M. NAÏ et al.: INFLUENCE OF STEAM-SIDE OXIDE SCALES ON THE CREEP LIFE OF A BOILER SUPERHEATER TUBE
43–46
INFLUENCE OF STEAM-SIDE OXIDE SCALES ON THE CREEP
LIFE OF A BOILER SUPERHEATER TUBE
VPLIV [KAJE NA PARNI STRANI CEVI IZMENJEVALNIKA
TOPLOTE NA DOBO TRAJANJA CEVI ZARADI LEZENJA
Martin Naï, Tomá{ Létal, Jiøí Buzík, Pavel Lo{ák
Brno University of Technology, Faculty of Mechanical Engineering, Institute of Process Engineering, Technicka 2, 616 69 Brno,
Czech Republic
nad@fme.vutbr.cz
Prejem rokopisa – received: 2017-07-05; sprejem za objavo – accepted for publication: 2017-09-22
doi:10.17222/mit.2017.106
Creep life is a limiting factor in the case of boiler tubes operating above the creep temperature and, for a given material, it is
generally determined by the actual operating temperature and stress. Standard approaches to the calculation of a tube creep life
mainly take into account the stress caused by pressure loading and the mean temperature in the tube wall. However, the
estimation of the temperature and stress may be difficult for water-tube boilers because oxide scales tend to form on the inner
surfaces of the tubes, exposed to steam and indirectly affecting the resulting creep life. As the scales increase the thermal
resistance of a tube wall and, consequently, the wall temperature, the creep life is reduced. Moreover, the presence of oxide
scales leads to a higher hydraulic resistance of a tube, which can cause further increase in the temperature in some tubes of the
bundle if the oxide-scale growth rate is not uniform. Additionally, the elements increasing the creep strength may leach out of
the surface should the scales be porous, thus inducing an even more pronounced decrease in the creep life. This paper
investigates the effects of oxide scales on the creep life of boiler tubes in a cross flow. Thermal and stress conditions in the tubes
are thoroughly investigated, using analytical as well as numerical approaches.
Keywords: oxide scale, creep, superheater, tube, CFD, FEA
Doba trajanja cevi super izmenjevalnikov toplote je omejena zaradi lezenja materiala pri visokih temperaturah obratovanja.
Kriti~na temperatura lezenja je odvisna od napetosti, ki nastopajo med obratovanjem izmenjevalnika toplote. Standardni pristop
za izra~un dobe trajanja zaradi lezenja materiala temelji na upo{tevanju napetosti, ki jih povzro~a delovni tlak prenasi~ene pare
in povpre~ne temperature na stenah cevi. Vendar je temperaturo in napetosti te`ko oceniti, ker se na notranjih stenah cevi, ki so
izpostavljene pari, nalaga oksidna plast oz. {kaja, kar neposredno vpliva na njihovo dobo trajanja. [kaja na stenah cevi zmanj{a
toplotno prevodnost in s tem poslab{a prenos toplote in posledi~no spremeni temperaturo sten cevi, kar zmanj{a njihovo dobo
trajanja. Poleg tega {kaja pove~a hidravli~no upornost cevi, ki nadalje povi{a temperaturo v nekaterih snopih cevi, ~e rast {kaje
ni enkomerna. Dodatno se lahko iz sten cevi zaradi {kaje izlu`ijo legirni elementi, ki pove~ujejo odpornost materiala proti
lezenju. To lahko povzro~i poroznost oksidne plasti na ceveh, kar {e dodatno zmanj{a njihovo dobo trajanja. V ~lanku avtorji
obravnavajo vpliv oksidne {kaje na `ivljensko dobo cevi izmenjevalnikov toplote v toku ogrevalnih plinov pre~no na pretok
pare. Analizirali so termi~ne in napetostne razmere v ceveh izmenjevalnika toplote z analiti~nimi in numeri~nimi metodami.
