See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/326622636

Buried flexible pipes: Design considerations in applying as2566 standard

Article  in  Australian Geomechanics Journal · June 2018

CITATION READS

1 8,314

2 authors:

Burt Look Donald Anthony Cameron

Foundation Specialist Group University of South Australia
52 PUBLICATIONS   94 CITATIONS    71 PUBLICATIONS   873 CITATIONS   

SEE PROFILE SEE PROFILE

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

Model for prediction of resilient modulus incorporating matric suction for recycled unbound granular materials View project

Accomodating trees in a suburban landscape on clay soils View project

All content following this page was uploaded by Burt Look on 17 April 2019.

The user has requested enhancement of the downloaded file.

 BURIED FLEXIBLE PIPES:
DESIGN CONSIDERATIONS IN APPLYING AS2566 STANDARD

Burt G. Look1 and Don A. Cameron2

1
Foundation Specialists Pty Ltd, Brisbane, 2University of South Australia

ABSTRACT
AS/NZS 2566.1 (1998) standard for buried pipes neither represents the current state of practice nor is it accurate with
respect to ground parameters provided therein. The values referred to as soil modulus in this Standard provides confusion
in the industry. The standard incorrectly equates an embedment modulus (which is a soil – pipe modulus reaction) and a
native soil modulus. These are then combined into an effective soil modulus. Typically for flexible pipes, over 85% of
its calculated performance is dependent on the installation procedure which affects the soil stiffness. The soil-pipe
stiffness is secondary to the soil stiffness. The equation used in calculating the pipe deflection has limitations. The
Standard is also currently dated in terms of guidance on allowable deflection and properties for the types of pipes currently
available in the market.
1 INTRODUCTION
Standards are used to guide a given process and to provide uniformity in the product. However, standards need to be
updated to represent current state of practice. When this does not occur, the designer may be exposed legally in either
trying to apply an inappropriate standard or use a “non-standard” approach which although based on sound engineering
principles may be viewed as non-conforming. The current standard for buried flexible pipelines (AS/NZS 2566.1, 1998)
is one such example which presents these issues.
In buried pipe design, stresses and deflections are caused by factors including soil dead loads, live loads, soil variability
and differential settlements. Strength is used to resist stress, while pipe ring stiffness (RS) is the ability of the pipe to
resist deflection. Rigid pipes are designed to accommodate the stresses imposed based on a class of pipe and/or haunch
and side support provided during installation, while flexible pipes are designed to limit deflections. Flexible pipes are
those that can deflect at least 2 percent of pile diameter without structural distress while a pipe is considered rigid if it
deflects less than 2 percent. An appropriate rigid or flexible standard is then applied for design, although some pipes may
fall into either category (a semi-rigid pipe). The design and installation of buried pipes then varies based on:
• The strength and stiffness of the pipe material selected (rigid vs flexible);
• The live and dead load applied to the pipe, which vary with the shape, width and depth of the embedment;
• The foundation material below the pipe;
• The embedment materials (embedment zone);
• The native soil supporting the embedment soils laterally.
This paper discusses the issues associated with the current Australian and New Zealand Standard for flexible pipes.
However, it should not be assumed that the standards applicable for rigid pipes are without limitations. For example,
Matyas (1983) examines the validity of existing rigid pipe standards and compares various methods for predicting vertical
earth loads on pipes. The various methods give a wide range for the load factor, with the variability increasing for the
height of fill to pipe width (H /B) ratio. A case study relating to the current Australian standard for rigid pipes was
presented by Look and Tandjiria (2012). They showed that the Australian Standard for rigid pipes provides little guidance
on the loads for rigid pipelines placed on rigid supports, when surrounded by settling soft ground. The increased rigidity
of the system relative to the soft soil attracts greater load to the buried pipe as the adjacent soft ground settles, while the
pipe on a rigid support experiences negligible movement.
For flexible pipes, Cameron (2005) provides the background of various methods for pipe design and highlights a number
of uncertainties including the applicability of pipe stiffness and the acceptable deflection criteria for pipes.
It is essential that the principles of pipe - soil interaction are understood and correctly applied in the case of flexible pipes,
as there is a significant scope for the Standards for buried pipes to be applied inappropriately. This paper provides the
relevant background, and illustrates some current issues for the Standards for flexible buried pipes, that have had legal
implications.
Comparison between designs from applying the Australian Standard directly with outcomes from the use of alternative
“standard” methods is provided in this paper to illustrate the limitations of the AS/NZS 2566.1 (1998), the current
Australian and New Zealand Standard for flexible pipes.

