Engineering Failure Analysis 92 (2018) 405–417

Contents lists available at ScienceDirect

Engineering Failure Analysis

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

Localized pitting corrosion of API 5L grade A pipe used in

T
industrial fire water piping applications
Chidambaram Subramanian
Numaligarh Refinery Limited, Assam, India

A R T IC LE I N F O ABS TRA CT

Keywords: In this present paper, failure analysis is conducted for a pipe used in fire water applications. In
Internal corrosion petroleum industries many failures were observed in API 5 L grade steel pipes. The past in-
Carbon steel spection history records were observed, which reveals that failed pipeline is poorly maintained
Water quality since construction. Visual inspection, macroscopic and microscopic examinations were con-
Corrosion voids
ducted for the failed pipe. The microscopic analysis reveals that the pitting corrosion is a root
Recirculated water
cause of failure which perforates the wall thickness at 6′o′ clock pipe position. The mechanism of
pitting corrosion is correlated with ferrite matrix dissolution. The Cu, Co & Zn ions were detected
in scale deposits reveals that the fire water is severely contaminated by various makeup water
sources which results in under deposit corrosion. The various sources of foreign elements are
discussed in detail. The copper deposits identified at suspected location attributed to galvanic
corrosion. Moreover, the severity of pitting is aggravated by grain boundary attack which acts as
an anode during corrosion process. Based on various experimental evidences, the conclusions
were drawn and further recommendations are discussed to mitigate the corrosion failure.

1. Introduction

The corrosion of carbon steel is a critical concern for various major chemical processing industries particularly in fire piping
network system. Although various pipe line materials are available in the market but still carbon steel is a candidate material for
transporting chemicals, acids, hydrocarbon products and water. In processing industries 98% of pipelines are constructed from
variety of tubular carbon steels. The allowable stress for carbon steel pipe is designed based on pipe wall thickness and it will be 0.4
times of specified yield strength (SYS) considering (FOS = 2.0) factor of safety [1]. After certain period of time, the wall thickness of
the pipe gradually decreases uniformly throughout the pipe or in localized area due to corrosion leading to increase of hoop stress.
These hoop stress mechanically react with pipe wall to produce bulging, buckling, deformation and rupture [2, 3]. The various causes
for reduction in wall thickness of the pipe after certain time frame are due to corrosion, erosion, wear, mechanical dent or any of the
above combinations. Many literatures were published regarding various pipeline grade materials. However, still carbon steel pipe-
lines failures are pertaining in petroleum and petrochemical industries. These pipelines undergo several failures such as uniform
corrosion [1], galvanic corrosion [2], erosion corrosion [3], pitting corrosion [2] and microbiological corrosion [4–6]. Also many
authors have shown that corroded portion of the pipe was obtained in reduced wall thickness. Although, welding defect [7, 8], stress
corrosion cracking and hydrogen cracking [9] cause serious damages in the pipe; they were not observed to reduce wall thickness.
The failure of fire water pipe line during an emergency may lead to uncontrolled situations resulting to social-economic losses.
The fire water piping is a part of firefighting systems in petroleum refineries and petrochemical plants. For example, the fire water
pipeline supplies huge amount of water which mixes with polar solvent compound in ratio of 97:3 to form foam at room temperature.

E-mail address: chidambaramselva@gmail.com.

https://doi.org/10.1016/j.engfailanal.2018.06.008
Received 22 March 2018; Received in revised form 9 June 2018; Accepted 15 June 2018
Available online 18 June 2018
1350-6307/ © 2018 Elsevier Ltd. All rights reserved.
C. Subramanian Engineering Failure Analysis 92 (2018) 405–417

