The Structural Performance of Flexible Pipes
The Structural Performance of Flexible Pipes
The Structural Performance of Flexible Pipes
2002
Recommended Citation
Suleiman, Muhannad Taher, "The structural performance of flexible pipes " (2002). Retrospective Theses and Dissertations. 1033.
https://lib.dr.iastate.edu/rtd/1033
This Dissertation is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University
Digital Repository. It has been accepted for inclusion in Retrospective Theses and Dissertations by an authorized administrator of Iowa State University
Digital Repository. For more information, please contact digirep@iastate.edu.
INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI films
the text directly from the original or copy submitted. Thus, some thesis and
dissertation copies are in typewriter face, while others may be from any type of
computer printer.
In the unlikely event that the author did not send UMI a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
by
DOCTOR OF PHILOSOPHY
Ames, Iowa
2002
UMI Number: 3061868
UMI6
UMI Microform 3061868
Copyright 2002 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
Graduate College
Iowa State University
major Professor
Signature was redacted for privacy.
Co-major Professor
Signature was redacted for privacy.
For the Ma r Program
iii
TABLE OF CONTENTS
LIST OF FIGURES
LIST OF TABLES
ABSTRACT
1. GENERAL INTRODUCTION
1.1 Objectives and Scope
1.2 Literature Review
4. CONCLUSIONS
5. INTRODUCTION
5.1 General Background
5.2 Objectives and Scope
5.3 Modeling
7. CONCLUSIONS 91
9. CONCLUSIONS 107
REFERENCES 138
ACKNOWLEDGMENTS 143
VI
LIST OF FIGURES
Figure 1.14 Parallel plate test for 450 mm HDPE pipe (Sargand et al. 1998). 29
Figure 1.15 Parallel plate results for 450 mm PVC pipe (Sargand et al. 1998). 29
Figure 1.16 Stress-strain results of compression test (Zhange and Moore 1997). 30
Figure 1.17 Results of stress relaxation test on PVC pipe (Sargand et al 1998). 31
Figure 1.18 Results of stress relaxation test on HDPE pipe, (Sargand et al 1998). 32
Figure 1.19 Creep test results subtracting the instantaneous strain (Zhange and
Moore 1997). 34
Figure 1.20 Trench cross section showing the terminology used by ASTM 02321. 39
Figure 2.3 MN. DOT specifications for plastic pipe storm sewer installations. 48
Figure 3.7 Pipe wall deflection between "1 and 3 o'clock" in a 15 inch diameter
pipe. 54
Figure 3.10 Heaving of 48 inch diameter pipe to above the ground surface. 59
Figure 6.1 Plastic pipe crown deflection with respect to height of soil cover for
elastic soil. 73
Figure 6.2 Plastic pipe crown vertical deflection with respect to position on pipe
circumference for different height of soil covers for the case of elastic soil. 74
viii
Figure 6.3 Vertical stress in a soil column above the pipe for 4.6 m soil height. 76
Figure 6.4 Horizontal stress vs. horizontal distance from pipe spring line for 4.6 m
soil height. 76
Figure 6.5 PE pipe vertical deflection vs. height of soil cover for elastic soil case. 77
Figure 6.6 PE pipe vertical deflection percent vs. height of soil cover for elastic
soil case. 77
Figure 6.9 Plastic pipe vertical deflection using hyperbolic tangent soil modulus and
power bulk soil modulus for both ANSYS and CANDE. 82
Figure 6.10 Plastic pipe vertical deflection percent using hyperbolic tangent soil
modulus and power bulk modulus for both ANSYS and CANDE. 82
Figure 6.11 PE pipe vertical deflection using hyperbolic tangent and bulk soil moduli
models for both ANSYS and CANDE. 83
Figure 6.12 PE pipe deflection percent using hyperbolic tangent and bulk soil moduli
for both ANSYS and CANDE. 83
Figure 6.13 PE vertical deflection percent using hyperbolic tangent soil modulus with
power bulk modulus for both ANSYS and CANDE. 84
Figure 6.14 Finite element mesh used to model Utah State University soil-pipe cell. 86
Figure 6.15 Vertical deflection percent with respect to soil cover for PE pipe
compared with Moser (1994). 86
Figure 6.16 Vertical deflection percent with respect to soil cover above spring line. 87
Figure 6.17 Deflection of 1200 mm diameter PE pipe vs. construction steps using
both ANSYS and CANDE. 87
Figure 6.18 Vertical deflection percent for 1200 mm PE pipe vs. construction step
for ANSYS and CANDE. 90
Figure 8.1 Normalized stress strain compression test results on HDPE material. 97
Figure 8.3 Change of normalized stress-strain lines slope with strain rate. 98
Figure 8.4 Intersection of normalized stress-strain lines for 10"1 and 10"2 /sec
strain rates. 99
Figure 8.6 Change of slope of normalized stress-strain lines with strain rate using
the focus point. 100
Figure 8.7 Change of HDPE tangent modulus with strain for five different strain rates. 101
Figure 8.9 Normalized time vs. creep strain relationship for six different stress levels. 102
Figure 8.10 Normalized creep strain curves using the focus point. 103
Figure 8.11 Change of the slope of focus point lines with stress level for creep tests. 103
Figure 8.12 Vertical deflection percent with respect to soil cover using constant and
strain rate dependent HDPE modulus. 106
Figure B-5 Joint separation of 0.4 inch at 170 feet from MH 201. 119
Figure B-6 Two rods penetrating the pipe wall at 133 feet from MH 201. 119
Figure B-7 Schematic diagram of the pipe surveyed at Xylite Street, Blaine. 120
Figure B-8 Schematic diagram of the pipe surveyed at Ottawa Street, Le-Center. 122
X
Figure B-9 Deflection in 30 inch diameter pipe at a joint 227 feet from MH 3. 122
Figure B-10 Joint separation and water infiltration due to deformation of the
piece connecting two pipes. 123
Figure B-l3 Schematic diagram of the pipe surveyed near Interlaken Road,
Fairmont. 125
Figure B-17 Schematic diagram of the pipe surveyed at Railroad Avenue, Gaylord. 127
Figure B-20 Schematic diagram of the pipe surveyed at Jonquil Lane, Maple Grove. 129
Figure B-21 Schematic diagram of pipe 1 surveyed at 77th Place Lane, Maple Grove. 130
Figure B-22 Schematic diagram of pipe 2 surveyed at 77th Place Lane, Maple Grove. 130
Figure B-23 Schematic diagram of the pipe surveyed 96th Place North,
Maple Grove. 132
Figure B-24 Pipe wall deflection in the 15 inch diameter pipe. 133
Figure B-25 Vertical joint offset in the 24 inch diameter pipe. 133
xi
LIST OF TABLES
Table 1.1 Summary of the buckling results of the loading tests (Moser 1998). 17
Table 1.3 Design values of E' (psi) after Hartley and Duncan (1988). 21
Table 1.4 Pipe stiffness of parallel plate test at two different loading rates. 30
Table 1.5 Pipe stiffness results of stress relaxation tests of parallel plate test. 32
Table 1.7 Recommendation for installation and use of soils and aggregate for pipe
foundation embedment, and backfill (ASTM D2321 ). 40
Table 6.1 Pipe and elastic soil properties used in the analysis. 72
Table 8.1 HDPE modulus change with stress level and time. 104
Table B.1 Performance evaluation for the sites surveyed in July, 2000. 135
Table B.2 Performance evaluation for the sites surveyed December, 2000 and
January, 2001. 136
Table B.3 Performance evaluation for the sites surveyed March, 2001. 137
xii
ABSTRACT
High density polyethylene pipes have been used for over a decade but little
information is available on the in-service behavior of these pipes. At the request of the
HDPE was undertaken. The objective of the first part of this dissertation was to investigate
the field performance of in-service HDPE pipes using visual information obtained from a
remote, motorized video camera. Ten projects with a total length of 12,006 feet were
were cross sectional deformations, wall buckling, wall crushing, wall cracking, joint
separation, and sediments. Few major structural problems were noticed due to the effect of
CANDE is one of the commonly used programs for buried pipe analysis; however,
the limitations of CANDE include application only to small deflections, and neglect the time
effects. The recent tendency of using thermoplastic pipes for deep applications, which
increased the need for investigating the effect of large deflections, and the dependency of the
properties of such pipes on strain rate and time led to the use of ANSYS. The main advantage
of CANDE relative to ANSYS is the use of the nonlinear soil models while ANSYS has the
language was written to model the soil behavior using hyperbolic tangent modulus with both
power and hyperbolic bulk modulus models. CANDE and the small and large deflection
theories of ANSYS were compared with Moser's (1994) results. This comparison showed
that CANDE over-predicts the pipe deflections as the soil approaches the shear failure and
that ANSYS better describes the pipe behavior. CANDE can be used as long as the shear
failure of the soil was not reached. Considering large deflections for flexible pipe analysis
Since the parallel plate test deflection rate is not related to practical loading rates and
the time effect is not considered, new mathematical constitutive models were developed for
HDPE material using the compression tests results performed by Zhang and Moore (1997).
These constitutive models consider the effect of strain rate and time on the HDPE modulus.
These mathematical models were programmed in the code written in ANSYS. A finite
element analysis was used to validate the use of the programmed equations and to compare
the pipe deflection using small deflection theory with linear elastic HDPE modulus and large
deflection using strain rate dependent HDPE modulus for the case of SM soil. This
importance of considering large deflection and strain rate pipe modulus for deeply buried
HDPE pipes.
1
I. GENERAL INTRODUCTION
The broad objective of this research is to improve and expand knowledge of the
geostatic load response of high-density polyethylene (HDPE) pipe. Although this relatively
new construction material has been used for over a decade, few data are available on its field
performance. Also, current numerical methods of analyses are limited to small deflections,
and existing constitutive models for the plastic pipe assume elastic behavior and ignore time
effects. To address these issues, this study includes three separate components, each with its
own specific objective: 1) evaluation of the field performance of in-service HDPE pipe, 2)
expansion of finite element method (FEM) analyses of soil-HDPE pipe system to include
application of ANSYS with regard to large deflections and 3) improvement of HDPE pipe
constitutive models to include strain rate and creep effects. This tripartite approach has
resulted in a dissertation that is organized in three distinct and somewhat independent parts.
Part I of this the dissertation deals with the field performance of HDPE pipe. At the
request of the Minnesota Concrete Pipe Association (MCPA) an investigation to evaluate the
field performance of HDPE was undertaken. One objective of this study is to investigate the
field performance of in-service HDPE pipes using visual information obtained from a
remote, motorized video camera. Ten projects were surveyed in Minnesota between July,
2000 and March, 2001. The performance characteristics that were considered in this
evaluation were cross section deformations, wall buckling, wall crushing, wall cracking, joint
separation, and sediments. These characteristics and their locations along the length of the
The finite element analysis of buried flexible pipes is presented in Part II of this
thesis. The recent trend of using the plastic pipes for deep applications increased the need for
investigating the effect of neglecting large deflections. Culvert ANalysis and DEsign
(CANDE) is one of the most commonly used programs for buried pipe analysis; however, the
limitations of CANDE include neglecting the time effect and its applicability only to small
deflections. These limitations lead to the consideration of another analysis tool. In this study
ANSYS, a general finite element program, was used to model the soil-pipe system. The main
advantage of CANDE relative to ANSYS is the use of hyperbolic and power soil models,
while ANSYS has the advantage of modeling large deflections. One objective of the
theoretical analyses is to compare the results of CANDE with small and large deflection
theories of ANSYS for elastic soil case with geostatic applied loads. A second objective is to
write a code using ANSYS programming language to model the soil behavior using
hyperbolic tangent modulus with both power and hyperbolic bulk soil modulus models.
Using this code, the effect of large deflections on the behavior of polyethylene (PE) and
Polyvinyl Chloride (PVC) flexible pipes is investigated. The written code was also improved
Part III of the dissertation addresses the constitutive models of HDPE material. The
objective of Part III is to develop mathematical models to describe the tangent modulus of
HDPE material considering the effect of strain rate and time. These mathematical models
were also programmed and included in the code written in ANSYS. The mathematical
models, their derivation, ANSYS programming, and a finite element case study on the effect
Nearly all pipes can be classified as either flexible or rigid, depending on how they
perform when installed. Flexible pipes take advantage of their ability to move, or deflect,
under loads without structural damage. Common types of flexible pipes are manufactured
from polyethylene, polyvinyl chloride, steel, and aluminum. Rigid pipes are classified as
pipes that can not deflect more than 2% without significant structural distress such as
cracking.
Both flexible and rigid pipes require a proper backfill to allow the load transfer from
the pipe to the soil, although the pipe backfill interaction differs. When a flexible pipe
deflects against the backfill, the load is transferred to and carried by the backfill. When loads
are applied to rigid pipes, on the other hand, the load is transferred through the pipe wall into
In this section, the soil structure interaction, pipe performance limits, pipe
performance parameters, plastic pipe installation, and previous analytical and experimental
Marston (1913) published his original theory (now known as Marston's load theory)
about how to determine the vertical loads carried by a ditch conduit. According to this
theory, the backfill material tends to consolidate and settle downward relative to the native
soil around the ditch. This relative movement creates shear (friction) acting upward. Based
on the free body diagram of a ditch conduit shown in Figure 1.1, considering a thin
horizontal soil element of a thickness dh within the ditch, the forces acting on the soil
element are the vertical load at the upper boundary (V), the vertical load at the lower
boundary (V+dV), and the side friction forces. Marston neglected the effect of cohesion
between the backfill material and the native soil and assumed that the ratio of active earth
pressure is applied to calculate the horizontal stresses. The firictional force is calculated using
the coefficient of sliding friction between the fill and the native soil (g).
The value of the vertical force V at any depth can be calculated by solving the
equilibrium equation of vertical forces shown in Figure 1.1. This value of the vertical force V
1 _ p ~1WB
v = r-B-^r (ID
Marston (1930) formulated the differential equations for other types of pipe
installations. For shallow embankments, depending on the relative movement between the
soil column above the pipe (interior prism) and the embankment material around it (exterior
prism), the shear forces can act downward or upward. If the exterior prism settles relative to
the interior prism, the firictional forces on the interior prism will act downward and the load
on the conduit is greater than the prism load (projection condition). On the other hand, if the
interior prism settles relative to the exterior one, the frictional forces will be directed upward
and the load applied on the pipe is less than that of the prism load (ditch condition). A neutral
condition can also be considered where the top of the conduit settles the same amount as the
exterior prism. In this case the load applied on the conduit is the prism load. Considering the
direction of the frictional forces, the derivation of the equation of loads due to embankment
installations is the same as shown in Figure 1.1. The value of the vertical force V at any
5
depth in an embankment installation is given by Equation 1.2. The positive sign is used for
projection conditions while the negative sign is used for ditch conditions.
„±2Kflh/B
v = y.D (1.2)
±2 Kfi
£ Ground Surface
7/
V
• _y_
dh * Ww AKgdhV/B
T
V+dV
Notes:
o
B : Trench width.
D : Pipe Diameter.
W : Weight of soil element= y.dh.
K : Ratio of active earth pressure.
The horizontal stress acting on the sides of the backfill is K V/B.
The horizontal force acting on a soil element with height dh is K Vdh/B.
The frictional vertical force acting on a soil element with height of dh is K n dh V/B.
The weight of soil element with height of dh is yBdh.
Equation 1.1 shows that as the trench width increases, the vertical load V increases.
This is true as long as the calculated vertical load is less than that calculated by Equation 1.2
for the projecting conduit case. The width of the trench at which the calculated vertical loads
using both equations, 1.1 and 1.2, are equal is called the transition width.
If the height of the fill is sufficiently increased, the shear stresses do not extend to the
top of the embankment but cease at a horizontal plane within the fill. This plane is called the
plane of equal settlement. In this case the load carried by the conduit depends on the prism
load between the top of the conduit (pipe crown) and the plane of equal settlement. Equation
1.3 shows the value of the vertical force as a function of the height of the plane of equal
settlement (He). The positive sign is used in case of projection condition, while the negative
HlKiihlB , TT tl
K= <,3>
forces around a buried flexible pipe as shown in Figure 1.2. This hypothesis is based on the
elastic ring theory and the experimental work performed on metal flexible pipes at Iowa
theory and assumed that the load is uniformly distributed over the bedding width of the pipe
which is equal to the applied vertical load, and 2) the passive horizontal pressure on the pipe
sides is distributed parabolically over 100° and the maximum value of pressure is equal to the
modulus of passive resistance of the side-fill material (e) multiplied by half the horizontal
deflection. This stress distribution was used to derive the original Iowa formula given in
Equation 1.4.
7
K Wcr
AX = (1.4)
£7 +0.061er4
where:
AX = Horizontal deflection or change in diameter, in.
Di = Deflection lag factor.
K = Bedding constant.
Wc = Marston load per unit length of pipe, lb/in.
r = Mean radius of the pipe, in.
E = Modulus of elasticity of the pipe material, lb/in2.
I = Moment of inertia of the pipe wall per unit length, in4/in.
e = Modulus of passive resistance of the side fill, lb/(in2)(in).
