You are on page 1of 9

Tunnelling and Underground Space Technology 51 (2016) 1119

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology


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

Structural performance of buried prestressed concrete cylinder pipes


with harnessed joints interaction using numerical modeling
Masood Hajali a, Ali Alavinasab a, Caesar Abi Shdid b,
a
b

Pure Technologies, 3420 State Route 22 West, Suite 130, Branchburg, NJ 08876, USA
Lebanese American University, Department of Civil Engineering, 211 E 46th St., New York, NY 10017, USA

a r t i c l e

i n f o

Article history:
Received 7 October 2014
Received in revised form 13 September 2015
Accepted 8 October 2015
Available online 22 October 2015
Keywords:
Buried concrete pipes
Pipe joints
Prestressed concrete cylinder pipes
Numerical modeling
Earth pressure
Finite element analysis

a b s t r a c t
Broken prestressing wire wraps are the main cause of failure in buried prestressed concrete cylinder pipes
(PCCP), which form the backbone of water and wastewater infrastructure networks in North America.
Advanced numerical modeling using non-linear finite elements is used to model the effect of the number
and location of broken wire wraps on the structural performance of Class 125-14, 96-in. PCCP. The modeling technique used is unique in that it considers full interaction between adjacent pipes with harnessed
joints, as well as combined internal and external loading with full soilpipe interaction. Performance indicators in the various components of PCCP are monitored as internal pressure is increased. A sensitivity
analysis is presented for how manipulating the severity of the damage affects the failure pressure of
the pipe. The results show that the internal fluid pressure required to cause failure can be as much as
34% lower when the damage is at the barrel of the pipe, and that the internal pressure that causes yielding
of the wire wraps decreases by 66% as the damage worsens from 5 to 100 wire breaks.
2015 Elsevier Ltd. All rights reserved.

1. Introduction
Prestressed concrete cylinder pipe (PCCP) originally appeared in
1942 as Lined Cylinder Pipe (LCP). After a decade, Embedded Cylinder Pipe (ECP) was developed as another type of PCCP with concrete encasement of the steel cylinder on both sides and
prestressing wire wrapped around the outer concrete core. The
design and manufacturing standards of PCCP in the United States
are now published by the American Water Works Association
(AWWA) in the AWWA C301 Standard Specifications for Reinforced
Concrete Water Pipe Steel Cylinder Type, Prestressed (AWWA
C301-52), with the latest revision released in 2007 (AWWA
C304-2007). While PCCP is widely used nowadays in underground
water and wastewater transmission networks, the understanding
of its behavior under combined internal and external loading is still
being gradually developed.
PCCP consists of a concrete core, a steel cylinder, high tensile
prestressing wires, and an outer mortar coating layer. The concrete
core is the load-bearing component with the steel cylinder acting
as a water barrier between inner and outer core concrete layers. Prestressing wires produce a uniform circumferential compressive
pressure in the concrete core that balances tensile stresses
Corresponding author.
E-mail addresses: masood.hajali@puretechltd.com (M. Hajali), ali.alavinasab@
puretechltd.com (A. Alavinasab), caesar.abishdid@lau.edu.lb (C. Abi Shdid).
http://dx.doi.org/10.1016/j.tust.2015.10.016
0886-7798/ 2015 Elsevier Ltd. All rights reserved.

developed in the pipe from internal fluid pressure. The mortar coating protects the prestressing wires from physical damage and external corrosion. Rupture of prestressing wires around the concrete core
is common in PCCP and can be the result of damage due to corrosion,
hydrogen embrittlement, overloading, or manufacturing defects. As
a result of this loss of circumferential compressive load around the
pipe, tensile stresses will develop that can lead to possible cracking
of the concrete core and cause leak or damage in the pipe.
While the structural condition of underground water and
wastewater mains can be assessed using ultrasonic tomography
methods (Abi Shdid and Hajali, 2014; Yang et al., 2010), satellite
detection of ground movement (Arsnio et al., 2014); direct inspections by remotely-controlled closed circuit television (CCTV), and
more recently sewer scanner and evaluation technology (SSET)
cameras, remains to be the most accurate and widely used method
to detect damage in buried infrastructure elements such as PCCP.
Such inspections need to be conducted on a regular and systematic
basis in order to monitor deterioration rates and perform on-time
replacement of pipes prior to their failure. However, the large size,
underground nature, and operating conditions of these facilities
make it prohibitively expensive to do so in a manner that mitigates
the serious effects of their failure. Owners have therefore resorted
to the use of risk curves that are developed based on numerical values assigned by inspectors that place any pipe in one of five internal condition grades (ICG) according to a subjective assessment of