Furthermore, the foam will be supplied in compressed manner to tank externals by sprinkler system. While spraying, the foam forms
aqueous film layer on hydrocarbon source and insulate the fire from an environment containing oxygen, cooling the fuel to protect
and control the fire in tank. The solvent compound is alcohol resistant in nature and fire water is supplied at room temperature. These
tanks primarily contain any intermediate or final hydrocarbon products and fire source from these combustible hydrocarbon or
petroleum based products is classified as class B fire. The recent advancements in automation system automatically supplies water
and solvent to form foam in foam tank, and releases the foam if smoke detector detects any fire or smoke in or nearby tank location.
Therefore, the pipeline carrying fire water mix with solvent to form foam will protect and control fire, and availability of these fire
piping systems is extremely important to hydrocarbon industries round the clock. Frequent inspections and testing of these pipelines
is a fundamental requirement in any asset integrity programmes of hydrocarbon industries. J Sorbel et al. proposed that frequent
inspection of firefighting equipment's reduces the failure [10]. Yulong et al. established functional analysis system technique (FAST)
through which he concluded that the fire equipment frequently failed during an actual emergency operation. In his paper he classified
the firefighting equipment's as minor, major and significant emergency failures [11].
This paper focuses on metallurgical failure analysis due to poor maintenance of fire water pipeline made of carbon steel. In this
case the cause of corrosion damage observed at the base metal of pipe is thoroughly investigated. The leak was observed at 6’o’ clock
position of the pipe. The water flow was temporarily stopped after detecting leakage and water flushing was terminated.
Subsequently the damaged pipe portion was replaced with new carbon steel material and the pipeline was again put to service. The
sectioned portion of the failed pipe was examined further in detail through systematic root cause analysis for characterizing the
damage features and to prevent similar failures in future. The detailed literature survey showed that corrosion product scale on top of
pitting corrosion contains hematite (Fe2O3), magnetite (Fe3O4), goethite (FeOOH), wuestite (FeO), and calcite (CaCO3) [2]. The
literature survey also shows few reports on pit depth and its characterization using scanning electron microscopy. Zhiyong et al.
investigated an isolated pit depth measurement in a failed gas pipeline [12]. However, literatures are limited for understanding a
correlation between pitting mechanism across a transverse section of a pipe in fire water applications.

2. Background information

The water pipe line has failed for the first time since it was commissioned twenty years back. The entire fire water piping network
was constructed with API 5 L Grade A carbon steel material. The material chemical composition and mechanical properties of the
pipe were collected from a bill of material as a part of inspection verification programme given in Table 1 and Table 2 respectively.
This seamless pipe was supplied from hot strip mill (hot rolled and normalized subsequently) of a steel plant. This pipe was produced
via basic oxygen furnace (BOF) steel making route with a carbon equivalent (CE) approximately 0.253 max and then it was hot rolled.
The carbon equivalent (CE) of a line pipe is obtained as stated in Eq. 1.

CE = C + Mn/6 + (Cr + Mo + V)/5 + (Ni + Cu)/15 (1)

The total pipeline engaged in fire water service within refinery facilities is approximately 50 km in length, with a dimensions of
φ304.8 x t 6.02mm2. Welding procedure specification & qualification record of a pipe shows that the pipe edges are prepared for
double V groove butt joint, subsequently welded by E-6010 electrode for root pass and E-7018 for remaining passes. Thereafter, each
weld joints were accentuated to non-destructive tests (dye penetrant and radiography) and reports were stored. Further entire piping
was subjected to hydro-test at 25 kg/sq.cm to assure a structural (welding) integrity and it was confirmed that neither leakage nor
pressure drop was recorded during hydro test. Subsequently, a pipe was externally coated by zinc ethyl silicate primer of 160 μm,
epoxy zinc phosphate of 160 μm and acrylic polyurethane of 120 μm serially to obtain 440 μm of total coating thickness. Similar
coating was applied on internal surface of pipe. An external coating was applied to protect the pipe from atmospheric corrosion in an
industrial atmosphere. Similarly, an internal coating was applied to avoid corrosion of pipe from fire water.
Each phase of pipe line construction complied with quality standards and there was no evidence of abnormality. However, poor
maintenance was evidenced and inspection was never carried out on aforementioned piping resulting to leakage. On reporting of
failure, it was confirmed from inspection history records that no prior inspection was carried out on entire pipeline of fire water
network. The preliminary inspection on a leak pipe reveals that localized corrosion has attacked the base metal of the pipe but there
was no abnormality on or near weld joint. Schematic representation of failed pipe within refinery complex is shown in Fig. 1. The
corrosion allowance for the pipe is 1.5 mm and the design standard suggests replacement of the pipe if the thickness of the pipe
is < 4.52 mm (critical wall thickness tcr≤4.52 mm). In a suspected portion of the pipe the thickness was measured to be < 1.5 mm
and hence replacement of failed pipe is recommended.

Table 1
Chemical composition of API 5L Grade A [1].
Elements weight percentage (%) Carbon Manganese Phosphorous Sulfur Iron

ASTM Specifications 0.22 max 0.90 0.020 max 0.020 max balance

406
C. Subramanian Engineering Failure Analysis 92 (2018) 405–417

Table 2
Mechanical property requirement of API 5 L Grade A [1].
Mechanical properties Yield Strength Tensile Strength Elongation Hardness

ASTM Specifications 207 MPa 330 MPa 22 187 BHN

Fig. 1. Schematic representation of fire water piping and grain boundaries along with pitting directly impact service life of pipeline.