The modulus of passive resistance (e) was investigated by Watkins and Spangler
(1958). They noticed that it does not have the dimension of a modulus and thus could not
represent a soil property. The modulus of soil reaction (E') was derived as the product of the
modulus of passive resistance and the mean radius of the pipe. Substituting the E' into
Equation 1.4, a new formula called the "modified Iowa formula" was derived (Equation 1.5).
Marston's Load
(1.5)
EI + 0.061£>
where:
E' = Modulus of soil reaction, lb/in2.
The fact that buried flexible pipes continue to deflect over time under a constant load,
in the case of consolidated side-fills, led to the consideration of time dependent deflection. A
deflection lag factor, Dl, was introduced to magnify the short term deflection as shown in
Equation 1.6. The Dl value of 1.5 was found reasonable for long term deflection, but a value
of 1.0 was to be used in case the prism load theory was used to calculate the applied loads. A
more popular and practical form of this equation is shown in Equation 1.7 where the pipe
stiffness, PS, can be measured using ASTM D-2412 parallel plate test.
DLKWcr3
(1-6)
£/ + 0.061£>
AX _ lOOP^f
(1.7)
d 0.149(PS) + 0.061£'
where:
d = pipe diameter, in.
PS = pipe stiffness (lb./in./in.) = P/AY.
The problem with the soil stress distribution around flexible pipes is that the stiffness
of the pipe was assumed equal to the stiffness of the surrounding soil. Sargand et al. (1998)
stated that the stiffness of the profile wall plastic pipe is often (1/10) or less of the soil
stiffness, specially when installed in a dense granular material. So the uniformly distributed
vertical load along the horizontal surface passing through the pipe crown is questionable.
9
Moser (1990) indicated that the effective load on the flexible pipe lies somewhere between
Greenwood and Lang (1990) stated that the pressure distribution proposed by
Spangler may not apply to all pipes but it is widely accepted together with the Iowa formula.
Spangler and Handy (1982) noted that the Iowa formula has been derived for pipes made of
elastic materials such as steel, so it may or may not apply to plastic pipes, especially those
which experience creep and relaxation with time under constant loading.
Schluter and Caposselo (1999) compared the results of the modified Iowa formula
with the actual field performance for both very stiff (semi-rigid) and very flexible pipes. It
was shown that the modified Iowa formula excessively overestimates the deflection for very
stiff pipes and underestimates the deflection for very flexible pipes.
Watkins (2001) stated that Spangler used the theory of elasticity to derive the Iowa
Formula to predict ring deflections but it was not intended for design. Consequently, Iowa
formula is limited to elastic behavior for both the pipe and the soil. Therefore, it is limited to
the yield stress and small ring deflections. Watkins also stated that the performance limit of
flexible pipes is not limited to the elastic range and the ring deflection at the performance
limit is not small. Performance limit is in the plastic range and ring deflection based on the
yield stress should not be the performance limit for buried flexible pipe. Nevertheless, ring
deflection should be controlled because of its effect on the other performance limits.
Earth pressure exerted on a buried structure depends on the condition of the response
of the structure to the earth pressure. There are three lateral earth pressure conditions:
geostatic, active, and passive. Rankine's (1857) earth pressure coefficients can be used to
calculate horizontal stresses. When the buried pipe is geostatic, with zero horizontal
10
deflection, the soil at the sides will be in a state of elastic equilibrium. The side fill will be in
the active case when the horizontal diameter decreases and passive when the horizontal
diameter increases.
Moser (1990) stated that performance limits are usually established by the designer
based upon the pipe material and the required performance. He also suggested that the
performance limits could be divided into deflection, strain, stress, and buckling.
1.2.2.1 Deflection
According to Moser (1990) the performance limits of flexible pipes are usually
deflection related. Excessive deflection reduces the flow capacity and causes joint leakage.
Deflection can happen during transportation, construction, and due to imposed service loads.
Flexible pipes can deflect (vertical diameter shortening) due to applied loads and can elongate
(horizontal diameter shortening) as a result of the compaction process as shown in Figure 1.3.
Deflections can be controlled by the method of installation and the backfill type. The buried
pipe deflection should always be equal or less than the design deflection limit for a specific
product. Deflection is usually expressed as a vertical deflection percent defined as the change
in the vertical pipe diameter divided by the original pipe diameter. In a soil box, Spangler
observed excessive steel pipe ring deflection up to 20%, so he recommended, with a factor of
Spangler (1941) specified different stages of deflection for metal flexible pipes as
shown in Figure 1.4. When the load applied on the metal pipe increases, the pipe deflects into
the soil, and the passive soil pressure develops. If the load was increased and the soil was
well compacted, the flexible pipe flatten at the pipe crown starts to develop. The shape of the
11
flexible pipe in this stage is described as heart shaped. The pipe crown forms an upward
concave shape which result from increasing the load to greater levels.
Rogers et al. (1996) studied the influence of the installation procedure on pipe-soil
structure interaction for PVC pipes using a visual method. Laboratory installations were used
to model the field installation for two types of field practices which are called 'good site
practice' and 'poor site practice'. The 'good site practice' was performed to model and study
the effect of compaction on the pipe shape during construction. The 'poor site practice' was
performed to model and study the effect of different haunch support conditions on the pipe
deflection during construction. During the installation of PVC pipes, different cross-sectional
configurations were observed. These configurations are elliptical, heart shaped, inverted
Elongation
— Undeformed pipe
Deflection
technique to limit deflection and, thus, the strain. The strain limit is used to prevent strain
corrosion which is defined as an environmental degradation of the pipe material after the pipe
wall strain is greater than a threshold strain. Total circumferential strain consists of bending
strain, ring strain, and strain due to Poisson's ratio as defined in Equations 1.10, 1.11, and
1.12.
£* = 6(—)(-^-) (1.10)
(ui>
£v = -V.£, (1.12)
where:
Sb = Bending strain.
£c = Compression strain.
eu = Poisson's circumferential strain.
e/ = Longitudinal strain.
t = wall thickness, in.
D = Pipe diameter, in.
Ay = Vertical deflection, in.
P = Vertical soil pressure, psi.
E = Young's modulus of pipe material.
v = Poisson's ratio.
1.2.2.3 Buckling
Buckling phenomena may govern the design of the flexible pipes subjected to a high
applied loads or soil pressure. The more flexible the pipe, the less is its buckling resistance.
The buckling of flexible pipes does not only depend on the pipe material properties but also
on the pipe geometrical properties and the surrounding soil stiffness. An exact solution for
14
the buckling of a pipe buried in soil is not warranted because soil behavior is not very
predictable. The critical buckling pressure formula developed by Meyerhof and Baike (1963)
for a buried circular pipe is given by Equation 1.13. Actual test showed that this equation
(1.13)
where:
EI = Pipe stiffness.
R = Pipe radius.
the depth of soil cover using the Utah State University pipe-soil cell shown in Figure 1.6.
Pipes were loaded until local and general buckling were noticed in full scale tests. Moser
also described the difference between the classical buckling and the buried HDPE pipe
buckling. The classical structural buckling is a catastrophic sudden failure once the critical
load is applied. Increasing the load applied on HDPE pipe buried in soil results in the first
stage of the wall local buckling which is called dimpling. The pipe then can sustain more
load before the second stage of the local wall buckling, called waffling, takes place. If the
load is increased, a general buckling will develop. These local buckling patterns are shown
in Figures 1.7and 1.8, while in Figure 1.9 the general buckling of HDPE pipe illustrated.
Table 1.1 presents a summary of the buckling results of the tests described in Moser (1998).
The data in this table indicate that the dimpling pattern, which represents the beginning of the
local buckling, was observed in most of the tests where local buckling was reported. These
(a) (b)
Table 1.1. Summary of the buckling results of the loading tests (Moser 1998).
Wall crushing occurs if the compression stress reaches the pipe yield strength.
Generally wall crushing is a primary performance limit for most rigid or brittle pipes (see
Figure 1.10). This may also be reached for stiff flexible pipes. The ring compression stress is
given by:
where:
P - Prismatic soil load plus the effect of live load, psi.
D = Outside pipe diameter, in.
A = Pipe cross sectional area per unit length, in2/in.
18
Flexible buried pipe performance under applied loads is dependent on both the pipe and the
Greenwood and Lang (1990) suggested that soil stiffness is the most important
parameter that affects the flexible pipe performance. Soil stiffness is the ability of the soil to
assist the pipe to withstand the applied loads, to retain the pipe's structural integrity, and to
Watkins and Spangler (1958) introduced the empirical modulus of soil reaction, E', to
account for the restraint developed by the soil at the sides of the pipe. Hartley and Duncan
(1988) mentioned that because of the empirical nature of the modulus of soil reaction (E'), it
may introduce a large uncertainty in deflections calculated using the modified Iowa formula.
installed pipes, and back-calculating the modulus of soil reaction using the modified Iowa
formula to calculate the modulus of soil reaction values for pipe design. Howard (1977)
collected deflection and installation data on over 100 buried pipelines. The modulus of soil
reaction, E', was back-calculated and presented as a design recommendation for a variety of
soil types and compacted densities as shown in Table 1.2. Table 1.3 shows the values of E'
revised after considering the effect of burial depth published by Hartley and Duncan (1988).
The scatter of the available field data and the cost of carrying out field experiments
led to the theoretical studies of modulus of soil reaction. Many studies have related Ez to
constrained soil modulus (Ms) which was determined by performing one-dimensional tests on
representative soil samples. The design values of E' are found by multiplying Ms by a
constant, k, whose value lies between 0.7 and 1.5 as found by elastic analysis (Chambers et
al., 1980). Burns and Richard (1964) stated that the value of (k) depends only on the
normalized pipe-soil stiffness, (M$R3/(EI)pipe), and the Poisson's ratio for the soil (vs). For
most flexible pipe installations, they found that the value (k) is very close to unity. This
means that the modulus of soil reaction, E', is approximately equal to the constrained soil
modulus, Ms, as shown in Equation 1.15. As a result of this conclusion, Chambers et al.
(1980) and Krizek et al. (1971) suggested that Ms should be used in place of E' in the
where:
k = Constant with a value between 0.75 and 1.5 (typically =1.0).
Es = Young's modulus of the soil at the springline.
vs = Poisson's ratio of the soil at the springline.
20
relative 40%-70% , .
System). , , . relative
density relative ,
density
density
Fine-grained Soils (LL<50)
Soils with medium to no
plasticity CL, ML, ML-CL, 50 200 400 1000
with less than 25% coarse
grained particles
fine-grained Soils (LL<50)
Soils with medium to no
plasticity CL, ML, ML-CL,
with more than 25% coarse
100 400 1000 2000
grained particles, or Coarse
grained Soils with fines GM,
GC, SM, SC contains more
than 12% fines
Coarse-grained Soils with
Little or No Fines GW, CP,
200 1000 2000 3000
SW, SP contains less than
12% fines
Crushed Rock 1000 3000 3000 3000
21
Table 1.3. Design values of E' (psi) after Hartley and Duncan (1988).
It is necessary to recognize that the soil modulus varies with stress or strain level and
can be determined by various types of laboratory tests. Several types of soil moduli such as
initial, tangent, and secant moduli determined from triaxial tests are applied to solve soil-
structure interaction problems. Experimental studies by Janbu (1963) have shown that the
initial tangent modulus can be expressed in terms of confining pressure as shown in Equation
1.16. Kondner (1963) proposed the use of a hyperbola to describe the soil stress strain
r \n
<73
et=kpc (1.16)
where:
22
where:
ai = Major principal stress.
(g\. as)u = Ultimate deviator stress.
Duncan and Chang (1970) stated that soil behavior is nonlinear, inelastic, and
dependent upon the magnitude of the confining pressures. Duncan and Chang (1970)
combined Kondner and Janbu models to develop a soil tangent modulus model. This was
done by linearizing the hyperbolic equation proposed by Kondner (1963) to find the values of
the initial modulus and the ultimate deviator stress as a function of confining pressure. Then
they suggested defining the ultimate deviator stress in terms of the soil strength parameters,
and substituting the soil strength parameter for the ultimate deviator stress in the derivative of
Equation 1.17 to get the tangent modulus shown in Equation 1.18, where Ei is defined by
Equation 1.16.
where:
Et = Tangent modulus.
Ei = Initial tangent modulus, referred to Equation 1.16.
Rf = Ratio between the asymptote to the hyperbolic curve and the maximum shear
strength.
0 = Soil friction angle.
c = Soil cohesive strength.
23
Duncan et al (1980) proposed a power model for the soil bulk modulus given in
Equation 1.19. The limitations of this model are: 1) the use of only one point on the stress
strain curve since the bulk modulus was assumed to be independent of the deviator stress (Ci-
03) and only dependent on the confining stress (03) which is constant for a given test; and 2)
( ^
<73
B, = K
b P. (1.19)
x Pa y
where:
Kb = A bulk modulus number.
m = The exponent determining the rate of variation of Bj with (%.
Selig (1988) suggested another mathematical model for the soil bulk modulus based
on the hydrostatic compression test to be used for the design of buried pipes. The soil was
compressed under an increasing confining pressure applied all around the soil sample. The
curve relating the mean applied stress and the volumetric strain was found to be reasonably
represented by a hyperbola as shown in Equation 1.20. The tangent bulk modulus equation
and the parameters were determined using the same method Duncan and Chang (1970) used
to define the tangent modulus. The bulk modulus equation is given in Equation 1.21.
(1.20)
(1.21)
where:
B = Tangent bulk modulus.
24
Musser (1989) stated that using the hydrostatic soil parameters, A, andfu, developed
by Selig (1988) resulted in larger deflections than those based on field observations. Selig
(1990) published soil stress strain parameters of hyperbolic tangent and bulk modulus models
for plastic pipe installations at different compaction levels for different soils.
Pipe stiffness can be determined using both the parallel plate test and the curved beam
test while pipe material modulus can be determined using the compression test. Plastic pipe
stiffness is strain rate and time dependent. In this section, different pipe tests and the
dependency of the plastic pipe stiffness and modulus on loading rate and time will be
discussed.
Plastic pipe stiffness is the measurement of the load capacity of the pipe itself
subjected to loading conditions. Pipe stiffness is a function of the material type and the
geometry of the pipe wall. Plastic pipe stiffness can be determined using the ASTM D-2412
parallel plate load test shown in Figure 1.11. A 6-inch long pipe specimen is loaded at a
constant rate of 0.5 in./min at a constant temperature of 23° C. The pipe stiffness (PS) is
defined as the ratio of the applied force (F) in pounds per linear inch over the measured
change of pipe inside diameter (Ay). Pipe stiffness can also be defined as the slope of the
load deflection diagram. The stiffness factor (SF), which is the value of pipe modulus
multiplied by moment of inertia, is defined as shown in Equation 1.22. The pipe stiffness at
25
5% vertical deflection, i.e. the change in vertical pipe diameter divided by the original pipe
diameter, is typically used as the design value of stiffness. This represents the secant pipe
stiffness at 5% deflection. ASTM D-2412 stated that the stiffness of pipes with larger sizes
made from relatively low modulus material may be affected by creep due to the time taken to
reach the 5% deflection. Both pipe stiffness and stiffness factor are highly dependent on the
degree of deflection. For a large magnitude of vertical deflection percent a correction factor
F 3
El = 0.0186—D (1.22)
ay
where:
E = Flexural modulus of elasticity.
I = Moment of inertia.
D = Mean diameter.
F = Load applied to the pipe ring.
Ay = Measured change in inside diameter in the direction of applied load.
Gabriel and Goddard (1999) stated that pipes in service are affected by the soil
passive action at the pipe spring line. This reduces the pipe wall bending moment and
increases the wall ring compression. In parallel plate laboratory tests, all lateral restraint is
absent, which maximizes wall bending and minimizes ring compression. This is a significant
departure from the anticipated field service conditions. Gabriel and Goddard (1999) proposed
a curved beam test to measure the pipe stiffness. The curved beam responds to loads with
less wall bending moment than that of full ring. Therefore, greater proportion of the wall
compression dominates the response of the curved beam than that of a full ring. Hence, the
curved beam more closely approximates the field conditions of a buried pipe, Gabriel and
Goddard (1999).
26
The curved beam test was performed by Gabriel and Goddard (1999) using 90° arcs
cut from pipes of diameters ranging from 12 to 48 inches as shown in Figure 1.12. The pipe
arcs were loaded in a short period of time and the stress relaxation of HDPE pipe material
was investigated. Gabriel et al. (2002) proposed a time independent stiffness which was
Zhange and Moore (1997) studied the behavior of HDPE material using a compression test
on cylindrical samples with a height of 1 inch and a diameter of 0.5 inches as shown in
Figure 1.13. Tests were performed at different conditions including constant strain rates,
abrupt strain rate, and creep. All of these tests were performed at a constant temperature of
23° C.
27
À.
< 90
T
Figure 1.12. Flexible pipe curved beam test.
Vi inch
1 inch
Schluter and Shade (1999) performed the parallel plate test at three different loading
rates using PVC and HDPE pipes. These rates were 0.05 inVmin, 0.5 in./min, and 5 in./min.