12

M. Hajali et al. / Tunnelling and Underground Space Technology 51 (2016) 1119

covariates such as pipe age, material, and extent of damage. Such


risk models are then used to predict failure pressures and
remaining life of pipe elements. A more accurate and efficient
approachthan using subjective ICGto producing failure data
for developing such risk curves is the use of accurate numerical
modeling such as the FEM method described by Jung et al.
(2014) that models full interaction between buried pipes and the
overburden soil.
2. Background
The behavior of PCCP under combined internal and external
loading has been under investigation since the middle of the twentieth century, and studies of PCCP failures have led to the development of the design standards (AWWA Research Foundation, 2007).
Despite standardizing the design and manufacturing of PCCP, the
inclusive understanding of the structural behavior and performance of damaged PCCP is still in its infancy.
Advances in the field of numerical modeling techniques and
finite element analysis have led to considerable benefits to many
engineering industries (Mahendran, 2007; Ovesy et al., 2015) and
not only allows the introduction of innovative and efficient products, but also the development of accurate design methods.
Xiong et al. (2010) used a nonlinear FEM to study the correlation
between the degree of prestressing stresses during manufacturing
of PCCP and their associated resultant stresses in the outer concrete core and prestressing wires. The study also compared the
resultant stress obtained with another FEM model that replaces
the effects of tensile stresses in prestressing wires with an equivalent radial pressure around the pipe. The results obtained from the
equivalent radial pressure model were within a 10% deviation of
the proposed model (Xiong et al., 2010).
Rauniyar (2013) conducted full-scale experimental tests and
numerical modeling of ECP under three-edge bearing load. The
study utilized three-dimensional nonlinear finite element analysis
for the numerical modeling, and used composite material with
complex stress phenomenon due to prestress and interaction
between the various component layers of the ECP. The model
accounted for the contribution of each component, the manufacturing process, and the simultaneous effects of shrinkage, creep
and relaxation (Rauniyar, 2013). More advanced models have been
used in recent studies to investigate the behavior of PCCP using an
extended form of the finite element method (XFEM). Alavinasab
et al. (2011a) used XFEM to study crack initiation, growth, and life
prediction analysis of damage in Prestressed Concrete Noncylinder
Pipe (NCP). Nonlinear FEM has been used not only in static analysis
of PCCP, but also in the dynamic response of PCCP. Alavinasab et al.
(2010, 2011b) used FEM to evaluate the natural frequencies and
mode shapes for Structural Health Monitoring (SHM) of PCCP.
The structural performance of damaged PCCP is dependent on
the number of broken prestressing wire wraps as well as on the
location of such break regions along the length of the pipe.
Alavinasab et al. (2013) studied the effect of the location of broken
wire wraps on the strength of PCCP using advanced computational
modeling. The study compared three different locations for the
defect: at the spigot joint, at the bell joint, and in the barrel of
the pipe. The study however consider no interaction between the
bell end and spigot end of adjacent pipes, which leads to overly
conservative results that assume complete disjointedness of adjacent pipes. The results found that strength reduction for a PCCP
with low to medium number of wire wrap breaks at a joint was
about 20%. The paper argued that cracking in the pipe will occur
much sooner when the defects occurred at the joint rather than
in the barrel of the pipe. Alavinasab and Hajali (2014) studied
the effects of broken wire wraps at the joint in the safe operation
of an adjacent pipe.

Kang and Davidson (2013) used finite element analysis to study


the effect of concrete lining on the structural performance of buried concrete-lined steel pipes. The approach utilized detailed soil
modeling and presented a design method for the concrete lining
of pipes (Kang and Davidson, 2013). Allouche et al. (2014) examined the performance of liners for the rehabilitation of cured-inplace pipes (CIPP). The study evaluated the usefulness of various
types of testing for tracking the deterioration of CIPP liners inservice and concluded that they are more cost effective than
replacement of damaged concrete pipes (Allouche et al., 2014).
Artificial Neural Networks (ANN) were used by Amaitik and
Amaitik (2008) to develop a PCCP wire breaks prediction model.
The ANN was trained on real-world acoustic monitoring data.
The ANN takes the monitoring period, pipe age, soil resistivity,
design pressure, design soil density, design soil cover, type of
pre-stressing wire wrap, wire diameter, and wire pitch as inputs;
and predicts the number of wire breaks (Amaitik and Amaitik,
2008). ANNs were also used successfully to predict the compressive strength of concrete and its degradation under corrosion
attack in buried concrete wastewater collection pipes (Hewayde
et al., 2007). Another approach was developed by Kleiner et al.
(2004) to model the deterioration of buried PCCP using a fuzzy
rule-based, non-homogeneous Markov process. The model yielded
possibility of failure at every point along the life of the pipe. However, adequate and sufficient data to validate the model were not
provided (Kleiner et al., 2004).
Other studies have attempted to use probabilistic models for
investigating the trends of structural deterioration of concrete
pipes. Younis and Knight (2010) used a new ordinal regression
model for the deterioration of reinforced concrete wastewater
pipelines based on cumulative logits. The model was presented
using the Generalized Linear model formulation and incorporated
the interaction effect between the explanatory variables (Younis
and Knight, 2010). However all such models do is predict the probability of a buried concrete pipe falling in one of the five ICGs, as
compared to the others that give a numerical value for the ICG of
a pipe. They hence fall short of accurately modeling the failure
based on an exact measurement of deterioration variables such
as broken wires.

3. Problem statement and contribution


3.1. PCCP joints
The rigid nature of PCCP makes the joint a very important component of the pipeline. The function of a pipe system generally
determines the performance requirements of the joints, but in general, joints make construction a lot easier. Joints are also designed
so that when pipe sections are laid together they will make a continuous line of pipe with an interior free from irregularities. Joints
are normally designed to provide soil-tightness, water-tightness,
the ability to accommodate lateral and longitudinal movement,
and the strength to handle shear that induces vertical deformation.
Joints in PCCP consist of a spigot ring, a bell ring, a rubber gasket, a steel spigot ring, a steel bell ring, grout on the exterior of the
joint between the two pipes, and a harness clamp. Spigot and bell
steel rings are welded to the steel cylinder from the ends. Joints are
designed to allow a pipe to deflect during installation and operation while maintaining a watertight seal. After the spigot steel ring
is pushed home into the bell steel ring, the clamp halves of the harness are tightened using bolts for sealing of the joint. When joined,
the bell and spigot ends compress the rubber gasket into a groove
to form a high pressure seal. A schematic of the PCCP joint is shown
in Fig. 1; such a joint provides high shear strength, excellent water
tightness, and flexibility. A layer of mortar or cement paste is