3. Investigation

3.1. Visual examination

On observation of leakage at one location of fire water service piping network, the failed portion and nearby area of the pipe was
visually inspected and damages were photo documented. The leakage is found apparent on the seamless base pipe and the failed pipe
was sectioned by cold cutting without affecting the damaged area (Refer Fig. 2). The chemical composition of pipe line material was
analyzed by x-ray fluorescence (XRF) by using Oxford instrumentation X-Met 7500 model which is primarily well known technique
for positive material identification. In addition, chemical composition was analyzed to confirm the material of construction as
specified in original isometric drawing and it was found that the chemical analysis of failed pipe is API 5 L Grade A carbon steel (refer
Table 3).

3.2. Mechanical characterization

The tensile specimen sample was prepared from nearby portion of failed pipe to determine the mechanical properties of the
material. The conventional tensile test was carried out using Instron universal tensile testing machine at a strain rate of 10−3 to
evaluate the tensile properties of the pipe. Hardness of pipe was measured using MIC 201A GE model diamond pyramid indenter
supplied by general electric. It is depicted from Table 4 that maximum yield strength, maximum tensile strength and elongation
obtained from conventional test is higher than minimum specified value (refer Table 4) of design codes and standards. Moreover, the
hardness of the failed sample was found less than as specified in design codes and standards. Therefore, from evaluation of me-
chanical characterization, it was concluded that mechanical integrity of failed pipe is in sound condition which is essentially required
for fire water services.

407
C. Subramanian Engineering Failure Analysis 92 (2018) 405–417

Fig. 2. Internal localized pitting corrosion in pipe and failed portion sectioned for failure characterization.

Table 3
Chemical composition of failed pipe measured by XRF.
Elements weight percentage (%) Cu Mn P S Ni Al Ti V Fe

Failed pipe 0.05 0.44 0.030 0.020 0.05 0.02 0.02 0.01 balance

Table 4
Mechanical properties obtained from conventional test.
Mechanical properties Yield Strength Tensile Strength Elongation Hardness

Failed pipe 234 MPa 342 MPa 27 163 BHN

3.3. SEM and EDX analysis

3.3.1. As-received sample condition

An elemental analysis was conducted on failed samples using energy dispersive X-ray technique which was attached with a
scanning electron microscope Vega2 LSU model. Samples of (l) 2 mm X (w) 2 mm X (t) 2 mm were prepared without affecting
corrosion product deposits. An elemental and topographical analysis shows that corrosion product scale (Refer Fig. 3(a) & (b)) was
formed in bottom portion of the pipe and also contained high amount of oxygen. Significant amount of chloride ions, sulfur ions,
silica ions and calcium ions were accumulated on the bottom portion of pipe. Moreover in the bottom portion of the pipe small
amounts of cobalt and zinc ions along with significant amount of copper ions were observed on the corrosion product scale. Con-
trastingly, the corrosion products in the top portion of the pipe contained iron and carbon ions with an oxide scale (Refer Fig. 3(c)).
The morphology of corrosion product scale in bottom portion of the pipe is adherent and porous in nature. But the top portion of the
pipe surface shows that no heavy deposits corrosion product scale is visible unlike bottom surface (Table 5).
The SEM/EDX of the bottom portion of the pipe clearly reveals that corrosion product scales formed on top of the pits contain
significant amount of corrosive ions resulting to pitting corrosion. Further, those pits were grown towards the direction of gravity.
The EDX examination also confirms that the fire service water was contaminated by foreign particles namely Cu, Co & Zn ions. From
an elemental analysis, it was inferred that paint or coat damage in the internal portion of the pipe might have created void initially,
altering the localized electrochemical behavior in the bottom portion, and further initiated the pits on those suspected location.
Further, chlorine and sulfur ions in the scale deposit are continuously driving the corrosion process. The zinc and aluminum ions in
the corrosion product scale is attributed to the internal paint coating of the pipe as the various paint products in market contains zinc
and aluminum as primary elements.
The water reservoir located within the refinery was feed from nearby lake and this reservoir acts as a source of water for entire fire
water system. In addition to reservoir water, process treated water especially boot water from overhead piping and strip water from
sour water stripper section, and recirculated cooling water from cooling towers were also feed into fire water systems and it is called
as makeup water. The source of silica in our case might be sand particles in water and fire water containing silica tend to damage the

408
C. Subramanian Engineering Failure Analysis 92 (2018) 405–417

Fig. 3. SEM/EDX profile analysis of failed cross section at different locations of the sample (a), (b) bottom inner surface & (c) top inner surface.

pipe by erosion corrosion during fluid flow. The process leak or any upset in recirculated cooling water might be dissolved copper in
cooling water and source of copper is from brass based heat exchanger tubes. The cooling water treated with corrosion inhibitors also
contains calcium carbonate scales on heat transfer surfaces. The source of calcium ions might come from recirculated cooling water,
which used as make up water to fire water systems. The source of sulfur and chlorine ions detected at bottom 6 ‘o’ clock pipe positions
might have obtained from strip water and overhead boot water respectively. The source of cobalt might be obtained from catalyst
section from where the original sour water generated.

