1982) whereby the material is
heated up to 600°C and cooled down to 320°C in six
Effects of hydrogen are twofold: steps, each taking from l h up to 100 h (For more
Decarburization, both surface as well as internal. details, see Literature Cited.) Sometimes even test
Surface decarburization does not result in fissures, just blocks are installed within the equipment, so that prop-
the metal's surface layer is depleted with carbon as this er test pieces will be available for future destructive
reacts with the hydrogen in the synthesis gas to form testing during its lifetime.
CH4 (or CO when oxygen-containing gases are pres- Furthermore, current learned opinion suggests that
ent). Effects are slight losses both in strength as well as by limiting the amount of certain chemical elements in

AMMONIA TECHNICAL MANUAL 25 2002

the steel, by means of the so-called Watanabe "J"-fac- Remark: That is, of course, when all restrictions
tor (Watanabe et al., 1974) and the Bruscato "X"-factor applicable to considered material have been taken into
(Bruscato, 1970), and at the same time limiting the account. An example of such a restriction is note G18
hardness to 248 HV maximum, the danger of hydrogen in ASME II-D for plate A-204 grade B (Cr-V2 Mo
embrittlement has been eliminated as much as possible. steel); prolonged exposure to temperatures above
Embrittlement cannot be excluded from potential 875°F (470°C) may convert the carbide phase to
problems as may be encountered here, but by evaluat- graphite.
ing the Watanabe and Bruscato factors from the chem-
ical composition, as given in the original material cer- Practical Approach
tificates, important clues may become apparent as to
the likelihood of hydrogen embrittlement problems. If In normal ammonia plants the number of pipe lines
deemed likely, then metal hardness measurements at are far in excess of a thousand and the number of major
the inside surfaces of the equipment will have to be equipment are more than a hundred, therefore, the task
organized. of evaluating the remaining lifetime in a thorough man-
ner is enormous, as in principle all these items should
Aging be dealt with. It must be decided what effects are of
importance for each considered item and appropriate
Metastable microstructures like perlite may exhibit a data should be collected, either in order to put quanti-
precipitation process in order to get rid of unwanted tative figures behind these effects or to, at least, make
solute from the supersaturated solid solution. The start informed decisions. This could mean initiating certain
of this process can be triggered by heat treatment, plas- inspections (such as ultrasonic thickness measurement,
tic deformation or otherwise, but in the case of "natu- hardness measurement at internal metal surfaces, and
ral" aging no external influence is required, although so on), collecting operating histories for creep evalua-
elevated temperatures accelerate the process. tions and/or unearthing the original material certifi-
The resulting precipitation is mainly in the form of cates to establish the value of Watanabe's "J"-factor
carbides for ferritic alloy steels. As mentioned earlier, and/or Bruscato's "X"-factor.
the alloy usually suffers by becoming harder, less duc- To facilitate deciding what effects may have a bear-
tile and losing much of its impact strength. ing on the various pipe lines and equipment items,
Concerning the practical consequences of the phe- C.E. has developed decision diagrams for all media
nomenon of "aging" for lifetime assessment, a very present in ammonia plants. The intention is that for
down-to-earth approach may be followed here. If a each entry on a line list and an equipment list, one
material standard or specification provides long-term looks through the appropriate decision diagram in
creep values at elevated temperatures, then this materi- order to narrow down, in a structured manner, the
al must be regarded as being suitable at that tempera- risks/threats for each specific pipe line or equipment
ture for the specified period. item. Having established the relevant threats, it
For instance, EN 10028 Part 2 (1992) provides creep becomes clear what kind of testing and/or data collect-
values for the material 10 Cr Mo 9 10 (DIN material ing is appropriate and planning may go ahead for all
number 1.7380) up to 600°C and for up to 200,000 h. these inspections and/or tests. Tests which can be done
The mere fact that these values are given indicates that during normal plant operation are generally not a prob-
aging is not a problem for the considered material and lem, but inspections/tests inside of equipment need to
that special examinations are not required to monitor be planned ahead, as it is better to coincide with regu-
any aging effects. lar plant stops. In this way the last few years of the
Even when creep values are extrapolated to establish plant's normal lifetime can be spent producing as nor-
usage factors, as discussed earlier, in conjunction with mally as possible, while at the same time one may gath-
T 0102 beyond the longest listed creep values (usually er sufficient information to decide whether to extend
200,000 h) then aging effects need not be anticipated the plant's lifetime, by replacing/repairing those
for these extended periods. items/pipe lines beyond safe use and retaining /adapt-