13

M. Hajali et al. / Tunnelling and Underground Space Technology 51 (2016) 1119

end of the pipe, but also because of the relative displacement and
rotation between adjacent pipes. A Class 125-14, 244 cm (96-in.)
ECP was considered and modeled for this purpose, and the stresses
and strains developed in the prestressing wire wraps, concrete
core, and mortar coating were compared for the cases of wire
breaks at the joint and in the middle of the pipe. The structural performance of the PCCP was evaluated using four (4) different measurements: micro cracking, visible cracking, yielding of wire wraps,
and rupture stress of wire wraps.
4.1. Finite element modeling

HARNESS CLAMP

RUBBER GASKET

Fig. 1. Schematic of the joint in PCCP.

placed on the inner side of the bell ring and on the external side of
the spigot end of the pipe. The exterior of the joints are subsequently grouted to protect the steel joint rings from deterioration.
Concrete thrust blocks were commonly used to carry the unbalanced thrust forces at the elbows and other fittings of concrete
pipes. These blocks derive the major part of their resistance to
movement from the soil wedge behind them. While concrete
thrust blocks provide an acceptable equilibrant for unbalanced
thrust forces, the possibility of disturbance to the soil wedge by
street excavationsrequired for the installation and maintenance
of other utilitieshave lead concrete pipe manufactures to develop
harnessed joints. Such joints are capable of transmitting forces
through the pipe wall, which are then dissipated by friction to
the surrounding soil. Although the joint is harnessed, it still retains
the flexibility to permit slight differential settlement.
3.2. Theoretical and practical motivation
Once in tension, cracks in the concrete initiate, and with additional strain, its load carrying capacity decreases further. At high
strains, the load carrying capacity of the concrete core becomes
negligible, and the loads are transferred to other components of
the PCCP. This load transformation differs at the joints due to the
interaction between the pipes spigot and bell rings. The interaction between adjacent pipes thus affects the performance of PCCP
and needs to be considered and carefully modeled.
The effects of the location and number of broken prestressing
wire wraps on the performance of damaged PCCP have not been
thoroughly investigated, and the few studies available in the literature have not considered the effect of the interaction at harnessed
joints between adjacent pipes. Broken wire wraps in PCCP can occur
near the joint or in the barrel (middle) of the pipe. The objective of
this study is to investigate how the location of broken prestressing
wire wraps affects the strength of the PCCP when full interaction
exists between the bell and spigot ends of the adjacent pipes. Such
understanding is critical to owners and inspectors in evaluating
both the short- and long-term performance of a pipeline system.

A numerical model of a 244 cm (96-in.) diameter Class 125-14


ECP (rated for 14 feet underground depth) with a length of 6.1 m
(20 feet) is assembled using three-dimensional nonlinear finite elements that incorporate the creep and shrinkage effects of the concrete, as shown in Fig. 2a. The ECP section is modeled using a
composite element with five layers representing the inner concrete
core, the steel cylinder, the outer concrete core, the prestressing
wire, and the mortar coating. Care is taken when modeling the prestressing wire wraps and the joint rings to ensure that a realistic
behavior of the ECP was achieved. The model predicts the performance of a PCCP, utilizing the tensile strengths of the prestressing
wire, the steel cylinder, the concrete core, the mortar coating, as
well as a plasticity algorithm that models concrete crushing in
compression regions. The model assumes perfect interface
between the prestressing wires and the concrete. A four-node
quadratic shell element, in which each node has six degrees of freedom, is used in modeling the undamaged and damaged portions of
the pipe. An eight-node linear brick element, in which each node
has three translational degrees of freedom, is used for modeling
the joint ends of the pipe. Fig. 2b shows the finite element mesh
model in the PCCP and the details of the different layers in the pipe
wall thickness.
The ECP section was modeled based on the tensile strength of
concrete and a plasticity algorithm that facilitated concrete crushing in compression regions. The concrete was modeled using a
three-dimensional composite axi-symmetric element with additional adjustments made to predict the failure of brittle materials.
Cracking and crushing were determined by a failure surface, which
formed the boundary between the undamaged zone and failure
(damaged) zone. Once the failure surface was reached, cracking
or crushing occurred. The nonlinear elastic behavior of concrete
is defined by the multi-linear stressstrain relationships governed
by scalar damaged elasticity as shown in Eqs. (1) and (2) (Hajali
and Abi Shdid, 2013).

r Del : e  epl

Del 1  dDel0

4. Modeling approach
Modeling the structural behavior of a damaged underground
PCCP at a joint is a complex nonlinear problem due to the interaction between not only the broken prestressing wire wraps and the

Bell End

Spigot End

Fig. 2a. Spigot and bell ends of the PCCP numerical model.

14

M. Hajali et al. / Tunnelling and Underground Space Technology 51 (2016) 1119

Fig. 2b. Finite element mesh model and the layers of the PCCP wall thickness.

Fig. 3. PCCP with 5, 35, 70, and 100 broken wire wraps.

Table 1
Damaged pipe length corresponding to number of wire breaks (WB).
ECP class
designation

Internal
pressure
rating

Backfill
height
rating

Damage length, cm (in.)