3.4. Metallography

The metallography specimens were prepared by standard metallographic techniques via mechanical cutting, sectioning,
mounting, grinding and polishing. The samples were grinded using 240, 320, 400 and 600 grit papers initially and then fine polished
using 6 μ, 3 μ, 1 μ and 0.25 μ diamond abrasives finally. Samples were mechanically cut down using cutting machine of discotom 65

409
C. Subramanian Engineering Failure Analysis 92 (2018) 405–417

Table 5
shows results of elements analyzed in corrosion product scale.
Elements Spectrum 1 (6'o'clock pipe position) Spectrum 2 (6'o'clock pipe position) Spectrum 3 (12'o'clock pipe position)

Weight % Atomic % Weight % Atomic % Weight % Atomic %

OK 33.15 54.93 28.52 52.98 26.65 55.91

Mn K 0.69 0.33 0.42 0.23 0.01 0.01
Fe K 52.12 24.74 59.95 31.90 73.34 44.08
CK 7.25 16.01 4.30 10.65 – –
Al K 0.39 0.39 0.55 0.60 – –
Si K 1.06 1.00 0.74 0.78 – –
SK 0.36 0.30 0.25 0.23 – –
Cl K 0.18 0.13 – – – –
Ca K 0.58 0.39 0.43 0.32 – –
Co K 0.98 0.44 1.28 0.65 – –
Cu K 2.24 0.94 2.74 1.28 – –
Zn K 1.00 0.41 0.82 0.37 – –
Total 100 100 100 100 100 100

model of struers make. Samples were mounted using automatic mounting press of simplimet model 1000 of buehler make. Thereafter,
the samples were grinded and polished using metaserv 250 model of buehler make. Prepared samples were etched with freshly
prepared 2% nital solution for 5 s and metallographic images were captured and recorded by optical microscopy of zeiss axio Vert A1
model. Metallographic specimens were taken from A, B and C locations of failed pipe as shown in Fig. 2. The metallographic structure
at location A near perforated region is shown in Fig. 4 and nearby region is shown in Fig. 5. The transverse intact pipe materials at
representative locations are showing α-ferrite microstructure having no inter-phase boundaries or deformation characteristics across
grain structure. Near to leaked section of pipe the remaining wall thickness was found approximately 1.5 mm (refer Fig. 4).
At lower magnification, the α-ferrite phase appear as an equiaxed grain structure and shows that failed API 5 L Grade pipe was
produced and supplied via hot rolled and then normalized (air cooled) route (see Fig. 5). Higher volume fraction of an equiaxed un-
deformed α-ferrite structure apparently shows lower carbon equivalent of supplied pipe material. The specimen representing inner
bottom surface of pipe at location B (see Fig. 2) is shown in Fig. 6 and Fig. 7. It is observed that the corrosion product contains oxide
scale formation due to electrochemical reactions. The edges of corroded cavities are not smooth and close angle connections of α-
ferrite in the microstructure is electrochemical in nature with the presence of electrolyte (refer Fig. 6). A similar result was also
observed by other researchers for pitting corrosion characterization [2, 3]. The light phase (i.e. white layer) shows intact ferrite
grains in Fig. 6. However, an electrochemical reaction progress the intact ferrite grains entrained inside an oxide layer showing dark
phase in the microstructure (refer Fig. 6). Similar results were observed by other researchers for characterizing uniform corrosion,
pitting corrosion and galvanic corrosion [1–3].
The ferrite grain boundaries acts as heterogeneous anodic sites and grain interior acts as homogeneous cathodic sites for the pipe
material resulting to localized corrosion cell formation. Preferential corrosive attack on the ferrite grain boundary results in material
loss until wall thickness completely perforates on an inner bottom surface i.e. 6‘o'clock position of the pipe (see Fig. 7 and Fig. 9). The
copper ions detected at 6‘o'clock pipe position accelerated localized galvanic corrosion attack. Few concentrations of copper ions
reacted with steel surface and steel preferentially affected by galvanic corrosion.

Fig. 4. Nearly 1.5 mm remaining wall thickness of corroded pipe after 20 years of service.

410
C. Subramanian Engineering Failure Analysis 92 (2018) 405–417

Fig. 5. Corrosion of ferrite matrix phase in carbon steel subjected to fire water service.

