Int J Adv Manuf Technol

DOI 10.1007/s00170-008-1869-7

SPECIAL ISSUE - ORIGINAL ARTICLE

Damage assessment of boiler tubes subjected to erosion

and corrosion failures by non-invasive technique
K. Anbarasu & P. Sampathkumaran & M. Janardhana &
S. Yogesh & S. Seetharamu

Received: 5 March 2008 / Accepted: 16 September 2008

# Springer-Verlag London Limited 2009

Abstract Incompatible permanent strains in any engineer- Keywords Residual stress . Non-destructive testing .
ing component are caused by the formation of residual Erosion . Corrosion . Magnetic Barkhausen Noise .
stresses due to mechanical loading or thermal effect. They X-ray diffraction
get generated or modified at every stage of the service life
of the component. The stresses tend to increase due to the
type of damage occurring in any engineering component
such as fatigue, corrosion, erosion, or a synergistic effect 1 Introduction
under the influence of environmental conditions. The
present investigation highlights about the residual stresses The permanent strains caused an account of plastic
in a typical boiler tube component and the same is deformation results in locking up of residual stresses in
characterized by nondestructive methods viz., Magnetic any engineering components. These stresses are formed
Barkhausen Noise and X-ray Diffraction techniques to during any engineering process such as fabrication/manu-
identify the type of damage such as corrosion and erosion facturing, involving welding, casting, forging, rolling etc.
have been discussed. The correlation of the stress data with Generally, any in-service component will have higher level
respect to the hardness and microstructure of the in-service of residual stresses due to the mechanical or thermal
boiler tubes in comparison with an unused tube has been influence. It is important to know the nature and magnitude
attempted. The result reveals that there is a considerable of stresses to enhance the service life. The compressive
variation in the magnitude of the stress level as well as the stresses generally favors in service components increasing
hardness value in the damaged region compared to the the fatigue strength, whereas presence of tensile stresses
unaffected location. Furthermore, it is seen that the micro- beyond the acceptable limit may result in failure of a
structural features support the data. The damage assessment component.
methodology thus serves as a tool to quantify the type and This work presents the case study on corrosion and
extent of damage of in-service tubes and forms a part of the erosion problem affecting the stress levels in piping systems
quality assurance practice. especially in boiler tubes. For residual stress measurements
both X-ray diffraction (XRD) and Magnetic Barkhausen
Noise (MBN) methods have been employed and the data
K. Anbarasu have been generated while analyzing the data the corrosion
National Institute of Technology,
and erosion damages have been identified based on the
Trichirapalli 620015 Tamil Nadu, India
stress pattern. The stress data has been correlated with
P. Sampathkumaran (*) : M. Janardhana : S. Yogesh : hardness values in respect of in-service tube in comparison
S. Seetharamu with that of new or virgin tube. Furthermore, the metallur-
Material Technology Division, Central Power Research Institute,
Bangalore 560080 Karnataka, India
gical features using optical microscope reveal that the
e-mail: psampathkumar29@yahoo.com microstructure of in-service tube is different from that of
e-mail: sampath@cpri.in the virgin tube.
 Int J Adv Manuf Technol

