Proceedings of CREEP 8

Eighth International Conference on Creep and Fatigue at Elevated Temperatures

July 22-26, 2007, San Antonio, Texas

PVP2007-26827

REMNANT LIFE ASSESSMENT OF PLATFORMER HEATER T9 TUBES USING

API 579 OMEGA METHOD

Ohgeon Kwon and Charles Thomas

MPT Solutions

David Knowles and Andrew Saunders-Tack

Shell Global Solutions International

ABSTRACT procedure.1, 2 The accuracy of the results increase from Level 1

Tubes operated in furnaces are known to suffer from to Level 3, but the level of data requirements and the cost also
damage as a result of time dependent strain, i.e. creep. As such increases at the same time1.
there is a need to predict remaining life. The MPC Omega Level 1 or 2 “desk top” life assessment using an API 5303
Method has been adopted by API (API 579 “Fitness for creep rupture-based method provides plant operators with
Service”) as the standard method of life assessment and has relatively quick and simple answers. It is an attractive method
been used to perform a remnant life assessment on a platformer when the plant has to decide whether the component can be put
furnace. Samples of T9 tubes were removed from the heater back into operation within a short period time such as the
based on site inspection to determine the levels of strain. This inspection period. However, it is often found that current
inspection was undertaken after 224,000 hours operation at a service life has already exceeded the minimum design life or
process gas outlet temperature of 540°C. The uncertainty API 530 mean material data optimistically predicts longer lives
caused by the difference in life based either on minimum and than may be expected.4, 5 Furthermore it is critically dependent
mean material properties, indicates the necessity of proceeding upon having a good history for stress and thermal exposure of
to a Level 3 assessment in order to determine the actual the tube materials which often is not available. As an
material properties. The results of this approach using the API alternative, API 579 strain based life assessment6, the so-called
579 Omega method demonstrated that the hot face material had MPC Omega method, can be employed in a complementary
properties within the scatter band, mid-way between the mean way using short term tests to gain an accurate assessment of
and minimum data lines and provided confidence as to the current material condition.
remaining life. The Omega approach successfully allowed the In recent years, the MPC Omega method has been applied
impact of variations in future operating conditions to be to many applications in the petroleum and power industries.4, 7
explored providing information to set the operational envelope This approach allows the use of assessment specific material
and define inspection and replacement strategies. properties obtained from laboratory tests but still in a rapid and
timely manner.
INTRODUCTION In this paper the procedure of phased approach remnant
Reliable and accurate life assessment of components life assessment will be demonstrated for platformer heater tubes
operating in creep range enables plant operators and owners to using API 530 and 579 methods. Application of the results
substantially extend component life. Significant economic from the Omega creep testing method to the life assessment
benefit can be achieved by making correct decisions to will be discussed in terms of significance of using actual
repair/replace components, continue operation with revised material properties evaluated from the test on the assessment.
operating conditions and determine the inspection schedule. The effect of the parameters such as temperature, pressure and
A phased approach (three-level approach) has been widely corrosion rate on the remnant life can be examined by carrying
used as the most logical and cost-effective life assessment out a sensitivity analysis. This will enable plant operators to

1 Copyright © 2007 by ASME

better manage the future integrity of the heater tubes and safely Table 1 Thickness measured from the ring samples (unit:
extend the tube life with higher confidence level. mm, nominal thickness: 8.55mm)
12 o’clock 3 o’clock 6 o’clock 9 o’clock

C0 →C4 =Material coefficients for strain rate and omega

NOMENCLATURE Sample 1 6.83 7.64 6.78 6.5
Sample 2 5.68 6.67 6.43 6.88
parameter
Δsr =Adjustment factor to account for the material scatter
band, a range of -0.5 for the lower bound of creep The minimum wall thickness measured was 5.68mm at the
strength to +0.5 for the upper bound of creep strength 12 o’clock position of the Sample 2, which suffered most wall
Δcd = Adjustment factor for creep ductility, a range of +0.3 thinning due to corrosion during operation. This is consistent
with the fireside of the tube. The maximum wall thickness was
ε&o
for brittle behaviour and -0.3 for ductile behaviour
=original strain rate measured at the 3 o’clock position of Sample 1, indicating that
ε&co
this area had been exposed to relatively low temperatures. The
=current strain rate maximum reduction from the nominal wall thickness (8.55mm)
tr =total rupture life was 34%.