Fig. 6. Dark ferrite phase entrained inside oxide corrosion product.

Fig. 7. Microstructure reveal grain boundary corrosion which filled with oxide scales.

411
C. Subramanian Engineering Failure Analysis 92 (2018) 405–417

Fig. 8. Failed pipe at higher magnification shows corrosion on ferrite grain boundaries.

It is clearly inferred from microstructure that the pipe was not soaked for a long time homogeneously above an upper critical AC1
temperature which leads to absence of triple junction in grain boundary. Therefore, subsequently obtained ferrite grain size is same as
prior austenite grain size and this could further lead to absence of triple junction grain boundary. The triple junction grain boundaries
forming during uniform soaking treatment for a long time have a better corrosion resistance than the material having absence of
triple junction. An absence of triple junction in grain boundary acts as potential heterogeneous anodic sites favoring an electro-
chemical reaction. The preferential dissolution of ferrite grain in grain boundaries filled with iron oxides is shown in Fig. 6, Fig. 7 and
Fig. 8. The size effect i.e. grain boundary area (anodic sites) is comparably smaller than grain interior area (cathodic sites) and it also
plays a vital role in electrochemical reaction. Each localized galvanic corrosion cell propagates through the wall thickness resulting in
excessive wall loss of metal on or near pit region. Similar results were also observed by other researchers [2, 3] [12, 13].
The thickness of the pipe reaching lower than the critical thickness was detected by ultrasonic thickness gauging measurements.
The corrosion rate for this pipe is not known due to unavailability of measured thickness data at various time intervals i.e. the
thickness was not measured since construction. The damaged pipe wall might lost its mechanical strength in localized pitted region
within corroded zone as it underwent pitting corrosion along with corroded grain boundary. As lower wall thickness was not in-
spected frequently, the fire water suddenly released from the ruptured pipe. Perforation appeared on transverse direction of a pipe
wall and perforated edges apparently appeared in irregular fashion showing complete ferrite grains dissolution as shown in Fig. 9
which was clearly witnessed through photomicrograph. This class of corrosion is called as localized internal corrosion as specified by
many chemical processing, petroleum and petro-chemical industries.
The pipe internal surface wall was observed with significant number of pits as shown in Fig. 10 & Fig. 11. The service water
directly in contact with internal wall surface produces significant number of corrosion pits due to electrochemical reaction. The pits

Fig. 9. Through thickness wall perforated shown by arrow resulted in leakage.

412
C. Subramanian Engineering Failure Analysis 92 (2018) 405–417

Fig. 10. Large number of internal corrosion pits at internal wall surface directly contact with water.

Fig. 11. Isolated pits on internal wall surface contact with water.

were mostly formed in varied shapes and sizes depending upon localized electrochemical behavior of ferrite grain and its boundary
structure. It could be very well understood from Fig. 10 and Fig. 12 that pits of different shapes namely sub surface, elliptical,
shallow, undercut, irregular, circular, shallow, and narrow were formed at bottom section (6 ‘o’clock position) of pipe wall. Few
isolated corrosion pits were also detected and all pits were observed only on inner diameter of the pipe. It is well depicted from
Fig. 12 that the pits were grown downwards, in the direction of gravity.
The size effect increases the kinetic process of an electrochemical reaction producing pit clusters nearby sub surfaces. Therefore,
the faster electrochemical reaction rate increases the growth causes of pit clusters as shown in Fig. 10. However, it is difficult to
predict an exact time when the pit was originally initiated. Other researchers have shown that the pits were initiated long back during
piping operations but in many cases leakages are known only after material rupture of the pipelines [1, 2] [12]. In this present case
study, the photomicrographs revealing pitting corrosion was initiated and propagated long time back, and not monitored by available
nondestructive methods resulting to leakage. Even though ultrasonic gauging was conducted, there might be a chance of not placing
the probe at exact suspected location to monitor wall losses and this localized area might be missed in an ultrasonic examination. The
ferrite grain boundary acts as anodic sites with respect to ferrite grain interior which completely undergoes dissolution. Similar
mechanism was also reported in many literatures for uniform corrosion. Therefore, it is important to mention that sometimes me-
chanism of pitting corrosion may overlap the mechanism of a uniform corrosion.