2 Effect of residual stress in water wall tubes signals generated during magneto elastic interaction (flux
density changes) is dependent on the characteristics of the
Corrosion attack either due to pitting or caustic effect occurs stress prevailed in a material. These signals are strong
either at the water side or fire side which is prevalent in most functions of stress condition and hence the stress levels can
of the coal-fired thermal power plants. The type of corrosion be assessed. The measurements are possible on complicated
Robert et al. [1] generally takes place are flow-accelerated geometries, coarse grained, and texture materials, etc.
corrosion, oxygen pitting, cohelant corrosion, caustic corro- Furthermore, the stresses measurements up to a depth of 1.5
sion, acid corrosion, organic corrosion, phosphate corrosion, mm can be done. The MBN signals are quite sensitive to
hydrogen damage, and corrosion-assisted cracking. Toribio microstructural variations in the material. The MBN mea-
[2] has studied the effect of induced stresses due to stress surement is independent of magnetic permeability variations.
corrosion cracking in high strength steel to understand the
role of residual stress on the material behavior. In the case of 3.2 X-ray Diffraction measurement
erosion, the tubes experience metal loss on the OD side,
which might occur due to the impact of solid particles In the XRD method, the scattered X-rays of a chromium
involving soot blower erosion. The ultimate failure results in target from the irradiated atoms/crystals are reinforced in a
rupture due to increase strain as the tube material erodes specific direction of an extremely narrow range and
away. The type of damage that happens in such tubes may propagated. The phenomenon is known as diffraction. The
give rise to lower compressive stresses, which may be condition of diffraction is met using Bragg’s law nλ=2d
ascertained by experimental means. It is reported in the Sin θ. Where d is the inter atomic spacing, θ is the Bragg
literature of Capriotti et al. [3] that decrease in level of angle and λ is the wave length of characteristics of X-ray
compressive stress would reduce the mechanical properties radiation. The diffraction pattern so obtained is measured
due to the hydrogen embrittlement type of damage. for various = angles (angle between plane and surface
Furthermore, the type of damage would result in either normal) in terms of a shift in 2θ, which accounts for the
reduction in compressive stresses or presence of tensile change in inter planar spacing (d). The stress Withers [6]
stress. The trend in the stress level may tend to increase value is determined from the slope of the plot of 2θ vs. sin2
towards tensile side beyond the acceptable limit due to the = using the following equation.
service condition prevalent in such tubes/materials and
ð$2qÞ
finally end up in window type of opening failure. Investiga- sf ¼ K

tion on the influence of induced stresses in corrosion/erosion @ sin2 y
affected water wall tubes could not be cited in the literature
and hence this particular work addresses this issue. K
¼ 1=2½E=ð1 þ uÞ cot qðp=180Þ
Where K is the stress constant factor, E is the young’s
3 Methods of measurements modulus and υ=poission’s ratio of the material.

The residual stress measurements are done either by

destructive or non-destructive methods Chandra et al. [4]. 4 Experiments
Among them, nondestructive testing (NDT) is preferred. In
nondestructive testing method, Magnetic Barkhausen Noise The cut tubes are thoroughly cleaned by acetone prior to the
and X-ray Diffraction techniques are widely used for measurements. In the corroded tube, the sensor is made to
measuring the residual stress. scan the entire length of the tube to record the MBN
parameter by positioning it at different locations (i.e.
3.1 Magnetic Barkhausen Noise measurement corroded, uncorroded). At these specific locations, the
XRD measurements are done to know the magnitude and
MBN is based on Barkhausen effect Baldev et al. [5] and the nature of stress levels. The same procedure was
applicable to ferromagnetic materials. The effect takes place repeated for the virgin tube, which is taken for comparison
when a magnetic field is swept in the material along a purposes. Both the damaged and virgin tubes are examined
hysterisis loop. Magnetic Barkhausen noise is due to for microstructure and as well as for hardness profile
irreversible change in the magnetic wall domain move- measurements. The photographs of typical ID corrosion and
ments during the hysterisis. The MBN signals can be external erosion attacks are shown in Figs. 1 and 2,
acquired by a sensor coil or hall type of probe and the respectively.
Int J Adv Manuf Technol

Table 1 MBN and XRD parameters of in-service and virgin water

wall tubes

S. Sample Type of MBN XRD (stress)

No designation tube (mp) (MPa)

