Engineering Failure Analysis 18 (2011) 11931202

Contents lists available at ScienceDirect

Engineering Failure Analysis

journal homepage: www.elsevier.com/locate/engfailanal

Corrosion and fatigue failure analysis of a forced draft fan blade

H. Kazempour-Liacy a,, M. Mehdizadeh a, M. Akbari-Garakani b, S. Abouali c
a
b
c

Metallurgy Department, Niroo Research Institute (NRI), Tehran, Iran

School of Metallurgical and Materials Engineering, University of Tehran, Tehran, Iran
Department of Materials Science and Engineering, Sharif University of Tech., Tehran, Iran

a r t i c l e

i n f o

Article history:
Received 20 November 2010
Received in revised form 8 February 2011
Accepted 19 February 2011
Available online 25 February 2011
Keywords:
FDF blade
2014-T6 Al alloy
Failure analysis
Corrosion
Fatigue

a b s t r a c t
The present work was carried out to investigate the failure mechanism of a power plant
forced draft fan (FDF) blade made of 2014-T6 Al alloy, using visual examination, microstructural characterizations, study of the fracture surface and hardness testing. Based on
the obtained results, surface defects including corrosion pits due to the existence of Cl
ion and also erosion pits due to mechanical action of particles in the air entering from
the filtration system and abrading on the fan blades surfaces, increased the probability
of surface crack initiation. Cyclic loading and stress concentration on these surface defects
influenced the fatigue life of the blade due to elevating the mean stress and resulted in fatigue failure of the investigated FDF component.
2011 Elsevier Ltd. All rights reserved.

1. Introduction and background

The failure analysis of a forced draft fan (FDF) blade of a boiler unit in a fossil power plant is presented. The investigated
axial fan consists of two stages of blades which have been in service condition for about 200,000 h (according to the specifications given by the manufacturer, the expected life of these blades is 280,000 h). The fracture occurred in one of the first
stage blades 40 days after the last inspection while other blades in the both stages were damaged considerably.
In general there are four main types of fans used in fossil power plants including forced draft fans, induced draft fans,
primary air fans, and gas-recirculation fans. Draft fans are generally responsible for maintaining the flow of gases through
the boiler. A power plant can create draft through forced draft, induced draft, balanced draft, and natural draft. The forced
draft fan is placed upstream from the boiler and it applies a positive pressure to push the air and flue gases through the system [1,2]. Forced draft fans (FDF) supply the air necessary for fuel combustion by pushing the air through the combustion air
supply system and into the furnace. These fans must supply enough airflow to overcome any frictional resistance and airheater leakage problems and are typically the most efficient fans in the power plant [1]. Periodic inspections of the FDF
blades and the filtration system preventing entrance of the external objects play the most important role in the maintenance
of these fans.
2. Materials and experimental procedures
The investigated blades are made of 2014-T6 Al alloy and the manufacturing process includes forging and subsequently
solution heat treatment and artificially aging (T6 temper). 2xxx Al-alloys are a series of wrought alloys in which copper is the
principal alloying element, often with magnesium as a secondary addition. These alloys require solution heat treatment to
Corresponding author. Tel.: +98 21 88094320; fax: +98 21 888364620.
E-mail address: hkazempour@nri.ac.ir (H. Kazempour-Liacy).
1350-6307/$ - see front matter 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engfailanal.2011.02.014

1194

H. Kazempour-Liacy et al. / Engineering Failure Analysis 18 (2011) 11931202

obtain optimum properties. In the case 2014 Al-alloy, precipitation heat treatment (aging) is employed to further increase
mechanical properties [13].
Fig. 1 illustrates the general view of the damaged blades. Table 1 summarizes the operation conditions of the FDF under
evaluation and Table 2 shows the chemical composition of the blades in comparison with the chemical composition of 2014
Al alloy.
Visual examination was performed on the whole FD fan to determine the amount of damages. Also, the fracture surface of
the fractured blade was studied using both stereo optical microscope and scanning electron microscope (SEM). The energy
dispersive spectroscopy (EDS) microanalysis was performed to characterize the element concentration of the selected areas.
Also, to evaluate the microstructural changes in the fractured blade, samples were prepared and etched by a solution of
2.5 ml HNO3 + 1 ml HCl + 1.5 ml HF + 95 ml water and were studied by optical microscope.
Macro-hardness test was carried out to investigate the mechanical properties of the airfoil and root sections of the fractured blade. Samples were tested in an ESEWAY DV RB-M testing unit using HB method. At least three indentations were
made for each hardness measurement and the average hardness values are reported.
3. Results and discussion
As can be seen in Fig. 1, most of the blades have been damaged and these damages are more significant in the second
stage of blades. Besides, deteriorations are more severe on the leading edge rather than the trailing edge, in both first and

Fig. 1. The general view of the damaged FD fan blades.