5 WB

35 WB

70 WB

100
WB

Class 12514

0.86 MPa
(125 psi)

4.3 m
(14 ft)

5.5
(2.17)

38.6
(15.18)

77.2
(30.36)

110.2
(43.38)

where Del0 is the initial (undamaged) elastic stiffness of the material,


Del is the degraded elastic stiffness, and d is the scalar stiffness
degradation variable, which can take values from zero (undamaged
material) to one (fully damaged material). Damage associated with
the failure mechanisms of the concrete (cracking and crushing)
therefore results in a reduction in the elastic stiffness.

e_ e_ el e_ pl

where e_ is the total strain rate, e_ el is the elastic component of the


strain rate, and e_ pl is the plastic component of the strain rate.
The average wall thicknesses of the inner and outer concrete
core are 4.93 cm (1.94 in.) and 11.43 cm (4.5 in.), respectively.
The average thicknesses of the steel cylinder and mortar coating
are 0.15 cm (0.0598 in.) and 2.30 cm (0.904 in.), respectively. The
center-to-center spacing of prestressing wire wraps is set to be
1.10 cm (0.434 in.). The lengths of the bell and spigot ends are considered to be 14 cm (5.5 in.) and 19.05 cm (7.25 in.), respectively.
The average wall thicknesses for the bell ring and spigot ring are
0.635 cm (0.25 in.) and 0.95 cm (0.375 in.), respectively. By virtue
of symmetry, only half of each pipe section is modeled with fixed
displacement boundary condition in longitudinal direction

considered at the intact pipe side, thus allowing the damaged pipe
to move in longitudinal direction. In order to have a realistic behavior for the harnessed joint, interface elements are used at the joint
to allow small movement to occur. An efficient approach developed
by Mayer and Gaul (2005) is used to model the contact interfaces of
joints with segment-to-segment contact elements like thin layer or
zero thickness elements used between spigot and bell ends.
The model was subjected to loads corresponding to internal
fluid pressure, pipe and fluid weights, and external earth loads as
per the pipe dimensions and internal pressure and backfill ratings
defined by (AWWA C304-2007). The Soil was not directly modeled,
but rather the Marston Theory was used for calculating earth loads
on PCCP buried pipes. The formula is world-recognized as the Marston Load Equation for Rigid Pipes. The Marston Theory considers the
effect of vertical soil load and lateral soil support in its equations
(Moser and Folkman, 2001).
The number of broken wire wraps was varied over a broad range
of distress levels as shown in Fig. 3 in order to evaluate the cracking
of the concrete core and mortar coating, and the yielding of the prestressing wires with increasing internal fluid pressure as the number of broken wire wraps increases. The lengths of damaged ECP
sections corresponding to the various numbers of wire breaks considered for the joint and the barrel are shown in Table 1. Three (3)
different broken wire wrap scenarios were modeled: starting from
the spigot end and continuing along the pipe length, starting from
the bell end and continuing along the pipe length, and in the barrel
of the pipe. The scenario of broken wire wraps starting at the spigot
end of the PCCP is shown in Fig. 4a, and the scenario of broken wire
wraps at the barrel of the PCCP is shown in Fig. 4b.
A bilinear stressstrain relationship is used for modeling the
steel material. The values used to represent the performance of
the steel components are based on the yield and ultimate strengths

15

M. Hajali et al. / Tunnelling and Underground Space Technology 51 (2016) 1119

Fig. 4a. PCCP with broken wire wraps at the spigot end.

Fig. 4b. PCCP with broken wire wraps at the barrel.

provided in the pipe design specifications sheet or the standard


values in the relevant design standard. The yield strength of the
prestressing wire is typically 85% of its ultimate strength, while
the yield strength of the steel cylinder is either denoted on the pipe
design specifications sheet or taken from the design standard in
place at the time of production. The Yield Limit is reached when
either the steel cylinder or the prestressing wire reach its yield
strength. The ultimate strength of the prestressing wire is dictated
by the gage and class of the wire, while the ultimate strength of the
steel cylinder is determined by the grade of the steel. The Strength
Limit is exceeded when one of the PCCP components reaches its
ultimate strength, which, theoretically, will cause the failure of
the pipe.

Table 2
Material properties of PCCP based on AWWA C304 standard.

4.2. Material properties

where cc is the concrete density, taken as 2323 kg/m (145 lb/ft3);


fc0 is the 28-day compressive strength of concrete, taken as
31 MPa (4500 psi); cm is the mortar coating density, considered as
2242 kg/m3 (140 lb/ft3); and fm0 is the 28-day compressive strength
of mortar coating, taken as 37.9 MPa (5500 psi).
The gross wrapping stress, fsg, which is the stress in the prestressing wire during wrapping, is considered as 75% of the specified minimum tensile strength of the wire, as shown in Eq. (6). The
yield strength of wire, fsy, is taken as 85% of the specified tensile
strength of the prestressing wire, as shown in Eq. (7).

The material properties used in the PCCP model are obtained


from the current (AWWA C304-2007). The concrete core is modeled with a 28-day compressive strength of 31 MPa (4500 psi).
The tensile strength of the concrete core is modeled at 3235 kPa
(470 psi). The Youngs Modulus, density, and Poissons Ratio used
in the various components of the PCCP model are shown in Table 2.
The modulus of elasticity of the core concrete is calculated from Eq.
(4) and modulus of elasticity of the mortar coating evaluated using
Eq. (5). The stressstrain behavior of the concrete core and mortar
coating are modeled based on the (AWWA C304-2007).

