See discussions, stats, and author profiles for this publication at: [Link]

net/publication/230595285

Failure investigation of Platen Superheater Outlet Header

Article in Journal for Manufacturing Science and Production · January 2010

DOI: 10.1515/IJMSP.2009.10.1.17

CITATIONS READS

5 2,115

6 authors, including:

Himadri Roy Debasish ghosh Ghosh

Central Mechanical Engineering Research Institute Vimta Labs
85 PUBLICATIONS 804 CITATIONS 223 PUBLICATIONS 2,488 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Al-Mg2Si composite development View project

Automated failure detection View project

All content following this page was uploaded by Himadri Roy on 01 June 2014.

The user has requested enhancement of the downloaded file.

 Failure investigation of Platen Superheater Outlet Header

H. Roy1, *, D. Ghosh1, P. Roy1, A. Saha1, A.K. Shukla1, J. Basu2

1
NDT and Metallurgy Division, 2Human Resource and Development Group
Central Mechanical Engineering Research Institute, Durgapur-713209, India

ABSTRACT working fluid in the tubes is discharged at the inlet and

outlet ends into large diameter piping, to ensure mixing,
This investigation was primarily aimed to examine which is known as headers. In the header, tubes are
the failure of a platen superheater outlet header of a welded, spaced either axially and circumferentially and
210MW thermal power plant. Preliminary macroscopic considered to be critical components of the boiler [1].
examinations along with detailed scanning electron The spacing between the tubes is known as ligament.
microscopic and finite element modeling studies were Robert, Ellis and Viswanathan have made a detailed
carried out to understand the probable cause of failure. review on the damage mechanisms of headers [2]. They
In addition, chemical analysis, microstructural found that creep at weldments and thermal fatigue in
characterization, evaluation of mechanical properties are header body ligaments have emerged as the principle
the necessary supplements to understand the reason for damage mechanisms. The ligament cracking has been
the failure. Finally, it was concluded that start - stop found to be the major cause of rejection headers. Ryder
transients, temperature fluctuations during operation [3] first made a detailed description about the ligament
(minor in magnitude, but large in number), abnormal crack problems. He observed that out of about 157
thermal shocks (few but severe) and design deficiency secondary superheater outlet headers inspected, 28%
are considered as possible causes of formation of were found to fail because of ligament cracking. He
thermal fatigue cracks in the ligament region of the concluded that secondary superheater outlet headers are
investigated platen superheater outlet header. more susceptible to this form of cracking. But he could
not find any correlation between the age of headers and
Keywords: Failure; Header; Stress analysis; Ligament the susceptibility to cracking. Roland et al. [4] also
cracking; FEM. made detailed failure analysis of superheater outlet
header and found that the ligament cracking
phenomenon in the header occurred due to combined
1. INTRODUCTION effects of creep and fatigue.
The present investigation has been carried out on a
The function of a boiler is to convert water into failed platen superheater outlet header. The total
superheated steam which is further used in steam approximate running hours of the same has been around
turbine for electricity generation. Coal, oil and natural 1, 00,000 hours. Detailed FEM analyses conformed to
gas along with pre-heated air are used as fuel which is the fact that the main cause of failure is because of start
burnt in the furnace. The flue gas generated during stop transients resulting in thermal fatigue. Similar
combustion of the fuel crosses different sections namely observations have been obtained by earlier
water wall, superheater, re-heater and economizer. The investigators.

*
Corresponding Author: Himadri Roy
Scientist, NDT & Metallurgy Division, Central Mechanical Engineering Research Institute
Durgapur-713209, India; Tel.: +91-343-6452085/6510233; Fax: +91-343-2546745/2548204
E-mail address: h_roy@[Link], himadri9504@[Link]
 Failure of Platen Superheater Outlet Header

2. EXPERIMENTAL DETAILS of representative photographs of the damaged surface

and the fracture surface were taken using SEM [Make:
A failed superheater outlet header, as shown in Hitachi S 3000 N, Japan].
Fig. 1, was supplied for investigation. Apart from
supplying the failed header, the following technical Table 1
information was also provided by the concerned Chemical composition of the investigated steel
authority: (weight %).