Table 1
Specifications of the damaged FD fan.
230 m3/s
25 C
1589 mmws
1480 rpm
3960 KW
1122 mm

Air flow rate

Air temperature
Pressure of the exhausting air
Rotational speed
Engine power
Rotor diameter

Table 2
Chemical composition of the FDF blade under evaluation.

Sample
AA2014

Cr

Cu

Fe

Mg

Mn

Si

Ti

Zn

Al

0.0157
0.1 max

4.51
3.95

0.277
0.7 max

0.579
0.20.8

0.887
0.41.2

0.704
0.51.2

0.0375
0.15 max

0.14
0.25 max

Bal.
Bal.

H. Kazempour-Liacy et al. / Engineering Failure Analysis 18 (2011) 11931202

1195

second stages of blades. The Investigations were focused on the fractured blade which is shown in Fig. 2. The fracture surface
is approximately perpendicular to the longitudinal direction of the blade. Fig. 3 demonstrates the leading edge of the damaged blade after being cleaned. There are some holes on the pressure surface of the airfoil near the leading edge that are
similar to erosion holes. One of the most common damage mechanisms associated with power plant fan failure is erosion,
which is responsible for serious and costly maintenance. Observations show that there is a possibility of entering external
objects through the filtration system. Collision of these objects to the surface of the blades can result in erosion. The effects
of erosion can lead to extreme unbalance and high vibration. Erosion of fan units is characterized by the loss of fan material
through the mechanical action of particulate matter in the air flow impinging or abrading on the fan blades surfaces.
Fig. 4 shows the microstructure of the leading edge. It can be seen that a layer with a thickness of 5001000 lm and with
a different microstructure has covered the surface area and it has been disappeared on the tip of the leading edge (Fig. 5).
Moreover, Fig. 5 shows that some parts of the blade were separated from the suction and pressure surfaces and also from
the tip of the leading edge. Fig. 6 represents these surface holes in higher magnification. It seems that these holes with
the maximum depth of 200 lm formed due to corrosion pitting. Also, some signs of inter-granular corrosion were detected
in areas near the leading edge. The result of EDS analysis of the products in these pits is given in Fig. 7 which shows the existence of alloying elements oxide such as Al. Furthermore, the amounts of Cl and Cu are noticeable in the obtained result. The
most common types of corrosion problems found in fans units are inter-granular corrosion, general corrosion and erosion
corrosion. Microstructural studies of the samples surfaces approved the occurrence of pitting and inter-granular corrosion.
Regarding the location of the power plant and the results of EDS analysis, it seems that Cl ion is responsible for the corrosion
pitting which its initiation is generally controlled by the cathodic reaction kinetics and the propagation requires sufficient
Cl concentration in the solution contained within the pits. The electrochemical mechanisms of pitting corrosion are very
complex and not totally understood. The practical relevance of these theoretical models is still limited [46]. Pitting corrosion shows two distinct stages: initiation and propagation. It has long been known that pitting corrosion develops in the
presence of chloride. Cl are adsorbed on the natural oxide film [4], followed by the rupture of the film at weak points, with
formation of micro-cracks that are a few nano-meters wide. Many pits are initiated within a very short time, however most
pits will stop after a few days. Polarization studies have demonstrated that when pits stop growing, they will be re-passivated. When the metal is polarized once again, these passivated pits will not be reinitiated again, but pitting will start on fresh

Fig. 2. The general view of the fractured blade.

Fig. 3. Leading edge of the damaged blade. Signs of erosion can be seen in this region.

1196

H. Kazempour-Liacy et al. / Engineering Failure Analysis 18 (2011) 11931202

Fig. 4. Optical micrographs of the surface layer of the damaged blade.

Fig. 5. Optical micrographs of leading edge of the damaged blade.