Klju~ne besede: {kaja (oksidna plast), lezenje, super izmenjevalnik toplote, cevi izmenjevalnika, ra~unalni{ka dinamika fluidov
(CFD), analiza z metodo kon~nih elementov (FEA)
1 INTRODUCTION
Boiler-design life expectancy is usually in tens of
years and it is impossible to predict the exact operating
conditions for such a long time; therefore, various kinds
of damage must be expected. Although particular da-
mage mechanisms are intertwined and rarely occur
individually, they may be categorized according to pri-
mary causes:1,2
• temperature
• corrosion
• fluid flow
• operation
• manufacturing and installation
During manufacturing, different materials may be
selected. This error was also identified during the inves-
tigation of the boiler-tube damage where, instead of
high-grade 15020 steel, low-grade 12022.1 steel was
used, which was found out with a material analysis.3
The 15020 steel exhibits a high-temperature corro-
sion resistance up to 530 °C in a steam environment,
which makes it suitable for superheater tubes.4 On the
other hand, the 12022.1 steel just guarantees yield
strength at elevated temperatures and it is suitable for
basic piping and pressure vessels.5
1.1 Creep behaviour of steel
Creep is often considered as the limiting factor for
the boiler tubes exposed to high temperatures. There are
efforts to increase operating temperatures as they would
increase the boiler efficiency. However, this increase
may considerably decrease the life of exposed com-
ponents; therefore, it is important to assess the effect of
an increased temperature.
Materiali in tehnologije / Materials and technology 52 (2018) 1, 43–46 43
UDK 620.191:620.1:67.017 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 52(1)43(2018)
This assessment can be performed thorough a tran-
sient-creep simulation using a FEA (finite-element
analysis) that takes into account the creep-material
model as well as the history of operation conditions. As
this approach would be time consuming, simplified
methods are preferred when applicable.
According to many standards, the ultimate tensile
strength of a steel at a given temperature is guaranteed
only for a limited time. When a tube is exposed to tem-
perature and stress for a given time, the creep life
reduction can be estimated according to the procedure in
standard EN 12952–4.6
2 THERMAL ANALYSES
Since the creep behaviour of steels is strongly
affected by the temperature, a thorough thermal analysis
of tubes made of steel 12022.1 was performed. A tube
from the first row of the superheater was selected as the
most exposed. The tube cross-section was TR 38 × 5.6
and the dimensions as well as tolerances were set
according to standard ^SN 426710.7
The temperatures in the tube wall were between the
flue-gas temperature and steam temperature and it
depended on several factors:
• thermal conductivity and thickness of the steel;
• thermal conductivity and thickness of the oxide scale;
• flue-gas velocity distribution around the tube, its
temperature and other properties;
• steam temperature and velocity in the tube.
Temperature distribution was investigated using a 2D
model of the tube wall. Heat-transfer coefficients were
estimated using a 2D CFD analysis of the cross flow at
the flue-gas side and empiric calculations at the steam
side.
2.1 Heat-transfer coefficient at the steam side
The heat-transfer coefficient was assumed to be con-
stant across the circumference of the tube inner surface.
At the investigated spot, there was steam at a tempera-
ture of 369 °C and a pressure of 3.56 MPa. The steam
with a velocity of 11.7 m/s was in a turbulent regime
(Re ~ 170 000). The values were taken from operational
records.
The heat-transfer coefficient was calculated using a
formula based on the Nusselt number:



=
⋅Nu
d in
(1)
where:

 heat-transfer coefficient
Nu Nusselt number
 thermal conductivity
din internal tube diameter
The Nusselt number was estimated using the Gniel-
sinsky formula (Equation (2) from 8 for the tube side:
( )
Nu
d
l
=
⎛
⎝
⎜ ⎞
⎠
⎟ ⋅ ⋅
+ ⋅ ⋅ −
⋅ +
⎛
⎝
⎜
⎞
⎠
⎟


8
1 12 7
8
1
1
2 3
2Re Pr
. Pr /
in
tr
/
lg Re . )
3
218 15
⎡
⎣
⎢
⎤
⎦
⎥
⋅ − −  
(2)
where:
Re Reynolds number
Pr Prandtl number
din internal tube diameter
ltr tube length
The final value of the heat-transfer coefficient at the
inner wall of the tube was 709.69 W/(m2·K).
2.2 Heat-transfer coefficient at the flue-gas side
The heat-transfer coefficient at the flue-gas side was
assumed to be a function of the angle around the tube
due to cross-flow effects. Distribution of the coefficient
was estimated using a simplified 2D CFD analysis that
was designed to accurately describe von Karman vor-
tices. Figure 1 shows dimensions of the simulated
domain as well as the location of the boundary condi-
tions used.9 A mesh with 41302 elements was created
and its quality is described in Table 1.