Ωm
ts =service time
= multi-axial damage parameter (Omega factor)
L = Remnant life
σe =effective stress
σ1,σ 2 ,σ3 =maximum principal stresses
LOTISTM =Laser-Optic Tube Inspection System Sample 1

METALLURGICAL EXAMINATION
Sample tubes were removed from the hottest furnace cell

• Sample 1 – extracted from the upper section of

surveyed and consisted of (see Figure 1):
Sample 2

• Sample 2 – extracted from the middle section of

tube, where strain is lowest

tube, where strain is highest up to 1.45%

(measured by LOTISTM on the internal tube bore)
Figure 1 Tube samples removed from the hottest
The tubes were reportedly fabricated from 9Cr1Mo alloy steel furnace cell.
equivalent to ASTM A213 T9.
Metallurgical examination was carried out on the micros
Sample 2 tube shows significant oxide/corrosion product
taken from the ring sample to identify any creep damage and
and spalling of the oxide scale on the fireside where the tube
microstructural degradation.
was directly exposed to the flames. However, there was no
Low alloy ferritic steels are typically supplied in the
significant oxide spalling on Sample 1 from the cooler part of
normalised condition. The associated microstructure normally
the furnace. There was no evidence of distortion such as
consists of isolated islands of pearlite in a matrix of ferrite.
sagging or bulging in the tube samples.
During service at elevated temperature, the pearlite structure
Tube expansion was calculated by determining the
breaks down and the lamellar platelets of carbide spheroidise.
difference between the maximum and minimum diameters
At first, this occurs within the original islands of pearlite.
found within the tube samples supplied. 1.35% of diametrical
Eventually however, the iron carbide becomes evenly
growth was obtained. This was consistent with the strain
distributed amongst the ferrite matrix and the original islands of
information of the tube provided by LOTISTM measurement.
pearlite disappear. The extent to which these changes occur is
The LOTISTM method is an “internal” laser-based tube
used to judge the severity of service. In the original as supplied
inspection technique which measures tube damage in terms of
condition, the microstructure is given an “A” classification.
creep deformation.8
The most extensively modified structures are given an “F”
Ring samples were cut from the tube where the corrosion
classification which means the original pearlite islands have
damage was most severe. The area exposed to the maximum
disappeared and the individual carbides scattered evenly
heat flux was referenced as the 12 o’clock position in the ring
amongst the ferrite matrix have coarsened. Microstructures in
sample. Thickness measurements made are reported in Table 1.
between these extremes are rated accordingly. The extent of

2 Copyright © 2007 by ASME

microstructural modification provides an estimate of severity of considered consistent with the observed microstructure and the
service with respect to temperature and time. temperature operated.
Table 2 Microstructural degradation and creep damage
20 μm Microstructural Creep Damage Hardness, HV10
Classification Classification
Sample 1 C/D* A** 175, 179, 171, 176
Average=175
Sample 2 D/E A 163,171,161,163
Average=164
*, ** see APPENDIX for classifications of microstructural condition and creep damage

LEVEL 1&2 ASSESSMENT

Creep Rupture based assessment (API 530)
A creep rupture based remnant life assessment was
(a) Sample 1 (3 o’clock position: cold side)
undertaken by using stress rupture data applicable to the

20 μm
9Cr1Mo ASTM A213 T9 material as published in documents
such as API 5303. By input of the actual stress and
temperatures experienced in service, an estimate of the time to
rupture can be obtained.
Information in relation to the boiler operating history

• Estimated operating hours: 224,000 hrs.

applied for the assessment was:

• Operating pressure: 12 bar.

• Process gas outlet temperature: 540°C
• Peak tube skin temperature: 655°C
• Tube fabricated from 9Cr1Mo alloy steel