AMMONIA TECHNICAL MANUAL 26 2002

ing all others or to shut down the plant permanently. Note that other construction codes don't deviate on this
In the decision diagrams we have included two other point although in exceptional cases the time base of
aspects that are of interest for an assessment of remain- 100,000 h can be changed in a more appropriate value.
ing lifetime and/or major-overhaul associated with life- The difference is due to the fact that the actual oper-
time extension. ating temperature and actual operating pressure are sel-
dom equal to the design temperature and design pres-
External corrosion sure. An hour working at a temperature of say 20°C
(35°F) lower than the design pressure makes that such
Many years of unchecked external corrosion, either a "creep" hour counts for far less than an hour at the
unseen under insulation with damaged and/or broken design temperature, that is, while assuming that in both
sheeting, or simply due a lack in the upkeep of paint cases the pressures were equal. The following example
protection systems, may result in serious problems illustrates the fact that a pressure vessel after operating
(such as leakages). In more than one instance piping for 250,000 h is still capable of operating much longer.
and equipment had to be replaced either wholly or part-
Cylindrical Vessel.
• Material: 10 Cr Mo 9 10 (EN 10028-2:1992).
Asbestos • Design conditions: pressure 5 MPa (725 psi), tem-
perature 530°C (986°F), lifetime 100,000 h, corrosion
In the past, this excellent material both for insulation 3 mm (0.12 in.).
purposes as well as for gaskets was used quite exten- • Dimensions: O.D. 2,000 mm (783A in.); nominal
sive without knowing the health consequences. thickness 85 mm (3"/j2 in.), undertolerance 1 mm (0.04
Nowadays authorities and also workers unions insist on
in.).
removing asbestos altogether and replacing it with
Approximate estimates of the remaining life of a ves-
acceptable substitute materials.
sel or pipe can be obtained by using the "linear dam-
Figures 1-4 illustrate this process of going through
age" rule. The creep damage to the material by operat-
the enormous amount of pipe lines and equipment in
ing at a stress 0j and at a temperature 9j is assumed to
order to arrive at the most likely failure cause(s), for
which that particular pipe or equipment shall be be in ratio with the actual operating time (toper y) vs. the
checked and/or evaluated. (minimum) time to cause rupture temperature 9j with
stress 0j (tj-jj). The cumulative creep damage due to
Example of Creep Assessment operating at different conditions, characterized by pres-
sure (but read stress) and temperature, also known as
It is important to make a clear distinction between usage factor, should not exceed 1.0:
operating time and (creep) lifetime, these two entities
are certainly not identical. Although plants are usually u (usage factor) = oper,ij < 1.0
designed for a lifetime of 100,000 h, that is, the design
t
stresses in the creep range of the construction materials
are based on creep stress values for 100,000 h, this
doesn't mean that the operating time of the plant can't In order to evaluate the usage factor, the load history
be much longer. should be available. The load history is the total time at
For the design stress, see Appendix I of ASME H-D each relevant pressure/temperature-combination. The
where in the creep range for temperatures up to 1,500°F pressure range and temperature range are to be subdi-
(about 820°C) the design stress is limited by: (1) 100% vided into convenient intervals in order to arrive at a
of the average stress to produce a creep rate of distinct number of load combinations. All the load
0.01%/1,000 h; (2) 67% of the average stress to cause combinations together are sometimes called the "load-
rupture at the end of 100,000 h and 80% of the mini- ing matrix".
mum stress to cause rupture at the end of 100,000 h.

AMMONIA TECHNICAL MANUAL 27 2002

 J3ACKTODECJPEP1. _
OTHER POTENTIAL RISKS."

43BESIOS-
ET6RNM.
TC- > "Eä
CORROSION
RISK

/
NO
EXIERti*'.

RISK
l
. Jf

Figure 2. Decision diagram for steam condensate.

AMMONIA TECHNICAL MANUAL 28 2002

 Figure 3. Decision diagram for steam (42/50 barg.) and BFW.

AMMONIA TECHNICAL MANUAL 29 2002

 Figure 4. Decision diagram for process gas/liquid.