3.5. Water analysis

The water samples of failed pipe network system were collected at different locations in refinery for analyzing its chemical

413
C. Subramanian Engineering Failure Analysis 92 (2018) 405–417

Fig. 12. Transverse sample corrosion pattern and various shapes of pit namely sub surface@1, elliptical@2, undercut@3, circular@4, irregular@5,
shallow@6 and narrow@7 in ferrite matrix. Pits were close together beneath 0.5 mm internal wall surface.

constituents. Both physical and chemical properties of water were measured [4] and detailed results are shown in Table 6. Turbidity
is 50 Nephelometric Turbidity Units (NTU), pH is in range of 7.5 to 8.5 and oxidation reduction potential is 225 mV. A water sample
contains chlorine (Cl−) ion as low as 22 ppm, however calcium (Ca2+) and magnesium (Mg2+) ions are found much higher. All the
water parameters were maintained within specified limit suitable for fire water applications and no abnormality was noted. Similar
results were obtained by other researchers with some minor differences depending upon location of industry, type of water used and
recirculation systems etc… [2–4].

4. Discussions

A pitting is one among special form of degrading mechanism in metallurgical corrosion for various pipeline steel grades used in
petroleum industries. The differential cell produce localized pits which continue to grow by autocatalytic process called self-sti-
mulating and self-propagating. An anodic and cathodic reaction has occurred near each individual pit, and anodic surface of the
metal will self-dissolute within a pit as given in Eq. (2). Simultaneously in adjacent cathodic surfaces, occurrence of oxygen reduction
as given in Eq. (3) causes pit to stimulate and propagate. An overall electrochemical reaction maintains electro neutrality resulting to
positive charge within the pit and migrate electrons to form hydroxide product as schematically illustrated in Fig. 13.
Fe → Fe2 − + 2e− (2)

1
O2 + H2 O + 2e− → 2OH−
2 (3)

1
Fe + O2 + H2 O → Fe(OH)2
2 (4)
An overall electrochemical reaction given in Eq. (4) and in this present case shows that an electrochemical reaction favorably

Table 6
Results of water sample collected from pipe in different locations.
Parameters Units Water sample

Turbidity NTU 50
pH – 7.5–8.5
Conductivity μs/cm 141.0
Hardness ppm 92
Free residual chlorine ppm traces
Oxidation reduction potential mV 225
Anions
Chlorine (Cl−) ppm 22.90
Sulfate (SO42−) ppm 21.00
Silica (SiO2−) ppm 11.80
Cations
Calcium (Ca2+) ppm 50.00
Magnesium (Mg2+) ppm 42.00

414
C. Subramanian Engineering Failure Analysis 92 (2018) 405–417

Fig. 13. Schematic illustration of localized pitting corrosion damage in pipe.

depends on material heterogeneity. This is evidenced by appearance of corrosion morphology on inner wall surface of the pipe,
featuring oxygen corrosion along with cluster of pits appearing across the transverse/thickness direction at bottom portion of the
pipe. As a result of corrosion process few oxide scales formed on the top inner portion (12 ‘o’clock) might have fell down to the
bottom inner surface (6’o’clock) due to gravity. The corrosion product scale collected from those bottom surfaces contain high
amount of oxygen and iron resulting to form iron oxide which is evidenced by elemental analysis. The presence of calcium carbonate
in the corrosion scales obtained from corroded surface is attributed to the alkali treated water used in fire water service to maintain
its pH within specified limit range.
The foreign elements namely copper, aluminum, zinc, cobalt, silica, chlorine and sulfur evidenced in corrosion product scales
cause an under deposit corrosion especially on the restricted water flow or water stagnant areas. All these corrosion aggravating
favorable conditions leads to formation of pit clusters on localized area of the bottom inner pipe wall. Similar results were reported in
literatures on how under deposits lead to major and minor leakages [13–16]. The foreign particles (copper, aluminum, zinc and
cobalt) may carry over with fire water if fire water is chemically treated with di-oxide chlorine (ClO2) poly silicate treatment. The
silica particles in the fire water will remove the metal surface by impingement will result in gouging and corrosion may also initiate
from those gouges.
Frequent water circulation and water flushing in the fire water piping systems may wash away an under deposits particularly on
inner bottom surface of pipe. Scanning electron image reveals morphological evidences on localized attacks of the pipe. The pit
exhibits red brownish appearance on various scale formed species evidencing lack of water circulation inside the pipe. Corrosion in
an oxygen depleted area beneath under-deposits might have achieved anodic character when compared to its adjacent area.
Consequently, most favorable situation prevails for pitting corrosion in large number of aspects leading to aggressive attack on the
pipe. Internal epoxy or polypropylene coated carbon steel pipe is used for fire water piping systems which may prevent oxygen
reaction with line pipe.
Perforated pit would have taken a long time (may be several years) to grow downwards before it fails as it appeared in photo-
micrographs. There will be no indication of failure until leakage is observed with extreme suddenness in the field. Dissolved oxygen
content in water due to local conditions is also suspected for pinhole leak. Pit observed on internal diameter of the pipe undercuts the
surface layer during its growth as shown in Fig. 12. It was learned from the metallography that all the pits initiated from inner wall
(ID) surface and progressed out-ward. It is studied from the detailed metallographic investigation that failure of the pipeline material
affected by both anodic grain boundary uniform corrosion (local galvanic cell due to copper deposits) and pitting corrosion. The
dissolved oxygen content in water has kinetically driven the corrosion process and the presence of oxygen has led to an oxygen
reduction within the pit. The dissolved oxygen content or oxidizers like ferric and cupric chloride in water increased significantly as
the water recirculated for longer time. Oxygen enriched humid atmosphere formed inside the pipe favored localized rapid corrosion
attack. Similar results were observed by other researchers [17, 18]. The few ferrite phases which are relatively intact within material
must contain an oxide layer. This oxide and metal interface lead to aqueous corrosion in damaged area revealing ridges after many
years is shown in Fig. 6.
Various pit shapes are anti-symmetric, non-uniform and blunt in nature. However, other researchers have shown that pits were
found deep, narrow and sharp [3] [12]. The sharp pits rupture faster than the blunt pits as shown in many literatures. Other
researchers have also shown that the mechanical stressor plays a vital role at root of pit depth leading to rupture of material [13]. The
red rust scales formed initially due to corrosion were further oxidized resulting to brown Fe2O3oxide scale. Similar results were
obtained by other researchers [2] [8]. Schematic representation of pitting corrosion is shown (Refer Fig. 13). Appropriate inspection
programme must be carried and wall thickness must be monitored in suitable time intervals to obtain the corrosion rate. Addition of
suitable corrosion inhibitors to water will control under deposit scale formation and oxygen or oxidizers. These inhibitors directly
prevent autocatalytic process within pit and localized galvanic cell within interface of grain and grain boundaries.