Changing the loading rate by a factor of 100 resulted in a 6.5% stiffness change in PVC pipes
and 56% of HDPE pipes. Schluter and Shade (1999) also stated that the ASTM D-2414
deflection rate of 0.5 in./min does not relate to the real world deflection rate and that a
deflection rate of 0.05 inVmin is more realistic. It was concluded that both laboratory
measurements and theoretical calculations of ASTM D-2412 are too simplistic and that the
deflection rate effect on PVC pipes is minor but has a great influence on HDPE stiffness.
Sargand et al. (1998) stated that plastic pipes are rarely subjected to a deflection rate of 0.5
in./min which is specified in the parallel plate test and concluded, after a review of the field
data, that plastic pipes generally have a deflection rate of 0.01 to 0.06 in./min. Sargand et al.
(1998) also tested PVC and HDPE pipes using a variable rate parallel plate test at two
different rates of 0.5 and 0.05 in./min. Figures 1.14 and 1.15 show the load per unit length
(lb/in.) versus the vertical deflection percent results of variable loading rate parallel plate
tests performed on 18 inch HDPE and PVC pipes respectively. The pipe stiffness calculated
using Equation 1.22 is summarized in Table 1.4. These results showed that the loading rate
has little effect on the PVC pipe stiffness, while HDPE material is more sensitive to the
loading rate. The reduction of pipe stiffness was 3% to 6% for PVC and 25% for HDPE
pipes. Figures 1.14 and 1.15 and Table 1.4 clearly show that special treatment needs to be
120 i
100
c
x>
0 5 10 15 20
Vertical deflection percent
Figure 1.14. Parallel plate test for 450 mm HDPE pipe (Sargand et al. 1998).
200
160
je 120
"Sb
80
-•— Rate = 0.05 in 7min j
40 • Rate =0.5 in/min (•
0
0 5 10 15 20
Vertical deflection percent
Figure 1.15. Parallel plate results for 450 mm PVC pipe (Sargand et al. 1998).
30
Table 1.4. Pipe stiffness of parallel plate test at two different loading rates.
35
30
25
I20
"5
U 15
55 — — Strain rate= 10A-l/sec
10 Strain rate= 10A-2/sec
- * - - Strain rate= 10A-3/sec
5 Strain rate= 10A-4/sec
*—Strain rate= lO-S/sec
0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16
Strain
Figure 1.16. Stress-strain results of compression test (Zhange and Moore 1997).
Zhange and Moore (1997) performed various compression tests on HDPE pipe
material. These tests were constant strain rate, abrupt strain rate, and creep. The constant rate
tests were conducted at strain rates ranging from lO'Vsec to ÎO'/sec. These results are shown
in Figure 1.16. This Figure shows that for all strain rates the material is nonlinear with a
modulus independent of strain rate at approximately 0.04 (or 4%) strain. For later reference
31
in this dissertation, the region with strains less than 4% is called region A while that with
strains more than 4% is called region B. The strain abrupt test was performed by changing
the strain rate from 10"3/sec to 10"2/sec; after a brief period of rapid stress increase, the stress
attained the level it would have held if the new strain rate had been used from the beginning
of the test. Zhang and Moore (1997) concluded that HDPE material does not have strain rate
history dependency.
Sargand et al. (1998) studied the stress relaxation of PVC and HDPE pipes using the
parallel plate test. A variable load was applied over a period of one hour to maintain three
different vertical deflection percentages of 5%, 10%, and 15%. Figure 1.17 shows the stress
relaxation results for PVC pipes while Figure 1.18 shows the stress relaxation of HDPE
pipes. The time was extrapolated to estimate the pipe stiffness at 50 years. Table 1.5
summarizes the results of both tests. The percent reduction in stiffness was dependent on the
percent deflection for PVC pipes with a range of percent reduction between 12 and 32%. The
percent reduction in pipe stiffness is greater for HDPE pipes but it is less dependent upon the
vertical deflection percent. The percent reduction in stiffness for HDPE pipes was between
75 and 82%.
Goddard (1999) stated that parallel plate test results are not comparable from one
diameter to another for plastic pipes because of the time effect of the test. The time required
for a 12 inch pipe to reach the 5% deflection is 1.2 min, while it is 4.8 min for a 48 inch
diameter pipe. Figures 1.17 and 1.18 show that PVC and HDPE stress relaxation starts upon
the application of the load. Consequently, recorded loads of larger pipes are more heavily
affected by rapid stress relaxation than smaller pipes. Gabriel et al. (2002) introduced the
32
idea of time-independent stiffness using curved beam tests. HDPE pipes of 12 to 48 inches in
diameter were tested. A constant load was applied and the pipes deformed from 0% to 10%
chord displacement in just over _ second. This near instantaneous displacement was
200
180 1 ! —• 4% !
--10%|
e 160 4
- A• - - 15%!
| 140 4 4
120 h
rS
1 100
80 1
,
i
"O !
o 60 » qf J
u i ;
40
20 i !
0 1 !
10 20 30 40 50 60 70
Time (min)
Figure 1.17. Results of stress relaxation test on PVC pipe (Sargand et al 1998).
120
1 !
100 1 * J/O
e 1*A 10%
is i rn 1
80 4
S
<l> 60
1" — _
'
! 40
20
0 1
10 20 30 40 50 60 70
Time (min)
Figure 1.18. Results of stress relaxation test on HDPE pipe (Sargand et all 998).
33
Table 1.5. Pipe stiffness results of stress relaxation tests of parallel plate test
intended to decrease the stress relaxation. A regression line was fitted to the nearly linear
portion of displacement-stiffness curve between 5% and 10%. The regression line was
extrapolated backward to 0% displacement which represents the zero time. The value of
stiffness at this point is called the time independent stiffness. The plastic pipe stiffness is
time dependent, and back extrapolation to zero time does not yield a representative value of
stiffness since the pipe stiffness at time "t" after installation is what controls the pipe
behavior. The thermoplastic pipe stiffness is also depending on the loading rate which does
not relate to the field condition in the tests described in Gabriel et al. (2002).
Zhange and Moore (1997) conducted creep tests at different stress levels using the
compression test. It was noted that the secondary creep is almost constant and that the
primary creep showed a rapid increase in creep strain. The results of creep tests at different
stress levels are shown in Figure 1.19. The instantaneous strain was subtracted from these
Greenwood and Lang (1990) suggested that the field measured deflections increase
for a period of time and then stabilize to a constant level. By recognizing the long-term soil-
34
pipe behavior of creeping and relaxation, they expressed the pipe stiffness term as in
Equation 1.23.
iCTPE>
sp - 3
(1.23)
D
where:
Sp = Pipe stiffness term.
Ctp = Pipe stiffness retention factor.
E = Pipe modulus of elasticity.
I = Moment of inertia.
D = Pipe diameter.
0.12 -
Stress= 5.2 MPa
-- - Stress= 7.1 MPa
Stress= 9.2 MPa
Stress= 11.7 MPa
—*— St res s= 152 MPa
0.08 - St res s= 19.1 MPa
Sargand et al. (2001) studied the time dependency of thermoplastic pipe deflection in
the field using 12 HDPE and 6 PVC pipes with diameters between 30 and 60 inches buried
under 20 and 40 feet embankments. These pipes were buried in two different backfills of
35
crushed limestone and river sand. The vertical and horizontal deflections and the
circumferential shortenings were monitored for 8 months during and after construction. The
vertical and horizontal deflections, and circumferential shortening stabilized within two
months from the completion of the construction. Lars Janson (1996) reported that more than
eight years of constant pipe deflection gave no reason from a practical point of view to
change the long term pipe modulus that could be determined after only six weeks of testing.
Trantina and Nimmer (1994) stated that the objective of the graphical curve fitting technique
used to study the material time dependent properties was to obtain the most accurate fit while
Nimmer (1994) also stated that engineering judgment should be used concerning the
appropriate extrapolation in time and caution should be exercised when more than one order
Buried pipes are subjected to dead loads and live loads. Dead loads are the weight of
backfill materials carried by flexible pipes. For flexible pipes, the design dead load in the
trench can be determined either using the Marston load or the prism load, whichever is
larger.
Live loads are due to traffic, impact, construction, frost action, and expansive soil.
recommends pipe designers to use the wheel load of either H-20 or HS-20 as their design live
load because highways and bridges are usually designed for "worst case" conditions.
Katona (1990) studied the minimum depth of soil cover for HDPE pipes under traffic
36
loading. Different cases of truck loading and pipe diameters were tested. The diameters
investigated were 12, 15, 18,24, 30, and 36 inches. Two different levels of compaction (85%
and 100%) were used on the silty clayey sand. Minimum soil covers for the different
diameters of pipe investigated are presented in Table 1.6. These results were calculated
using the pipe short term properties for minimum and maximum moments of inertia in the
case of 85% compaction and only minimum moment of inertia for the case of 100%
compaction. Katona (1990) also noted that pipes with larger diameters require less depth of
soil cover than small diameter pipes and suggested that manufacturers make larger diameters
stronger. Klaiber et al. (1996) and Lohnes et al. (1997) studied the performance of HDPE
pipes using laboratory and full scale tests on different pipe diameters. Laboratory parallel
plate tests on different pipe diameters were completed to determine the pipe stiffness and the
stiffness factor according to ASTM D-2412. Parallel plate tests revealed that pipes with
smaller diameters have a greater pipe stiffness at 5% deflection than large diameter pipes.
The stiffness factor (EI) that considers the effect of the pipe diameter showed that pipes with
large diameters have greater stiffness factors than pipes with smaller diameters. Lohnes et
al.' s full scale field tests on pipes from the same manufacturer showed that 48 inch diameter
pipes, with the same backfill material, experienced less vertical deflection percent and less
strains than 36 inch diameter pipes at the same loads. Comparing the vertical deflection
percent does not necessarily imply that pipes with large diameters are stronger than pipes
with smaller diameters. Two pipes with different diameters may deflect the same amount but
this result in different values of vertical deflection percent where the effect of the pipe
diameter is included.
Construction loads resulting from heavy equipment and installation may also be a
excavator is heavier than highway trucks, so the design load may be underestimated if a
H-10 12 12 12
H-15 16 12 12
12 H-20 19 15 12
H-25 21 17 12
H-30 23 19 12
H-10 12 12 12
H-15 14 12 12
15 H-20 18 14 12
H-25 21 16 12
H-30 23 18 12
H-10 12 12 12
H-15 14 12 12
18 H-20 18 13 12
H-25 20 16 12
H-30 23 18 12
H-10 12 12 12
H-15 12 12 12
24 H-20 15 12 12
H-25 18 12 12
H-30 20 14 12
H-10 12 12 12
H-15 12 12 12
30 H-20 12 12 12
H-25 15 12 12
H-30 18 12 12
H-10 12 12 12
H-15 12 12 12
36 H-20 12 12 12
H-25 12 12 12
H-30 15 12 12
38
traffic load is chosen for design. During pipe installation, McGrath and Selig (1994)
observed that the impact resulting from compaction equipment on the pipe may lead to
flexible pipe distortions, especially when the compaction is operated too close to the pipe.
Therefore, it is necessary to design a pipe that can withstand not only the traffic load, but also
ASTM D-2321 provides recommendations for the installation of buried plastic pipes.
thermoplastic pipes under a wide range of service conditions. Figure 1.20 shows different
regions around an underground pipe as specified in ASTM D-2321. The space between the
pipe and the trench wall must be wider than the compaction equipment used in the pipe zone.
The trench width should be wider than the greater of pipe outside diameter plus 16 inches or
pipe outside diameter times 1.26 plus 12 inches. Backfill material particle size is limited to
material passing 1_ inch sieve. The recommendations for installation and use of different
soils and aggregates for foundation, embedment, and backfill around the plastic pipes are
shown in Table 1.7. The minimum soil densities were specified based on attaining an average
modulus of soil reaction, E', of 1000 psi. The moisture content of embedment material must
be within suitable limits to permit placement and compaction to the required levels with a
reasonable effort. This moisture content limit is set in the pipe zone to control the pipe
excessive deflection. A minimum depth of backfill above the pipe should be maintained
before allowing vehicles or heavy construction equipment to transverse the pipe trench,
which depends on the soil type. A key objective during installation of flexible plastic pipes is
to compact embedment material under pipe haunches to ensure complete contact with the
39
pipe bottom and to fill voids below the pipe. The lack of adequate compaction of embedment
material in the haunch zone can result in excessive deflection, since it is this material that
trench dimensions, the backfill envelop, and the minimum and maximum cover limitations.
The typical trench width is twice the pipe nominal diameter but not more than the minimal
diameter plus 2 feet. The height of the initial backfill shown in Figure 1.20 should be at least
6 inches above the pipe crown. The height of the initial and the final backfill should be at
least 1 foot over the pipe crown. The height of the bedding material is typically between 2
and 6 inches.
Final backfill
1 T
Initial backfill
Springline 6 to 12 in.
-y Pipe zone
c.
Bedding Haunch zone
Foundation
(may not be required)
Figure 1.20. Trench cross section showing the terminology used by ASTM D2321.
40
Table 1.7. Recommendation for installation and use of soils and aggregate for pipe foundation
embedment, and backfill (ASTM D2321).
Soil dass
Class IA Class IB Class II Class HI Class FV-A
General do not use if process to check gradation do not use if obtain geotechnical
Recommendation migration of obtain required in case of flow water may cause evaluation of
fines expected gradation to to minimize instability oproposed material
minimize fines fines migration
migration
Embedment use vibratory - min. density -min. density -min. density -min. density
Compaction compactor 85% Standard 85% Standard 90% Standard 95% Standard
Proctor Proctor Proctor Proctor
- use vibratory - use vibratory - use vibratory - use impact
compactor or compactor or compactor or compactor or
hand tampers hand tampers hand tampers hand tampers
Bums and Richard (1964) for an elastic circular pipe deeply buried in an infinite elastic
medium subjected to horizontal and vertical loading. Solutions for pipe deflection, thrust and
moment were obtained for full bonding and free slip at the pipe-soil interface. Moser (1997)
stated that the greatest shortcoming of Burns and Richard solution is that it assumes double
symmetry. That is, it assumes the soil-pipe system is symmetric about horizontal and vertical
axes. In the elastic solution, no assumption was made or needed for the pressure distribution
around the pipe. On the other hand, Sargand et al. (1998) stated that the elastic solution can
not account for different pipe installation modes and the results are not useful if the height of
Moore (1995) used a three-dimensional finite element analysis to model HDPE pipes
subjected to various burial depths considering the pipe geometrical properties, corrugations,
for various backfill materials using short and long term pipe material properties. The stresses
corrugations, cross section of the HDPE pipe. His study showed that the two dimensional
analysis can predict the compression and circumferential stresses of the pipe.
Taleb and Moore (1999) presented a two-dimensional finite element model to study
the response of metal culverts to the compaction process. The culvert deformation and
bending moment during the process of backfilling were predicted and compared with the
experimental results. AFENA was used to perform the two-dimensional finite element
analysis using elastic-plastic soil model with linear variation of soil modulus with depth. The
plastic soil behavior was assumed as the behavior of the soil beyond shear failure. This
42
analysis showed a good agreement between the measured and analytical pipe deflections and
bending moments.
Moore and Taleb (1999) used a three dimensional finite element analysis to study the
metal culvert response to live loading and compared the results with the experimental results
of Webb et al. (1998). Real truck load tests were performed by Webb et al. (1998) on metal
pipes where the truck passed over metal pipes with two different depths of soil cover (1 and 2
feet) at two different soil condition (well and loose compacted). Moore and Taleb (1999)
modeled the soil as a linear elastic material with a linear modulus variation with depth. The
pipe performance properties studied were the distributions of deflection, thrust, and moment
along the culvert axis. Culvert deflections show the correct pattern with a maximum
magnitude difference of the pipe crown of 20%. The measured thrust exceeds the predicted
values using a three-dimensional analysis with a maximum difference between measured and
predicted values of about 50%. The bending moment pattern is successfully predicted with a
magnitude of 50% to 70% difference from the measured values. Moore and Taleb (1999)
suggested that this may be due to soil shear failure. They also performed a three-dimensional
analysis using a reduced soil stiffness based on engineering judgment to take into
consideration the shear soil failure. This approach affects the whole soil zone but gives more
reasonable moment results. Their study showed that neglecting the shear failure of soil
results in a greater difference between the measured and the analytical results.
flexible pipes. The theoretical methods were compared with full scale testing results. These
theoretical methods were semi empirical equations (Iowa formula), closed form analytical
solutions (Burns and Richard elastic solution), and the finite element method. Moser (1997)
43
stated that the finite element method most closely presents the full scale test results and that
Hashash and Selig (1990) studied the long-term performance by monitoring a 24 inch
diameter and 576 feet long HDPE pipe under a soil fill up to 100 feet for 722 days. The
bedding, haunch, and the initial backfill materials were well graded crushed stone with 100%
compaction. A 77% of arching was determined from the measured field stresses, so the pipe
Sargand et al. (2001) measured the soil pressure at various heights above two 30 inch
diameter pipes of PVC and HDPE pipes. The measured geostatic pressure reached the pipe
crown was greater for the PVC pipe than the HDPE pipe. The soil pipe interaction zone
extended only about one pipe diameter above the pipe for both HDPE and PVC pipes.