Component

Youngs Modulus
(GPa)/(psi)

Density
(kg/m3)/(lb/ft3)

Poissons
Ratio

Concrete
Mortar coating
Prestressing wires
Steel cylinder

27.17/3.94E+06
25.1/3.64E+06
193.05/28E+06
206.84/3E+07

2322.61/145
2242.58/140
7832.8/489
7832.8/489

0.2
0.2
0.3
0.3

0 0:3

Ec 0:074c1:51
fc
c
0

fm
Ec 0:074c1:51
m

0:3

5
3

f sg 0:75f su

16

M. Hajali et al. / Tunnelling and Underground Space Technology 51 (2016) 1119

f sy 0:85f su

The prestressing wire used is a 6-gage, Class III wire, with an


ultimate strength, fu, of 1737.5 MPa (252 ksi). The Modulus of Elasticity of the wire, Es, after wrapping at fsg, for stress levels below fsg
is taken as 193,050 MPa (28,000 ksi). The stressstrain relationship
for the prestressing wire, after wrapping at fsg, is given in Eq. (8).

f s e s Es

es 6 f sg =Es
for es > f sg =Es
for

f s f su 1  1  0:6133es Es =f su 

2:25

where es is strain in prestressing wire.


The yield strength of the steel cylinder in tension is taken as
227 MPa (33,000 psi), and the strength of the steel cylinder at failure is taken as 310 MPa (45,000 psi). Density and Poissons Ratio of
the prestressing wire and steel cylinder used in the model are
shown in Table 2. The compressive strength of the mortar coating,
based on AWWA C304 standard, was 37.9 MPa (5500 psi) (AWWA
C304-2007).
5. Results and discussion
The effects of the number and location of broken wire wraps on
the failure of buried PCCP are evaluated by monitoring the stresses
and strains developed in the various components of the pipe. This
is done while modeling PCCP with full interaction at the harnessed
joint and considering the external loading around the pipe using
the Marston Theory. Three important pipe failure indicators are
considered: the onset of micro and visible cracking in the concrete
core, onset of micro and visible cracking in the mortar coating, and
the yielding and rupturing of the prestressing wire wraps.
5.1. Cracking of concrete core
The onset of micro and visible cracking in the concrete core of a
PCCP is the first indication of deterioration and the eventual failure
of the element. Micro cracking of the mortar coating or concrete
core is defined by the strain associated with a crack that is
P0.001 in. wide, and visible cracking of the mortar coating or concrete core is defined by the strain associated with a crack that is
P0.002 in. wide. Micro cracks appear in the concrete at a hoop
strain of 2.0E4, while visible cracks appear at a hoop strain of
1.46E2. Once concrete begins to crack in tension, its load carrying
capacity decreases with additional strain. The hoop strain developed in the inside concrete core of the damaged ECP at the barrel
of the pipe was compared to the strain developed at the spigot and
bell ends. Fig. 5 shows this comparison for a 244 cm (96-in.) diameter Class 125-14 ECP with five (5) broken wire wraps. The results
indicate that visible cracking in the concrete core in barrel of the

1.60E-03

pipe occurs at 1379 kPa (200 psi), while the visible cracking at
the spigot end and bell ends of the pipe occur at 1393 kPa
(202 psi) and 1482 (215 psi), respectively.
The results show that, under the same level of damage, the hoop
strain inside the concrete core at the invert of a Class 125-14 PCCP
reaches the limits of micro cracking (initiation of deterioration) at
lower internal fluid pressures when the damage occurs at the spigot end, compared with pipes having wire breaks at the barrel and
bell end of the pipe. However, the model shows a faster rate of
deterioration at the barrel of the pipe as the internal fluid pressure
increases, thus resulting in visible cracks occurring at the barrel of
the pipe prior to appearing at the joint. A possible explanation of
this behavior is the confining effect that the harness clamp imposes
on the hoop strain at the joints, thus preventing the rapid rate of
hoop strain increase.
5.2. Cracking of mortar coating
Similarly, the onset of visible cracking in the mortar coating in
the barrel, spigot end, and bell end of the PCCP with five (5) broken
wire wraps occurred at 1517 kPa (220 psi), 1593 kPa (231 psi), and
1655 kPa (240 psi) internal fluid pressure, respectively as shown in
Fig. 6. Also, the onset of micro cracking in mortar coating in the
barrel, spigot end, and bell end of the pipe with five (5) broken wire
wraps occurred at 1437 kPa (208 psi), 1530 kPa (221 psi), and
1600 kPa (232 psi) internal fluid pressure, respectively. The results
of this second failure indicator show thatunder the same level of
damage in PCCPboth micro and visible cracks would occur at the
barrel of the pipe sooner (at lower internal pressure) when compared to the spigot and bell ends of the same PCCP. This difference
from the behavior of the hoop strain in the concrete core can be
attributed to the fact that the harness does not provide the same
confinement effect at the joint on the mortar coating that it does
on the concrete core.
5.3. Yield and rupture of prestressing wires
As the internal fluid pressure increases inside a damaged PCCP
with broken prestressed wire wraps from corrosion or other causes
listed earlier, the stress level in undamaged wires increases beyond
the yield limit. Additional internal pressure will eventually cause
the stress level in the remaining prestressed wire wraps to go over
the ultimate strength limit, thus leading to fracture of additional
wires. This third failure indicator is studied by monitoring the
stresses in the prestressing wires and determining the internal
fluid pressures which will cause them to reach their yield and ultimate strength limits. The impact of increased damage through a
growing number of broken prestressing wire wraps on these failure stresses is also analyzed, and the results are summarized in
Table 3. Figs. 7 and 8 show the stresses in the prestressing wires

Visible Cracking

1.40E-03

1.60E-03
Barrel

1.20E-03

1.40E-03

6.00E-04

Strain

Strain

Spigot

8.00E-04

Bell

4.00E-04

Visible Cracking

1.00E-03

Micro Cracking

8.00E-04

Spigot

6.00E-04
Bell

4.00E-04

Micro Cracking

2.00E-04
0.00E+00

Barrel

1.20E-03

1.00E-03

2.00E-04
0.00E+00
0

200

400

600

800

1000

1200

1400

1600

Internal Pressure (kPa)


Fig. 5. Hoop strain in concrete core of PCCP with five broken wire wraps.