Table 1: CFD mesh quality
Min Max Average
Element quality 0.47 1 0.99
Aspect ratio 1 1.95 1.03
Skeweness 1.31e-10 0.56 1.49e-2
Orthogonal quality 0.68 1 0.99
The heat-transfer-coefficient distribution obtained
from the CFD analysis takes into account only convec-
tion. However, radiation also plays an important role in
this case.10 The radiation influence was estimated from
thermal-hydraulic calculations and it was added, as a
constant value of 15.9 W/(m2 K) across the tube circum-
ference, to the convection-heat-transfer-coefficient
distribution from the CFD analysis.
M. NAÏ et al.: INFLUENCE OF STEAM-SIDE OXIDE SCALES ON THE CREEP LIFE OF A BOILER SUPERHEATER TUBE
44 Materiali in tehnologije / Materials and technology 52 (2018) 1, 43–46
Figure 1: CFD analysis of flue-gas-flow-simulation boundary
conditions (A = 20·D, B = 3·D, E = 10·D, F = 12.5·D)
The vortices are apparent from the CFD simulation,
which was interrupted only after a steady vortex for-
mation was established. Results from the transient CFD
simulation for the last 0.0272 s, which corresponds to the
lifetime of one vortex, were processed. Using this
approach, the whole periodic cycle of the cross flow was
very accurately described.
The resulting overall heat-transfer-coefficient distri-
butions as well as symmetrical mean distribution are
shown in Figure 3. It shows that the highest mean value
of the heat-transfer coefficient is on the left side of the
tube where hot flue gasses have the biggest influence. On
the other side of the tube, the heat-transfer coefficient is
much lower because of the formation of vortices.
2.3 Tube temperature distribution
Probably due to the change of the material, there was
already an oxide scale on the inner surface of the inves-
tigated tube after 62,000 h of the operation as determined
with a metallurgical analysis.3 Since oxide has a lower
thermal conductivity than steel, higher mean tempera-
tures and a resulting decrease in the creep life may be
expected.
The oxide scale composed of Fe3O4 that was included
in the FEA simulation has a thickness of 0.5 mm and its
material properties were taken from11,12.
The resulting temperature distributions for the two
cases are shown in Figure 4. As can be seen, the mate-
rial temperature is a bit higher where the oxide scale
occurs (the top half), which may influence the tube life.
For the creep assessment in the next section, maximum
temperatures along the circumference (obtained from the
FEA simulation) are used. A comparison of the mean-
temperature distributions across the circumference is
shown in Figure 5.
3 TUBE CREEP-LIFE ESTIMATION
An uneven through-wall stress distribution may
change due to a creep process; therefore, the membrane
stress is considered to be the most significant factor
influencing the creep life. The calculated temperature is
conservatively considered as the maximum value
thorough the wall, which is 433.86 °C for the tube
without oxide scale and 438.54 °C for tube with oxide
scale.
M. NAÏ et al.: INFLUENCE OF STEAM-SIDE OXIDE SCALES ON THE CREEP LIFE OF A BOILER SUPERHEATER TUBE
Materiali in tehnologije / Materials and technology 52 (2018) 1, 43–46 45
Figure 5: Maximum wall-temperature distributions: steel only (solid
line), steel with oxide scale (dashed line) tube creep-life estimation
Figure 3: Overall heat-transfer-coefficient distributions across the
tube circumference at time points (grey) and the symmetrical mean
value (black)
Figure 4: Temperature-distribution comparison for the tube cross-
sections: with oxide scale on the internal surface (top half) and
without oxide scale (bottom half)
Figure 2: Flue-gas velocity around the investigated tube
3.1 Stress estimation
The membrane stress intensity in the tube wall is
driven by the pressure and it can be calculated using
formula (3) derived from 13:
f
p d e ve
veop
=
− +c os cs cs
cs
( )2
2
(3)
where:
pc calculation pressure
dos external tube diameter
ecs tube thickness
 thinning coefficient
Calculated stress intensity fop in the tube is
10.605 MPa.
3.2 Influence of the material characteristics
As shown in Figure 5, the maximum temperatures
are in ranges of 412–434 °C and 415–438 °C for steel
and steel with oxide scale, respectively. The creep
temperature range of the 15020 steel starts at 450 °C,
therefore, without a change of the material, the creep
would be of no concern. The current material – the
12022 steel – is in the creep mode, starting at 380 °C;
therefore, the creep must be checked.