i) Material : BS 3604-HFS 622.31 Elements C Cr Mo S P

specification % Weight 0.12 2.07 0.88 0.022 0.018
ii) Nominal diameter : 355.6 mm
iii) Nominal wall : 30 mm Hardness measurement was carried out using
thickness Hardness Tester 751 (Instron Wolpert, U.K) in Vickers
iv) Maximum working 158.2 Kg/cm2 hardness scale under 20 Kgf load. Tensile testing was
pressure carried out in 6mm diameter and 30 mm gauge length
v) Working metal : 464 0C specimens using an Instron testing machine
temperature [Model:5350]. In order to do stress analysis of the above
vi) Total running : Approx. 100,000 header unit subjected to thermo-mechanical loading,
hours hours. finite element method using ANSYS Code (Version 7.0)
has been used. Finite element model has been developed
using SOLID70/SOLID45 elements of the ANSYS
package program. SOLID70 is a thermal element
required for thermal analysis. SOLID45 is a structural
element used in the structural analysis.

3. RESULTS AND DISCUSSION

It has been observed that the failed header is a long

cylindrical vessel of an approximate length of 16 m. The
Fig. 1: Failed superheater outlet header after being outer diameter of the vessel is 355.6 mm and its
removed from service thickness is 30mm.
Number of stubs is attached with the header both
Visual examination of the failed superheater outlet longitudinally and circumferentially throughout the
header was made after removal from service. The header length. At each longitudinal location of the stubs
chemical composition of the failed header under in the header, either nine or eight stubs is arranged
investigation was analyzed using standard spectrometer circumferentially at an angular pitch of 22.5 degree in
and has been shown in Table 1. Small specimens were the bottom half of the header cross-section. Stub sizes in
cut from the header, polished up to 1000 grade emery Row 1 to Row 7 are equal; Row 8 and Row 9 stub sizes
paper followed by cloth polishing and then etched with are equal but are different from the previous set of stubs
freshly prepared nital (2 wt%). Scanning electron size. Row 9 stubs are absent in every alternate cross-
microscope (SEM) was used for microstructural section at the header locations along the length of the
analysis and fractographic studies. Samples were header unit. Besides this, it has been observed from the
properly cut to separate the failed region from the as site that number of hangers, at an approximate spacing
received header and these fracture surfaces were of 1720mm in the length direction is supporting the
properly cleaned using an ultrasonic cleaner. A number entire header unit. Fig. 1 shows the as received header

2
H. Roy et al. International Journal for Manufacturing Science & Production

removed from service. Metallographic specimens from the samples of

Careful visual inspection reveals that there are supplied pieces of platen super outlet header were
ligament cracks between the bore holes and along the selected, sectioned, polished and etched. All the
circumference of the header body (Fig. 2). The cracks specimens were examined under Optical Microscope
are found to be oriented axially in the tube-hole (MeF 3) and Scanning Electron Microscope (S – 3000
penetrating from inner surface. No major cracks were N) for evaluating micro-structural characteristics. The
observed on the longitudinal axis of the header. unetched sample from the transverse section with
Macroscopic examination of the specimens prepared respect to the bore hole revealed the presence of straight
from the failed superheater header using a research and unbranched oxide filled cracks propagating from
stereo-microscope (SZX – 9) also shows radial cracks in inner surface of the header body (Fig. 4). Similar
ligament region emanating from bore–holes of the observations were found in the longitudinal direction
header (Fig. 3). with respect to the bore hole (Fig. 5).

Fig. 2: Ligament cracks were observed between the

bore-holes and along the circumferential
direction of the header body. Fig. 4: The unetched sample revealed straight and
unbranched oxide filled cracks from inner
surface of the header body.

Fig. 3: Longitudinal section with respect to header

500μm
body along the diameter of the bore–hole
showing radial cracks in ligament region
emanating from bore–holes of the header Fig. 5: Longitudinal direction with respect to the bore
hole also revealed oxide filled crack.