• Nominal Inner Diameter: 97.2 mm

(b) Sample 2 (12 o’clock position: hot side) equivalent to ASTM A 213 T9.
Figure 2 Typical microstructure of cold and hot side of • Nominal Outer Diameter: 114.3 mm
• Nominal thickness: 8.55 mm.
tube samples.
• External wall thinning rate: 0.11mm/year
Similarly, microstructural examination can reveal creep
damage if present. Damage rating “A” means no damage The linear rate of wall thinning was determined based on
observed. Rating “E” means that the sample is heavily cracked the present measured tube wall thickness, an original tube wall
and life is effectively expended. Damage between these thickness and the hours that the tube has spent in service.
extremes is rated accordingly in terms of the extent of creep The measured minimum wall thickness of 5.68mm was
voids evident and their distribution through the microstructure. adopted as the current wall thickness. The difference between
The level of thermally induced microstructural degradation the original wall thickness and the minimum wall thickness
and the extent of creep cavitation damage were classified as measured was averaged to evaluate the wall thinning rate. An
shown in Table 2. The microstructures from the 3 o’clock estimate of the original wall thickness had to be made by
position of the Sample 1 and the 12 o’clock position of Sample assuming that all tubes initially had a nominal design wall
2, which represent cold side and hot side respectively, have thickness of 8.55mm. The evaluated wall thinning rate was
been recorded as shown in Figure 2. The microstructure was 0.11mm/year (1.25x10-5 mm/hour), assumed to be constant
classified as level “C and D” (on a scale of A to F) for the cold during the service life of the tube.
side and “D and E” for the hot side. There was no evidence of The assessment was undertaken by calculating a remnant
creep cavitation on the grain boundaries. life based on temperature, stress and wall thinning rate. The
Hardness tests were carried out on the micro specimens service and future life of the tube was broken down into one
taken from four positions, i.e., 12, 3, 6 and 9 o’clock positions. year intervals for the purposes of calculation. It was assumed
The results showed a hardness average of 175 HV10 for that wall thickness and hence stress was constant over each one
Sample 1 and 164HV10 for Sample 2, see Table 2. This is year period. The expected life under the conditions appropriate
consistent with typical hardness of T9 material having to each one year period was calculated and converted to a
undergone moderate softening in service. This was also damage/life consumption fraction. The damage fraction was

3 Copyright © 2007 by ASME

summed from year to year and failure was taken to occur when where, S l = log 10 (σ e ) , σ = 1 [(σ −σ )2 + (σ −σ )2 + (σ −σ )2 ]1/ 2
e 1 2 1 3 2 3
the accumulated damage fraction reached 1. 2

C0 →C4 : Material coefficients for strain rate and Omega

The minimum (lower bound) and average creep rupture
properties for 9Cr1Mo steel available from API 530 standard
Δsr : Adjustment factor to account for the material scatter band, a
parameter (available in API 579)
were used in the assessment. The results from the assessment
are listed in terms of remnant life of the tube in Table 3. range of -0.5 for the lower bound of creep strength to +0.5
for the upper bound can be used.
Table 3 Results of rupture based Level 1&2 remnant life
assessments Δcd : Adjustment factor for creep ductility, a range of +0.3 for
brittle behaviour and -0.3 for ductile behaviour can be used.
Tube material: 9Cr1Mo steel API 530 API 530
T: Temperature in °F
Temperature: 655°C minimum average
Pressure: 12 bar In order to determine the current strain rate and the Omega
parameter at operating conditions without knowing the actual
Corrosion rate: 0.11mm/yr 0.7 years 18 years
material properties, it is necessary to assume that the tube
Service time: 224,000 hrs
combination of the parameters, Δsr and Δcd.
material has lower bound or mean properties by adjusting the

Most of the creep life has already been exhausted when A strain based Level 1/2 remnant life assessment has been
evaluated using the API 530 method with minimum rupture undertaken by using strain rate and Omega data applicable to
properties. The difference in predicted remnant life between the 9Cr1Mo ASTM A213 T9 material as published in API 579
6
minimum and average properties with the API 530 method was . By input of the operating conditions, geometry and corrosion
almost 18 years. An immediate repair or replacement of the rate as used in the previous rupture based remnant life
furnace tubes is required if the decision has to be made based assessment, an estimate of the time to failure can be obtained
on the API 530 minimum property assessment. by employing equation (1). The life fraction was determined
based on the following equation which is an alternative form of
equation (1). The tube life is fully consumed when the life

t Ω ε&
Strain based assessment (API 579 Omega method) fraction, ts/tr becomes 1.
= = s m co
Tubes operated in furnaces are known to suffer from
t r ⎛ 1 ⎞ t s Ω mε&co + 1
ts ts (4)

⎜⎜ ⎟⎟
damage as a result of time dependent strain, i.e. creep.