Suppose the loading matrix consists of just four combinations:

Number of Pressure Temperature Holding

Load inMPa in°C Time in h
v- pc] Combination

610 «C 1 4.8 520 50,000

4M <C 2 4.8 500 50,000

«w«e 3 4.8 520 50,000

« «e
-
4 4.8 500 100,000
\T
41 4« 4B »tM>4

AMMONIA TECHNICAL MANUAL 30 2002

 The equivalent primary membrane stress occurring in In the material standard EN 10028 both the average
load combinations No. 1 and No. 2 is: stresses to cause rupture as well as the average stresses
to produce a creep of 1% are given for several time
OD-d 2000 - 81
a = pd = 4.8* bases. Unfortunately, the minimum stresses to cause
2zd 2 * 1.0 * 81 rupture are not provided. We will resort to the widely
accepted approximation of the minimum stress to cause
=56.9 MPa rupture being not less than 80% of the average stress to
produce rupture.
R
m,min(100'000 fc 520°C> = °-8 * Rmg (100,000 h;
The equivalent primary membrane stress for load 520°C) = 0.8 *103 MPa = 82.4 MPa.
combinations No. 3 and No. 4 (4.6 MPa) is: R
m,min(2°0>0°0 h; 520°C) = 0.8 * Rmg (200,000 h;
520°C) = 0.8 *91 MPa = 72.8 MPa.
R
OD - d 2000 - 81 m,mm(l°0>000 h; 500°C) = 0.8 * Rmg (100,000 h;
0- * = 4,6 *
= Pd 2zd 2*1.0*81 500°C) = 0.8 *135 MPa = 108. MPa.
Rm min(200,000 h; 500°C) = 0.8 * Rmg (200,000 h;
= 54.5 MPa2 500°C) = 0.8 *120 MPa = 96.0 MPa.
An example of the estimation of the time basis for
stress value of 56.9 MPa.at 520°C:
Here, in our case, we should have available the min- with 9 = 520°C
imum times to rupture for stresses of 56.9 MPa and
54.5 MPa respectively, both at 520°C as well as at tti
_ R.n.,»nl',;-9))- I°B ai> log 1,1 •
" (log gj) lOg I,;

> ~ log
500°C. These can be estimated realistically by assum-
ing a rectilinear relationship between the logarithm of
rupture time and the logarithm of minimum creep rup- Iog0i= log 56.9= 1.75511
ture stress. If one were to have two minimum rupture log &! = log 100,000 = 5
stresses for different time bases, but at the same tem-
perature, then the minimum creep rupture stress at the log tr2 = log 200,000 = 5.30103
considered temperature for any time basis, may be esti- log Rmimin(100,000 h 520°C) = 1.91593
mated by either linear interpolating or linear extrapo- log Rmmin(200,000 h 520°C) = 1.86213
lating. However, when extrapolating one should not go ===> t, (56.9 MPa, 520°C) = 1058997 = 794,000 h
beyond three times the longest available time base. In a similar way one establishes:
Denoting the minimum rupture stress with time basis tr(54.5 MPa, 520°C) = 1060043 = 1 010 000 h
IT and temperature 0 as Rm min^1"'®)'tnen two known
t,.(56.9 MPa, 500°C) = 10663M = 4 358 000 h
rupture stresses with time bases t^ and t^ are given as
t,.(54.5 MPa, 500°C) = 1067495 = 5 616 000 h
VminCtri;«) and VrainC^)' respectively. To esti- As extrapolation should not go beyond three times
mate the time basis ^ corresponding to a specific value the highest available time base, all estimates will be
of the minimum rupture stress Rm niiiA'®) =CT>a^ at reduced down to 600,000 h being three times 200,000
the same temperature 0, then one may use the follow- h.
ing formula:

