Buried Pipe Failures Dependent on Soil Stiffness, Too

By Agostino Napolitano, Pipeline Department Manager, and Angelina Parlato,

Onshore Pipeline Techical Manager, Saipem SpA

Performance limits and potential failures must be identified for buried pipe design.
Chief among those are excessive deformations of the pipe, wall buckling and
collapse.

Maximum allowable pipe deflection is usually not determined by structural failure

but by conditions such as clearance for pipe-cleaning equipment, special sections,
and pipe appurtenances (attachments). Pipeline designers classify pipes as “flexible”
or “rigid,” depending on how they perform after installation.

Flexible pipe can move, or deflect, under loads without structural damage, while
rigid pipes cannot deflect significantly without structural distress, such as cracking.
From the standpoint of rigidity, pipes were classified by Marston in 1930:

Rigid conduits: The cross-sectional shapes cannot be distorted sufficiently to change

the vertical or horizontal dimensions more than 0.1% without causing materially
injurious cracks.

Semi-rigid conduits: The cross-sectional shapes can be distorted sufficiently to

change the vertical or horizontal dimensions more than 0.1%, but not more than 3%
without causing materially injurious cracks.

Flexible conduits: The cross-sectional shapes can be distorted sufficiently to change

the vertical or horizontal dimensions more than 3% before causing materially
injurious cracks.

In reality, the behavior of a buried pipeline will depend on how its stiffness compares
with the stiffness of the soil in which it is to be buried. The soil-pipe system is
statically indeterminate. As such, the interface pressure between the soil and pipe
cannot be calculated by statics alone.

As soil and surface loads are placed over a buried pipe, the ring tends to deflect -
primarily into an ellipse with a decrease in vertical diameter and an almost equal
(slightly less) increase in horizontal diameter. The increase in horizontal diameter
develops lateral soil support which increases the load-carrying capacity of the pipe.
The decrease in vertical diameter partially relieves the ring of load since the soil
above the pipe takes more of the load in arching action over the pipe.

Soil-Pipe System

The amount of deflection induced by installation that will occur in any buried pipe
depends on three factors: pipe stiffness, soil stiffness, and earth load and surface
load due to construction equipment.

Measurements made by Marston and Spangler equations reveal the load on a

flexible pipe is substantially less than that on a rigid pipe. The level of lateral earth
load also depends on the nature of the backfill and its level of compaction, as well as
the stiffness of the side walls of a trench, if the pipe sits in a trench rather than in
an embankment fill.

Therefore, it should be readily appreciated that the backfill and its construction are
vital to the performance of a flexible pipe. Unfortunately, designers have placed too
much attention on the structural properties of the pipe rather than on the soil.

There are many pertinent variables in the complex interaction of pipe and soil. For
pipes, one widely recognized variable is pipe flexibility, D/t, which is an inverse form
of pipe stiffness, EI/D3, and may be used for analyses involving pipe stiffness of
plain steel pipe (no mortar linings or coatings).

Pipe flexibility (D/t) normally ranges between 30 and 100 for oil and gas onshore
pipelines. Another common variable is pipe deflection, Δx/D. For soil, the most
pertinent variables are the friction angle φ and the vertical strain of sidefill
embedment, ε. An approximation of strain is given by ε = σ/E’ where σ is the vertical
soil stress, and E’ is the soil stiffness modulus.

Soil modulus is not a modulus of elasticity and is not a constant. It is often found,
approximately, as the slope of a secant to the stress-strain diagram from laboratory
compression tests. The confined compression test is conservative since the
laboratory sample is subject to a confining pressure and compressed vertically.

Sidefill soil, on the opposite, is compressed vertically, compressed horizontally

(radially) and confined longitudinally (biaxial compression) for which the
compression (vertical strain) is less than it is for confined compression tests. To
define if the behavior of a pipe is flexible or rigid, it is convenient to consider the
stiffness ratio Rs as the ratio between soil stiffness and pipe stiffness:

Rs = E’ / (EI/D3) = E’D3/EI [Eq. 1-1]

Where:

E’ is the soil stiffness modulus of the soil (slope of a secant on the stress strain
diagram);

EI/D3 = E/12(D/t)3 is the pipe stiffness, where I is the transverse moment of inertia
per unit length of individual pipe wall components.