44
2.1 Objective
Although high density polyethylene (HDPE) drainage pipes have been used for over a
decade, little information is available on the in-service behavior of these pipes. At the request
of the Minnesota Concrete Pipe Association (MCPA) an investigation to evaluate the field
A list of potential sites was provided by MCPA for investigation. Ten projects were
selected from this list by the research team for evaluation based on a variety of variables
The objective of this study was to evaluate the field performance of in-service HDPE
pipes using visual information obtained from a remote, motorized video camera. The ten
projects were surveyed in Minnesota between July, 2000 and March, 2001. These projects
are located in the cities of Mankato, Blaine, Mounds view, Fairmont, Gaylord, Le Center, and
Maple Grove as shown in Figure 2.1. The total pipe length surveyed in these projects was
2.2 Methodology
The surveys were conducted with a remote, motorized television camera that moved
through the pipe. Figure 2.2 shows the type of camera used in the study. The interiors of
the pipes were televised and recorded on videotape. The tapes were sent to Iowa State
University where they were evaluated, and apparent structural and/or sediments problems
were noted.
The performance characteristics that were considered in the evaluation were cross
section deformations, wall buckling, wall crushing, wall cracking, joint separation, and
46
sediments. These characteristics and their locations along the length of the pipe were
recorded. Data sheets for recording these characteristics are shown in Appendix A. The data
SIBLEY
Tu El
BLUE
EARTH
presented in Figure 2.3. These specifications allow for a maximum pipe diameter of 36
inches and require well-graded granular backfill with 100% passing the 1 inch sieve. Well
graded backfill material should extend at least 12 inches above the pipe crown. The total
depth of soil cover above the pipe crown should be between 2 and 20 feet. The bedding
The trench width should be the larger of 24 inches plus the inside pipe diameter or two pipe
diameters.
Maple Grove uses the same specifications with a different type of required trench
which is shown in Figure 2.4. The improved pipe foundation material, more than 6 inches in
depth, is used in the case of poor native soil. The projects surveyed in this study were
reported to have been installed according to MN. DOT specifications except for three
installations where HDPE pipes with diameters larger than 36 inches were used and the sites
Top of Pavement
73 7Z 77 73 ^ 1—
i Cover
2' Min. to
12" Min. 20' Max.
Embedment
material O.D.
15% of O.D.
Bedding
1 Foundation
Notes:
Maximum nominal pipe diameter is 36 inches. Embedment material per Spec. 3149.20 modified to 100%
passing the 1 inch sieve. Construction requirements per Spec. 2451 modified so that the embedment material is
compacted in uniform layers 8 inches or less in thickness. Bedding shall be class B per. Spec. 2451.3.
j Trench width per ASTM D2321 except as modified to provide (6 nominal inside pipe diameter
— or 12 inches (whichever is greater) on each side of the pipe, to allow for compaction.
±_ The zone immediately beneath the pipe shall only be compacted sufficiently to
provide uniform support.
Figure 2.3. MN. DOT specifications for plastic pipe storm sewer installations.
Compacted
Backfill
Coarse
Filter
Improved Aggregate
Pipe
Foundation
The observations and problems of the sites surveyed in this study are summarized in
Appendix B. This chapter contains the evaluations of the surveyed projects and their
interpretations.
Table 3.1 presents a summary of the total lengths surveyed for each HDPE pipe
diameter. The total length of the surveyed HDPE pipes was 12,006 feet. As may be seen,
nine different diameters of the HDPE pipes were surveyed ranging from 12 to 48 inches.
Over 56% of the HDPE pipes surveyed were 30 inches or more in diameter.
It was observed that structural problems occurred in 19% of the length of the 48 inch
diameter pipe, 8% of the 36 inch diameter pipe, 6% of the 30 inch diameter pipe, 6% of the
24 inch diameter pipe, 33% of the 18 inch diameter pipe, 2% of the 15 inch diameter pipe
and 3% of the 12 inch diameter pipe. Figure 3.1 illustrates the weighted percent of structural
problems observed for each pipe diameter. The weighted average of structural problems is
defined as the ratio of the length with structural problems for each pipe diameter to the total
50
length surveyed for the same diameter all multiplied by the ratio of the length surveyed for
the pipe diameter to the total pipe length surveyed. Structural problems are defined as the
existence of local buckling dimpling pattern, cross sectional deformations, wall cracks,
and/or pipe sags. Figure 3.1 shows a general trend of increasing structural problems as the
diameter increases. Pipes with 27 and 42 inch diameters showed no cross sectional
deformations. The percent of structural problems in the 48 inch diameter pipe was dramatic
when compared with other pipe diameters. This indicates an increasing probability of
problems with pipes whose diameters are greater than 36 inches, which is the largest HDPE
I 2
0
-4
•
•
•
0 10 20 30 40 50 60
Pipe diameter (inch)
Elliptical cross sectional deformations with both horizontal and vertical shortenings
were noticed at several locations during this study. Figure 3.2 is an example of the horizontal
in the horizontal diameter is illustrated in Figure 3.3. Figure 3.4 shows the relationship of
51
40
35
Cfl
1 30
.c "o 25
i - 20
^ I 15
IV. 10
5
0 * » •• U
0 1 2 3 4 5 6
Height of soil cover (H)/ Pipe diameter (D)
percent of structural problems with the ratio of the height of the soil cover (H) to the inside
pipe diameter (D). This figure shows that pipes with small or large H/D ratios have a higher
percent of structural problems. The largest value of H/D indicated the highest percent of
structural problems (mainly cross sectional deformation with vertical shortening). This was
interpreted as the effect of geostatic loads on the pipe crown. The smallest value of H/D also
indicated a high percent of structural problems (mainly cross sectional deformations with
horizontal shortening). This was interpreted as the effect of installation with a shallow soil
cover and no surface loads to restore the pipe to 'round' again. Dimpling at small H/D ratios
was also observed at two locations. This was interpreted as the effect of traffic loads. Except
for one case (H/D= 2.7), all pipes with H/D ratios between 2 and 4 did not show any
structural problems.
The 24 inch diameter pipe cross-section shown in Figure 3.3 has 10 to 12 inches of
elliptical cross sectional deformations (vertical shortening) that extended for 10 feet. This
represents an average of 46% deflection of pipe diameter. Associated with the vertical
Figure 3.5. One inch dimples in 24 inch diameter pipe.
deformation were 1 to 2 inch dimples and pipe wall cracks at "2 o'clock"; these are shown in
Figures 3.5 and 3.6. A Pipe sag of 8 to 10 inches in a 24 inch diameter pipe that extended for
6 feet was also noticed in the pipe shown in Figure 3.3; no plans were available for this site.
The horizontal increase in the pipe diameter suggested that the soil around the pipe was not
compacted properly.
Dimpling patterns, pipe wall deflections, and wall cracks were noticed nine times in
the regions of "1 to 3 o'clock" and "9 to 11 o'clock" - four cases with cracks, three with wall
deformations, and two with dimpling. Figure 3.7 shows a 3 inch wall deformation between
"1 and 3 o'clock" in the 15 inch diameter pipe which represents 20% of the pipe diameter.
These match Moser's (1998) results that were shown in Table 1.1, where six out of nine local
Figure 3.7. Pipe wall deflection between "1 and 3 o'clock" in a 15 inch diameter pipe.
55
Dimpling was observed at four different sites in this survey. It was in pipes with 15,
24, 30, and 36 inch diameters. One example of the dimpling in a 24 inch diameter pipe is
shown in Figure 3.5. The dimples occur over 152 feet of the pipe length surveyed, which
represents 1.25% of the total pipe length. In two of the cases that showed dimpling, the pipes
were buried at a relatively shallow depth of 1 to 2 feet under surface traffic loads. The
minimum specified soil cover by the MN. DOT is 2 feet, as was shown in Figure 2.3. Table
1.4 showed that the minimum soil cover calculated by Katona (1990) is even less than 2 feet.
One of the two cases that showed dimpling with shallow soil covers has a 2 foot depth of soil
cover which satisfies the Minnesota DOT specifications, while the other has a 1.2 feet of soil
Pipe wall tearing was noticed twice in this survey. Figure 3.8 shows an example of a
wall tearing (pointed by the arrow) in a 15 inch diameter pipe. This wall tearing is expected
f
\
\
^ r:-:=7H
A summary of the structural performance for the pipe diameters that showed
structural defects is presented in Table 3.2. The overall length percent of dimpling for the 15
inch diameter is 0.25%, 0.37% for the 24 inch diameter, 5.8% for the 30 inch diameter, and
1% for the 36 inch diameter. The 36 inch diameter pipe that showed dimpling is buried at a
depth of 2 feet, while the 30 inch diameter pipe with dimpling was buried at a depth of 2.6
At one site, a 48 inch diameter pipe with 1.7 feet of average soil cover heaved to
above the ground surface. Associated with the heave was a joint separation and a
circumferential crack 75 inches long and about 0.04 inches (1 mm) wide as shown in Figures
3.9, 3.10, and 3.11. The location of the heave, about 30 feet from the outlet into a marsh,
suggests that buoyancy from a high water table combined with the shallow depth of cover
Pipes buried in sandy native soils with diameters 36 and 48 inches show cross
sectional deformations. The 48 inch diameter pipe, buried at a depth of 1.2 feet, showed
cross sectional deformations with horizontal shortening. The 48 inch diameter pipe buried at
6.5 feet of soil cover showed 1 inch elliptical cross sectional deformation over 22 feet, which
represents 0.9% of the 48 inch diameter pipe surveyed. Pipes buried in regions with a native
soil of till and lake sediments with 12, 15, 18, 24, 36, and 48 inch diameters showed cross
sectional deformations.
Pipes surveyed are buried in two different types of trenches. Pipes buried in a trench
like that shown in Figure 2.3 experienced a slightly higher percentage of structural problems
(9%) than pipes buried in a trench like that shown in Figure 2.4 which had 7% structural
problems.
Table 3.2. The structural performance of each pipe diameter.
Figure 3.10. Heaving of 48 inch diameter pipe to above the ground surface.
60
Pipes with diameters 12, 24, 30, 36, and 48 inches showed pipe sags. No information
was available on the foundation material. All of these cases occurred where the native soil is
till or lake sediments. This suggests that care should be exercised when using HOPE pipes in
regions where till or lake sediments may be the foundation material. Pipes buried in sandy
native soils showed no pipe sag. One of the cities specified that the sand native soil can be
were noticed at 37 locations which is 6.5% of the total number of joints surveyed. Joint
separations with root penetration were noticed at three locations where the pipe was buried in
a native peat soil (see Figure 3.12). Deformation of the connection between two adjoining
pipes was noticed at 10 locations. Cracks were noticed at five different sites with 27 total
cracks. Examples of these cracks are shown in Figures 3.6 and 3.13.
61
The relationship between the year of installation for a given project and the percent of
structural problems is shown in Figure 3.14. This figure shows that pipes recently installed
have more structural problems than pipes that have been in service for a longer period of
time. HDPE material elastic modulus decreases with time, so pipes with a short in-service
time period should have more stiffness to resist the applied load than old pipes (i.e. those in
service for several years). This was not observed in the locations surveyed in this study.
45
40
35
?S 30
S 25
« 20
|15
-§ 10
£ 5
0
1994 1995 1996 1997 1998 1999 2000
Year installed
Figure 3.14. Percent of structural problems vs installation year.
3.2 Sediments
Sediments and water were observed in many of the pipes surveyed. The percentage of
sediments in the pipes was plotted versus pipe slope in Figure 3.15. Though there is a wide
scatter, a general trend could be observed in the decreasing percentage of sediments with the
increasing pipe grade, especially when the grade is more than 0.4%. Locations with pipe
sags had no significant relation with the scatter of the data shown in Figure 3.15. This
suggests that the sediments observed in these pipes are caused by hydraulic factors and not
by structural deformations.
63
•y%«—• •
*.
*
•
•
•
•
# a# *»## # 1 • • • *—«—
0 1 2 3 4 5 6 7
Pipe grade (%)
Figure 3.15. Change of percent of sediments with pipe grade.
64
4.CONCLUSIONS
In-service HDPE pipes were surveyed for structural problems. These problems were
located along the pipe and then evaluated in this study. Ten projects with a total of 12,006
feet of pipe were studied and the following instances of structural problems were observed: 4
cases of dimpling or local buckling, 29 cross sectional deformations, 7 pipe sags, 27 wall
Wall buckling affected 152 feet of the pipe length with a vertical deflection percent
ranging from 3% to 7%. Wall buckling affected 1.3% of the total pipe length surveyed. At
Gaylord, MN vertical deflections as high as 46% were observed in a 24 inch diameter pipe
and could be included in the buckling category. Other vertical deflections ranged from 4%
to 23%. Vertical deflections affected 266 feet of the pipe length or 2.2% of the total pipe
length surveyed. In addition, 7 joints had vertical deflections from 4% to 12%. Most vertical
deflection percentages were about 10%, and 17 of the 19 observed vertical deflections
Horizontal deflections ranged from 3% to 13%, with the highest percent at Gaylord,
MN. Horizontal deflections occurred at sites with shallow backfill and affected 354 feet of
Pipe sags were observed at four sites, with a minimum of 3% and a maximum of
38%. The greatest sag was at the Gaylord site. Sag affected 145 feet of the pipe length or
1.2% of the total pipe length studied. Wall cracks were between 0.05 and 0.1 inch wide and
extended between 10 to 20 inches around the pipe circumferences. Joint separations were
usually an inch or less but at two projects in Maple Grove, MN there were 2 inches of
The results of this survey lead to the conclusion that the frequency of structural
problems increases as the pipe diameter increases. Pipes 48 inches in diameter experience
Our review of the ratio of the height of soil cover (H) to the inside pipe diameter (D)
revealed no relationship between H/D and the frequency of structural problems. The highest
H/D was associated with vertical shortening which resulted from the effects of geostatic
loads. Pipes with low H/D ratios also had a high percent of structural problems in the form of
horizontal shortening, likely resulting from side compaction during construction and
Dimpling, the precursor of general wall buckling, was observed at four different sites
in pipes of 15, 24, 30, and 36 inches in diameter. Two of the four pipes that experienced
dimpling were buried under two feet or less of soil cover and were subjected to traffic loads.
Pipes installed in 1999 showed a higher percentage of problems than pipes installed
in 1995, 1996, and 1997. Pipes buried in a trench with vertical walls experienced a slightly
higher percentage of structural problems (9%) than pipes buried in trenches with sloping
walls (7%). One pipe had been pierced with two steel rods.
A 48 inch diameter pipe under 1.7 feet of soil cover heaved above the ground surface.
The heave caused some joint separation and cracking. Buoyancy from a high water table and
Pipes with diameters between 12 and 48 inches buried in regions where the native soil
is glacial till or lake sediments showed cross sectional deformations, and some pipes showed
longitudinal sag. Although no specific information was available on the foundation material,
66
this suggests that care should be exercised when using HDPE pipes in regions where till or
It was reported that most of the sites investigated followed the MN. DOT
specification. This may explain the few major structural problems noticed in this
5. INTRODUCTION
The theoretical study of flexible pipes was started by Spangler's work on metal pipes
(1941). Since then, more flexible materials (i.e. Aluminum, Polyvinyl Chloride (PVC), and
Polyethylene (PE)) have been used in the pipe industry. Part I of this thesis indicated the
effect of the backfill material and soil structure interaction on HDPE pipe in-service
behavior. In Part II, finite element analysis, a useful analytical tool in the study of soil
structure interaction, will be used to study the flexible pipe behavior. One such analytical tool
is Culvert ANalysis and DEsign (CANDE), a 2-D finite element program commonly used to
analyze and design buried pipes that was developed by Katona (1976). The stated significant
limitations of CANDE are 1) small deflection theory, 2) neglecting out-of-plane effects such
(1976).
A variety of other finite element analyses have been used to model the response of
buried flexible pipes. Taleb and Moore (1999) performed a two dimensional finite element
analysis for metal pipes under earth loading with an elastic-plastic soil model, where the
modulus varied linearly with depth and the finite element procedure AFENA was used.
Moore and Taleb (1999) studied the three dimensional response of metal culverts to live
loading with a semi-analytic procedure based on the use of Fourier Integrals. The soil was
modeled using a linear elastic response with the modulus varying linearly with depth. El
Sawy et al. (1997) used the same approach to study the stability limit state of long span
shallow and deeply buried metal culverts. All of the analytical studies cited above used a
simplified soil constitutive law (elastic or elastic plastic) to model the soil.