AUSTRALIAN GEOMECHANICS | VOLUME 53: NO.2 JUNE 2018 101

BURIED FLEXIBLE PIPES: DESIGN CONSIDERATIONS IN APPLYING AS2566 STANDARD
LOOK AND CAMERO

2 DESIGN OF FLEXIBLE PIPES

The design of flexible pipes in Australia and New Zealand is governed by the standard AS/NZS 2566.1 – 1998 Buried
flexible pipelines Part 1: Structural design (Standards Australia, 1998). Design requires checks on deflection, strength
and buckling. The latter 2 considerations are considered structural considerations while the deflection analysis requires
geotechnical considerations and is discussed herein.
Matthews (2013) compares the differing design considerations and installation requirements contained in the relevant
Standards on buried pipes, as well as product acceptance and expectancy. Pipes are classified as rigid (e.g. steel reinforced
concrete pipes), semi-rigid (e.g. fibre reinforced concrete pipes) and flexible (e.g. plastic and metallic pipes). In rigid
pipes, strength is used to resist the wall stresses from the external and internal loads, while for flexible pipes stiffness is
used to resist ring deflection and possible buckling.
A flexible pipe typically supports only a fraction of an imposed load, and relies on adequate side wall earth pressure to
limit its deflection. A large portion of the imposed load is transferred from the pipe wall to the surrounding soil. This
limits the vertical defection while balancing with an associated horizontal deflection.
Often construction issues may govern the performance of buried flexible pipes. The installation of flexible pipes needs
to be carefully considered as excessive compaction of backfill over the pipe or too large a compaction equipment may
cause ovalling of the pipe (reduced vertical diameter as the lateral diameter increases). Compaction to the side of the pipe
is important to get right; under-compaction reduces lateral support when the pipe is in service, while too vigorous side
compaction causes reverse ovalling as observed by Cameron (2005), which can be beneficial if properly controlled. The
influence of side soil compaction on pipe deflection has been modelled by Elshimi and Moore (2013), albeit semi-
empirically.
External loads on pipes include dead and live loads. To determine dead load due to soil above the pipe, AS 2566 uses the
weight of the column of soil acting on the projected area of the pipe. This is the “prism load” and is a conservative
approach as it assumes the total weight of soil and other loadings is transferred to the pipe, thus ignoring the contribution
of friction between the trench walls and the trench. The effect of surface loads (live and dead loads) diminishes rapidly
as the depth of cover increases over a buried pipe.
The buried pipe deflects according to the ratio of the load on the pipe and the combined stiffness of the pipe and
surrounding soil. The predictive equation for deflection can be expressed simply as:
𝐿𝑜𝑎𝑑 𝑜𝑛 𝑃𝑖𝑝𝑒
𝑃𝑖𝑝𝑒 𝐷𝑒𝑓𝑙𝑒𝑐𝑡𝑖𝑜𝑛 =
𝑃𝑖𝑝𝑒 𝑆𝑡𝑖𝑓𝑓𝑛𝑒𝑠𝑠 + 𝑆𝑜𝑖𝑙 𝑆𝑡𝑖𝑓𝑓𝑛𝑒𝑠𝑠
This is a simplification of the factors in the Iowa formula and its variations, which are widely used for predicting
deflections of buried flexible pipes. Howard (1977) discusses the variation of soil stiffness, E', with soil type and degree
of compaction. This is not a traditional soil stiffness value. The E' used is a modulus of soil reaction which has a
complicated soil-structure interaction specific to pipes, and which is based on the geometry of the embedment zone, the
embedment soil surrounding the pipe and the native soil surrounding the embedment zone (soil + placed soil). Howard
(1977) provides back calculated values of E' based on data from over a hundred field installations. These results apply
only to the initial deflections measured soon after construction and so design standards generally factor E' in line with
expectations of long term performance.

3 ISSUES WITH THE DESIGN STANDARD OF FLEXIBLE PIPES

The predicted long-term deflection has to satisfy an allowable deflection which depends on the type of pipe. There are 3
fundamental issues in this standard which provides confusion in the assessment of pipe deflection:-
1) Allowable Deflection. The typical material characteristics provided in the Standard do not allow for many
current types of pipes used in industry.
2) Fundamental formula used to predict the deflection. Various researchers have found failings in the prediction
equation adopted as the basis for the deflection equation.
3) Predicted Deflection. This is derived from a “system” soil modulus (E) which is based on the “Native Soil”
Modulus (E'n) and the Embedment Modulus (E'e), as well as the geometry of the installation. AS2566 equates
E'n and E'e. This is incorrect, even from the first day of the publication in 1998, and has led to much confusion
in its application in practice

102 AUSTRALIAN GEOMECHANICS | VOLUME 53: NO.2 JUNE 2018

BURIED FLEXIBLE PIPES: DESIGN CONSIDERATIONS IN APPLYING AS2566 STANDARD
LOOK AND CAMERON

3.1 ALLOWABLE DEFLECTION

The first issue is derived from current pipe materials not being in the current standards. One has to rely on the
manufacturer of the product for specification of that material. However, the principal author was involved in a legal case
study where the civil designer used the parameters provided by the manufacturer. Following non-performance of the
buried pipe, the developer initiated proceedings on the various parties, including the designer for not carrying out their
own independent assessment. The engineer for the plaintiff argued the designer should have independently obtained the
various parameters to certify the design. Yet a designer typically relies on the information provided (as available in their
product brochures and on a manufacturer’s web site) to be correct.
Thus, composite pipes which have been in use for over a decade, are not explicitly stated as a product in AS 2566 (1988).
Without any “Standard” guidance, designers can face legal proceedings for any non-performance of the pipe even when
they use the product properties as supplied by the manufacture.
Another example of changing pipe products can be seen from the historical upgrades of CANDE (Culvert Analysis and
design). This is a design and analysis program for all types of buried structures which was first released in 1976. The
many upgrades had code restructuring software improvements in terms of processing and interface as well as soil models
(Mlynarski et al. 2007). However, additional design criteria for various types of new pipes were included. This has
continued with additions to the pipe type library in upgrade release in 2011, 2013, 2015 and 2017.
(http://www.candeforculverts.com/history.html).
These examples highlight the need for standards to be updated as many “new” products (in the past 20 years) are not
represented in the current standards. During review of this paper, an Amendment 1 to AS 2566 was released (AS/NS
2566.1/Amdt 1/2017). This provides additional pipe products with its characteristics.
3.2 FORMULAE FOR CALCULATION OF DEFLECTION
The second issue concerns the calculation of the predicted deflection. This is compared with the allowable deflection as
provided in AS 2566 (1988) for the type of pipe and /or the manufacturer. The predicted long-term deflection is derived
from:
∆𝑦 𝐾×10−3 ( 𝑤𝑔 + 𝑤𝑔𝑠 + 𝑤𝑞 )
= (1)
𝐷 8×10−6 𝑆𝐷𝐿 +0.061 𝐸′
where
∆𝑦
= predicted long-term vertical deflection to the diameter at the neutral axis of the pipe wall
𝐷

K = bedding constant (assumed to be 0.1 typically)

wg + wgs + wq = soil dead load + surface applied dead load + surface applied live load (kPa)
SDL = long-term ring-bending stiffness of the pipe = EI/D3 (N/m/m, since I is in m4/m length)
E' = effective combined modulus of soil reaction of embedment material and native soil modulus (MPa)
Equation 1 is a variation of the Spangler Iowa formula. Moser and Folkman (2008) provide the relevant background of
the derivation of the equation. The original formula was for horizontal deflections, but Spangler (1941) assumed the
horizontal (x) and vertical (y) deflections as equal for small deflections, as Δx = 0.913ΔY (Howard, 1977) for pipes of
various sizes experiencing deflections less than 10% of the original pipe diameter.
The values of the deflections used to back calculate E' represents the initial deflection after construction. A deflection lag
factor in the Iowa formula is incorporated in AS2566.1(1998) by taking the value computed by Eqn. 1 as the 30-day
value, and then adjusting for various time intervals when the short term performance deflection measurement is carried
out (AS2566.2, 2002).
The bedding constant varies with the bedding angle (Moser and Folkman, 2008) and its variation is provided in Table 1
and Figure 1.