3
 Failure of Platen Superheater Outlet Header

The microstructure near the major crack, adjacent to

the bore hole surface, reveals ferrite, complete
breakdown of bainite and coagulated alloy carbide in
association with few isolated creep cavities at the grain
boundaries (Fig. 6). The general microstructure away
from the ligament region shows ferrite, partial
breakdown of bainite and some coagulated alloy
carbides at the grain boundaries (Fig. 7). The chemical
composition (Table 1) mechanical properties (Table 2)
of the selected material as reported in confirms to the
grade ASME SA 213 P 22. The hardness measurements 25μm
and evaluated mechanical properties like yield strength,
ultimate tensile strength and percentage elongation Fig. 7: General microstructure away from the ligament
showed no abnormal deviation from normally expected region shows ferrite, partial breakdown of
value for the material construction of header. bainite and some coagulated alloy carbides at
the grain boundaries
(a)
Table 2
Tensile and average hardness properties of
the given material

Average
Sample YS UTS %
Hardness
No. (N/mm2) (N/mm2) Elongation
(HV)
1. 375 550 26
2. 380 555 24 155±3
3. 372 546 27
25μm

(b) As the full header unit is very long, for

simplification of the analysis and also to reduce the
finite element model size, it has been assumed that the
stress behavior that will be obtained from the analysis of
a portion of the header unit containing two hangers will
be able to indicate the critical stress locations and to
predict maximum stress magnitudes under the operating
loads. So, about 3000 mm length of the header model
has been considered for the analysis in order to find out
the overall stress pattern in the unit. The finite element
model of the unit has used two hanger supports at a
50μm distance of 1720 mm near the middle region of the
model. Absence of Row 9 stubs in alternate cross-
Fig. 6: Microstructure near the crack, adjacent to the section has been ignored in the model, since this will
bore hole revealed (a) ferrite with complete have negligible influence in the variation of overall
breakdown of bainite and coagulated carbide stress distribution of the unit.
(b) SEM photograph revealed few isolated The following physical and mechanical properties of
creep cavities. the header material and operating loading data are

4
H. Roy et al. International Journal for Manufacturing Science & Production

considered in the finite element analysis: Mises equivalent stress and axial stress distributions
• Working metal temperature = 4640 C near some critical areas of the header unit.
• Maximum working pressure = 15.8 MPa Fig. 9 shows Von Mises equivalent stress near
• Young’s modulus = 0.18e6 MPa different stubs at a distance of about 2 m from the right
• Poisson’s ratio = 0.3 end of the header model considered in the present
• Tensile strength = 490 MPa analysis. The figure shows variation of stress near the
• Yield strength = 275 MPa stubs along the circumferential direction. It is noticed
• Coefficient of linear expansion = 14e-6/ 0K that the maximum equivalent stress occurs near the
• Thermal conductivity = 33e-3 W/ mm 0K borehole edges of row 1 stub and the stress value is 290
• Assumed heat transfer coefficient = 0.2e-4 W/ mm2 MPa. Corresponding stress value near the row 9 stub is
0
K 269 MPa. The figure indicates a circumferential
• Minimum proof stress at 4640C = 220 MPa variation of maximum equivalent stress in the stubs of
different rows between 290 MPa and 140 MPa for the
Fig. 8 shows the three-dimensional view of the applied thermo mechanical loading in the finite element
model of the header unit considered for the present model.
analysis. The figure shows different stub locations in the
model. It may be mentioned here that the finite element
model of the header unit consists of 100399 elements
and 30848 nodes.

Fig. 9: Von Mises equivalent stress near the stubs

along the circumferential direction (2 m from
the right end of the model).

Fig. 8: Three dimensional view of the header body

considered for finite element analysis

Linear elastic stress analysis has been performed on

the super heater header model subjected to thermo-
mechanical loading. The analysis does not consider any
fluctuations in the operating loads. The theoretical stress
analysis work has been aimed to identify the critical
locations of maximum stress occurrence and also to
estimate the maximum stress magnitudes due to the
applied loadings. As the header component is subjected
to multiaxial stress state, Von Mises equivalent stress
magnitudes are obtained for the present investigations. Fig. 10: Von Mises equivalent stress near the stubs
Figs. 9 through 11 represent the enlarged view of von along the circumferential direction (1.33 m
from the right end of the model).