69
pipes. In his work, PE pipes were loaded using the Utah State University loading pipe-soil
cell shown in Figure 1.6 at different soil compaction levels in an SM soil and the pipe
Watkins and Anderson (1999) listed the differences between the analysis of soil-pipe
interaction system and a simple linear elastic continuum. This list included the nonlinear
stress-strain relationship of soil and large deflections that may be involved using very
flexible pipes. The recent trend of using plastic pipes for deep applications increased the need
for large deflections effect investigation. Although CANDE allows the use of non-linear soil
The main advantage of CANDE relative to ANSYS is its capacity to use hyperbolic
and power soil models, while ANSYS has the advantage of modeling large deflections. One
objective of the theoretical analyses presented later in this part is to compare the results of
CANDE with the small and large deflection theories of ANSYS for the case of elastic soil
properties under geostatic loading conditions. The second objective is to write a code using
the ANSYS programming language to model the soil behavior using hyperbolic tangent
modulus with both power and hyperbolic bulk soil modulus models and to study the effect of
large deflections on PE and PVC flexible pipes. The written code was also improved to
5.3 Modeling
CANDE uses three types of elements in pipe-soil structure modeling. These elements
are plane strain, beam column, and interface. The plane strain element is used to model the
70
soil and has two translational degrees of freedom in the X and Y directions. The pipe is
modeled with beam-column elements. This element has three degrees of freedom at each
node, two translations and one rotation. Plastic pipes are modeled as elastic materials.
The CANDE library has three solution levels, six soil constitutive models, and five
pipe types. CANDE solution level 2,"homogenous model", generates the finite element mesh
automatically, and is shown in Figure 5.1. This figure shows that the maximum soil cover
above the pipe crown is 3R, where R is the pipe radius. If the soil cover exceeds this limit,
the program automatically truncates the cover at the 3R level and applies the truncated soil
load as a distributed load on the soil surface. Katona (1976) and Musser (1989) have a
problems. It has a large library of element types, permits small and large deflection analyses,
and eight types of material nonlinearities. ANSYS also allows the user to program using the
Language (APDL)". This language allows the user to build the finite element model, repeat
the commands, use macros, if-then-else branching, do loops, and vector and matrix
operations. The file can be used to create the model, do the calculations, solve and generate
the results for each load step. Problems with different load increments can also be solved,
and the model parameters such as number of elements , can be updated for each load step.
71
<L
Quadrilateral element
Maximum
Beam-
column
element
6R
A plastic pipe 610 mm in diameter with four different soil covers (1.5, 3.05, 4.6, and
6.1 m) above the pipe spring line was modeled using both CANDE and ANSYS as shown in
Figure 5.1. In these models, the soil was assumed to be a linear elastic material, no interface
elements were used between the soil and pipe elements, and the pipe was assumed to be
elastic with a smooth "no corrugation" uniform thickness (T). The soil and pipe properties
used are given in Table 6.1. Plastic pipe properties included in this table are CANDE default
plastic pipe properties. PE, which also referred to as HDPE, pipe properties included in Table
6.1 are taken from Corrugated Polyethylene Pipe Association (CPPA) (1997) and meet
Table 6.1. Pipe and elastic soil properties used in the analysis.
* Assumed value
Soil y = 1,920 kg/m3
Figure 6.1 shows the computed vertical deflection of the plastic pipe crown with
respect to soil cover above the pipe spring line using small and large deflection theories of
ANSYS, and the small deflection theory of CANDE for the soil pipe system described above.
This figure shows that the three solutions result in nearly identical pipe crown deflections.
For the case of 1.5 m soil cover above spring line (1.22 m above pipe crown), the three
73
solutions give equal pipe crown deflections. This leads to the conclusion that for the systems
modeled, small deflection theory modeling adequately describes the pipe behavior and so
ê -4
§
5 -8 'V
c
5 -12 ANSYS, small
• » - - ANSYS, large
|-16 --—CANDE
-20
0 1 2 3 4 5 6 7
Height of soil cover above springBne (m)
Figure 6.1. Plastic pipe crown deflection with respect to height of soil cover for elastic soil.
Figure 6.2 shows the vertical deflection for different points on the plastic pipe
circumference for various depths of soil cover above the pipe spring line. This figure shows
the same trend of a maximum deflection at the pipe crown and minimum deflection at pipe
invert with virtually no difference in the results of all three methods for all cases of soil
cover. As an overall comparison, the three different methods showed good agreement for this
range of soil covers. The large deflection analysis using ANSYS has an insignificant effect
Figure 6.3 shows a comparison of the vertical stress above the plastic pipe for the
case of 4.6 m of soil height. The three different methods used resulted in a similar trend with
a maximum difference of 5% or less. The vertical stress above the pipe is equal to the
geostatic vertical stresses to a depth of 50 cm above the pipe crown which is almost one pipe
0
-2
-8
12
6.1 m soil cover
14
16
• ANSYS, small
- » - ANSYS, large
18
—*— CANDE
•20
20 40 60 80 100 120 140 160 180
a: Position on pipe circumference (deg)
Figure 6.2. Plastic pipe crown vertial deflection with respect to position on pipe circumference
for different height of soil covers for the case of elastic soil.
75
diameter. Sargand et al. (2001) measured the soil pressure above PVC and PE pipes and
found that the soil pipe interaction zone was about one pipe diameter above the pipe for both
PVC and PE pipes. Figure 6.4 shows the change in the horizontal stress with the change in
the horizontal distance measured from the pipe center for the case of 4.6 m soil height. All
three methods of solution resulted in the same trend and essentially the same values of
horizontal stresses at different distances from the pipe center. The horizontal stress decreases
as the distance from the pipe surface increases. The horizontal stress becomes constant at a
distance slightly greater than one meter. This distance is about twice the pipe diameter which
Results of the same pipe-soil system described above, but using the properties of the
PE pipe shown in Table 6.1, are shown in Figure 6.5. This figure shows that both CANDE
and ANSYS, using both small and large deflection theories of ANSYS, give almost the same
results for pipe invert and crown deflection with only slight differences for the case of 6.1 m
soil height above the spring line. The change of vertical deflection percent, which is the
change of the vertical pipe diameter divided by the inside pipe diameter, of PE pipe with soil
cover is shown in Figure 6.6. This figure shows the same trend, with differences in the
vertical deflection percent between the three methods increase as the soil height increases.
Figure 6.7 shows a typical soil stress-strain relationship and indicates the change of
soil modulus with confining pressure (03) according to Duncan et al. (1980). This response
has been modeled using a hyperbolic stress strain relationship, Kondner (1963), discussed in
chapter 1. Power bulk modulus and hyperbolic bulk modulus models were developed by
Duncan et al. (1980) and Selig (1988) respectively. Figure 6.8 shows a flowchart of the code
76
-»—ANSYS, small
•— ANSYS, large
i- CANDE
3 60
Pipe Crown
100
50 55 60 65 70 75 80 85 90
Vertical stress (kN/mA2)
Figure 6.3. Vertical stress in a soil column above the pipe for 4.6 m soil height.
- - • • • ANSYS, small
— - ANSYS, large
—6—CANDE
o
-2 — Pipe invert
-4 Pipe crown
-6
-8
I -10
-12
3 -14
•CANDE
- • - - ANSYS, small
- -A- - ANSYS, large
CANDE
£ -16 -
- -X- - - ANSYS, small
- ANSYS, large
-18
-20
0 1 2 3 4 5
Height of soil cover above springline (m)
Figure 6.5. PE pipe vertical deflection vs. height of soil cover for elastic soil case.
5 0.8
CANDE
--ANSYS, small
> 0.2 — -A- - ANSYS, large
0 1 2 3 4 5 6 7
Height of soil cover above spring line (m)
Figure 6.6. PE pipe vertical deflection percent vs. height of soil cover for elastic soil case.
78
for soil models, which was written in ANSYS using the APDL programming language. The
improved code works as a load step program where the stresses induced at the beginning of
the first load step used to calculate the soil modulus is based on the soil element depth. Soil
elements are considered as blocks of soil masses with an average soil modulus for each soil
block. These blocks will enable modeling different regions and different soil materials used
in the backfill. Each load step is solved using "n" substeps by applying a stepped load
incrementally within each load step. The results of each sub-step are used to calculate a new
average modulus for each soil block and re-solved again until the end of the load step. The
load step is solved using the average modulus calculated from all sub-steps for each soil
Higho3
Intermediate o3
Low o3
n substeps
END
because solving the load step using the average calculated moduli at the beginning and at the
end of each load step resulted in greater error when compared with CANDE results. The
stresses calculated using the average modulus are then used to calculate the soil element
modulus at the beginning of the next load step. This cycle will be repeated N times, where N
Figure 6.7 shows that as the shear stress applied on soil elements increases (i.e. [(or
<j3)/2]), the soil becomes weaker. Shear failure occurs when the tangent Young's modulus
approaches zero. The improved code limits the parameter, {Rf ((d- 03)(i-sin<p)/2(C. cos<p+ c%
simp)}, shown in Equation 1.18 to 0.95. If the value of this parameter is greater than 0.95, the
value of the tangent modulus is assumed to be 0.05 of the initial tangent modulus. This
avoids solving for the case of a zero modulus. The minimum and maximum soil bulk
modulus values were set as the tangent modulus divided by 3 and tangent modulus multiplied
by 8 (Et/3 and 8E,) respectively. This limits the value of the Poisson's ratio within the range
of zero and 0.48. A tension failure check was also included in the program. Tension failure
occurs when the confining pressure becomes tensile. In this case, the stress ratio (03/Pa) of
ANSYS utilizing the hyperbolic soil model with power bulk modulus was compared
with CANDE for a pipe soil system of a 610 mm diameter pipe with the plastic pipe
properties shown in Table 6.1. The buried pipe structure analyzed is shown in Figure 5.1.
The soil model constants used in this analysis for an SM soil with 90% compaction (Moser
(1990)) are shown in Table 6.2. Moser (1994) described the SM soil used in his work as
"lesser quality than most soils specified as backfill and so it is a worst case test". Five cases
81
of different heights of the soil above the pipe spring line were used for comparison (1.5, 3.05,
Figure 6.9 shows the plastic pipe crown and invert deflections for all cases of soil
cover using both CANDE and the small deflection theory of ANSYS. This figure shows that
the results of ANSYS with the improved code show good agreement with the results of
CANDE for both the pipe invert and the pipe crown. The vertical deflection percent of the
pipe for the same plastic pipe-soil system is presented in Figure 6.10. This figure also shows
The same pipe-soil system was analyzed for the cases of PE pipe with 1.5, 3.05, 3.8,
and 4.6 m of soil height above the pipe spring line. The soil was modeled using the
hyperbolic bulk modulus model for an ML soil with parameters shown in Table 6.2 (Musser
(1989)). The PE pipe invert and crown deflections using both ANSYS and CANDE are
shown in Figure 6.11. The results of ANSYS with the improved tangent and bulk hyperbolic
modulus code show a good agreement with the results of CANDE for both the pipe invert
and the pipe crown. The vertical deflection percent of the PE pipe is presented in Figure 6.12.
This figure also shows a good agreement between both ANSYS and CANDE. The
comparison of the cases above validates the use of ANSYS with the improved code to model
c
E -10
c
o
u
ë -20
•o — - Pipe crown, CANDE x.* -
- - • - - Pipe crown, ANSYS
f-30 — -a — Pipe invert, CANDE
Pipe invert, ANSYS
-40
0 2 3 4 5
Height of soil cover above spring line (m)
Figure 6.9. Plastic pipe vertical deflection using hyperbolic tangent soil modulus and power
bulk soil modulus for both ANSYS and CANDE.
= 2.5
W
•o
—•—CANDE
* ANSYS, small
> 0.5
0 1 2 3 4 5 6 7
Height of soil above springline (m)
Figure 6.10. Plastic pipe vertical deflection percent using hyperbolic tangent soil modulus
and power bulk modulus for both ANSYS and CANDE.
83
-2
-8
• Pipe invert, CANDE
-10 * Pipe crown, CANDE
— -A — Pipe invert, ANSYS
-12 Pipe crwon, ANSYS
-14
0 1 2 3 4 5
Height of soil cover above spring line (m)
Figure 6.11. PE pipe vertical deflection using hyperbolic tangent and bulk soil moduli
models for both ANSYS and CANDE.
0.9
c 0.8
0.7
§ 0.6
3 0.5
5 0.4
I 0.3 —•—CANDE
> 0.2 -1- ANSYS
O.l
0 l 2 3 4 5
Height of soil cover above spring line (m)
Figure 6.12. PE pipe deflection percent using hyperbolic tangent and bulk soil moduli for
both ANSYS and CANDE.
84
Because PE pipes are flexible and the properties used are documented in CPPA
(1997), the large deflection theory was used to compare the results of both CANDE and
ANSYS. Figure 6.13 shows the vertical deflection percent of PE pipes using CANDE and the
small and large deflection theories of ANSYS. The large deflection theory for PE pipes has
little effect on the results for a soil cover of 3.05 m. Increasing the fill height from 4.6 to 6.1
m increases the effect of the large deflection theory. These figures show that both ANSYS
with the small and large deflection theories and CANDE have a good agreement for soil
covers up to 6.1 m.
I 2.5
ë
= 2
f ,.5
V
"O
« 1
y •—CANDE
- - ANSYS, small
§. 0.5
•A- - ANSYS, large
0
0 1 2 3 4 5 6 7
Height of soil above spring line (m)
Figure 6.13. PE vertical deflection percent using hyperbolic tangent soil modulus with power
bulk modulus for both ANSYS and CANDE.
Moser (1994) studied the behavior of 1200 mm diameter PE pipes using the Utah
State University pipe-soil cell shown in Figure 1.6, with an SM soil at different compaction
levels. The results of 90% compaction were reported in Moser (1994) and the associated soil
model parameters in Moser (1990). Figure 6.14 shows the finite element model for soil-pipe
system used in the Utah State University soil-pipe cell. This soil pipe system was analyzed
85
using ANSYS and CANDE level 3 analyses. Figure 6.15 shows the results of these analyses
compared with Moser's (1994) experimental results. According to this figure, both ANSYS
and CANDE results showed a good agreement with Moser's experimental results up to a soil
depth of cover of 9.0 m. CANDE overestimated the pipe deflection as the applied load
increased above 9.0 m. This is due to the fact that one soil element around the pipe reached
the shear failure. The difference between CANDE and ANSYS in dealing with the soil
element modulus may cause this difference. CANDE deals with a different modulus for each
soil element, so any local soil failure will affect the pipe results more than that of ANSYS,
where each soil block that consists of a number of elements has an average modulus. ANSYS
results showed a good agreement with the results of Moser with a negligible large deflection
theory effect up to 12.2 m of soil cover. The large deflection theory showed a better
agreement with Moser's experimental results than the small deflection theory of ANSYS for
the soil cover of more than 12.2 m, where the vertical deflection percent is between 3% and
4%. The maximum difference between ANSYS small and large deflection analyses was 12%.
ANSYS small and large deflection theories were used to compare the performance of
both PE and PVC pipes for deeply buried pipes. A 610 mm diameter flexible pipe with
different soil covers up to 18 m above the pipe spring line was modeled using ANSYS with
both small and large deflection theories. The material and cross sectional properties for
The PVC pipe properties used are specified in ASTM F-678-89, while the PE pipe
properties used are taken from CPPA (1997) and meet ASSHTO M252 and M294. The soil is
modeled using the hyperbolic tangent soil modulus with power bulk modulus model. The SM
86
S m
-V
-T.T, .
-•
R= 600mm 5.5 m
mm.
- •:.-<>:-= -v", _ _ ,
~'h:*2 j?fi
%mm
u: ::-tc^
-rr^-rrK
1R:'
•:-: :i-:5S3W% v\ ---'-
3.7 m
Figure 6.14. Finite element mesh used to model Utah State University soil-pipe cell,
10.0
9.0 — - — Moser (1994)
8.0 - CANDE
- - • - - ANSYS, Large
I 7.0
6.0
- O- - ANSYS, Small
J 5.0
1 4.0
1 3.0 7W2
S 2.0
3^"
£ 1.0 rr#
0.0
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0
Height of soil cover (m)
Figure 6.15. Vertical deflection percent with respect to soil cover for PE pipe compared with
Moser (1994).
87
soil parameters for these models are shown in Table 6.2. The pipe-soil system shown in
Figure 5.1 was used to investigate the effect of the large deflection theory as the height of
soil cover increased for both PE and PVC pipes. The results of the pipe vertical deflection
percent are presented in Figure 6.16. This figure shows that the PE pipe exhibited a greater
deflection percent than the PVC pipes. The difference between the small and the large
deflection theories is greater for PE pipes than for PVC pipes. PE pipes are more sensitive to
the consideration of the large deflection theory than PVC pipes. Figure 6.16 also shows that
the large deflection theory effect becomes significant, more than 10%, for vertical deflection
percents of 4% or more.
9.0
_ 8.0
|7.0
6.0
|4.0
« 3.0
|2.0
1.0
0.0
0 5 10 15 20
Soil height (m)
Figure 6.16. Vertical deflection percent with respect to soil cover above spring line.
construction process. If any material is added to the finite element model, ANSYS option of
"element birth and death" can be used. The model with all elements and nodes was generated
88
and the elements which were added to the system in the subsequent load steps were set to the
inactive mode. The status of an element in ANSYS, active or inactive, can be changed from
one load step to another but not within a load step. The improved code reads the number of
active elements for each load step and calculates the modulus for the active elements only. In
this improved code, NN is the total number of construction lifts, nn is the soil lift under
construction for the current load step, and nn-1 is the lift added to the system during the last
load step. At the beginning of the construction of the nn soil lift, the CANDE construction
technique assumes that displacements exist for the nn-1 lift but no displacement is considered
for the nn lift. So lift nn does not experience any pre-existing settlement. This creates a
displacement discontinuity between the soil lifts. To avoid this discontinuity, large deflection
theory was used in ANSYS to consider the effect of the deflection of soil elements during the
nn-1 lift construction. Therefore using the large deflection theory, the new soil elements
(reactivated elements) added to the system as soil lift number nn will not be in their
originally specified geometry but will be influenced by the deflection of the previous
construction steps.