Rm,min - (logRmiinin (^,0) - log O) * log t^

logt r =
Rm.min -6) ~ log Rm, min

AMMONIA TECHNICAL MANUAL 31 2002

 Having estimates for the (minimum) times to pro- operating time! Other aspects may prove to be larger
duce rupture for the relevant stress values and temper- problems, but they are surmountable too.
atures, together with the actual occurring holding times The key to success is, or so we believe and have
one may establish the numeric value for the usage fac- experienced, to start rather early with making a sys-
tor tematic inventory of all potential problems for all pip-
ing and equipment in the plant. Our decision diagrams,
YV ty 50,000 50,000 50,000 100,000
\\~~~ (t If J mmH^^ ""•"
•-
Jv
i
,±.
i •-
I
— r are helpful in this respect as they assist in getting a
i j ^y 600,000 600,000 600,000 600,000 "systemic" and "objective" inventory. Having done so,
one can start planning and organizing with confidence
= 0.42 all required specific investigations and testing in the
coming regular shutdown(s) of the plant. The results of
In spite of an overall operating time of 250,000 h all these inspections together with the numeric values
with an original design lifetime of only 100,000 h, of the "creep" usage factors based on the operating his-
there is apparently no danger for creep rupture in this tories of the relevant items will make clear what piping
example. Having spent only about 42% of the "creep" and/or equipment has to be replaced, repaired or does-
lifetime, there is no reason not to continue with operat- n't need any further attention for prolonged (safe) use.
ing this piece of equipment unless off course severe
corrosion and/or other forms of severe material detori- Literature Cited
ation have occurred.
American Petroleum Institute, Characterization Study
Conclusion of Temper Embrittlement of Chromium-Molybdenum
Steels, API Publication 959 (1982).
Faced with the question "what to do with a 25-year Bruscato, R., 'Temper Embrittlement and Creep
old plant?", then one possible answer is: Embrittlement of 2_ Cr - 1 Mo Shielded Metal Arc
"Revamp, that is, a major overhaul coupled with a Weld Deposits," Welding J. 49 (4) (1970).
lifetime extension program." De Wit, J. J., and A. Riezebos, "Upgrading a 25-year old
This is in our experience a sensible answer but Ammonia Plant," Ammonia Plant Safety & Related
requires a thorough investigation beforehand into the Facilities, Vol. 39, AIChE, New York (1999).
status of the plant, whether it is feasible with regards to Riezbos, A., "Pre-Reforming, A Revamp Option," pre
creep and/or the number of pipelines and equipment sented at the Sud-Chemie Seminar, Lisbon, Portugal
that have to be replaced. The final "go-ahead" decision (May 10-12, 2000).
Verduijn, W. D., and J. J. de Wit, "Energy Conservation:
will depend partly on the outcome of this investigation.
Key to Survival for Fertiliser Producers," Proc. No.
Creep is certainly not the major problem for piping 479, presen- ted at the International Fertilizer
and equipment that most of us would fear it to be. That Society Meeting, London (May 10,2001).
is evidently demonstrated with the given example eval- Watanabe, J. et al.,'Temper Embrittlement of 2_ Cr -1
uation of a vessel operating within the creep range for Mo Pressure Vessel Steel," ASME 29th Petroleum
250,000 h which corresponds with over 30 years of Mech. Eng. Conf, Dallas, TX (1974).

AMMONIA TECHNICAL MANUAL 32 2002

 QUESTIONS AND ANSWERS
Question: Start/stop cycles of components operating in accommodating differential thermal expansions.
the creep regime is the main damage mechanism for Start/stop cycles have in principle nothing to do with
these components. How is this effect taken into creep, and, if the lifetime of a component shows a
account? marked dependence on the number of (thermal) load
C. J. Dekker, Continental Engineering B.V.: Creep is cycles, then the advice is to review the design with
the slow but continuous increasing elongation of mate- respect to thermal gradients and adapt the design
rial due to sustained (primary) stresses in the material. accordingly.
Total duration of the sustained stress, the stress level, Question: Low melting residuals to calculate the F or
and, of course, the temperature of the material are X factors are not available on certificates of compo-
important for creep, not the number of interruptions in nents built in the 1960s or 1970s. How to determine
stress and/or temperature. However, for fatigue, the them?
number of startups and shutdowns, or taken together as Dekker: To evaluate the so-called J-factor of Watanabe
start/stop cycles, are of prime importance. Evaluating or the X-factor of Bruscato, one needs to know the
fatigue requires knowing the alternating stress ampli- amount of Si, Mn, P, Sb, As and Sn in the 2_Cr-lMo
tude which includes the secondary and peak stress steel. Indeed, these elements will rarely be mentioned
catagories. Often, during startups and shutdowns, large in the chemical analyses of steels dating back from the
temperature gradients may occur in constructions 1960s and 1970s and, consequently, one has to resort to
which give rise to large secondary stresses. This is other means. For instance, these days, it is possible to
especially true during emergency shutdowns and for determine such trace elements in situ with portable
constructions that are poorly designed with respect to (micro)spectrometric equipment (working on the prin-
ciples of gaschromatography).

AMMONIA TECHNICAL MANUAL 33 2002