⎝ Ω mε&o ⎠
Therefore, there have been attempts to understand the
relationship between creep damage in terms of strain, time and

ε&o = ε&co /(ε&co Ω m t s + 1) =original strain rate, and ε&co is current

life consumption. These parameters have been incorporated in where, ts=service time, tr=total rupture life,
the Omega Method for creep life assessment, developed by
Material Properties Council (MPC) and adopted by API (API
The values of parameters, Δsr and Δcd to provide minimum
strain rate
579 “Fitness for Service”6) as the standard method of life
assessment. In this methodology, a current strain rate at the and average creep properties for 9Cr1Mo steel available from
operating condition and a multi-axial damage parameter the API 579 and the results from the assessment are
(Omega) are used to predict remaining life to failure using the summarised in Table 4.
following relationship: The 18 years difference in predicted remnant life between
L=
Ω mε&co
1 (1) API 530 lower bound and mean creep properties decreased to
11 years in the case of lower bound and mean properties
Knowing the Omega factor and the current strain rate published in API 579. The immediate action for repair or
therefore provides a method of estimating remaining life. An replacement of the furnace tubes is not required and a typical
advantage of this methodology is that many of the material inspection interval, for example, 6 years interval could be
constants have been well characterized and are published in achieved. However, if operating history (temperature/pressure
API 579 in the form of constitutive equations using time history) is not well defined, for example, the furnace is
temperature parameters such as Larson-Miller parameter as operated at higher temperature in future operation, it is highly
follows; likely to have a premature failure within the established

[ ]
inspection period as demonstrated in Table 4.

log10 ε&co = −⎢(C0 + Δ sr ) + ⎢

⎡ ⎡ 1 ⎤ ⎤
C1 + C2 Sl + C3 Sl2 + C4 Sl3 ⎥ (2)
In order to increase credibility of using the minimum
⎥
⎣ 460+ T ⎦
⎣ ⎦
properties in API 579, the accuracy of the assessment and the

[ ]
credibility of material properties need to be improved by

log10 Ω = (C0 + Δcd ) + ⎢

⎡ 1 ⎤
C1 + C2Sl + C3Sl2 + C4Sl3
conducting Omega creep testing on the tube material in
⎣ 460+ T ⎥⎦
(3)
concern. Test specimens machined from the sample tubes were
subjected to a creep test, the objective of which was to measure

4 Copyright © 2007 by ASME

the instantaneous strain rate without having to proceed to would not result in significant thermally induced
sample failure. This information is then used with database microstructural changes. After the initial strain rate was
knowledge of the strain rate and Omega factor to provide an determined, both temperature and stress were raised to achieve
accurate estimate of remaining life. at least 3x10-5 hr-1 strain rate.

Table 4 Results of strain based Level 1&2 remnant life

assessments (T=655°C, P=12bar, 0.11mm/year)
Δsr Δcd Remnant Life,
years
API 579 minimum -0.5* 0.3* 6 (2)**
API 579 mean 0 0 17
∗Δsr = -0.5 for lower bound of creep strength, Δcd = 0.3 for creep
brittle , **Assessment at 675°C for a demonstration

LEVEL 3 ASSESSMENT: OMEGA METHOD

Omega creep Testing
Two circumferential creep specimens (HG1 and HG2) Figure 4 Typical testing data showing changes in
were machined from the hot side of Sample 2 so that the gauge temperature and stresses during test (CG1)
length of the specimen encompassed service related degraded -8
material. One specimen (CG1) was prepared from the cold side
of Sample 1 as a reference test. -8.5
The preparation of the creep specimen is illustrated
schematically in Figure 3. In order to prepare circumferential
ln ( strain rate), h-1

-9

specimens, appropriate weld beads were laid on the outside y = 3.579E+01x - 1.025E+01
edge of the tube to allow the ends to be of sufficient size to -9.5 R2 = 9.967E-01

accommodate threaded ends and to ensure that the welds and

-10
associated heat affected zones were not part of the specimen
gauge section.
-10.5
HG1, 731C, 18.6MPa

Weld bead Linear (HG1, 731C, 18.6MPa)

-11
0.015 0.02 0.025 0.03 0.035 0.04 0.045
Creep strain

Figure 5 Natural log of the strain rate versus strain

yielding the Omega parameter.