The initial soil stresses at the first sub-step are dependent upon whether the element
is part of the initial pipe soil system (pre-existing soil) or part of the new construction
increment. If the element belongs to a new construction, the initial 03/Pa is assumed to be
To validate the use of the improved program, an embankment construction of the pipe
soil system shown in Figure 5.1 was modeled and solved using both ANSYS and CANDE.
The numbers shown on soil elements in Figure 5.1 are the construction process steps. Five
load steps of soil layers were added to the embankment as numbered in Figure 5.1. The soil
89
was modeled using the hyperbolic tangent modulus model with the power bulk modulus
model for SM soil parameters shown in Table 6.2. Figure 6.17 shows the pipe invert and
crown deflections as the soil is added to the embankment using both ANSYS and CANDE.
This figure shows a good agreement between ANSYS and CANDE results. Figure 6.18
presents the vertical deflection percent for different load steps using both CANDE and
ANSYS. This figure shows that both ANSYS and CANDE give the same trend and vertical
deflection percent. This comparison further validates the improved soil models used in
ANSYS and the procedure used in ANSYS to model the construction process. The effect of
the displacement discontinuity in CANDE has a negligible effect on the pipe deflection
during construction.
ANSYS
- CANDE
0
Load Step
Figure 6.18. Vertical deflection percent for 1200 nun PE pipe vs. construction step for
ANSYS and CANDE.
91
7. CONCLUSIONS
CANDE is one of the most commonly used programs for buried pipe analysis;
however, the limitations of CANDE such as neglecting large deflections, leads to the
program used in structural engineering, has been modified to provide a better analysis tool.
There is little difference in the results calculated by both ANSYS and CANDE for the case of
The improved code in ANSYS used to model the soil material behavior showed a
good agreement with CANDE results for pipes with depth of burial up to 6.1 m. This,
therefore, validates the use of improved soil models in ANSYS for buried pipe analyses.
For cases of 9 m or more of soil cover, using CANDE for deeply buried pipes, with
SM soil which is described as having less quality than most soils used for pipe applications,
did not show a good agreement with either ANSYS or the results provided by Moser (1994).
As the soil elements approach shear failure, CANDE over-predict the pipe deflections.
ANSYS large deflection theory produces a better agreement with 1200 mm diameter
PE pipes whose vertical deflection percent is around 4%. The maximum difference between
ANSYS small and large deflection theories is 12%. The difference between ANSYS small
and large deflection theories for PVC and PE pipes started to be significant at 4% vertical
deflection. Both analyses shown in Figures 6.15 and 6.16 suggest that considering large
The written code was further modified to accommodate the modeling of the
construction process. This was accomplished by using the features of "element birth and
death" in ANSYS and by reading the number of active elements for each load step within the
92
improved code to calculate the modulus for the active elements only. This also showed a
good agreement with CANDE. The effect of displacement discontinuity of CANDE during
In general, CANDE is adequate for shallow buried pipes (up to 6 m for SM soil).
93
8.1 Introduction
improve the constitutive models for HDPE pipe to include strain rate and time effects. This
part seeks to develop the appropriate constitutive models for HDPE material.
independent of loading rate and time duration of the applied load, provided that the load is
sufficiently small to maintain a linear stress-strain response. For other than linear elastic time
materials respond to loads with a significant departure from that of linear elastic time
independent materials. Thermoplastic pipes show creep and stress relaxation under constant
the direction of the applied load. Stiffness may be considered to have two components: that
due to the material and that due to the form of shape of the element. The material component
of stiffness that acts to restrain deformation is the modulus. The geometric component of
stiffness also acts to restrain deformation. This component can be moment of inertia, cross
sectional area, and/or length. For time independent material, flexural compression and ring
compression elastic moduli are close in value. It is inappropriate to assume that this applied
Since the parallel plate test deflection rate is not related to the practical loading rate,
the time effect is not considered, and flexural compression dominates the pipe behavior in the
95
parallel plate test, compression tests conducted by Zhange and Moore (1997) were used to
develop the nonlinear strain rate and time dependent HDPE tangent modulus models.
The objective of this part is to develop mathematical models using the data available
in Zhange and Moore (1997) discussed in chapter 1. The mathematical models describe the
tangent modulus of HDPE material considering both strain rate and time effects. The
mathematical models were programmed in ANSYS and used in a finite element analysis. A
case study of the effect of decreasing HDPE modulus on pipe deflection was performed to
validate the use of the ANSYS program with the HDPE moduli models and to investigate the
effect of using both small and large deflection theories of ANSYS on HDPE pipe buried in
SM soil.
The hyperbolic mathematical model suggested by Kondner (1963) and Duncan and
Chang (1970), used to model the time independent nonlinear soil response, is used in this
chapter to develop mathematical models to describe the strain rate and time dependent
response of the HDPE material. The stress strain curves and the creep curves reported by
Zhange and Moore (1997) provided the raw data used to develop these mathematical models.
The stress strain relationship shown in Figure 1.16 was modeled as part of this
research using the hyperbolic mathematical model given in Equation 8.1. The derivative of
the stress-strain equation is the slope or the HDPE tangent modulus which is shown in
Equation 8.2. Equation 8.1 can be linearized, as shown in Equation 8.3. This relationship
represents a line with an intercept (a) related to the initial modulus and a slope (b) related to
96
the strain rate, as shown in Figure 8.1. The change of initial modulus (EO with strain rate was
modeled using a power function, as shown in Figure 8.2 and Equation 8.4. The slope (b) of
the normalized stress strain fits was mathematically modeled as shown in Equation 8.5 and
Figure 8.3. By substituting the initial modulus (1/a) and the slope (b) shown in Equations 8.4
and 8.5 respectively into Equation 8.2, the tangent modulus model of HDPE pipes as a
function of stress level and strain rate can be determined as shown in Equation 8.6. This
<7 = (8.1)
a + be
—= E a _ (l-6cr)-
(8.2)
de ' {a + be) 1 a
— = a + be (8.3)
a
E i = - = 29053Pa(—)-°1207 (8.4)
a Ae
\ -0.0537
6 = 0.0243(—) (8.5)
At
where:
a = the intercept of the normalized stress-strain line,
b = the slope of the normalized stress strain line.
£ =strain.
a = stress, MPa.
As/ At = strain rate.
Et = tangent modulus at different stress levels, MPa.
Ej = initial tangent modulus, MPa.
Pa = reference atmospheric pressure.
97
strain relationship shown in Figure 8.1 can also be modeled using lines intersecting at a focus
point. The focus point was found by determining the point of intersection of the normalized
stress strain lines of ÎO'/sec and 10"2/sec strain rates and assuming it as an initial guess for
the focus point of all normalized stress-strain lines, as shown in Figure 8.4. Then, using the
MS Excel solver to minimize the least square error between the measured data and the new
intersected lines, the common focus point of all lines was found as shown in Figure 8.5. The
statistical analysis of the focus point method showed a coefficient of correlation (R2) of
different lines in the range of 0.99. Equation 8.7 describes the linear relationship of stress to
strain ratio versus strain for the case of strain rate of 0.1/sec using the focus point technique.
The slope of the lines constructed using the focus point versus the strain rate was fit to a
power mathematical model as shown in Figure 8.6. Equation 8.8 shows the mathematical
model of the HDPE material tangent modulus using the focus point approach. It is
0.009
Strain rate
0.008
0.007
«= 0.006
S
S 0.005
|0.004
# 0.003
0.002
0.001
0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16
Axial strain
Figure 8.1. Normalized stress strain compression test results on HDPE material.
98
4.4
4.3
4.2
£
B 4.1
3.9
3.8
0 1 2 3 4 5 6
Log (1/Strain rate)
0.05
g 0.04 ir
•O
N
IS 0.02
I
û 0.01
0
0.0E+00 2.0E-02 4.0E-02 6.0E-02 8.0E-02 1.0E-01 1.2E-01
Strain rate
Figure 8.3. Change of normalized stress-strain lines slope with strain rate.
99
5.E-03
4.E-03 • Strain rate = 10M
n Strain rate = 10*-2
3.E-03
£ 2.E-03 J3*
l.E-03
C/3
-l.E-03
-2.E-03
-0.1 -0.05 0 0.05 0.1 0.15
Strain
Figure 8.4. Intersection of normalized stress-strain lines for 10"' and 10"2 / sec strain rates.
0.006
Focus poult
(-0.0542,-0.001084)
-0.001
-0.002
-0.1 -0.05 0 0.05 0.1 0.15
Strain
hypothesized that the focus point is related to the residual stresses and strains during the
(1 -[0.02447(—)"°05256 a]}2
E. = Af (8.8)
{-0.001084 + 0.05418[0.02447(—) ~0 05256 ]}
At
Equation 8.8 was used to calculate the tangent modulus value for HDPE material
under compression at different strain rates, as shown in Figure 8.7. This figure shows that the
tangent modulus of HDPE at high strain rate is greater than that at less strain rates. This
figure also shows that the tangent modulus is decreasing linearly but slowly and independent
x> 0.03
co 0.02
Figure 8.6. Change of slope of normalized stress-strain lines with strain rate using the
focus point.
101
3000
• Strain rate= 0.1/sec
2500
— — Strain rate= 0.01/sec
2000 + Strain rate= 0.001/sec
- - Strain rate= 0.0001/sec
"g 1500 - — — Strain rate= 0.00001/sec -
1000
06
H 500
The parallel plate test results shown in Figure 1.14 can also be used to fit a hyperbolic
model. The normalized deflection load diagram of Figure 1.14 is shown in Figure 8.8. Since
only two different rates were used, the change of slope and intercept with deflection rate
A hyperbolic model was also used to fit the creep strain curves shown in Figure 1.19.
A linear function was used to fit the normalized creep strain data as shown in Figure 8.9.
Figure 8.10 shows the results of using the focus point technique explained above which
results in a coefficient of correlation (R2) ranging between 0.97 and 0.99. The slope of the
focus point lines versus the stress level was fit to a power mathematical model, as shown in
Figure 8.11. The tangent modulus considering the time effect was derived using the same
102
technique described in section 8.3.1. The resulting tangent modulus considering time is given
by Equation 8.9.
0.3
0
0 5 10 15 20
Deflection percent
Figure 8.8. Normalized stress-strain results of parallel plate test.
1.E+06
Stress (MPa)=
l.E+06
1.E+06
c
| 8.E+05
CO
13
E 6.E+05 *15.2
e 19.1
4.E+05
2.E+05
O.E+OO
0 2000 4000 6000 8000 10000 12000
Time (sec)
Figure 8.9. Normalized time vs. creep strain relationship for six different stress levels.
103
1.4E+06
Stress (MPa)=
1.2E+06
1.0E+06
8.0E+05 15.2
E 6.0E+05
4.0E+05
2.0E+05 ; j/1-' *
(-1150.67i14387)
0.0E+00
-4000 -2000 0 2000 4000 6000 8000 10000 12000
Time (sec)
Figure 8.10. Normalized creep strain curves using the focus point.
160
140
120
100
0 5 10 15 20 25
Stress (MPa)
Figure 8.11. Change of the slope of focus point lines with stress level for creep tests.
104
where:
A: Y coordinate of the focus point of the time creep strain curve.
B: X coordinate of the focus point of the time creep strain curve.
Go,: initial strain at the current stress level.
Sargand et al. (2001) and Lars Janson (1996) concluded, based on field tests, that the
pipe deflection stabilizes in a period of six to eight weeks. Based on these results, in this
research the HDPE modulus was extrapolated to a time of six weeks instead of 50 years.
Table 8.1 shows the extrapolated values of the modulus considering the time effect and using
Equation 8.9. This exploration showed an average decrease of the HDPE modulus of 77% at
six weeks.
Table 8.1. HDPE modulus change with stress level and time.
The nonlinear strain rate and time dependent HDPE tangent moduli developed in
section 8.3 were programmed in the ANSYS code which is used to model the nonlinear time
independent soil properties and which was developed in Part II of this thesis. For each sub-
105
step, the HDPE modulus is calculated for each pipe element based on the element
compression stresses and strains at the end of the previous sub-step. The sub-step is solved
using the average modulus of all pipe elements. At the end of each load step an average value
of the pipe modulus for all sub-steps is calculated and used to solve for the pipe response.
To validate the use of strain rate and time dependent HDPE tangent modului
programmed in the ANSYS code, the pipe soil system shown in Figure 5.1, which was used
to study the effect of pipe material shown in Figure 6.16, was used to compare the results of
HDPE pipe deflection using a constant and variable HDPE pipe modulus. The soil used in
this study is an SM soil with parameters given in Table 6.2. This soil is described as having
lesser quality soil than most soils used for plastic pipe applications. The pipe modulus
calculated using Equation 8.8 was employed in this analysis. The applied distributed load
was increased linearly with time for each load step. The results of ANSYS small and large
deflection theories using constant and strain rate dependent pipe moduli are shown in Figure
8.12. For both cases, this figure shows that considering the strain rate dependent HDPE
modulus increased the pipe deflection. The difference between pipe vertical deflection
percent using constant and strain rate dependent HDPE pipe modulus varies between 6% and
11%. Differences up to 32% were noticed between the small deflection theory solution using
elastic pipe properties and the large deflection theory solution considering the strain rate
dependent pipe properties. This shows that both large deflection theory and strain rate
dependent pipe properties need to be considered for deeply buried HDPE pipes. The large
deflection solution using the strain rate dependent HDPE modulus did not converge for the
case of 18.25 m soil cover. This is due to the fact that increasing the load and decreasing the
Small, constant E,
Small, variable E
A Large, constant E
X Large,variable E
0 10 15 20
Soil height (m)
Figure 8.12.Vertical deflection percent with respect to soil cover using constant and
strain rate dependent HDPE modulus.
The axial strains in the beam elements (i.e., the pipe walls) were also investigated to
determine where they lie on the stress strain diagram for HDPE material. The maximum axial
strains in the beam elements were observed at the pipe spring line. The calculated average
percent strain, which is the average of strains at the end of each load step, varies between
0.07% at shallow depths and 1.1% for deeply buried pipes. These strains correspond to
vertical deflection percents of 0.6% and 7% respectively. The calculated strains in the beam
elements are below the 4% limit which defined region A in section 1.2.3.2.4 of this
dissertation with reference to the compression test results of Zhange and Moore (1997)
9. CONCLUSIONS
Since the parallel plate test deflection rate is not related to the practical loading rate
and the time effect is not considered in parallel plate tests, compression tests conducted by
Zhange and Moore (1997) were used to develop strain rate and time dependent HDPE
tangent modulus models. Two different methods were used to model the dependency of the
tangent modulus on the loading rate. These two methods were the independent slope-
intercept method and the focus point method. The effect of modulus time dependency was
also included in the focus point model. The pipe modulus as a function of time was
extrapolated to six weeks instead of 50 years. This extrapolation showed a modulus reduction
between 71% and 82%. The mathematical models were programmed in ANSYS and used in
a finite element analysis. The effect of decreasing HDPE modulus on pipe deflection was
studied using both small and large deflection theories of ANSYS. Comparing the results of
both constant and variable HDPE pipe modulus for the case of SM soil, a maximum
difference of 11% of the vertical deflection percent was noticed at soil cover of 15 m.
Differences up to 32% were noticed between the small deflection theory solution considering
elastic pipe properties and the large deflection theory solution considering the strain rate
dependent pipe properties. This shows that both large deflection theory and strain rate
dependent pipe properties need to be considered for deeply buried HDPE pipes.
108
This research consists of the following tasks that were completed: a literature review,
a nonlinear strain rate and time dependent tangent modulus of HDPE material using the
results of compression test by Zhange and Moore (1997) which were programmed in
ANSYS. The following conclusions can be formulated from the investigation of in-service
HDPE pipes, finite element analysis of flexible pipes, and the development and study of the
strain rate time dependent modulus models for HDPE material. It should be noted that some
different locations, and 12,006 feet in length was completed; the structural
diameter increases.
3. HDPE pipes buried in regions where the native soil is glacial till or
The investigation showed that in most cases the MN. DOT specifications
were followed. The limited significant structural problems are due to the
compare the pipe behavior in the elastic soil case with CANDE. The results
of these analyses were well compared with CANDE. The large deflection
theory did not show a significant effect on the response of flexible pipes for
hyperbolic bulk modulus models. The small and large deflection theories of
ANSYS showed good agreement with CANDE. Two different soils, two
different pipe materials, and two different soil models were used in the
comparison. This validates the use of ANSYS with the improved code to
model pipe soil systems. Large deflection theory did not show a significant
effect in case of nonlinear time independent soil modulus for soil heights
CANDE and the small and large deflection theories of ANSYS were
CANDE can be used as long as the shear failure of the soil was not
reached. The small and large deflection theories of ANSYS showed better
agreement with Moser results for soil depth of 9 m or more. ANSYS large
theory with Moser results in cases where vertical deflection percent of 4%.