5. Conclusions

1) API 5 L Grade A is right choice of material chosen for fire water applications. No metallurgy up-gradation is required;

415
C. Subramanian Engineering Failure Analysis 92 (2018) 405–417

2) The localized corrosion rate in internal bottom surface is many fold higher than top internal surface of the failed pipe;
3) The bottom inner surface of this damaged pipe was affected by under deposit corrosion, localized pitting corrosion and anodic
grain boundary corrosion. The pipeline is corroded for several years since installation resulting in complete rupture of material;
4) This localized attack is continue to grow in gravitation direction featuring pitting corrosion characteristics;
5) Ferrite grain boundary and ferrite grain interior act as a heterogeneous anodic site and cathodic site respectively;
6) The corroded voids is filled with corrosion product scales contain iron oxides;
7) In addition the presence of chloride and sulfur above a threshold concentration at scale deposits has autocatalytic effect on
corrosion reactions;
8) The corroded ferrite phase also entrained inside an oxide layer and stagnation of fire water or lower water flow is evident from
corrosion product deposit morphology.

6. Recommendations

The following recommendations are proposed for corrosion protection of entire fire water piping services

• There appears to be shortcoming in maintenance over a period of time in entire fire water pipeline operation. The periodic water
circulation and cleaning is proposed to avoid scale formation at the inner surface of the fire water piping systems. The corrosive
elements in under deposits will also be washed away along with cleaning water.
• The proper water circulation and cleaning process will remove an under deposits formed on inner wall surfaces of the pipe.
• The water used for circulation and cleaning shall be drained completely, and recommended not be reuse in fire water piping or
similar piping applications unless otherwise such water is treated to remove corrosive elements in it.
• Phosphate and nitrate based anodic inhibitors shall be used in effluent water treatment to deactivate foreign elements present in
boot water and strip water. Anodic inhibitors will reduce the possibility of chloride and sulfur ions respectively in boot and strip
water.
• Cathodic inhibitors like azoles shall be added to recirculated cooling water which will prevent steel from galvanic corrosion by
avoiding copper deposit formation.

Future scope to improve carbon steel durability in fire water service applications

• Carbon steel pipe with proper protective coating appear to be the future material for piping applications;
• Utilizing waste water and regenerated water like boot water and stripper water in fire water requires new generation of corrosion
inhibitors called ambiodic inhibitors (effective to control both anodic and cathodic reactions) are needed in petroleum industrial
environment.

Acknowledgement

The author gratefully acknowledges the anonymous reviewers for many valuable comments that helped to substantially improve
this article.

References

[1] American Society of Metals ASM handbook Vol. 13 Corrosion pp 113–114.