Soil stiffness E’ can vary from 100 kPa for damped soil to 50-100 Mpa for well-
compacted, coarse-grained soil. Pipe stiffness EI/D3 is inversely related to pipe
flexibility D/t and reduces progressively to zero for high D/t values. Usually pipe
stiffness contributes significant resistance to pipe deflection if Rs is less than about
200, or D/t less than 50 (Figure 2). Conversely, soil stiffness is crucial for flexible
pipes and Rs is usually greater than 200.

Figure 1: Pipe stiffness vs. D/t, assuming E = 2.1E+8 kPa

Figure 2: Stiffness ratio vs. pipe flexibility

Pipe Deflection

In order to predict pipe deflection of buried circular pipes, designers normally use
the modified IOWA Formula. The formula is given by the ratio between the load
acting on the pipe and the soil-pipe stiffness as per equation 1-2:

#$%& $( )*)+ 12 345 6

∆𝑥 = = ;
)*)+ ,-*..(+,, / ,$*# ,-*..(+,, 78/9.9;<5 6 7 =

therefore pipe deflection is

∆? 12 3@A
= BC [eq 1-2]
1 /9.9;<7 =
D6

Where in SI units:

Δx = Horizontal deflection [m]

DL = Deflection lag factor
K = Bedding constant
W = Earth load on pipe = γHD [kN/m]
r = Mean pipe radius [m]
E = Modulus of tensile elasticity of the pipe material [kPa]
I = Moment of inertia per unit length, [m3]
E’ = Modulus of soil reaction, [kPa]

Under most soil conditions, flexible pipe tends to deflect into a nearly elliptical shape
and the horizontal and vertical deflections may be considered equal for small
deflections (Δ).

The relationship between horizontal deflection and vertical deflection in buried

flexible conduit was determined in the laboratory for pipe of various sizes at
deflections under 10% by Spangler in the following equation:

Δx = 0.913 Δy

where:
Δx = Horizontal deflection
Δy = Vertical deflection

The deflection lag factor (DL) accounts for the fact that, in soil-pipe systems, the soil
consolidation at the sides of the pipe continues at an ever-decreasing rate with time,
after the maximum load reaches the buried pipe. Experience demonstrates that
deflection of buried flexible pipe will continue for a period of time after completion of
pipe installation before final equilibrium is achieved.

Spangler originally said, “The deflection lag factors observed in the experiments
range from 1.38 to 1.46, and in no instance was equilibrium completely attained.”

His experience had shown deflections could increase by as much as 30% over 40
years. For this reason, he recommended the incorporation of a deflection lag factor
of 1.5 as a conservative design procedure.

It should be noted, however, that the load proposed by Spangler in the formula is
the Marston load for a flexible pipe, which is smaller than the prism load (often 1.5
times smaller). For a pipe with good flexibility, the long-term load will not exceed the
prism load. The increase in load with time is the largest contribution to increasing
deflection. Therefore, for design, the prism load can be used to compensate for the
increased trench consolidation load with time and resulting increased deflection.
When deflection calculations are based on prism loads, the DL should be 1.0.

The bedding constant K accommodates the response of the buried pipe to the
opposite and equal reaction to the load force derived from the bedding under the
pipe. The bedding constant varies with the width and angle of the bedding achieved
in the installation. Table 1 provides values determined, theoretically, by Spangler
and published in 1941.

Bedding Angle
K
(Degrees)
0 0.110
30 0.108
45 0.105
60 0.102
90 0.096
120 0.090
180 0.083
Table 1: Bedding constant

The modulus of soil reaction E’ is not the same as defined in equation 1-1. Many
research efforts have been attempted to measure E’, without much success. The
most useful method has involved the measurement of deflections of a buried pipe for
which installation conditions are known, followed by a back calculation through the
IOWA Formula to determine the real value of E’.

Hartley and Duncan, in 1987, gave a range of E’, varying with soil types, compaction
levels and cover depths. Their work also suggested to assume E’ equal to the soil-
constrained modulus, which can be based on laboratory tests. Jeyapalan and
Watkins have recalculated the values of E’, published by Howard (1977), finding that
E’ can vary several hundred percent from the one predicted, and showing how
difficult it is to predict that value.