110
8. The response of PVC and HDPE pipes were compared using small
and large deflection theories of ANSYS. PE pipes are more sensitive to the
10. The written code in ANSYS was also improved to consider the
construction process. The new code was compared with CANDE. The
insignificant.
11. The mathematical models improved for the HDPE material were
deflection theory using constant and variable HDPE pipe modulus resulted
theory solution considering elastic pipe properties and the large deflection
properties using SM soil as a backfill. This shows that both large deflection
Ill
theory and strain rate dependent pipe properties need to be considered for
The global significance of the findings of this research are: 1) based on observations
of in-place installations, the granular backfill materials around the HDPE pipes resulted in
significant for vertical deflection percent of 4% or more, 3) CANDE is adequate for shallow
buried pipes and over predicts the pipe deflections if soil elements approached the shear
failure, 4) linear elastic pipe material properties and small deflections are not adequate to
APPENDIX A.
Project Information:
No. of segments
Site Characteristics:
Native soil
Backfill material
Pipe characteristics:
Installation:
Surface loads
Pipe Grade:
Design grades
Other observations:
Segment Grade:
Other observations:
oo
oo
Location 0- 30- 40- 50- 60- 70- 80- 90- 100- 110 -
1
Wall crushing
Wall cracking
Joint
separation
Sediments
APPENDIX B.
This project which has a 24 inch diameter dual wall ADS HDPE pipe, 329 feet in
length with a slope of 0.23%, is buried at an average depth of 11.3 feet. The native soil is a
limestone rock. This pipe was installed in 1999 and surveyed July 24, 2000 from MH 37-3
to MH 37-6. Figure B-l shows a schematic diagram of this project. Slightly elliptical cross
section deformations with vertical shortening were observed over 125 feet that started 170
feet from MH 37-3. Water infiltration was observed at two joints 36 and 132 feet from MH
37-3 respectively, as shown in Figure B-2. The arrow in Figure B-2 is pointing to the
infiltration location. Sediments and water were observed over 49% of the pipe length.
MH37-3 MH37-6
O O
Length: 329 feet
Diameter: 24 inch
Slope: 0.23%
Figure B-l. Schematic diagram of the pipe surveyed at Woodland Avenue, Mankato.
This project which has a 24 inch diameter dual wall ADS HDPE pipe, 295 feet in
length with a slope of 0.42%, is buried at a depth of 3.1 feet. The native soil in this region is
peat. The pipe was installed in 1996 and surveyed July 24, 2000 from MH B-30 to MH A-9.
Figure B-3 shows a schematic diagram of this project. Figure 3.11 shows one of the two 12
inch cracks (about 0.04 inch (1 mm) wide) between "9 and 12 o'clock" that were observed at
180 and 190 feet from MH B 30. Three joints (at 63, 103, and 207 feet from MH B 30) of
the 14 joints have root penetration. Sediments and water were observed over 83% of the pipe
length.
MH B-30 MH A-9
o o
Length: 295 feet
Diameter: 24 inch
Slope: 0.42%
Figure B-3. Schematic diagram of the pipe surveyed at Thompson Street, Mankato.
This project which has 24, 42, and 48 inch diameter ADS HDPE pipes, 1,886 feet in
length with slopes ranging from 0.2% to 0.44%, is buried at depths ranging from 1.2 feet to
11.5 feet. The native soil is fine sand of glacial outwash origin. These pipes were installed
in 1997 and surveyed July 26, 2000 from MH 119 to MH 202. In addition, observations
were made at the ground surface between MH 201 and MH 202. Figure B-4 shows a
schematic diagram of the pipe surveyed at Bronson drive and Belle lane. The video survey
revealed significant elliptical deformation with horizontal shortening of about 2 inches in the
118
48 inch diameter pipe 28 feet from MH 109 to 360 feet toward MH 202. Other problems
noted are about 0.4 inch (10 mm) joint separation at 170 feet from MH 201 toward MH 202
as pointed by the arrows in Figure B-5, and two steel rods penetrating the pipe wall at 133
feet from MH 201 toward MH 202 as shown in Figure B-6. Water and sediments were
observed over 86% of the 48 inch diameter pipe. No significant problems were noted in the
Diameter: 24 24 42 48 48 inch
a. MH 119 to MH 112.
, No pavement ,
h H
Length: 43 100 360 267 feet
Diameter: 48 48 48 48 inch
b. MH 112 to MH 202.
Figure B-4. Schematic diagram of the pipe surveyed at Bronson Drive, Mounds view.
119
Figure B-5. Joint separation of 0.4 inch at 170 feet from MH 201.
Figure B-6. Two rods penetrating the pipe wall at 133 feet from MH 201.
Surface observation of the 48 inch diameter pipe between MH 201 and MH 202 with
1.7 feet of soil cover revealed a pipe heaved above the ground surface. Associated with the
heave was a joint separation and outside circumferential crack 75 inch long and about
120
0.04 inch (1 mm) wide as shown in Figures 3.9, 3.10, and 3.11. The location of the heave,
about 30 feet from the outlet into a marsh, suggests that buoyancy from a high water table
combined with shallow depth of cover most likely caused the heave.
This project which has 24 and 30 inch diameter ADS HDPE pipes, 1,354 feet in
length with slopes varying between 0.15% and 0.27%, is buried at depths ranging from 3.6
feet to 5.6 feet. This pipe was installed in 1995 and surveyed on July 26, 2000 from MH 7 to
MH 1. Figure B-7 shows a schematic diagram of the surveyed pipe. The native soil is fine
sand with the water table 3 to 4 feet deep. According to Blaine City specifications, this soil
can be used as a foundation for the pipe after dewatering. Localized wall buckling of less
than 1 inch at the pipe crown was noted over 118 feet of 460 feet of the 30 inch diameter
pipe. These deformations were between 110 to 160 feet and 200 to 218 feet measured from
occurred as a dimpling pattern similar to that shown in Figures 1.7 and 3.5. Sediments were
MH 7 CB 6 CB 5 CB 3 CB 2 MH 1
o o o o o o
Length: 364 316 214 218 242 feet
Diameter: 24 24 24 30 30 inch
Figure B-7. Schematic diagram of the pipe surveyed at Xylite Street, Blaine.
121
This project which has 30 and 36 inch diameter HDPE pipe, 1,528 feet in length was
surveyed December 27, 2000 from MH 1 to MH 6 and 251 feet beyond toward the outlet of a
lake. No plans were available for this site. Figure B-8 shows a schematic diagram of the
Ottawa Street pipe. The survey was stopped at 251 feet from MH 6 toward the lake outlet
because of ice ("No access") as shown in Figure B-8. Joint separations of 1 inch wide and 7
inch long at "7 o'clock" were observed 7 feet from MH 1 and also at 256 feet from MH 2.
Pipe sag of 1 to 2 inch in the 30 inch diameter pipe was observed over a 2 feet length starting
at 54 feet from MH 3. Figure B-9 shows elliptical cross sectional deformations with vertical
shortening at a joint 227 feet from MH 3 toward MH 4 which has a 30 inch diameter pipe.
Another elliptical cross sectional deformation with vertical shortening of 3 to 4 inches over
105 feet in length was observed in the 36 inch diameter pipe; this deformation started 50 feet
from MH 5. Horizontal diameter shortenings were noticed at two locations: at a joint 277
feet from MH 2 toward MH 3 that has 30 inch diameter pipe and also in the 36 inch diameter
pipe (1 inch horizontal shortening over 22 feet) starting at 312 feet from MH 5. Heart shape
pipe deformations with 1 to 2 inch crown deformation in the 36 inch diameter pipe was noted
at a joint 23 feet from MH 5. Deformations of the plastic connection between two pipe
shown in Figure B-10. All of these connections were between MH 5 and the end of the
survey toward the lake outlet. In this region, an average of 2 to 3 inches of water above 1
inch of sediments was noticed. A layer of ice was also observed in the region between MH 6
and the end of survey toward the lake outlet. The ice started at 214 feet from MH 6 and
122
extended for 37 feet toward the lake outlet. Sediments and water were observed over a
length of 1,314 feet, which is 86% of the total pipe length surveyed at this site.
MH 1 MH 2 MH 3 MH 4 MH 5 MH 6 lake
o o o o—o O—fO
No access
Length: 108 311 290 30 538 251 feet
Diameter: 30 30 30 36 36 36 inch
Slope: NA NA NA NA NA NA
Figure B-8. Schematic diagram of the pipe surveyed at Ottawa Street, Le-Center.
Figure B-9. Deflection in 30 inch diameter pipe at a joint 227 feet from MH 3.
123
Figure B-10. Joint separation and water infiltration due to deformation of the piece
connecting two pipes.
This project which has 12, 24, 27, 30 and 36 inch diameter HDPE pipe, 2,231 feet in
length with a slope ranging from 0.32% to 6.6%, buried at depth ranging from 1.65 feet to
8.36 feet was surveyed December 21,2000 from MH OR 215 to MH 7. This project consists
of four different installations. Schematic diagrams of these installations are shown in Figures
B-ll, B-12, B-13, and B-14. Cross sectional deformation with horizontal shortening of 1
inch in the 36 inch diameter pipe was noticed at a joint 132 feet from MH OR 211 toward
MH OR 210. Dimpling (localized wall buckling pattern) of less than 1 inch at "1 o'clock"
was noticed over 20 feet of 1,143 feet of the 36 inch diameter pipe. These deformations were
located between 425 and 445 feet measured from MH OR 210 toward MH OR 208. Pipe sag
of 1 to 2 inch in the 36 inch diameter pipe was noticed between 12 and 22 feet measured
from MH OR 215 toward MH OR 214. Two dents and two joint separations were also
observed. The two dents were 1 inch deep at 33 feet from MH OR 215 toward MH OR 214
124
MH 2 MH3
O O
Diameter: 24 inch
Slope: 0.32%
Figure B-ll. Schematic diagram of the pipe surveyed at Shoreacres Drive, Fairmont.
Diameter: 36 36 36 36 36 36
a. MH OR 204 to MH OR 214.
Diameter: 36 30 30 30 27 inch
b. MH OR 214 to MH OR 221.
Figure B-12. Schematic diagram of the pipe surveyed at Highland Avenue, Fairmont.
125
while the two 1 inch joint separations were at 266 feet from MH OR 210 and 88 feet from
shortening of 2 inches in the 12 inch diameter pipe were observed over 10 feet between 30
and 40 feet from MH 5 toward MH 6 as shown in Figure B-15. An average of a half inch of
water was observed over 80% of the total pipe length of this site, while 11% of the length
Slope: 0.38%
Figure B-13. Schematic diagram of the pipe surveyed near Interlaken Road, Fairmont.
MH 4 MH 5 MH 6 MH 7
o o o o
Length: 120 152 75 feet
Diameter: 12 12 12 inch
Wlope: NA NA NA
These installations which have 24, 30, and 48 inch diameter HDPE pipes, 1,941 feet
in length with a slope ranging from 0.03% to 0.39%, buried at depths ranging from 1.6 feet to
6.5 feet were surveyed January 17, 2001 from MH A to MH 29 and continued to a lake
outlet. This project consists of four different installations. Schematic diagrams of these
installations are shown in Figures B-16, B-17, B-l8, and B-19. The survey was stopped as
shown in Figure B-16 for unknown reasons and no plans were available. A 15 inch
circumferential crack, 1 inch wide, extended between "9 and 3 o'clock" at 30 feet measured
inches in the 24 inch diameter pipe, which represents an average of 46% deformation of pipe
diameter, were observed over 10 feet length that started 30 feet from MH A as shown in
Figure 3.3. Associated with the deformations, were 1 to 2 inch dimples and pipe wall cracks
at "2 o'clock" as shown in Figures 3.5 and 3.6. Pipe sag of 8 to 10 inches in the 24 inch
127
diameter pipe starting at 34 feet measured from MH A extended for 6 feet. Elliptical cross
sectional deformation with horizontal shortening of 1 inch in the 48 inch diameter pipe was
noticed over 22 feet starting 190 feet from CB 22 toward CB 20. A 4 inch pipe sag in the 48
inch diameter pipe extended over 104 feet between CB 23 and CB 24 and also a 2 inch sag
over 5 feet starting 440 feet from CB 20 toward CB 19 were noticed. Sediments, water, and
snow were observed over 1,750 feet, which is 90% of the total pipe length.
MH A No access MH B
a -O
No pavement
Length: 40 feet
Diameter: 24 inch
Slope: 0.29%
CB 19 CB 20 CB 22 CB 23 CB 24
o o o o o
Diameter: 48 48 48 48 inch
Figure B-17. Schematic diagram of the pipe surveyed at Railroad Avenue, Gaylord.
128
MH 7 MH 6 MH5B
o <
o •
o
No pavement
Diameter: 30 30 inch
MH 31 MH 30 MH 29 outlet
o o o o
< •
No pavement
Diameter: 48 48 48 inch
This project which has 15 and 18 inch diameter HDPE pipe, 346 feet in length with a
slope ranging from 0.76% to 4.73%, is buried at an average depth of 4.0 feet. This pipe was
surveyed March, 2001 from MH 170A to MH 167. Figure B-20 shows a schematic diagram
of the pipe surveyed at this project. The survey was stopped twice in this survey because of
129
ice ("No access") as shown in Figure B-20. Elliptical cross section deformations with
horizontal shortening of 1 inch were observed over 129 feet of the 18 inch diameter pipe
from MH 169 to MH 168. Elliptical cross sectional deformations with horizontal shortening
were also observed at two joints, one in the 15 inch diameter pipe and one in the 18 inch
diameter pipe. These joint deformations were located 21 feet from MH 169 toward MH 170,
and 35 feet from MH 168 toward MH 167, respectively. Pipe wall deflections of 1 to 2
inches between "9 and 11 o'clock" were also observed over 1 foot at 104 feet from MH 169
toward MH 168, and at a joint 14 feet from MH 170A toward MH 170B. Sediments and/or
Diameter: 15 15 15 18 18 inch
Figure B-20. Schematic diagram of the pipe surveyed at Jonquil Lane, Maple Grove.
This project which has 15 and 18 inch diameter HDPE pipe, 687 feet in length with a
slope ranging from 0.5% to 6.66%, is buried at average depth of 4.5 feet. This project was
surveyed March, 2001 from MH 106 to an exit manhole. A schematic diagram of the pipe
surveyed at this site is presented in Figures B-21 and B-22. The survey was stopped because
of ice ("No access") at the location shown in Figure B-21. Elliptical cross sectional
deformations with vertical shortening of 1 to 2 inches were noticed at two joints in the 15
130
inch diameter pipe, 101 feet from MH 104 toward MH 105 and 116 feet from MH 104
toward MH 103. Pipe wall deflections of 3 to 4 inches between "1 and 3 o'clock" and 1 to 2
inches at the pipe crown were observed in the 15 inch diameter pipe. These deflections were
over 2 feet, each starting at 84 and 159 feet from MH 102 toward MH 101, respectively.
Figure 3.7 shows a pipe wall deflection in the 15 inch diameter pipe. As shown in Figure 3.8,
wall crushing (tearing) associated with a hole and two cracks (both 0.08 inch (2 mm) wide
and 10 inch long) were noticed between "9 and 12 o'clock " in the 2 foot region starting 112
feet from MH 102 toward MH 101. A vertical offset of 2 inches in the 18 inch diameter pipe
was observed in a joint 40 feet from MH 101 toward the exit manhole. Water 1 to 2 inches
Diameter: 15 15 15 15 15 18 inch
Figure B-21. Schematic diagram of pipe 1 surveyed at 77* Place Lane, Maple Grove.
MH 104 MH 107
O O
No Pavement
Length: 97 feet
Diameter: 15 inch
Slope: 0.5%
Figure B-22. Schematic diagram of pipe 2 surveyed at 77th Place Lane, Maple Grove.