[2] Saud Al Subai, et al., Failure analysis of firewater pipeline, Case Studies in Engineering Failure Analysis, vol. 2, 2014, pp. 144–149, , http://dx.doi.org/10.1016/
j.csefa.2014.07.002.
[3] En-Na Yang, et al., Failure analysis of a leaked oil pipeline, Case Studies in Engineering Failure Analysis, vol. 4, 2015, pp. 88–93, , http://dx.doi.org/10.1016/j.
csefa.2015.09.004.
[4] S. Chidambaram, et al., IOP Conference Series: Material Science and Engineering, 314 (2018) 012027, , http://dx.doi.org/10.1088/1757-899X/314/1/012027.
[5] Hendrik Venzlaff, et al., Accelerated cathodic reaction in microbial corrosion of iron due to direct electron uptake by sulfate-reducing bacteria, Corrosion Science
66 (2013) 88–96, http://dx.doi.org/10.1016/j.corsci.2012.09.006.
[6] SSh Abedi, A. Abdolmaleki, N. Adibi, Failure analysis of SCC and SRB induced cracking of a transmission oil products pipeline, Eng. Fail Anal 14 (1) (2007)
250–261, http://dx.doi.org/10.1016/j.engfailanal.2005.07.024.
[7] Mohamed Hanafy El-Sayed, Grooving corrosion of seam welded oil pipelines, Case Studies in Engineering Failure Analysis, vol. 2, 2014, pp. 84–90, , http://dx.
doi.org/10.1016/j.csefa.2014.05.002.
[8] Peng Cheng-hong, et al., Perforated mechanism of a water line outlet tee pipe for an oil well drilling rig, Case Studies in Engineering Failure Analysis, vol. 4,
2015, pp. 39–49, , http://dx.doi.org/10.1016/j.csefa.2015.07.002.
[9] M.B. Djukic, V. Sijacki Zeravcic, G.M. Bakic, A. Sedmak, B. Rajicic, Hydrogen damage of steels: a case study and hydrogen embrittlement model, Engineering
Failure Analysis 58 (5) (2015) 485–498, http://dx.doi.org/10.1016/j.engfailanal.2015.05.017.
[10] J. Sobral, L.A. Ferreira, Availability of fire pumping systems under periodic inspection, Journal of Building Engineering Vol. 8 (2016) 285–291, http://dx.doi.
org/10.1016/j.jobe.2016.09.001.
[11] D.U. Yulong, M.A.O. Xing, X.U. Dajun, R.E.N. Fuchun, Analyzing the effects of failure on fire equipment in building by FAST, International Symposium on Safety
Science and Technology Procedia Engineering, vol. 45, 2012, pp. 655–662, , http://dx.doi.org/10.1016/j.proeng.2012.08.218.
[12] Zhiyong Liu, Wenju Liao, Wei Wu, Du Cuiwei, Xiaogang Li, Failure analysis of leakage caused by perforation in an L415 steel gas pipeline, Case Studies in
Engineering Failure Analysis, vol. 9, 2017, pp. 63–70, , http://dx.doi.org/10.1016/j.csefa.2017.07.003.
[13] M.N. Ilman, Kusmono, Analysis of internal corrosion in subsea oil pipeline, Case Studies in Engineering Failure Analysis, vol. 2, 2014, pp. 1–8, , http://dx.doi.
org/10.1016/j.csefa.2013.12.003.
[14] G. Bianchi, G. Fiori, P. Longhi, F. Mazza, Horse Shoe corrosion of copper alloys in flowing sea water: mechanism, and possibility of cathodic protection of

416
C. Subramanian Engineering Failure Analysis 92 (2018) 405–417

condenser tubes in power stations, Corrosion 34 (1978) 396–406, http://dx.doi.org/10.5006/0010-9312-34.11.396.

[15] Anees Malik, Abdelkader Meroufel, Saleh Al-Fozan, Boiler tubes failures: a compendium of case studies, Journal of Failure Analysis and Prevention (2015)
246–250, http://dx.doi.org/10.1007/s11668-015-9923-x.
[16] R.E. Hargrave, Unusual failures involving copper deposition in boiler tubing, Corrosion 47 (7) (1991) 555–567, http://dx.doi.org/10.5006/1.3585292.
[17] Y.H. Huang, T.C. Zhang, Effects of dissolved oxygen on formation of corrosion products and concomitant oxygen and nitrate reduction in zero-valent iron
systems with or without aqueous Fe2+, Water Res. 39 (2005) 1751–1760.
[18] S.L. Grise, B.J. Saldanha, Effects of oxygen, temperature and salinity on carbon steel corrosion in aqueous solutions, NACE International, 2008 CORROSION.

417