200

400

600

800

1000 1200 1400 1600 1800 2000

Internal Pressure (psi)


Fig. 6. Strain in mortar coating of PCCP with five broken wire wraps.

17

M. Hajali et al. / Tunnelling and Underground Space Technology 51 (2016) 1119


Table 3
Internal pressure (kPa) required to yield or rupture the prestressing wire wraps.
Yield
Barrel
5 WB
35 WB
70 WB
100 WB

1900
1320
920
720

Rupture
Spigot
2110
1410
970
730

Bell
2218
1850
1135
895

Barrel

19%

1.60E+06
Spigot

2300
1490
1000
755

Strength Limit State

1.70E+06

2380
1650
1120
900

2500
2150
1500
1100

Yield Limit State

1.50E+06

Bell

Stress (kPa)

No. of broken wire wraps

1.80E+06

1.40E+06
1.30E+06

Spigot

Bell

1.20E+06
1.10E+06

Barrel

1.00E+06
9.00E+05
8.00E+05

Strength Limit State

200

400

15%

600

800

1000

1200

1400

Internal Pressure (kPa)


Fig. 9. Stress in prestressing wires with seventy (70) broken wire wraps.

Yield Limit State

Barrel
Bell

Stress (kPa)

Spigot

Fig. 7. Stress in prestressing wires with five (5) broken wire wraps.

1.80E+06
1.70E+06
1.60E+06
1.50E+06
1.40E+06
1.30E+06
1.20E+06
1.10E+06
1.00E+06
9.00E+05
8.00E+05

20%

Strength Limit State

Yield Limit State


Spigot

Bell
Barrel

200

400

600

800

1000

1200

Internal Pressure (kPa)


1.80E+06

Strength Limit State

Fig. 10. Stress in prestressing wires with one hundred (100) broken wire wraps.

29%

1.70E+06
1.60E+06
Yield Limit State

1.80E+06 Strength Limit

1.40E+06

62%

Barrel

1.30E+06

1.60E+06

Spigot

1.20E+06
1.10E+06

Stress (kPa)

Stress (kPa)

1.50E+06

Bell

1.00E+06
9.00E+05
8.00E+05
0

500

1000

1500

2000

2500

Internal Pressure (kPa)


Fig. 8. Stress in prestressing wires with thirty-five (35) broken wire wraps.

Yield Limit
70WB

1.40E+06
100WB

35WB

1.20E+06

5WB

1.00E+06
8.00E+05

500

1000

1500

2000

2500

Internal Pressure (kPa)


for PCCP with five (5) and thirty-five (35) broken wire wraps,
respectively. Figs. 9 and 10 show the stress in the prestressing
wires for PCCP with seventy (70) and hundred (100) broken wire
wraps, respectively. Figs. 710 compare the stresses in the prestressing wires when the damage occurs at various locations of
the PCCP: spigot end, bell end, and barrel.
It is clear from Figs. 710 that when considering harnessjointed ECP with full joint interaction, the stress in the prestressing
wire wraps reaches the yield limit at lower internal fluid pressures
when the wire breaks are at the barrel than when wire breaks are
at the spigot or bell ends of the pipe. The spigot and bell locations
do not have the same effect on the yield pressure and the ultimate
pressure of the PCCP. Fig. 7 shows that for 5 wire breaks, yielding of
the prestressing wires occurs at an internal fluid pressure of
2110 kPa (306 psi) when the damage is at the spigot end and at
2218 kPa (322 psi) when the damage is at the bell end, compared
to only 1900 kPa (276 psi) when the damage is at barrel of the pipe.
In effect, the internal fluid pressure that will result in yielding of
the prestressed wires is almost 15% lower when the defect is in

Fig. 11. Stress in the prestressing wirebroken wire wraps at the barrel.

the middle of the pipe than when the defect is near the spigot
end. When considering the complete rupture of the prestressing
wires, this same difference becomes 8.0%. It can also be argued
here that the confinement effect that the harness provides at the
joint is the main reason for this delayed yielding of the prestressing
wires near the joint when compared to those at the barrel of the
pipe.
Fig. 8 shows that the prestressing wire wraps of a PCCP with 35
wire breaks reach their yield stress limit state at an internal fluid
pressure that is as much as 29% lower when the defect is in the barrel
than when it is at the spigot end of the pipe. This difference in internal pressure continues to diverge due to a higher rate of deterioration at the barrel, until it reaches about 31% at the ultimate
(rupture) stress. The same can be said about PCCP with 70 and 100
wire breaks where the yield-inducing internal pressures are 19%
and 20% lower at the barrel, respectively, and the rupture-causing

18

M. Hajali et al. / Tunnelling and Underground Space Technology 51 (2016) 1119

1.80E+06

Strength Limit
60%

1.60E+06

Stress (kPa)

Yield Limit

1.40E+06

100WB

70WB
5WB

1.20E+06

35WB

1.00E+06
8.00E+05

500

1000

1500

2000

2500

Internal Pressure (kPa)


Fig. 12. Stress in the prestressing wirebroken wire wraps at the bell end.