Table 1: Bedding constant varies with bedding angle (Typical value = 0.1 used with the Modified Iowa Formula).

Bedding Angle ⁰ 0 30 45 60 90 120 180

Bedding Constant K 0.110 0.108 0.105 0.102 0.096 0.090 0.083

AUSTRALIAN GEOMECHANICS | VOLUME 53: NO.2 JUNE 2018 103

BURIED FLEXIBLE PIPES: DESIGN CONSIDERATIONS IN APPLYING AS2566 STANDARD
LOOK AND CAMERO

Figure 1: Bedding angle and trench material

E' is the modulus of soil reaction of the embedment soil in non-trench construction, or in a trench system, which takes
into account the stiffness of the native soil adjacent to the sides of the pipe. E' is not “the effective combined soil modulus”
which is the terminology used in AS2566. E' is the product of the modulus of the enveloping soil (e, MPa/mm) by the
mean radius of the pipe (r, mm). Therefore E' is an empirical parameter and is not an inherent soil property, but
unfortunately it is also referred to as a “soil modulus” or soil stiffness, thereby causing confusion. It is not Young’s
modulus; E' cannot be measured simply either in the laboratory or in the field.
In a trench system, E, the modulus of soil reaction of the embedment soil, and the Young’s modulus of the native soil,
En, affect the modulus of soil reaction, depending on how close the trench wall is to the pipe relative to the pipe diameter.
Commonly, Leonhardt’s correction factor () is applied to evaluate E in these cases, and the correction factor has been
incorporated in AS2566. Leonhardt’s correction factor varies from 0.1 to 5.0 and is equal to 1.0 when E'n = E'e.
Even before the publication of AS2566 (1988), Jeyapalan et al. (1987) found that using the Spangler equation gave
unrealistic predictions of deflections for very flexible pipes embedded in soft backfills when compared with finite element
analysis and field data. Smith and Watkins (2004) also found the Iowa formula over-predicts ring deflection and state
“Designers obtaining values using the Iowa Formula are being subsidized by numerous factors of safety, both in the
equation and in the allowable deflections of standards”.
An alternative and relatively simple approach to estimate deflections of flexible pipes was provided by Dhar, Moore and
McGrath (2002). Dhar et al. (2002) recommended that pipe deflections should take into account both flexure and pipe
shortening. The Iowa or similar equation can provide the estimate of flexural pipe deflection, y, however pipe deflection
due to the developed hoop forces, H, must also be considered. The two pipe deflections are summed to give the vertical
deflection, while the lateral deflection is the hoop force deflection less the deflection due to bending. The Iowa equation
was re-arranged into:
𝐷𝑙 𝐾𝑤
𝜀𝑦 = 𝐸𝐼
(2)
( ⁄ 3 ) +0.061𝑀𝑠
𝑟
where y = Δy /(2R) = vertical diametric strain (deflection) due to flexure
Δy = vertical pipe deflection (mm)
r = pipe radius (mm)
Dl = deflection lag factor = 1 for short term loading, > 1 otherwise
w = the vertical pressure at the spring line (kPa)
Ms = one dimensional soil modulus (MPa)
EI = flexural stiffness of the pipe (N m2)
In this equation, the constrained modulus of the soil, Ms, replaced Spangler’s modulus of soil reaction, E, an issue that
will be discussed later in this paper. Constrained modulus is determined from measuring settlement of soil as vertical load
is applied to a sample in an oedometer and, as such, is a simple laboratory test. As Ms varies with applied stress, the secant
modulus is used in Equation 2, starting from zero stress to the expected vertical stress adjacent to the pipe. Constrained
modulus is related to Young’s modulus by the following equation:

104 AUSTRALIAN GEOMECHANICS | VOLUME 53: NO.2 JUNE 2018

BURIED FLEXIBLE PIPES: DESIGN CONSIDERATIONS IN APPLYING AS2566 STANDARD
LOOK AND CAMERON

𝜎𝑣′ 𝐸 (1−𝜈 )
𝑀𝑠 = = (1+𝜈𝑠 ) (1−2𝜈
𝑠
(3)
𝜀 𝑠 𝑠)

where, v = Effective vertical stress

v = Vertical soil strain
Es = Young’s modulus of the soil
and νs = Poisson’s ratio of the soil
The average deflection of the pipe due to hoop force includes an arching term according to Dhar et al. (2002), which
expresses how much load reaches the pipe. The overall expression for the deflection or strain due to hoop force is:
𝐹𝑉𝐴 𝑤
𝜀𝐻 = (𝐸𝐴)𝑝 (4)
[( ⁄𝑟 )+ 0.57𝑀]

where, H = Average diametric pipe strain due to hoop force

(EA)p = Hoop stiffness per unit length (N/mm)
FVA = Vertical arching factor, given by:
𝑆𝐻 − 1.17
𝐹𝑉𝐴 = 0.76 − 0.71 ( ) (5)
𝑆𝐻 + 2.92

where, SH = Hoop stiffness factor = M r / (EA)p

For flexible pipes, a full slip interface condition is normally assumed as in AS2566 (1998), while a no slip interface is
more common for rigid pipes (McGrath, 1999). For rigid pipes, FVA is approximately 1.4 while FVA can be less than or
equal to 1.0 for flexible pipes. The Australian Standard for buried flexible pipes recommends the full prism load (γ H Do)
to the crown be applied in design (Figure 2), i.e. FVA = 1, but neglects the soil load between the spring line and crown.