Figure 3 Specimen preparations from the tube Figure 4 shows a typical temperature and displacement
sample. curve. It should be noted that the testing conditions were
changed to determine the Omega values after determination of
Each sample was placed into a constant load creep the initial creep rate.
machine. Thermocouples were attached to the samples, one at The creep curve exhibits a primary regime of
the top of the gauge length and the other at the bottom of the approximately 100 hours. The initial creep rate was determined
gauge length. Samples were wrapped in steel wool; steel wool by taking the slope of the creep curve for the next 100 ~ 200
acts as a sacrificial material and helps to protect the sample hours as illustrated in Figure 4. When the curvature of the
from oxidation along with the argon atmosphere provided. strain time plot (strain rate acceleration) was measurable within
Each rig was set up to enable the growth of the sample to an acceptable time frame, the Omega parameter was
be measured throughout the test using extensometers. determined by plotting the natural log of the strain rate against
The testing stress was initially based on the service stress strain as shown in Figure 5.
The material scatter parameters Δsr and Δcd were adjusted
(mean diameter hoop stress). The testing temperature was
chosen to yield a strain rate of approximately 1x10-5 hr-1. The
temperature was predicted by setting the Δsr parameter in to match the predicted creep rate and the Omega factor re-
equation (2) to zero which represents average creep strength. calculated based on the actual testing conditions. The actual
However, in order to minimise any microstructural changes due testing conditions and measured initial creep strain rate and
Omega parameters are summarised in Tables 5 and 6. The
adjusted parameters, Δsr and Δcd, for the tube material tested
to high temperature exposure, the testing stress was raised
above typical operating values to ensure the test temperature

5 Copyright © 2007 by ASME

with respect to the API 579 mean property are also shown in
Tables 5 and 6. The positive values of the Δsr and Δcd,
Table 7 Initial strain rate and Omega parameter at
service conditions (655°C, 12 bar)

ε& _uniaxial, h- Ω_uniaxial ε& _tube, h-1 Ω_tube

parameters indicate the material exhibits higher strength and
creep brittle behaviour relative to the mean properties presented
1
in the API 579.

HG1 1.328E-08 111 6.535E-09 420

and parameter Δsr
Table 5 Testing conditions and measured initial creep rate
HG2 1.911E-08 68 9.402E-09 226
Δsr
°C
ID Temperature, Stress, Predicted Measured
CG1 1.205E-08 50 5.932E-09 155
MPa creep creep parameter
rate, h-1 rate, h-1
(adjusted)
A Level 3 strain based remnant life assessment was
HG1 717/722* 16 6.58E-06 7.93E-6 -0.081
undertaken by using strain rate and Omega measured from the
HG2 720 16.6 7.16E-06 1.24E-5 -0.239 Omega creep testing on the specimens extracted from the hot
and cold side of tube samples. The operating conditions,
CG1 719 17 7.89E-06 8.64E-6 -0.039 geometry and corrosion rate used in the above assessment were
applied. The remnant life of the furnace tube was calculated
Note: * temperature increased to 722°C during the testing
based on the results obtained from each Omega creep test and
Table 6 Testing conditions for Omega parameter and Δcd
are listed in Table 8.
determination Table 8 Remnant life calculation (T=655°C, P=12bar,
Δcd 0.11mm/year)
°C
ID Temperature, Stress, Predicted Measured
MPa Omega Omega Remnant life
parameter
(adjusted) HG1 14 years (122,640h)

HG1 731 18.61 25.3 35.8 0.143 15 years (131,400h)

HG2
HG2 740 18.11 24.9 20.3 -0.07
CG1 20 years (175,200h)
CG1 739 19.34 22.9 15.2 -0.199

As shown in Table 8, the remnant life assessment based on

Lower creep strength and brittle creep behaviour were
the result of the Omega testing on HG1 sample produced the
present in the specimens taken from the hot side of the tube
shortest life, although there was no significant difference
compared to the cold side material. The difference in creep
between the test samples from the hot side tube (Sample 2).
strength and ductility of the HG specimens relative to the CG1
The results from the remnant life assessment based on HG1 and
sample was considered consistent with the microstructural
CG1 specimens were compared with the predictions from API
changes that have occurred in tube samples. It is emphasised
530 (rupture based) and API 579 (strain based) assessments.
that the Omega creep testing produces not only remnant life but
The CG1 sample which came from cold side of the tube
also the current condition of the material concerned with
satisfies the mean property requirement defined in API 530 and
respect to the original condition.
579 as shown in Figure 6.
25