1.80E+06 Strength Limit


65%

1.60E+06

Stress (kPa)

Yield Limit

1.40E+06
100WB

70WB
5WB

1.20E+06

35WB

1.00E+06
8.00E+05

500

1000

1500

2000

2500

Internal Pressure (kPa)


Fig. 13. Stress in the prestressing wirebroken wire wraps at the spigot end.

internal pressures are 34% and 32% lower at the barrel, as shown in
Figs. 9 and 10. For 35 broken wire wraps, as can be seen from Fig. 8
and Table 3, the yielding of the wires occurs at 1320 kPa (191 psi)
for the barrel location, 1490 kPa (216 psi) for the spigot location
and 1850 kPa for the bell location, which is 29% lower when the
defect is at the barrel but only 6% when the defect is at the spigot.
For 35, 70 and 100 broken wire wraps, the barrel location and the
spigot location give quite the same internal pressure at yield.
It can also be seen from the numbers of Table 3 and plots of
Figs. 710 that these internal pressures that cause yielding and
the eventual rupture of the wire wraps drop as the damage in
the PCCP worsens and the number of respective wire breaks
increase from 5 to 100. Figs. 1113 therefore, comparefor each
location of defecthow the internal pressure that causes failure
of the prestressing wires varies with the length of the defect, which
corresponds to a different number of broken wire wraps. When the
wire breaks occur at the barrel of a PCCP, the prestressing wire
wraps reach their yield stress limit at internal fluid pressures of
1900 kPa (276 psi), 1320 kPa (191 psi), 920 kPa (133 psi), and

720 kPa (100 psi) for 5, 35, 70, and 100 wire breaks, respectively.
Similarly, when the damage is at the barrel, the onset of rupture
for the prestressing wire wraps occurs at internal fluid pressures
of 2300 kPa (334 psi), 1490 kPa (216 psi), 1000 kPa (145 psi), and
755 kPa (110 psi) for 5, 35, 70, and 100 wire breaks, respectively.
These translate to a 62.2% and 67.2% drop in the internal fluid pressures required to both yield and rupture the prestressing wires, as
the defect worsens from 5 to 100 wire breaks at the barrel of a
PCCP. Similar numbers are obtained for when the wire breaks
occur at the bell end and the spigot end of the PCCP, as shown in
Figs. 11 and 12, respectively.
Fig. 14 shows, as an example, the stress developed in the prestressing wire for a PCCP with thirty-five (35) broken wire wraps
at the spigot end of the pipe. Von Mises stress is used since it is
a good representative of the hoop stress developed circumferentially around the pipe (rH) and longitudinal or axial stress in the
pipe (rL). The color gradient indicates the calculated range of stress
for each element in the pipe model. Note that in Fig. 14 the stress is
reported in imperial units. It is interesting to observe from the
results that the highest amount of stress occurs near the location
of the damage. This is expected since the breakage of wire wraps
in a particular region of the pipe will result in more stress concentration on the remaining undamaged wire wraps in the vicinity of
the damage.
The external load was considered, based on a real case study, to
consist of 4.3 m (14 feet) of soil. Changing the height of earth cover
will change the internal pressure required to: (a) yield or rupture
the steel cylinder and prestressing wires, and (b) cause visible
cracking and micro cracking in the concrete core and mortar coating; but will not significantly change the difference between the
results of the bell, the spigot, and the barrel.

6. Conclusions and recommendations


In this paper, the effect of broken wire wraps on the structural
performance of a damaged PCCP was investigated using an
advanced three-dimensional numerical response model that
accounted for full interaction between the spigot and bell ends of
adjacent pipes and subjected the pipe to both internal fluid and
external soil loads. A sensitivity analysis was presented to give a
better understanding of how the number and respective location
of broken prestressing wire wraps affect the failure limit of the
pipe. Three different locations of defects were considered: spigot
joint, bell joint, and barrel; as well as four different numbers of
broken wire wraps that varied between 5 and 100. The failure of
the pipe was investigated by examining three critical indicators:
cracking of the concrete core, cracking of the mortar coating, and
yielding of prestressed wire wraps.
The results showed that wire wrap breaks at the barrel of the
pipe decrease the overall pressure capacity of PCCP more so than

Fig. 14. Von Mises stress (psi) in prestressing wires for PCCP with 35 WB at spigot end.

M. Hajali et al. / Tunnelling and Underground Space Technology 51 (2016) 1119

wire breaks located at the joints. The results indicate over 30%
strength reduction for a low to medium number of wire wrap
breaks occurring at the barrel when compared to the bell joint. This
strength reduction increases over 33% for medium to high number
of wire breaks. Additionally, the model results showed that if wire
wrap breaks occur at the barrel, it is anticipated that cracking in
the pipe will occur at lower internal pressures than if the breaks
occurred at the joints. These internal pressure limits can also drop
by as much as 65% with increased number of wire breaks. It is
therefore argued that the location and not just the extent of damage have a significant effect on the structural performance of PCCP.
The conclusions arrived at are significant for owners, operators,
and inspectors of underground PCCP pipeline systems alike. The
results presented will help these stakeholders develop risk models
to better access the risk of failure that certain pipe defects pose.
This in-turn will serve the purpose of more efficient and on-time
replacement of these facilities. While the conclusions arrived at
in this study can be expected to remain valid for other classes of
PCCP, the results remain specific to Class 125-14 96-in. PCCP and
would not extend to cover the entire range of PCCP classes without
proper validation studies. Furthermore, while several issues
regarding the structural performance of damaged PCCP with harnessed joints were discussed in this paper, there are many open
questions and issues that need to be researched. Future research
into the structural performance and monitoring of PCCP should
include full-scale experimental models of PCCP as well as developing risk curves for these infrastructure systems. Comparing risk
models developed based on numerical simulation as the one presented here to ones based on ANN.
References
Abi Shdid, C., Hajali, M., 2014. Detecting location of construction defects in drilled
shafts using frequency tomography analysis of cross-hole sonic logging. In:
Proceedings of the 2014 International Conference of Computing in Civil and
Building Engineering (ICCCBE), ASCE, Orlando, FL, pp. 17231730.
Alavinasab, A., Hajali, M., 2014. Do broken prestressing wire wraps at the joint affect
the safe operation of a pipe? In: Pipelines 2014, pp. 13791391.
Alavinasab, A., Padewski, E., Holley, M., Jha, R., Ahmadi, G., 2010. Damage
identification based on vibration response of prestressed concrete pipes. In:
Proceedings of Pipelines Conference: Climbing New Peaks to Infrastructure
ReliabilityRenew, Rehab, and Reinvest, ASCE, Keystone, CO, pp. 909919.
Alavinasab, A., Padewski, E., Higgins, M., Holley, M., Jha, R., Ahmadi, G., 2011. Crack
propagation in prestressed concrete noncylinder pipe using finite element
method. In: Proceedings Pipelines 2011: A Sound Conduit for Sharing Solutions,
ASCE, Seattle, WA, pp. 5564.
Alavinasab, A., Jha, R., Ahmadi, G., 2011. Damage identification based on modal
analysis of prestressed concrete pipes. In: Proceedings of Pipelines 2011: A
Sound Conduit for Sharing Solutions, ASCE, Seattle, WA, pp. 1223.