131
This project which has 15,18, and 24 inch diameter HDPE pipes, 1,409 feet in length,
is buried at an average depth of 4.5 feet. No plans are available for this project. This project
was surveyed March, 2001 from MH 7 to MH 1. Figure B-23 shows a schematic diagram of
the pipe surveyed at this site. The survey was stopped because of large amount of sediments
("No access") shown in Figure B-23. Elliptical cross sectional deformations with horizontal
region 10 feet long) was 27 feet from MH 7 toward MH 6 in the 15 inch diameter pipe as
shown in Figure 3.2; associated with this deformation, were a 2 inch joint vertical offset and
feet from MH 4 toward MH 3 in the 18 inch diameter pipe and at a joint 31 feet from MH 2
toward MH 1 in the 24 inch diameter pipe. Elliptical cross sectional deformations with
vertical shortening of 2 inches were observed at two joints (54 and 92 feet from MH 2 toward
MH 1) in the 24 inch diameter pipe. Deflections at the pipe crown of 1 to 2 inches over 4 feet
followed by 1 inch dimpling over 4 feet at the pipe crown were observed starting 280 feet
from MH 6 toward MH 5 in the 15 inch diameter pipe. Pipe wall deflections were observed
between "9 and 11 o'clock" at three different locations. These deflections were over 4 feet
of 2 to 3 inch deflection in the 15 inch diameter pipe as the arrow points in Figure B-24, over
1 foot of 1 inch deflection in the 18 inch diameter pipe associated with a 0.08 inch (2 mm)
wide and 10 inch long crack, and over 2 feet of 1 inch deflection in the 24 inch diameter pipe
associated with two cracks; each crack was 1 inch wide and 10 inches long. These
deformations were located at 266 feet from MH 6 toward MH 5,121 feet from MH 4 toward
noticed at three locations. These sags in the 24 inch diameter pipe were over 3 feet starting
76 feet from MH 3 toward MH 2, over 3 feet starting 15 feet from MH 2 toward MH 1, and
over 12 feet starting 166 feet from MH 2 toward MH 1. Associated with the pipe sag at 166
feet, were a hole and two cracks 0.12 inches (3 mm) wide and 15 inches long, between "9
and 12 o'clock". Joint separation of 1 inch was observed at 146 feet from MH 5 toward MH
4 associated with a hole at the pipe crown. Two vertical joint offsets of 2 inches were
observed at 30 feet from MH 7 toward MH 6 and at 100 feet from MH 3 toward MH 2. One
of these vertical joint offsets is shown in Figure B-25. A utility line penetrating through the
pipe wall was observed between "10 and 1 o'clock" at 202 feet from MH 5 toward MH 4.
Water and/or sediments were observed over 100% of the pipe length.
CB 7 MH 7 MH 6 MH 5 MH 4
o—o o——^-o o
No access "
Length: 51 94 288 424 feet
Diameter: 15 15 15 15 inch
a. CB 7 to MH 4.
MH 4 MH 3 MH 2 MH 1 Exit
o o o o o
Length: 97 164 235 56 feet
Diameter: 18 24 24 18 inch
b. MH 4 to exit.
Figure B-23. Schematic diagram of the pipe surveyed at 96lh Place North, Maple Grove.
133
B.ll Summary
Data from all the sites surveyed are summarized in Tables B-l, B-2, and B-3. In these
tables the site properties of pipe diameter, pipe length, pipe depth, design grade, structural
problems, and sediments observed for all projects surveyed are presented. The structural
problems noted in these tables are also summarized in Table 3.2. Sediments are mentioned in
more detail in these tables. The manholes numbers in these tables correspond to the manholes
numbers shown in the various site schematic diagrams presented in this appendix.
Table B.I. Perfomunce evaluation for Ihe titci lurvcycd in July, 2000.
Project MI I-Mil Diameter Length Ave. depth Design Cross -section Pipe sag Joint seperated Wall cracking Wall buckling Sediments and/or Water
grade dcfoimation in pipe
(inch) (feet) (feel) (%>
Mankato MH B30A-MH A9 24 295 3.1 0.42 3 Two 12 in. 60ft, 0-lin. sed, below 1.5 water
Thompson roots going cracks less or 1600,1-2 sed., below 0.5 water
street through equal 1 mm 25ft,0-lscd,below 0.5 water
Minksto MH 37-3-MH 37-6 24 329 11.3 0.23 •1", 125 ft 2 60ft, 0-lin. Sed
Woodland Av. water Infiltration 80ft, 0-lin. sed, below lin. water
80ft, l-2ln. sed, below lin. water
Moundsview MH1I9-MHII8 24 473 8 0.44 240,0-1 sed.
Belle lane
and M1I1I8-MHII5 42 342 11.5 0.3 38ft,1-2 in. sed.
Branson drive
Mil 1IS-MII113 48 280 6.8 0.2 273 ft, 1-2 in.sed.
MH 109-Mll 202 48 627 1.2 0.218 ••2", 332 ft 2 900,0-1 sed below 1 in water
900,2-3 sed U>
100ft,l-2 sed
Ui
1400,2-3 sed below 1 in water
450,1-2 sed
950,2-3 sed, below 1-2 In water
Blaine MH 7-CB 6 24 364 5 0.27 870, 1-2 in sed,
Xylite Street 1150, 2-3 in sed,
20ft, 0-1 in sed.
Project Mll-MH Diameter Length Ave depth Design Cross -section Pipe sag Joint seperated Wall cracking Wall buckling Sediments and/or Water
grade deformation in pipe
(inch) (feet) (feel) (14)
Le Center MHI-MH2 30 108 N/A N/A 1 8 ft,0-1" sed
Otaawa street 100 ft 0-1" sed below 0-1' water
MH2-M1I4 30 601 N/A N/A •3-4" at joint 1-2",2ft 1 110 ft.0-1 "water
••1"at a joint 20ft, I'sed.below I' water
90 (1,0-0 5' water
137 ft, 0 1" sed. below 0-1" water
64ft.l"aed below 2" water
MH4-MH6 36 568 N/A N/A • 3*4",105 ft 5(1) 180ft,I'sed below 12"water
••I".22 ft 300 ft.0-1 "sed
1-2" at joint (3) 88 M"sed below 1" water
MH6-outlet 36 251 N/A N/A 5(0 180 ft, 1" sed below 2-3'water
37 ft,4"watcr below 1 " ice
Gaylord MI1A-MHB 24 40 N/A 0 29 10-12",ion 8-10',6 ft 1-2",18 cracks 1-2", I0ft«* 33 ft, 1 in sed
1"-15" length 6 ft. 8-10 in. sed
MH7-MH6 30 351 5.45 0.36 351 It0-1'sed. and water
MH6-MH3B 30 148 46 03 50ft, I'sed
9811.0 5" sed and water
Railroad Avenue CB23-CB24 48 104 N/A N/A 4",104ft 104ft. I'sed below0 5" water
CB23-CB22 48 58 16 005 58 ft. I'sed belowl'water
CB22-CB20 48 258 25 0.39 "I",22 ft 127 ft, 1" sed below 1" water
127 ft,0 5" sed below 0.5" water
CB20-CBI9 48 510 65 0031 l-2",5 ft 445 ft, 1" sed below 1" water
65(1.0 5" sed and water
MH30-MH3I 48 286 N/A 017 35 ft,0 5" sed and water
95 ft, 1" aed below 1" water
66(1.1" sed below 2" water
MH30-oullet 48 186 N/A 024 90 ft, 1-2" snow
* Eliptical déformation with vertical shortening (1) Deformation of the connection piece between two pipes
** Elliptical deformation with horizontal shortening (2) Ml12 different than the one above
Dimling local buckling (3) heart shaped pipe
Table B 3. Performance evaluation for the sites surveyed March, 2001.
Project MH-MH Diameter Length Ave depth Design Cross -section Pipe Sag Joint scperatcd Wall crocking Wall buckling Wall crushing Sediments and/or Water
grade deformation in pipe
(inch) (feet) (feet) <%)
Maple Grove MHI70A-MH I70B IS 33 4 0.74 •1-2" at joint 33 (1,3-4* sed
Jonquil Line MHI70A-MH 170 IS 23 3.5 259 23 ft,3-4" sed.
MHI69-MHI70 IS 40 4 4.21 •• 1"at joint 33 ft, 0 5* water
7 ft, 2-3"sed below 0 5" water
REFERENCES
American Society of Testing Material, 1992. "Standard Practice for Underground Installation
of Thermplastic Pipe for Sewers and other Grsvity Flow Applications". ASTM D2321.
American Association of State and Highway Transportation Officials, 1992. " Standard
Specifications for Highway Bridges". American Association of State and Highway
Transportation Officials, Washington D C.
Bums, J. O., and Richard, R. M., 1964. "Attenuation of Stresses for Stresses for Buried
Cylinders". Proceedings, Symposium on Soil-Structure Interaction, pp. 378-392. University
of Arizona, Tucson, Ariz.
Chambers, R. F., McGrath, T. J., and Heger, F. J., 1980. "Plastic Pipe for Subsurface
Drainage of Transportation Facilities". NCHRP. Report 225, Transportation Research
Board, pp. 122-140. Washington, DC.
Corrugated Polyethylene Pipe Association, 1997. "Structural Design Method for Corrugated
Polyethylene Pipe".
Duncan, J. M. and Chang, C. Y., 1970. " Nonlinear Analysis of Stress and Strain In Soils".
Journal of the Soil Mechanics and Foundations Division. Proceedings of the American
Society of Civil Engineers, Vol. 96, No. SM5, pp. 1629-1654.
Duncan, J.M., Byrne, P., Wong, K.S., and P. Mabry, 1980. "Strength, Stress-Strain and Bulk
Modulus Parameters for Finite Element Analysis of Stresses and Movement in Soil Masses".
Report No. UCB/GT/80-01. University of California, Berkeley.
El-Sawy, K., Moore, I D., and B. Taleb, 1997. "Analysis of Stability Limit States for Long
Span Culverts". 5Cfh Canadian Geotechnical Conference of the Canadian Geotechnical
Society, pp. 178-184.
Gabriel, L.H, Ster, J. F., and Anthony, B. 2001. "A Test apparatus for Time Independent
Stiffness of Thermoplastic Pipe". PREPRINT. Transportation Research Board. Washington
DC.
Gabriel, L.H. and Goddard, J. R., 1999. "Curved Beam Stiffness for Thermoplastic Gravity
Flow Drainage Pipes". Transportation Research Record 1656, pp 51-57. Washington D C.
139
Goddard, J. 1999. "Curved Beam Stiffness Test and Profile wall stability". PREPRINT.
Transportation Research Board. Washington D C.
Greenwood, Mark E. and Lang, Dennis C., 1990. "Vertical Deflection of Buried Flexible
Pipes". Buried Plastic Pipe Technology. Philadelphia, PA: ASTM.
Hartley, James D. and Duncan, James M.1988 "E' and Its Variation with Depth," ASCE
Journal of Transportation Engineering, pp. 538-533, Vol. 113, No 5.
Hashash, N., and E. T. Selig, 1990. "Analysis of the Performance of a Buried High Density
Polyethylene Pipe". The First National Conference on Flexible Pipes, Columbus, Ohio, pp
95-109.
Howard, Amster K. 1977. "Modulus of Soil Reaction Values for Buried Flexible Pipe".
Journal of the Geotechnical Engineering Division, pp 33-43.
Katona, M. G., 1976." CANDE: A Modern Approach for the Structural Design and Analysis
of Buried Culverts". Report FHWA-RD-77-5. FHWA, U.S. Department of Transportation.
Katona M.G., 1990. "Minimum Cover Height for Corrugated Plastic Pipe Under Vehicle
Loading". Transportation Research Record 1288, ppl27-135. Washington D C.
Klaiber, F.W.,Lohnes, R.A., Wipf, T.J., and Phares, B.M., 1996. "Investigation of High
Density Polyethylene Pipe for Highway Applications: Phase I", Iowa Department of
Transportation Project HR-373, Ames, Iowa.
Krizek, R. J., et al. 1971. "Structural Analysis and Design of pipe Culverts," Report No. 116,
National Cooperative Highway Research Program, Highway Research Board, Washington,
DC., 155 pp.
Lohnes, R.A.,Wipf, T.J.,Klaiber,F.W., Conard, B E., and Ng, K.W 1997. "Investigation of
Plastic Pipes for Highway Applications: Phase II", Iowa Department of Transportation
Project HR-373A, Ames, Iowa.
140
Marston, Anson, and A. O. Anderson, 1913. " The Theory of Loads on Pipes in Ditches
and Tests of Cement and Clay Drain Tile and Sewer Pipe," Bulletin 31, Iowa Engineering
Experiment Station, Ames, Iowa.
Marston, A. and A. Anderson 1930. "The Theory of External Loads on Closed Conduits in
the Light of the Latest Experiments," The Iowa State College bulletin, No. 96, volXXVIH,
Iowa Engineering Experiment Station, Iowa State College.
McGrath, Timothy J. and Selig, Ernest T. 1994, "Backfill Placement Methods Lead to
Flexible Pipe Distortion." PREPRINT. Transportation Research Board, Washington, D. C.
Meyerhof, G. G. and Baike, L. D. 1963. "Strength of Steel Culvert Sheets Bearing against
Compacted Sand Backfill". Highway Research Board Proceedings, vol. 30.
Moore, I.D, 1995. "Three Dimensional Response of deeply Buried Profiled Polyethylene
Pipe". Transportation Research Record 1514, pp 49-58. Washington D C.
Moore, I D., and B. Taleb, 1999. " Metal Culvert Response to Live Loading Performance of
Three Dimensional Analysis". Transportation Research Record 1656, pp. 37-44. Washington
DC.
Moser, A. P.1990. "Buried Pipe Designs". McGraw Hill, Inc., New York.
Moser, A.P, 1994. "The Structural Performance of Buried 48 inch diameter N-12 HC
Polyethylene Pipes". Buried Structural Laboratory, Utah State University, Logan, Utah.
Moser, A.P, 1997. "Analytical Methods for Predicting Performance of Buried Flexible
Pipes". PREPRINT. Transportation Research Board, Washington D C.
Moser, A.P., 1998. "The Structural Performance of Buried Profile-Wall HDPE Pipe and The
Influence of Pipe Wall Geometry". Transportation Research Record 1624, pp 206-213.
Washington D C.
Morgan, B.C., Lohnes, R.A., Klaiber F.W., and B. H. Kjartanson, 1996. " Design Method for
Preventing Longitudinal Uplift of Corrugated Metal Pipe". Transportation Research Record
1541, pp. 107-111. Washington D C.
Musser, S.C, 1989. "CANDE-89 User Manual". Report FHWA-RD-89-169. FHWA, U.S.
Department of Transportation.
Rogers, C.D.F., Fleming, P.T., and Talby R„ 1996. "Use of Visual Methods to Investigate
The Influence of Installation Procedure on Pipe-Soil Interaction". Transportation Research
Record 1541, pp 76-85. Washington DC.
Sargand, S.M., Hazen, G.A., and Masadda, T. 1998. "Structural Evaluation and Performance
of Plastic Pipe". Report FHWA-OH-98-Ol 1. FHWA, U.S. Department of Transportation.
Sargand, S.M., Hazen, G., White, k., Moran, A., 2001. "Time-dependent deflection
ofthermoplastic pipes under deep buried burial". Transportation Research Record 1770, pp.
236-242. Washington D C.
Sargand, S.M., Masada, T., and Schehl, D, 2001. "Soil pressure measured at various fill
heights above deeply buried thermoplastic pipes". Transportation Research Record 1770, pp.
227-235. Washington D C.
Schluter C. J. and Shade J.W, 1999, " Flexibility Factor or Pipe Stiffness, Significant
Stiffness Considerations. Transportation Research Record 1656, pp 45-50. Washington D C.
Selig, E.T., 1988. "Soil Parameters for design of buried pipelines". Pipeline infrastructure,
proceeding of pipeline infrastructure conference of ASCE, ASCE, NY.
Selig, E. T. 1990. "Soil Properties for Plastic Pipe Installations," Buried Plastic Pipe
Technology, ASTM STP109. American Society for Testing and Materials, Philadelphia.
Spangler, M.G, 1941."The Structural Design of Flexible Pipe Culverts". Iowa Engineering
Experiment Station. Bulletin 153. Ames, Iowa.
Spangler M. G., and Handy R. L. 1982. Soil Engineering. Harper and Row, New York, NY,
4 Edition.
Taleb, B., and I D. Moore, 1999. "Metal Culvert Response to Earth Loading- Performance of
Two Dimensional Analysis". Transportation Research Record 1656, pp. 25-36. Washington,
DC.
Watkins, Reynold K. and Spangler, M.G. 1958. "Some Characteristics of the Modulus of
Passive Resistance of Soil: A Study in Similitude," Highway Research Board
Proceedings,Vol. 37, pp. 576-583.
Watkins, R. K., Anderson, L.R, 1999. "Structural Mechanics of Buried Pipes". CRC Press,
New York.
Watkins, R. K., 1999. "Performance Limits of Buried Profile Wall Pipe" PREPRINT,
Transportation Research Board, January, 2001. Washington D C.
142
Webb, M.C., Sussmann, J., and Selig, E.T., 1998. "Large Span Culvert Field Test Results".
NCHRP Project 12-45. Department of Civil and Environmental Engineering, University of
Massachusetts, Amherst.
Zhange C and Moore, I.D, 1997. "Nonlinear Mechanical Response of High Density
Polyehtylene. Part I.: Experimental Investigation and Model Evaluation". Polymer
Engineering and Science, Vol. 37, No.2 pp 404, 413. Washington D C.
143
ACKNOWLEDGMENTS
I would like also to thank my major professors, Robert Lohnes and Terry Wipf for
their guidance and input during the course of this project. Thanks to Professor Wayne
Klaiber for his help. I also would like to thank Professor Brian Coree for his continuous help,
for serving on my program of study and teaching me "Theory of Elasticity" where I got an
A!.
This research was funded by the Minnesota Concrete Pipe Association (MCPA). I
would like to thank MCPA for funding this project and providing the opportunity to
investigate the performance of in-service HDPE pipes. I would also like to thank the city
officials in Mankato, Moundsview, Blaine, Le-Center, Fairmont, Gaylord, and Maple Grove
(all in Minnesota) for allowing us to video-tape the in-service pipes and providing the design