Remnant life assessment 20

API 530 mean
Remnant life, years

The remnant life of the furnace tubes was determined

based on equation (1) which enables calculation of remnant life 15 API 579 mean
by using the initial strain rate and the Omega parameter
obtained for the operating temperatures and pressure. The 10
multiaxial parametric expressions for strain rate and Omega API 579 Lower
were evaluated to take account of state-of-stress for the tube 5
following the approach described in API 579. The correlated
API 530 Lower
multiaxial parameters are listed in Table 7 for the service 0
conditions. HG1 CG1

Figure 6 Comparison of remnant life results with API 530

and 579 prediction.

6 Copyright © 2007 by ASME

 The remnant life estimated from the HG1 test result falls 1,000,000
0.1 mm/year
between the mean and lower bounds of both API data sets,
0.15 mm/year
indicating the remnant life assessments based on the lower
0.2 mm/year
bound properties in the API standard were conservative.

Remnant life, hours

CG1

Future operation analysis 100,000

HG1
The future operation is likely to be represented by a range of
operating temperatures, pressure and corrosion rates.
Sensitivity analysis was undertaken to determine the effect of
temperature, pressure and corrosion rate on the remnant life of Pressure:
the furnace tubes. The results of the sensitivity analysis can 12 bar
enable units to be operated in a more flexible manner without 10,000
600 620 640 660 680 700 720
compromising integrity of the components. Two sets of o
Temperature, C
analysis were carried out based on the most likely future
operating conditions. Figure 7 Effect of temperature and corrosion rate on
1. Temperature ranges from 620°C to 700°C at constant remnant life of the heater tubes (shaded area indicates
pressure of 12 bar with three different corrosion rates operating conditions resulting in less than 5 years remnant
(0.1mm/year, 0.15mm/year and 0.2mm/year). life)
2. Pressure ranges from 10 bar to 16 bar at constant 1,000,000
temperature of 655°C with three different corrosion 0.1 mm/year
rates (0.1mm/year, 0.15mm/year and 0.2mm/year). 0.15 mm/year
0.2 mm/year
The remnant life of the platformer heater tubes was Remnant life, hours
determined for each condition based on the material properties CG1
evaluated from the HG1 test result. The results are shown in 100,000 HG1
Figures 7 and 8 along with the remnant life results from HG1
and CG1 test.

• The analysis described above showed a large

The results of the sensitivity analysis indicate: Temperature:
655°C
10,000

• The platformer heater tubes could be operated at a

sensitivity to temperature, pressure and corrosion rate. 8 10 12 14 16 18
Pressure, bar
maximum temperature of 675°C at 12 bar pressure for
Figure 8 Effect of pressure and corrosion rate on remnant
• At the lower temperature ranges (below 640°C), the
at least 5 years regardless of corrosion rate.
life of the heater tubes (shaded area indicates operating
effect of corrosion rate was greatest. However, the conditions resulting in less than 5 years remnant life)
remnant life became less sensitive to the corrosion rate
when temperature was higher than 680°C for the range 1.0
Remnant life_0.2/Remnant life_0.1

of corrosion rates employed. This is also illustrated in 0.9

Figure 9 showing sensitivity changes with 0.8

0.7
temperatures by comparing of the remnant life at
0.6

• As long as the temperature of the tube remains at

0.2mm/year to that of 0.1mm/year assessment.
0.5
0.4
655°C, the remnant life of the tubes exceeds 5 years
0.3 12 bar
and is independent of pressure and corrosion rate up to
0.2

• The sensitivity analysis displayed with remaining life

16 bar and 0.2mm/year, respectively.
0.1
0.0
versus temperature and pressure curves make it 600 620 640 660 680 700 720
possible to raise the level of confidence in the decision Temperature, o C
for repair/replacement or inspection scheduling.
Figure 9 Changing effect of corrosion rate on remnant life
with temperatures demonstrated by comparison of remnant
life of 0.2mm/y case to that of 0.1mm/y case.