Figure 2: Soil prism load over buried pipe

Equations (2) and (4) represents the hoop compression (circumferential shortening) and the conventional bending
deformations, respectively. Dhar et al. (2002) reported good correspondence between three laboratory pipe tests under
simulated embankment loading, and this approach. However, it could be argued that the method, although an
improvement on the Iowa equation, is not rigorous enough. Moore (1993) stated that hoop deflection is controlled by the
isotropic pressure around the pipe, while the flexural deflections are generated by the deviatoric pressure (Figure 3).
However, his approach was considered too complex to implement in routine practice.

AUSTRALIAN GEOMECHANICS | VOLUME 53: NO.2 JUNE 2018 105

BURIED FLEXIBLE PIPES: DESIGN CONSIDERATIONS IN APPLYING AS2566 STANDARD
LOOK AND CAMERO

Vertical stress, pv

Soil

Pipe
Horizontal
stress, ph

Vertical stress, 0.5(pv + ph) Vertical stress, 0.5(pv + ph)

wo -wd

-0.5(pv + ph)
(pv + ph)

wd
= + 

ISOTROPIC DEVIATORIC

Figure 3: Partitioning of stresses and deflections on a buried flexible pipe (after Moore, 1993)

3.3 GEOTECHNICAL CONSIDERATIONS

The more direct geotechnical issue with AS2566 is the input parameters adopted. Table 3.2 in AS 2566.1 (1998) as
summarized by Table 2 in this paper, applies to both E'n and E'e, which is implausible as the first term is a modulus of soil
reaction, while the latter is Young’s modulus. Duncan and Bursey (2013) recognised the confusion caused by using
modulus as a term by itself and, in particular relation to this paper, recommended that the term modulus of soil reaction
should be replaced with “coefficient of soil reaction”.
At first glance the values provided in the Table are confusing to a geotechnical engineer, who is often asked for advice
on likely input parameters. A loose coarse sand and a very dense, well-graded gravel can be expected to have Young’s
moduli of over 5 MPa and 100 MPa, respectively. A very stiff clay with SPT > 50 and a dry density ratio of 100% would
be expected to have a Young’s modulus of over 50 MPa – while no reliable modulus is implied by AS2566. An attempt
to consider the values representing an effective or softened value still leaves one no closer to the values in the Table.
The embedment modulus depends upon the level of compaction achieved while backfilling, but native soil modulus would
need to be established based on tests such as the Standard Penetration Test (SPT). SPT blow counts equivalent to the
various compaction limits are provided in Table 3.2 of the Standard. Firstly, an SPT of 4 – 10 is considered loose for
granular soils with a relative density of 15% to 35%, while Table 3.2 suggests an ID = 50% which is medium dense.
Additionally, similar SPTs in recent fills and native soils do not correspond to the same relative density. Skempton (1986)
provides values of the ratio (No)60 / Dr 2 as 40 and 55 for recent fills and natural deposits, respectively, where (No)60 is the
SPT N value corrected for overburden and energy and Dr is the relative density.

106 AUSTRALIAN GEOMECHANICS | VOLUME 53: NO.2 JUNE 2018

BURIED FLEXIBLE PIPES: DESIGN CONSIDERATIONS IN APPLYING AS2566 STANDARD
LOOK AND CAMERON

Table 2: Embedment and native soil – materials and moduli. Simplified from Table 3.2 of AS2566.1 (1998).

Modulus E'e and E'n (MPa)

for Dry Density Ratio (RD, %) and Density Index (ID, %)
Materials
RD = 85 90 95 100
Un-
USC compacted ID = 50 60 70 80
Description Classification
SPT ≤ 4 5 - 14 15 - 24 25 - 50 >50

Gravel – Single Size GU 5 7 7 10 14

Gravel - Graded GW 3 5 7 10 20

Coarse Grained with GP, SP, SW,

1 3 5 7 14
< 12% Fines GC, SC
Coarse Grained with GM, SP, GC,
> 12% Fines SM, SC NR 1 3 5 10

Fine Grained with > CL, ML

NR 1 3 5 10
25% coarse, LL < 50%
Fine Grained with < CL, ML
NR 1 3 7
25% coarse, LL < 50%
Fine Grained with CH, MH, OH
NR – No Reliable Modulus Value
LL > 50%

‘NR’ implies “no reliable modulus” and this is where geotechnical advice is usually sought. For example, NR is used for
all medium to high plasticity (CI to CH) soils. In such soils, it is often necessary to use first principles in order to assess
the appropriate design value of E'n. It is clearly incorrect to suggest that a hard CH clay has “no reliable modulus”.
However, the intent of not using such clays as embedment materials is consistent with good engineering practice. The
first author’s review of various designs and legal work has found the geotechnical engineer then tries to rationalise a value
relative to the values provided in the Table in AS 2566.1. Some of this rationalisation is discussed below.
Firstly, the inexperienced geotechnical engineer mistakenly assumes “NR” implies a low value, and a lower value than
what has been provided in the Table. This can be far lower than the drained value of the modulus of elasticity as an
equivalent to E'n. A rationalisation may be made to then further downgrade the modulus value to allow for the potential
for future softening of the soil due to wetting from rainfall or changes in water table (including due to leakage from the
pipe itself). Even with these “rationalisations”, the values are still far larger than provided by AS2566. The softening
rationalisation conveniently neglects that a drained modulus is already low and that AS2566 states that the values provided
do already contain a reduction in modulus allowing for groundwater above the pipe for cohesive soil. Another
rationalisation observed by the first author, uses the modulus of subgrade reaction and factoring for the pipe diameter to
arrive at a “modulus”. This also results in too high a value as compared with the values of Table 2.
The explanation is simpler than most can rationalise. The modulus terminology used by AS2566 is misleading and Table
2 must only be applied to provide values of the modulus of soil reaction of the embedment material. Young’s modulus of
the native soil should be chosen from values familiar to geotechnical engineers for long-term loading conditions. AS 2566
is in error by implying that E'e = E'n. Neither Howard (2006, 2015) nor the AWWA standards (1995, 2005) have E' e =
E'n, while AS 2566 implies such equality.
A further geotechnical issue arises with the recommended values of E'e in the standard. Modulus of soil reaction values
for embedment materials were originally based on Howard’s (1977) research, which back-calculated the horizontal pipe
deflections in a given soil. This modulus (which is not Young’s modulus) includes both a pipe deflection and a soil
deflection, and so is not readily obtained. Howard (2006) updated his values to account for vertical and not horizontal
deflections. Consequently, recommended values increased. His recommendations were adopted by the Bureau of
Reclamation (2013) for determination of short term pipe deflection for design purposes. A design reduction factor, Fd,
was incorporated in the design values of embedment modulus of soil reaction, Ee, as Howard’s values corresponded to
average pipe deflections. The Bureau of Reclamation report states, “Using the design factor values shown…., there is a