5
 Failure of Platen Superheater Outlet Header

concentration occurs. It is also found that the stress

magnitude reduces considerably at some distance away
from the borehole edges of the header. Maximum
equivalent stress values are found to be highly localized
in nature, which is evident from the above figures
shown for stress distribution. So, redistribution of stress
will take place and sizes of plastic zone will increase
near the critical areas of stub locations. Redistribution
will cause increase of strain concentration at the critical
zones i.e. near the borehole edges where stubs are
attached with the header. The resulting conditions are
susceptible to creation and growth of cracks during
long-term service of the unit near these critical regions.
Fig. 11: Von Mises equivalent stress near the stubs It may be mentioned here that the chances of crack
along the axial direction (1.33 m from the nucleation and propagation are higher in these regions,
right end of the model). if thermal transients and fluctuation in pressure and
thermal loadings during operation are considered which
Fig. 10 depicts the Von Mises equivalent stress are not taken into account in the present analysis.
magnitudes near different stubs at a distance of 1.33 m In the actual operating environments, super heater
from the right end of the model. At this cross-section of outlet header components operate at elevated
the header unit, maximum equivalent stress is found to temperatures and are often subjected to transient
be in the order of 220 MPa and it occurs near the temperature gradients due to start-up and shutdown
borehole edges of row 8 stub. Here, circumferential cycles. In such situations, much higher stresses than the
variation of maximum equivalent stress has been noted above calculated maximum stresses near the stub bore
between 220 MPa and 92 MPa in the stubs of different hole edges will occur. Hence, stress redistribution will
rows. result in more strain accumulation in the critical regions.
Fig.11 represents axial stress variation near different Stress fluctuations also occur due to the fluctuations in
stubs at a distance of 1.33 m from the right end of the pressure loading. All the above phenomena together
model considered. Maximum axial stress value in the may cause thermo-mechanical fatigue crack growth in
model is noted as 250 MPa and this occurs near the the unit near the critical locations, which are identified
borehole edge of row 8 stubs shown in Fig. 11. The in the present analysis. Cracks may radiate from tube
figure shows a variation between 250 MPa and 22 MPa borehole edges and linking up of adjacent tube cracks
near the boreholes at the stub locations. As this axial may develop ligament cracks. Header component are
stress value is found to be more than the yield stress often subjected to high magnitude of thermo mechanical
value of the material at the operating temperature of cyclic loadings as mentioned above and operate at high
4640C, chances of nucleation and growth of cracks near temperature for long duration such that both fatigue and
the borehole edges are there due to the long time use of creep damage may occur. Therefore, each must be taken
the component under the actual operating conditions. It into account for the remaining life assessment of the
has been noted that the estimated maximum equivalent header unit.
stress in the boiler header unit obtained from the above
analysis are more than the yield stress value of the
material at the operating temperature. So, the header 4. CONCLUSIONS
unit is actually subjected to elasto-plastic stress state
conditions. The present study shows that the stubs All the major cracks are radial with respect to the
create stress concentrations. Obviously, the critical bore-hole in the header body and extended across the
stress regions are at the locations where stress ligament region between the bore-holes. Micro-

6
 H. Roy et al. International Journal for Manufacturing Science & Production

structural examination revealed that the major ligament REFERENCES

cracks are transgranular, straight, un-branched and
contained oxide typical of thermal fatigue crack. The 1. Viswanathan R, Damage mechanisms and life
microstructure close to the ligament crack revealed assessment of high temperature components, ASM
material degradation in the way of isolated creep International, Metals Park, Ohio, 1989.
cavities at the grain boundaries possibly due to creep 2. Roberts BW, Ellis FV, Viswanathan R. Utility
fatigue interaction. Start-stop transients, temperature survey and inspection for life assessment of elevated
fluctuations during operation (minor in magnitude, but temperature headers. American Power Conference,
large in number), abnormal thermal shocks (few but TIS 7795, Chicago, 1985.
severe) and design deficiency are considered as possible 3. Ryder RH. Draft status report on the ligament
causes of formation of thermal fatigue cracks in the cracking problem, EPRI, Palo Alto,CA, 1987.
ligament region. Local temperature in the tube–hole 4. Roland R, King JP. Failure analysis of a cracked
ligament area exceeds the normal temperature by 500C superheater outlet header, 4th Conference on fossil
or more during transients and imposes large cyclic plant inspections, San Antonio, Texas, 1994.
thermal stresses in the ligament region in addition to the
steady state pressure stress leading to formation of
thermal fatigue / creep - fatigue cracks in the ligament
area.

View publication stats