19

Alavinasab, A., Padewski, E., Higgins, M., 2013. Effects of the location of broken
prestressing wires wraps in structural integrity of a damaged PCCP. In:
Proceedings of Pipelines 2013, ASCE, Fort Worth, TX, pp. 767774.
Allouche, E., Alam, S., Simicevic, J., Sterling, R., Condit, W., Matthews, J., Selvakumar,
A., 2014. A pilot study for retrospective evaluation of cured-in-place pipe (CIPP)
rehabilitation of municipal gravity sewers. Tunn. Undergr. Space Technol. 39,
8293.
Amaitik, Nasser M., Amaitik, Saleh M., 2008. Development of PCCP wire breaks
prediction model using artificial neural networks. In: Proceedings of Pipelines
2008, ASCE, Atlanta, GA, pp. 111.
Arsnio, A.M., Dheenathayalan, P., Hanssen, R., Vreeburg, J., Rietveld, L., 2014. Pipe
failure predictions in drinking water systems using satellite observations.
Struct. Inf. Eng., 110 (ahead-of-print)
American Water Works Association (AWWA), 2007. AWWA C301 Standard for
Prestressed Concrete Pressure Pipe, Steel-Cylinder Type, Denver, CO.
American Water Works Association (AWWA), 2007. AWWA C304 Design of
Prestressed Concrete Cylinder Pipe, Denver, CO.
AWWA, 2007. AWWA C-304 Standard for Design of Prestressed Concrete Cylinder
Pipe, Denver, CO.
AWWA Research Foundation, 2007. Failure of Prestressed Concrete Cylinder Pipe,
Denver, CO.
Hajali, M., Abi Shdid, C., 2013. Determination of the two-dimensional plastic zone
shape and SIF at the crack tip using RKPM. J. Iron Steel Res. 20 (12), 103114,
Elsevier.
Hewayde, E., Nehdi, M., Allouche, E., Nakhla, G., 2007. Neural network prediction of
concrete degradation by sulphuric acid attack. Struct. Inf. Eng. 3 (1), 1727.
Jung, J.K., Koo, D.H., Zhang, K., 2014. Verification of the pipe depth dependent model
using a finite element analysis. Tunn. Undergr. Space Technol. 39, 3440.
Kang, J.S., Davidson, J.S., 2013. Structural effects of concrete lining for concrete-lined
corrugated steel pipes. Struct. Inf. Eng. 9 (2), 130140.
Kleiner, Y., Sadiq, R., Rajani, B., 2004. Modeling failure risk in buried pipes using
fuzzy Markov deterioration process. In: Proceedings of International Conference
on Pipeline Engineering and Construction, ASCE, San Diego, CA, pp. 112.
Mahendran, M., 2007. Applications of finite element analysis in structural
engineering. In: Proceedings of International Conference on Computer Aided
Engineering, Chennai, India, pp. 3846.
Mayer, M.H., Gaul, L., 2005. Segment-to-segment contact elements for modelling
joint interfaces in finite element analysis. Mech. Syst. Signal Process. 21 (2),
724734.
Moser, Alma P., Folkman, Steven L., 2001. Buried Pipe Design. McGraw-Hill, New
York.
Ovesy, H.R., Asghari, M.M., Kharazi, M., 2015. Post-buckling analysis of delaminated
composite plates using a novel layerwise theory. Thin-Walled Struct. 94, 98
106.
Rauniyar, S., 2013. Development of Finite Element Model for Analysis of Prestressed
Concrete Cylinder Pipe-embedded Cylinder Pipe with Experimental
Comparison. M.S. thesis. The University of Texas at Arlington, Arlington, TX.
Xiong, H., Li, P., Lia, Q., 2010. FE model for simulating wire-wrapping during
prestressing of an embedded prestressed concrete cylinder pipe. Simul. Model.
Pract. Theor. 18 (5), 624636.
Yang, Y., Polak, M.A., Cascante, G., 2010. Nondestructive evaluation of the depth of
surface-breaking cracks in concrete pipes. Tunn. Undergr. Space Technol. 25 (6),
736744.
Younis, R., Knight, M.A., 2010. A probability model for investigating the trend of
structural deterioration of wastewater pipelines. Tunn. Undergr. Space Technol.
25 (6), 670680.