AUSTRALIAN GEOMECHANICS | VOLUME 53: NO.2 JUNE 2018 107

BURIED FLEXIBLE PIPES: DESIGN CONSIDERATIONS IN APPLYING AS2566 STANDARD
LOOK AND CAMERO

95-percent probability that the actual pipe deflection will not exceed the predicted value by more than 0.5 percentage
points.” Comparisons of recommended Ee values are provided in Figure 4 for coarse grained soils based on Table 2 and
the Bureau of Reclamation (2013) values. In the legend to Figure 4, “AS2566-GW” refers to “well graded” coarse grained
soils. It can be seen that the AS2566 values of Ee are conservative, particularly as compaction level, or dry density ratio,
increases. There is better agreement for fine-grained soils.
The Ee' values as currently advocated by Howard (2015) are provided in Table 3.

30
Modulus of Soil Reaction (MPa)

25

20
AS2566
15
AS2566-GW
10 USBR

0
85 90 95 100
Dry Density Ratio (%)

Figure 4: Comparisons of Ee values for coarse grained soil groups

Table 3: E'e values after Howard (2015): converted to metric equivalent.

Compacted Ee (MPa)

Soil Group Uncompacted Ee (MPa)
Moderate High
Crushed Rock 6.9 41.4
GW, GP, SW, SP 3.5 13.8 27.6
GC, GM, SC, SM 1.4 6.9 17.3
CL, ML 0.7 2.8 10.4
CH, MH, OH, OL, Pt Not recommended for embedment

3.3.1 Some guidance on native soil Young’s modulus

Many Australian pipe manufacturers are silent on the use of the AS 2566 native soil values, although some do state their
assumed native soil values (which are higher than AS 2566). Howard (1996) proposed typical values which have been
adopted by AWWA M45 (2005), and are given in Table 4 as a function of SPT blow count or unconfined compressive
strength. Thus, AS2566 is in direct contradiction with the international literature in implying E'e = E'n.

108 AUSTRALIAN GEOMECHANICS | VOLUME 53: NO.2 JUNE 2018

BURIED FLEXIBLE PIPES: DESIGN CONSIDERATIONS IN APPLYING AS2566 STANDARD
LOOK AND CAMERON

Table 4: Values of E'n for native soils (Howard, 1996 and AWWA M45): Converted to metric equivalent.

Granular Cohesive Native in Situ soil

SPT Unconfined
Description Compressive Strength Description E'n (MPa)
Blows / 300mm (kPa)
0–1 Very very loose 0 – 12.5 Very very Soft 0.34
1–2 12.5 - 25 Very Soft 1.3
Very loose
2–4 25 – 50 Soft 4.8
4–8 Loose 50 - 100 Medium 10.3
8 – 15 Slightly compact 100 – 200 Stiff 20.7
15 – 30 Compact 200 – 400 Very Stiff 34.5
30 - 50 Dense 400 – 600 Hard 69.0
>50 Very dense > 600 Very Hard 137.9

3.3.2 The substitution of constrained modulus, Ms, for E

E' as currently defined is a back-calculated value in an empirically derived equation. E' is not directly measured.
McGrath (1998) proposed that the modulus of soil reaction, E can be directly replaced by the constrained soil modulus
(Ms). This practice is now advocated by AWWA M45 (2005). It seems contradictory to the definition of E', although it
is convenient admittedly, to have a measurable stress-dependent parameter to replace a far less readily measurable one.
McGrath (1998) supported the Ms values recommended for the standard installation direct design (SIDD) method. Of all
the Ms values that had been reported by various authorities, the SIDD values came closest to Howard’s values which had
been derived from extensive field data. The SIDD values of Ms can be found in McGrath et al. (2002). M s depends on
soil type and level of compaction, and increases with stress level. Interestingly, Howard and Howard (2008) reviewed
available field data and concluded that a stress-dependent modulus could not be validated for installations backfilled to
7.62 m, so caution is needed if it is decided to apply stress-dependent Ms in design.

4 ILLUSTRATION OF USING THE VALUES OF En INFERRED IN THE STANDARD

A case study is used first to illustrate the inadequacy of using the AS2566 values for native soil and factoring for the
implied low values. The case presented is a pipe in a well compacted embedment material in different native soils.
Figure 5 compares the calculated deflection by using either the AS2566 values or typical geotechnical Young’s modulus
values for native soils. The example is for a 450-mm diameter with a pipe stiffness of 5,000 N/m/m with 450 mm of
compacted sand (19 kN /m3) on both sides of the pipe. The width of the trench is 1.35 m, which corresponds to a B/D
ratio of 3. The sand has been compacted to 60% density index (medium compaction) and therefore an embedment
modulus, Ee of 5 MPa applies according to AS2566. The imposed load is for 3 metres of soil above the pipe only. A
maximum surface traffic load of 18 kPa applies at that depth based on AS 2566.
4.1 EFFECT OF NATIVE SOIL MODULUS VALUE USED
Typical native modulus values for cohesive soils from Table 4 are compared with the low modulus values implied by AS
2566. Given that AS 2566 shows “NR” and the lowest tabulated value of 1 MPa, then that value has been applied as that
soil’s upper bound value.