Maximum through-wall circumferential bending stress due to ovalization can be

determined from Eq. 1-2 as:

∆I -
𝜎F = 4𝐸 [eq 1-2a]
1 1

This stress equation is used in the ASCE ALA guidelines to account for the stresses
due to ovalization on a buried pipelines. Pipe deflections and pipe through wall
bending stress are, therefore, related.

Figure 3 shows how pipe deflections and trough wall-bending stress (assuming a
steel pipe with E=2.1E+8KPa, X60, soil density γ=20KN/m3, DL=1; K=0,1) ranges as
a function of D/t and E’. It can be noted that pipes with low pipe flexibility D/t do
not require soil compaction since deformation limits are not likely to be reached for
standard cover depths. On the opposite, higher D/t implies the need of soil
compaction to limit pipe deflections and stress.

Figure 3: Pipe deflections and wall-bending stresses for steel pipelines buried at
depths 1, 2 and 3 meters.

Estimating Modulus of Soil Reaction

All of the evaluation starts from the basic idea that modulus of soil reaction E’ is a
known priori by the designer. Reality is quite different since E’ is often selected
among available literature tables (Table 2).

To understand the influence of E’ on pipe deflections, the IOWA Formula has been
rewritten in terms of Stiffness Ratio Rs (Equation 1-1), assuming:

W = (γH)D = (E’εD where ε is the soil strain and E’ is soil stiffness modulus
r3 = (D/2)3 = D3/8
K = 0.1
Dl = 1
d = (Δ/D) = pipe deflection

𝐷# 𝐾𝑊𝑟 N 0.1𝐸′𝜀𝐷𝑟 N 0.1𝐸′𝜀𝐷𝑟 N

∆𝑥 = = =
𝐸𝐼 + 0.061𝐸′𝑟 N 𝐸𝐼 + 0.061𝐸′𝑟 N 𝐸𝐼 + 0.061𝐸′𝑟 N
& 9.<7 = 5 6 9.< 9.< ],
= = BC = Z = eq [1-4]
V 78/9.9;<7W5 6 X /9.9;< X /9.9;< X9/9.;<],
B=Y6 [\

Equation 1-4 represents the IOWA Formula as a relation between two dimensionless
variables: d/εx= pipe deflection term; Rs = stiffness ratio term.

It can be noted (Figure 4) that such equation approaches a horizontal asymptote at

d/ε greater than 1 for high values of Rs and:

For Rs > 200, the pipe deflection term d/ε> 1, which implies that pipe deflection is
greater than vertical sidefill compression (flexible pipes);

For Rs < 200, the pipe deflection term d/ε < 1, which implies that pipe deflection is
less than vertical sidefill compression (rigid pipes).

Reynold K. Watkins & Albert B. Smith said empirical pipe deflections do not exceed
vertical compression of the sidefill soil and maximum d/ε should be 1. Therefore,
they proposed a formula (USU empirical) that does not exceed unity:

& ],
= [Eq. 1-5]
V N9/],

1.80
1.60
1.40
Ring Deflection Term d/ε

1.20
1.00
0.80
0.60
Iowa Formula
0.40
0.20 USU Empirical

0.00
0 100 200 300 400
Rs

Figure 4: A comparison between Iowa Formula and USU Upper-Limit Empirical

Formula.

Conclusion

Experience has shown that pipe deflections are limited for standard pipeline burial
depths; however, under some circumstances, excessive deflections have been
measured even at shallow depths, causing delays during construction. Starting from
the modified IOWA Formula, the following conclusions can warn the designers in
terms of pipe flexibility D/t:
• Pipes with D/t < 50 can be considered rigid since they are stiffer than the
backfill material surrounding it, and do not present excessive ovalization and
are not influenced by sidefill stiffness. A proper bedding is important since
they transmit most of the load through the pipe wall into the bedding.

• Pipes with D/t ∈ [50; 100] can be normally considered flexible, offering
significant structural benefits to the project. That is because if properly
installed, they can be buried much deeper than a similarly installed rigid pipe
because of the soil-pipe interaction, and will have a maximum deformation
similar to the sidefill compression. Therefore, the control of deflection is
obtained through the control of soil deformation by verifying backfill
compaction level.

• Pipes can experience high deflections and wall-bending stress even at shallow
depth if a proper soil backfilling compaction is not done. Soil stiffness E’ value
(in kPa) and modulus of soil reaction can be roughly estimated through the
relation 1,000H (where H is the cover depth in meter) to limit pipe deflections
to 3%.