3.3 Remaining creep life
An estimation of the remaining creep life was
performed using the procedure from standard EN
12952-4.6 Its results can be seen in Figure 6. Due to the
low stress intensity, the creep life of 9.64e8 h and
3.65e8 h for pure steel and the steel with oxide scale,
respectively, well exceeds the design life of the boiler.
However, as continuous thinning of the tube wall, due to
corrosion and erosion, increases the stress intensity, it
can be expected that the creep life will be significantly
reduced.
4 CONCLUSIONS
Heat-transfer tubes in water-tube boilers are exposed
to high pressures and temperatures. Oxide scales forming
on the inner surface of a superheater tube may increase
the tube temperature. In addition, the accompanying
reduction in the tube thickness leads to an increase in the
membrane stress induced by pressure. Even in the case
where these changes are slight, a combined effect on the
creep life can still be significant as shown in Figure 6.
In the cross-flow arrangement, there is unequal distri-
bution of the temperature along the tube circumference.
This factor should be considered in creep-life estima-
tions due to creep’s sensitivity to temperature. CFD and
FEA analyses may be required for an accurate estimation
of the temperature distribution.
Acknowledgement
The outcomes of project NETME CENTRE PLUS
(LO1202) were co-funded by the Ministry of Education,
Youth and Sports within the support programme
"National Sustainability Programme I".
5 REFERENCES
1 D. J. Flynn, R. D. Port, Nalco Chemical Company, The nalco guide
to boiler failure analysis, 2 nd ed., New York: McGraw-Hill, 2011,
599
2 M. Asgaryan, N. Simms, S. M. Wu, Prediction of the Remaining
Service Life of Superheater and Reheater Tubes in Coal-Biomass
Fired Power Plants, Adv. Mater. Res., 856 (2014), 343–8
3 J. [temberk, M. Hala{, Diagnostic inspection of boiler pressure
system, NDT servis s.r.o., May 2011, Report No. NDT/419/2011
4 ^SN 41 5020 – Steel 15 020 molybdenum alloy, Prague, Czech
republic, 1981, 12
5 ^SN 41 2022 – Steel 15 022, Prague, Czech republic, 1987, 8
6 European standard: EN 12952-4 Water-tube boilers and auxiliary
installations - Part 4: In-service boiler life expectancy calculations,
September 2011, AFNOR
7 ^SN 42 6710 – Tubes seamless cold formed with conventional
deviations from grades 11 to 16, Prague, Czech republic, 1979, 8
8 Verein Deutscher Ingenieure, VDI heat atlas, 2nd ed., Heidelberg:
Springer, 2010, 1585 p. (Springer reference)
9 J. Buzík, T. Létal, Comparison of force responses in a tube bundle
under different boundary conditions, 2015
10 Z. Jegla, J. Vondál, J. Hájek, Standards for fired heater design: An
assessment based on computational modelling, Appl. Therm. Eng.,
89 (Supplement C) (2015), 1068–78
11 M. Takeda, T. Onishi, S. Nakakubo, S. Fujimoto, Physical Properties
of Iron-Oxide Scales on Si-Containing Steels at High Temperature,
Mater. Trans., 50 (2009) 9, 2242–6
12 Mechanical properties of magnetite (Fe3O4), hematite (á-Fe2O3)
and goethite (
-FeO·OH) by instrumented indentation and molecular
dynamics analysis, https://www.researchgate.net/publication/
257079662_Mechanical_properties_of_magnetite_Fe3O4_hematite_
a-Fe2O3_and_goethite_a-FeOOH_by_instrumented_indentation_and
_molecular_dynamics_analysis, 22.06.2017
13 CEN. EN 12952-3: Water-tube boilers and auxiliary installations,
Part 3: Design and calculation for pressure parts of the boiler, 2012
M. NAÏ et al.: INFLUENCE OF STEAM-SIDE OXIDE SCALES ON THE CREEP LIFE OF A BOILER SUPERHEATER TUBE
46 Materiali in tehnologije / Materials and technology 52 (2018) 1, 43–46
Figure 6: Relative comparisons of stress intensities, temperatures and
creep lives for pure steel and steel with oxide scale (fop – membrane
stress at operating conditions, tc – mean temperature, Tal – time to
reach the theoretical rupture by creep)