1 WWT 1 Corroded tube 180 to 190 −30 to −50

Eroded tube 160 to 175 −80 to −110
2 WWT 2 Virgin tube 110 to 140 −250 to −280

It is noticed from the Table 2 that the tubes at the

damaged locations (corrosion/erosion) show lower hardness
compared to the virgin tube. Furthermore, the stress
measurements done at these locations also give good
indications of damage affected areas (refer Table 1),
Fig. 1 Corrosion at the ID of the tube wherein the corroded, eroded, and the virgin tubes show
compressive stresses in the range 30 to 50, 80 to 110, and
250 to 280 MPa, respectively. Here also, it is noticed that
5 Results and discussion
the stress values of the new tube is much higher compared
to the damaged tubes. Also the nature of stress prevailed is
The chemical composition of the water wall tube under
compressive type.
investigation is conforming to the medium carbon steel
In the work investigated Toribio et al. [2] on the hy-
grade SA 209 T1. The data in respect of MBN and XRD
drogen embrittlement detection on high strength steel by
parameters pertaining to the damaged (corrosion and
means of XRD residual stress technique, it is reported that
erosion affected) water wall tubes (WWT1) are shown in
the hydrogen embitter zones show lower compressive stress
Table 1. The same data for the virgin i.e., new tube
compared to the unaffected area. The reason for this is
(WWT2) is also provided in the same table (Table 1). The
attributed to the service condition prevailing over there
microstructure of the corroded and the virgin tubes are
resulting in the formation of nascent hydrogen reacting with
displayed in Figs. 3 and 4, respectively.
carbon to form methane and act as crack growth initiation
It is observed from the table (Table 1) that the MBN
at that point of time. Finally, the tube opens up due to
signals show high readings at the affected regions (both
hydrogen embrittlement. It is noticed from the table (Table
corroded and eroded regions) compared to the virgin tube.
2) that the MBN and XRD parameters does seem to depend
In fact, the readings are much higher in corrosion-affected
on the severity of the damage that has occurred. Thus, the
area compared to eroded zone. The hardness profile
present finding of lower compressive stresses and higher
measurements (Table 2) and the microstructural features
MBN readings in the damaged areas of in-service tubes
shown in Fig. 3 (compared with Fig. 4) of the affected
compared to new tubes is in line with the reported findings.
region supports the MBN data. Microstructurally, the virgin
In summary, it may be inferred from this work that the
tube displays ferrite with pearlite with equi-axial grains,
MBN and XRD method of measurements serve as a good
whereas the affected region reveals elongated ferrite grains
with pearlite.

Fig. 2 Tube depicting external erosion Fig. 3 Microstructure of corrosion affected tube (WWT1)
 Int J Adv Manuf Technol

Table 2 Hardness data of in-service and virgin water wall tubes

S. No Type of tubes Hardness value (HB)

1 Corroded/eroded 125–133
2 Virgin 147–156

& The methods used serve as a quality assurance tool to

predict the type of damage that has occurred in the
tubes.

Acknowledgements The authors wish to thank the CPRI manage-

Fig. 4 Microstructure of virgin tube (WWT2)
ment for having accorded permission to publish this paper. The
authors wish to acknowledge the assistance rendered by Shri.
tool to characterize the damaged and undamaged tubes Balakrishna Sake, Mr. S. Ezhilarasu and Mr. M. Biru Anand Prakash
effectively. in conducting the experiments.

6 Conclusion
References
Based on the above study, the following conclusions may
be drawn. 1. Port RD, Herro HM (1990) The nalco guide to boiler failure
analysis. McGraw Hill, New York, 57
& The XRD and MBN parameters measured are quite 2. Toribio J (1998) Residual stress effects in stress-corrosion cracking.
J Mater Eng Perform 7(2), 173
sensitive to the type of damage occurred and the
3. Capriotti R, Colavita M, De Paolis F, Silvestri P (2000) Hydrogen
characterization has effectively been done to assess the embrittlement detection on high-strength steel by means of XRD
corrosion and erosion damage of water wall tubes. residual stress determination technique, 15th World Conference on
& Stress levels measured by XRD method tend to Nondestructive Testing, Roma (Italy), 15–21
4. Chandra U (1997) Control of residual stresses, ASM hand book,
decrease with increase in exposure of tubes to corrosion
Material Selection and Design, 20, Materials Park, OH, USA, 811
or erosion. 5. Baldev R, Jayakumar T, Moorthy V, Vaidyanathan S (2001)
& MBN parameters, on the other hand, shows high Characterization of microstructure, deformation, and fatigue dam-
readings for the in-service tubes compared to virgin age in different steels using Magnetic Barkhausen Emission
Technique, 37, 789
tubes.
6. Withers P. J., Bhadeshia HKDH (2001) Residual stress Part-1
& Hardness profile measurements and microstructural “Measurement techniques” and Part-2 “Nature and origins. Mater
features give good support the XRD and MBN data. Sci Technol 17, 355