7 Copyright © 2007 by ASME

DISCUSSION difference in creep properties and remnant life
The significant difference in creep properties and between cold and hot side of tubes was observed.
consequently remnant life between the cold and hot side of the This highlighted the importance of the selection of the

•
tube highlights the importance of the selection of samples to tube sample.
increase the accuracy of the life assessment. The main The creep life fraction consumed and the remaining
objective of the selection of sample tubes was to remove tube life was determined using the material property of
samples that are representative of the worst components which HG1 (one of the hot side specimens) sample. A
will remain in service. The selection procedure involves visual further 122,000 hours or 14 years operation can be
examination along with review of inspection and operation realized based on the assumption that the tube is
histories, especially wall thinning rates and diameter operated at 655°C and 12 bar with a wall thinning rate
measurement provided by techniques such as LOTISTM, if
•
of 0.11mm/yr throughout that period of time.
available. The results of the sensitivity analysis made it possible
The fact that there was no evidence of creep voids on both to raise the level of confidence in the result for
tube samples suggests the creep void damage mechanism was repair/replacement scheduling.
not a factor in controlling the life of the tube in their current
condition and the degradation in the creep resistant properties
of the tube, for example, creep strength and ductility was
caused by the microstructural modification during service at ACKNOWLEDGEMENT
elevated temperature. However, creep damage may be likely to The Authors wish to acknowledge the Shell Global
dominate life as stress levels increase with wall thinning in Solutions International and Shell Refining (Australia) Pty Ltd –
future operation. Geelong Refinery for funding this work and providing
The outcome of the remnant life assessment allows materials and technical data.
strategic decisions on repair/replacement or inspection
schedules to be made. The Level 1 and 2 “desk top”
assessments undertaken for the platformer heater tube produced
different conclusions depending on material property data and REFERENCES
methodologies used. For example, the assessment based on 1. R. Viswanathan, “Damage Mechanisms and Life Assessment
API 530 minimum property data suggested immediate of High-Temperature Components”, ASM International.
repair/replacement. Conversely, use of mean property data
would suggest that reasonable remaining life existed and that 2. W.P. McNaughton, R.H. Richman, C.S. Pilar, and L.W.
the tube life could be managed by on going inspection. This Perry, “Generic Guidelines for the Life Extension of Fossil
inconsistency in outcome of the assessment can be resolved Fuel Power Plants”, EPRI CS 4778, Electric Power Research
once the results from the Omega creep testing on the selected Institute, Palo Alto, CA, Nov 1986.
tube samples taken from the furnace are available. The relative
position of the current material property within the scatter band 3. API Standard 530 “Calculation of Heater-Tube Thickness in
of the minimum and mean data in API 579 provides a clear idea Petroleum Refineries”, Oct, 1996.
of how the material degradation has been progressed during
service so far. The credibility of using the calculated remnant 4. M. Prager and S. Ibarra, “Approaches to Long Term Life
life based on the Omega creep test calculation relies largely on Prediction of Furnace and Boiler Tubes”, PVP-Vol. 359,
the future operating conditions which can be affected by Fitness for Adverse Environments in Petroleum and Power
process requirements and business drivers. Integrity operating Equipment, ASME 1997.
windows must be established to ensure that confidence limits
are not exceeded. 5. Jonathan D. Dobis and Ling Huang, “Assessment of
Graphitized Carbon Steel Tubes in Fired Heater Service”,

• A Level 1 and 2 remnant life assessment was

CONCULSIONS Corrosion/2005, TEG-205X Symposium, Paper No. 05559.

undertaken on a platformer heater tube using API 530 6. API Recommended Practice 579 “Fitness For Service”, Jan,
and 579 published data. A significant difference 2000.
(maximum of 18 years) in remnant life between
minimum property and mean properties available in 7. Martin Prager, “The Omega Method-An Engineering

• In order to improve the accuracy of the life

the API standard was obtained. Approach to Life Assessment”, Journal of Pressure Vessel
Technology, August 2000, Vol. 122, pp. 273-280
assessment, the Omega creep testing was undertaken
on circumferential creep specimens machined from the
hot and cold side of the heater tubes. A significant

8 Copyright © 2007 by ASME

8. Richard D. Roberts and John Brightling, “Maximize Tube
Life by Using Internal and External Inspection Devices”,
Process Safety Progress, Vol. 24, No. 4, December 2005.

APPENDIX

9 Copyright © 2007 by ASME

View publication stats