AUSTRALIAN GEOMECHANICS | VOLUME 53: NO.2 JUNE 2018 109

BURIED FLEXIBLE PIPES: DESIGN CONSIDERATIONS IN APPLYING AS2566 STANDARD
LOOK AND CAMERO

12% 2.3%

Soft Using AS 2566 Table for Native soil 2.2%

10% 2.2% 10.0%
using AS2566 Modulus of Soil Reaction E'

Using Native Soil Modulus from literature

using Native Soil Modulus from literature

2.1%
8.3%
Calculated Defletion (%)

8%

Calculated Deflection (%)

2.0%

6.3%
6% 1.9%
1.9% 5.2%
1.8%
4.3%
4%

1.7% 1.7%

2% 1.6%

1.6% 1.6%

Soft Firm Stiff Very Stiff Hard

0% 1.5%
0 0.2 0.4 0.6 0.8 1 1.2
0 10 20 30 40 50 60 70
Native Soil Modulus (MPa)

Figure 5: Comparison of deflection of a 450 mm diameter pipe using AS2566 and typical geotechnical modulus
values

For a 7.5% allowable deflection for a plastic pipe, and using the AS2566 values, the calculations of pipe deflection suggest
that a pipe with compacted sand will always be unsuitable in a soft to firm clay. Yet in practice, buried pipes with suitable
compacted backfill are performing adequately in most geomaterials. Thus “AS 2566 theory” does not match practice.
The E' values calculated for the soil-pipe system using Leonhardt’s correction factor and native soil moduli from Table 4
were 5.0 MPa and 7.1 MPa for the soft and hard clay native soil, respectively. If AS 2566 was adopted instead for the
native soil moduli, the corresponding calculated E' values would be just 0.6 and 2.2 MPa.
4.2 EFFECT OF PIPE STIFFNESS
The contribution of the pipe stiffness to the overall denominator (pipe stiffness + soil stiffness) is compared in Table 5.
This shows that the soil properties govern and that the pipe stiffness provide only a minor contribution to the estimated
pipe deflection. The “incorrect” calculated contributions using the AS2566 values are also shown in Table 5.
Table 5: Contribution of pipe stiffness to overall pipe + soil stiffness for example calculated

Soft Firm Stiff Very Stiff Hard

Native Soil (Clay)
Pipe Stiffness Contribution
Using native soil modulus 12% 10% 9% 9% 8%
from common literature
Using native soil modulus 53% 44% 34% 32% 23%
from AS2566

4.3 EFFECT OF HEIGHT OF FILL

The comparison was extended to varying heights of fill that could be sustained by the 450-mm pipe in the same trench,
given the values of native soil modulus used. The native soil in this case was a stiff sandy clay; for the AS2566 estimate
just 1 MPa was assumed for En, while a more reasonable value of 15 MPa was applied, based on the literature.

110 AUSTRALIAN GEOMECHANICS | VOLUME 53: NO.2 JUNE 2018

BURIED FLEXIBLE PIPES: DESIGN CONSIDERATIONS IN APPLYING AS2566 STANDARD
LOOK AND CAMERON

Adopting an allowable long-term deflection of 7.5% as used for polyethylene pipes (AS2566.1: 1988; Table 2.1), Figure
6 shows that a 450 mm pipe cannot sustain a depth of fill over 5 metres if the native modulus of a stiff clay (1 MPa) is
used according to the guidance of AS2566.1. If a native modulus of 15 MPa is used, then a depth of fill of up to 15 metres
can be placed over the pipe. AS2566 becomes more unreasonable when the pipe size or depth of fill increases.

Figure 6: Calculated deflection, using the AS2566 values and typical geotechnical modulus values.
4.4 EFFECT OF PIPE DIAMETER
Table 6 illustrates the influence of pipe diameter using similar conditions as in the initial case study (Figure 5). The native
soil modulus using Table 4 values (as used in the common literature) were again compared with the values provided in
AS2566 – Table 3.2.
Table 6 was constructed by extending the example of a 450-mm pipe to diameters of 300, 450, 900 and 1200 mm and
B/D ratios were 4, 3, 2, 1.75, respectively, for these pipe diameters. The distance from the pipe to the trench wall was
kept constant at 450 mm, which is a typical minimum width for hand compaction equipment. All pipes adopted the same
“typical” stiffness value (5000 N/m/m) and all other factors being the same, including the height of fill (3 m). Note that
the nominal pipe stiffness typically decreases with pipe diameter, and depends on pipe type. A pipe of 1200mm diameter
could be half the pipe stiffness value as compared with that of 30mm. Section 4.2 showed the relative contribution of
pipe stiffness to the calculated pipe deflection.
Table 6: Variation of deflections of pipes of different diameters under the same height of fill, with soil type and
chosen native soil modulus

Native Soil (Clay) Soft Firm Stiff Very Stiff Hard

Pipe Size (mm)
modulus value Calculated Pipe Deflection (%)
300 From literature 2.2 2.0 1.9 1.9 1.9
From AS2566 7.3 5.9 4.6 3.9 3.3
450 From literature 2.2 1.9 1.7 1.6 1.6
(Refer Figure 5) From AS2566 10.0 8.3 6.3 5.2 4.3
900 From literature 2.2 1.6 1.3 1.2 1.1
From AS2566 12.3 10.6 8.3 6.9 5.6
1200 From literature 2.2 1.5 1.2 1.1 0.9
From AS2566 12.9 11.2 8.9 7.4 6.0

AUSTRALIAN GEOMECHANICS | VOLUME 53: NO.2 JUNE 2018 111

BURIED FLEXIBLE PIPES: DESIGN CONSIDERATIONS IN APPLYING AS2566 STANDARD
LOOK AND CAMERO

Based on the calculated deflection, the 300-mm pipe would generally be acceptable for all native soils when the surround
material is moderately compacted, irrespective of values of En adopted.
For a 1200 mm pipe in a compacted gravel and having a very stiff native clay forming the trench, the deflection is marginal
at a fill depth of 3 m, according to AS2566, and the predicted deflection exceeds the performance limit for less stiff soils.
However, predicted deflections based on native moduli from common literature indicate acceptable performance for all
soil types.
Table 6 shows that the pipe diameter (and B/D ratio) does not have much influence, if common literature values are
adopted for the native soil. This is in stark contrast to AS2566, where deflections increase significantly with pipe
diameter. A 900mm or 1200 mm diameter pipe is predicted to exceed an allowable deflection of 7.5% even, when placed
in a stiff native soil with a well compacted embedment material.
4.5 SUMMARY OF EFFECT OF VARIOUS INPUTS USED TO CALCULATE PIPE DEFLECTION
For most pipes used in practice (which are small pipe diameters), the inconsistency in AS 2566 may go unnoticed. As
the pipe sizes or depth increases, this inconsistency becomes more noticeable. These considerations are summarised in
Table 7 for the case study presented herein.