sites. Oxygen will be reduced slowly in cathodic areas. These areas seem to be inter-metallic phases underneath the oxide
layer that more or less covers them. Where the film cracks, aluminum will oxidize rapidly and a complex intermediate
chloride AlCl4 will be formed [4]. Only a minute fraction of initiated pits will continue to propagate according to the
two electrochemical reactions [4]: oxidation at the anode formed by the pits bottom and reduction at the cathode outside
the cavity. The corrosion mechanism of aluminum alloys is depicted schematically in Fig. 8. Reboul et al. [6] proposed a
10-step mechanism in detail for the pitting corrosion of aluminum in the presence of chloride ions.
Inter-granular corrosion is the selective attack of the grain boundary zone, with no appreciable attack of the grain body or
matrix and corrosion can be caused by impurities at the grain boundaries, enrichment of one of the alloying elements or
depletion of one of these elements in the grain boundary areas [5,7]. The mechanism is electrochemical and is the result
of local cell action in the grain boundaries. Cells are formed between second-phase micro-constituents and the depleted
aluminum solid solution from which these micro-constituents formed. The micro-constituents have a different corrosion
potential than the adjacent depleted solid solution. In some alloys, such as the aluminummagnesium and aluminum
zincmagnesiumcopper families, the precipitates Mg2Al3, MgZn2, and AlxZnxMg are more anodic than the adjacent solid
solution. In other alloys, such as aluminumcopper, the precipitates (CuAl2, and AlxCuxMg) are cathodic to the depleted solid
solution. In either case, selective attack of the grain boundary region occurs [5]. In AlCu alloys containing CuAl2 precipitates,
substantial potential differences between the copper-depleted areas and adjacent material have been demonstrated [810].
According to chemical composition of the corrosion products in the surface pits, it seems that Cu depletion in grain boundaries caused inter-granular corrosion in a cyclic loading.
The degree of inter-granular susceptibility is controlled by fabrication practices that can affect the quantity, size, and distribution of second-phase inter-metallic precipitates. Resistance to inter-granular corrosion is obtained by heat treatments
that cause precipitation to be more general throughout the grain structure, or by restricting the amount of alloying elements
that cause the problem. Alloys that do not form second-phase micro-constituents at grain boundaries, or those in which the
micro-constituents have corrosion potentials similar to the matrix (e.g., MnAl6), are not susceptible to inter-granular corrosion [5].

1197

H. Kazempour-Liacy et al. / Engineering Failure Analysis 18 (2011) 11931202

Fig. 6. Optical micrographs of damages in different regions on the surface of the blade (Fig. 5).

element
Wt%

Al

Si

Cl

Ca

Mn

32.94

43.56

1.23

1.56

6.89

1.07

0.58

Cu
12.17

Fig. 7. Chemical composition of the products existing in the surface holes of the damaged blade.

Figs. 9 and 10 represent the EDS analysis results of the edge and the central part of the sample respectively. These results
indicate that chemical compositions of these two zones are similar to the chemical composition of 2014 alloy. The alloys in
the 2xxx series, do not have a good corrosion resistance as most other Al-alloys and under certain conditions, they may be
subjected to inter-granular corrosion. Therefore, these alloys are usually clad with high purity aluminum or with a magnesium-silicon alloy of the 6xxx series, which provides galvanic protection of the core material and thus greatly increase
resistance to corrosion. EDS analysis showed that the coating layer of the fan blades under evaluation, has the same chemical
composition with the matrix. So it seems that selecting this layer as a corrosion resistant layer is not suitable in this case.
Fig. 11 indicates the fracture surface of the damaged blade that is divided into three zones. Figs. 1216 show the stereo
and scanning electron micrographs of these three zones on the fracture surface. Striations are clearly visible in the first and

1198

H. Kazempour-Liacy et al. / Engineering Failure Analysis 18 (2011) 11931202

Fig. 8. Pitting corrosion mechanism of aluminum in the presence of chloride ions [4].

element

Mg

Al

Si

Mn

Cu

Wt%

0.27

92.73

0.46

0.83

5.72

Fig. 9. Chemical composition of the edge of the fractured blade (surface layer).

second zone (Figs. 1315). In the third zone (Fig. 16), dimples are detected which are attributed to a ductile fracture occurred
in the last stage. The beach marks indicate the cyclic loading during the service and show the different steps of fatigue crack
growth. Excessive vibration due to corrosion and erosion is a common damage mechanism associated with fan unit fatigue
failures. When an axial fan operates near the peak of the pressure-curve at a particular blade angle corresponding to the minimum flow rate, the fan can operate without separation of airflow over the blades [1]. Operation in this condition can result
in significant reduction in fatigue life of the component. Cyclic fatigue is another common damage mechanism associated
with boiler fan units. It is the failure of a material under cyclic loading. Failure occurs due to local strains repeatedly exceeding the yield point of the material, eventually leading to plastic flow failure [1,1113]. If the material surpasses the yield
point, the material becomes strain hardened due to local distortions in the lattice structure, but if the yield point is repeatedly exceeded small cracks appear due to the largely distorted lattice. This occurs early in the materials fatigue life, usually
<5% of the number of cycles to failure [1]. These small stress fractures begin to propagate eventually leading to material failure. Also, Pitting due to erosion or corrosion has a considerable influence on fatigue life under tensile dynamic loading which
may result in fatigue crack initiation in a corrosive medium due to the local stress concentration at a corroded surface [14].
According to what can be inferred from the above explanation, erosion, corrosion, vibration and fatigue are the sequential
events that led to the final rupture. The fracture mode of the component, determines the mechanism which leads to the
failure of the blade. The existence of striations on the fracture surface approves the occurrence of fatigue fracture [14,15].
On the other hand, it is well known that corrosion and fatigue damage can occur simultaneously and they have a combined
effect much more severe than each one occurring on its own. In a number of publications, the effect of corrosion pits on
fatigue behavior has been studied. Since fatigue cracks will initiate preferentially from any surface pits, including corrosion
pits, the presence of corrosion reduces the fatigue life of the component [16].