Table 7: Summary of differences using the native soil parameters from AS2566 and from the literature

Consideration in case Differences using Native soils values and compacted embedment material
study example
From literature From AS2566
Calculated deflection in Native soil has minor effect Native soil has significant effect even when embedment
3m trench for 450mm soil is compacted. Pipes exceed allowable deflections in
diameter pipe soft, firm and some stiff soils.
Pipe stiffness Pipe stiffness has minor Pipe stiffness has significant contribution
contribution
Calculated deflection of For 7.5% allowable deflection For 7.5% allowable deflection then maximum height of
450mm diameter pipe then 16 m maximum height of fill is 6 m
with varying height of fill fill applies
Pipe diameter 300 mm to 1200 mm dimeter Pipes of 300mm dimeter is fine in all native soils.
pipes acceptable in all native Pipes of 450mm unacceptable in soft / firm native soils
soils provided embedment
Pipes of 900 / 1200mm unacceptable in soft / stiff soils
material is compacted

4.6 SENSITIVITY OF INPUT PARAMETERS

The sofware @RISK (Palisade 2016) was used to run simulation models based on the alowable deflection equation of
AS2566. A “simulation” combines the uncertainties identified for the model, thus avoiding a single number calculation.
Instead, data variability is used to analyze every possible outcome in “what-if” scenarios. “Simulation” refers to a method
whereby the distribution of possible outcomes is generated by recalculating over and over again, using different randomly
selected sets of values for the probability distributions / range of valid combination of input values.
The tornado graph shows a sensitivity analysis for various inputs, with longer bars at the top representing the most
significant input variables (Figure 7). As previously, the distance from the pipe to the trench wall was kept as a constant
at 450 mm. All other parameters and the variability of each parameter are shown on the Tornado graph. Each likely value
of each parameter was varied by a third. Using a simulation (10,000 analyses), the output values were calculated with the
iterations associated with each variable. A “double-sided” tornado graph is shown in the Figure. The shading indicates
how the input associated with a bar changes in value when the output statistic increases or decreases. The darker section
of the bar shows where input values are "low", and the brighter section where input values are "high".
Thus, when pipe stiffness, embedment modulus or native soil modulus, is “high”, the deflection decreases (Figure 7).
However, when the pipe diameter or height of fill is “high” then the calculated pipe deflection increases. Aside from the
geometric properties of height of fill and pipe diameter, the native soil modulus then governs the design when the values
provide in AS2566 are used. The pipe stiffness provides the least effect based on the outputs rank shown in Figure 7.
The previous case study (Table 5) showed that the pipe stiffness contributes 34% for a stiff clay when using the AS2566
values of native soil modulus. When native soil modulus values from the literature are applied (Table 4), the stiffness
contribution is significantly less (9%).

112 AUSTRALIAN GEOMECHANICS | VOLUME 53: NO.2 JUNE 2018

BURIED FLEXIBLE PIPES: DESIGN CONSIDERATIONS IN APPLYING AS2566 STANDARD
LOOK AND CAMERON

Figure 7: Calculated deflection Tornado Graph using the inferred native soil modulus of AS2566

The relative ranking using soil modulus values from the literature is shown in Figure 8. All other values being the same
as in Figure 7 simulation.

Figure 8: Calculated deflection Tornado Graph using Young’s soil modulus (Table 4)

Figure 8 indicates a similar ranking of inputs when using higher values of modulus from the literature. The main
calculated change in the simulation output is the much lower relative ranking of the native soil modulus input, which is
no longer the governing soil parameter. The embedment soil modulus is the soil input parameter which most affects the
calculated deflection. The pipe stiffness again provides the least effect based on the ranking of outputs shown in Figure
8. The calculated deflection is also shown to be acceptable in all cases for polyethylene pipes which has an acceptable
deflection of 7.5% at 6.7m height of fill.

AUSTRALIAN GEOMECHANICS | VOLUME 53: NO.2 JUNE 2018 113

BURIED FLEXIBLE PIPES: DESIGN CONSIDERATIONS IN APPLYING AS2566 STANDARD
LOOK AND CAMERO

5 CONCLUSIONS
The E′ modulus of soil reaction is a hybrid modulus and is neither a soil modulus in the usual geotechnical sense, nor is
it a parameter which can be directly measured. E is the response of a soil-pipe system as it is loaded and so depends on
embedment geometry, surrounding soil properties and pipe stiffness.
Part of the confusion is the initial reference to both E'n and E'e as “soil modulus” in the Australian Standards AS2566.
Further confusion is the terminology used, as E'e is not a fundamental soil parameter. To avoid confusion in the future, E'
and E'e should be termed coefficients of soil reaction, with the latter term applying to the embedment material.
AS2566 provides a table which equates the embedment zone soil and native soil from soil type and relative compaction.
The E' values represent a combined pipe and soil response, back-calculated from experimental deflections, while E'n is
external to the direct influence of the pipe, and is the native soil Young’s modulus, which is not influenced by the pipe
deflection – but can influence the overall coefficient of soil reaction. Where AS2566 prescribes “NR” to a soil group and
compactive effort, this is commonly interpreted as a very low value. Yet the native soil modulus value for these soils is
several times higher than the value of the embedment coefficient of soil reaction. This is part of the inconsistency that
geotechnical engineers try to rationalise, while still clinging to the values in the Tables provided in AS2566, as they
assume that a Standard cannot be wrong.
Where finite element analyses (FEA) are used to examine soil-pipe interaction problems, values other than those provided
in AS2566 must then be used. Indeed, FEA require Young’s modulus and Poisson’s ratio for both embedment material
and native soil.
For low height fills and small diameter pipes then the native soil modulus value issues discussed herein are not material
to a design based on a calculated deflection criterion.
Typically, for flexible pipes, over 85% of the pipe’s “calculated” performance is dependent on the installation procedure.
The pipe stiffness has the least effect, while the height of fill governs the calculated deflection. Sensitivity analyses show
that if soil parameter values in AS2566 are used, then the calculated deflection is more sensitive to the native soil external
to the trench wall, as compared to the placed and compacted soil surrounding the pipe within the trench wall. This is
intuitively incorrect.
Finally, AS2566 (1998) currently does not represent the state of practice, with some pipe types that are being used not
currently represented within the standard.