1199

H. Kazempour-Liacy et al. / Engineering Failure Analysis 18 (2011) 11931202

element

Mg

Al

Si

Mn

Cu

Wt%

0.32

92.35

0.22

0.71

6.40

Fig. 10. Chemical composition of the central region of the damaged blade.

Fig. 11. Macrograph of the damaged blade fracture surface. 3 zones are marked on the fracture surface.

Fig. 12. Stereo micrograph of the fracture surface.

1200

H. Kazempour-Liacy et al. / Engineering Failure Analysis 18 (2011) 11931202

Fig. 13. SEM micrographs of center of the zone 1 (Fig. 11) on the fracture surface.

Fig. 14. SEM micrographs of edge of the zone 1 (Fig. 11) on the fracture surface.

Fig. 15. SEM micrographs of zone 2 (Fig. 11) on the fracture surface.

1201

H. Kazempour-Liacy et al. / Engineering Failure Analysis 18 (2011) 11931202

Fig. 16. SEM micrographs of zone 3 (Fig. 11) on the fracture surface.

Table 3
Results of hardness test of the fractured blade.
Macro-hardness (HB)
Location

Measurements

Root
Airfoil

148
142

Ave.
145
146

145
147

146
145

Table 3 summarizes the hardness test results of the fractured blade. The macro-hardness of the root and also the airfoil of
the blade, were determined to be about 145 HB. Therefore, it seems the investigated failure was not due to the mechanical
properties degradation of the blade. In the present case, it seems that the cyclic loading and stress concentration on surface
defects due to erosion and pitting corrosion influenced the fatigue life of the blade by elevating the mean stress and resulted
in fatigue failure of the investigated FDF component.
4. Conclusion
The failure analysis of a fractured forced draft fan blade reveals that since the power plant is located near a sea, Cl ion
presence in the atmosphere caused pitting in the surface of the blades. Besides, existence of surface defects due to erosion
through the mechanical action of particulate matter in the air flow, resulted in fatigue cracks initiation on the surface of the
investigated blade due to stress concentration. Furthermore, it is well known that corrosion and fatigue damage can occur
simultaneously and they have a combined effect much more severe than each one occurring on its own. Finally, the growth
of fatigue cracks, leads to the failure of the blade. The results of hardness testing showed that service conditions have not
affected the mechanical properties of the alloy.
Acknowledgements
The authors wish to express appreciation to Neka Power Plant Generation Management for supporting of this project. Mr.
B. Boroumandi and Mr. A. Haghshenas are gratefully acknowledged for their assistance in providing information.
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]

EPRI, Fan repair guideline. Technical report 1007082; 2002.

NRI, Neka Power Plant Generation Management Co., Technical report for a FDF blade failure analysis NRI-FNM 525/3399; 2009.
Crane FAA, Furness J. Selection and use of engineering materials. London: Elsevier Science & Technology Books; 1997.
Vargel C. Corrosion of aluminium. 1st ed. Amsterdam: Elsevier; 2004.
Winston RR. Uhligs corrosion handbook. 2nd ed. New York: John Wiley & Sons; 2000.
Reboul M, Warner T, Mayet H, Baroux B. Corros Rev 1997;15:47196.
Chlistovsky RM, Heffernan PJ, DuQuesnay DL. Int J Fatigue 2007;29:19419.
DuQuesnay DL, Underhill PR, Britt HJ. Int J Fatigue 2003;25:3717.
Kemp RMJ, Wilson RN, Gregson PJ. J Aerosp Eng 1993;207:97104.
Shiong Lin FU, Starke EA. Mater Sci Eng 1979;39:2741.
Jones K, Hoeppner DW. Corros Sci 2006;48:10922.
Jones K, Shinde SR, Clark PN, Hoeppner DW. Corros Sci 2008;50:258895.

1202
[13]
[14]
[15]
[16]

H. Kazempour-Liacy et al. / Engineering Failure Analysis 18 (2011) 11931202

Sankaran KK, Perez R, Jata KV. Mater Sci Eng A 2001;297:2239.
Pao PS, Gill SJ, Feng CR. Scripta Mater 2000;43:3916.
Bagnoli F, Bernabei M. Eng Fail Anal 2008;15:65465.
Liao M, Renaud G, Bellinger NC. Int J Fatigue 2007;29:67786.