6 REFERENCES
AWWA (1996). “Fiberglass Pipe Design”, AWWA Manual No. 45, American Water Works Association, Denver CO.
AWWA (2008). “PE Pipe – Design and Installation,” AWWA Manual No. 55
Bureau of Reclamation (2013). “Methods for prediction of flexible pipe deflections”, M-25 2nd Edition
Cameron, D.A. (2005). “Analysis of Buried Flexible Pipes in Granular Backfill Subjected to Construction Traffic”, PhD
Thesis, University of Sydney.
Dhar, A.S., Moore, I.D. and McGrath, T.J. (2002). “Evaluation of simplified design methods for buried thermoplastic
pipe”. Proc. Pipelines 2002, “Beneath our feet: Challenges and Solutions”, ASCE, Cleveland, OH, July, 11pp.
Duncan, J.M. and Bursey, A. (2013). “Soil modulus correlations”, Proc. Geo Congress 2013, ASCE, San Diego, CA, pp
321-336.
Elshimi, T.M. & Moore, I.D. (2013). “Modelling the effects of backfilling and soil compaction beside shallow buried
pipes”. Journal of Pipeline Systems Engineering and Practice, V4(4), pp 1-7.
Jeyapalan, J.L., S.W. Ethiyajeevakaruna and B. A. Boldon (1987). “Behaviour and design of buried very flexible plastic
pipes”, Journal of Transportation Engineering, Vol 113, No 16, pp 642 – 657.
Jeyapalan, J.K. and Britto, A.M. (2014). “Magical charts for effective E’…Removing the smoke and mirrors” ASCE,
Pipelines 2014: From Underground to the Forefront of Innovation and Sustainability pp 699 – 711.
Howard, A. (1977). “Modulus of Soil Reaction (E') values for buried flexible pipe”, Engineering and Research Centre,
Bureau of Reclamation.
Howard, A. (1996). “Pipeline Installation”, Relativity Publishing.
Howard, A. (2006). “The reclamation E’ table, 25 years later”, Proc. Plastics Pipe XIII International Conference,
Washington DC, USA.
Howard, A. and Howard, B. C. (2008). “A discussion of E′ versus depth”, Proc. ASCE Conference Pipelines 2008,
Atlanta, GA,
Howard, A. (2011), “Constrained Modulus of Crushed Rock for Pipeline Embedment”, Proc. ASCE Conference
Pipelines, Seattle Washington, pp. 300-311
Howard, A. (2015). “Pipeline Installation 2.0”, 2nd edition, Relativity Publishing

114 AUSTRALIAN GEOMECHANICS | VOLUME 53: NO.2 JUNE 2018

 BURIED FLEXIBLE PIPES: DESIGN CONSIDERATIONS IN APPLYING AS2566 STANDARD
LOOK AND CAMERON

Look, B. and Tandjiria, V. (2012). “Foundation considerations for buried rigid pipes on soft ground”, Proc. 11th
Australia New Zealand Conference in Geomechanics, Melbourne, pp 142 – 147.
McGrath T. (1998). “Replacing E' with the constrained modulus in flexible pipe design”, Proc, ASCE Pipeline
Division. Conference, San Diego, CA, pp 28-40.
McGrath, T., Moore I., Selig E., Webb, M. and Taleb, B. (2002). “Recommended Specifications for Large-Span
Culverts”, National Cooperative Highway Research Program, Report M473, Transportation Research Board.
Mosman, A.P. and Folkman S. (2008). “Buried Pipe Design”, 3rd Edition, McGraw Hill Publishers.
Mlynarski, M, Katona M.G., McGrath T.J. (2007). “Modernize and Upgrade CANDE for Analysis and LRFD Design
of Buried Structures”, National Cooperative Highway Research Program Report 619. Transportation Research
Board, Washington.
Matthews D.J. (2013). “A comparison of the design & installation requirements for rigid pipe and flexible pipe”,
Concrete Pipe Association of Australia.
Matyas, E.L. and Davis, J.B. (1983). “Prediction of vertical earth loads on rigid pipes”, ASCE Journal of Geotechnical
Engineering, V109(2), pp 190 – 201.
Moore, I.D. (1993). “Structural Design of Profiled Polyethylene Pipe. Part 1 – Deep Burial”. Research Report,
University of Western Ontario, Geotechnical Research Centre, GEOT-8-93, March.
Moore, I.D. (2001) “Buried Pipes and Culverts”. In: Rowe R.K. (ed), Geotechnical and Geoenvironmental Engineering
Handbook. Springer, Chapter 18, pp 541-568.
Palisade Corporation (2016). “@ Risk – Risk Analysis and Simulation”, Version 7.5
Skempton A.C. (1986), “Standard penetration test procedures and the effects in sands of overburden pressure, relative
density, particles size aging and over consolidation” Geotechnique 36, No. 3, pp 425 – 447.
Smith G and Watkins R (2004). “The Iowa formula – It’s use and misuse when designing flexible pipe”, ASCE Pipeline
Engineering conference, pp 1-7.
Spangler, M. G. (1941). “The Structural Design of Flexible Pipe Culverts,” Bulletin No. 153, Iowa State Engineering
Experiment Station.
Standards Australia (1998). AS/NZS 2566.1:1998 Buried flexible Pipelines Part 1: Structural design, Homebush,
Australia.
Standards Australia (1998). AS/NZS 2566.1:1998 Buried flexible Pipelines Part 1: Structural design – Commentary,
Homebush, Australia.
Standards Australia (2002). AS/NZS 2566.2:2002 Buried flexible Pipelines Part 2: Installation, Homebush, Australia.
Standards Australia (2017). AS/NS 2566.1/Amdt 1/2017: Amendment No. 1 to AS/NZS 2566.1:1998 Buried flexible
Pipelines Part 1: Structural design, Homebush, Australia.

AUSTRALIAN GEOMECHANICS | VOLUME 53: NO.2 JUNE 2018 115

View publication stats