Ground Movement or

Construction? How to Identify

Clear Ground Movement
Signatures in Inertial Measurement

Casey Dowling1, Pete Barlow1, Joel Van Hove1, Teko Hanvi2, Jim Hart3
1
BGC Engineering, 2Enbridge, 3SSD, Inc.

Pipeline Pigging and Integrity

Management Conference

February 12-16, 2024

Organized by
Clarion Technical Conferences

1339 https://doi.org/10.52202/072781-0075
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

Proceedings of the 2024 Pipeline Pigging and Integrity Management Conference.

Copyright ©2024 by Clarion Technical Conferences and the author(s).
All rights reserved. This document may not be reproduced in any form without permission from the copyright owners.

https://doi.org/10.52202/072781-0075 1340
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

Abstract

I nertial measurement unit (IMU) bending strain data has been recognized as a crucial tool for
detecting ground movement impact along operating pipelines. In Line Inspection (ILI) vendors
produce bending strain and movement reports which often include impact from landslides and
sinkholes, but also include a host of other causes. While these bending strain reports serve as a useful
initial screening tool, previous studies have found that often over 90% of these reported features are
related to pipeline construction or operation, not ground movement impact. It is not uncommon
for hundreds of bending strain features to be reported on a single 100-mile pipeline segment, and it
is important for pipeline operators to be able to identify ground movement caused bending strains.
Unlike construction-related bending strains, ground movement tends to increase strain demand over
time and often have a larger longitudinal strain component because of axial loading or pipe
elongation. Often, landslides and sinkholes produce signatures in IMU data that are evident during
a cursory review by subject matter experts (SMEs) with experience in geohazard mechanisms. Drawing
upon 10 years of experience using IMU data to characterize and monitor geohazard impact along
pipelines and geotechnical assessment of more than 6,500 bending strain features, this paper provides
examples of common IMU signatures indicative of landslide and sinkhole impact with the intent
that operators can learn to understand ground movement mechanisms that produce IMU bending
strain and how SMEs prioritize these for further assessment or action. Examples of signatures from
single run and run-to-run IMU data are presented along with a discussion of the basic mechanisms
that produce the signatures. Key ground-movement signatures within other IMU outputs such as
pitch, heading, and out-of-straightness (OOS) are also discussed. Construction-related IMU
signatures are provided to help operators understand how these differ from ground movement
signatures.

Introduction

Inertial Measurement Units (IMU) have become standard tools included in In Line Inspection (ILI)
runs performed by pipeline operators. Data from IMUs provide the x, y, and z position of the pipeline
over time and can be used to calculate pipe curvature and from that, bending strain (Hart et al, 2019).
These tools are used to provide accurate pipeline positional data, as well as identify areas of
anomalous bending strain induced by either ground movement or from pipeline construction. In
conjunction with traditional geotechnical assessment techniques, IMU bending strain data can be a
useful tool for geohazard identification and characterization as it can indicate pipeline impact from
ground movement as well as provide information on the ground movement pattern and extents of
the impacted pipe. Essentially, the IMU data enables the pipeline to become a horizontal slope
inclinometer, a typical in-ground installation that is utilized by geotechnical engineers to measure
slope movement and rates. ILI vendors have been able to distinguish intentional bends (bends
formed either in the field with a pipe bending machine or in the mill through induction) from
unformed bends (bends unintentionally induced on the pipeline either during construction or from

1341 https://doi.org/10.52202/072781-0075
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

ground movement) by assessing the location and length of the bending strain and the strain
magnitude. However, the majority of these reported bending strain features have been found to be
construction related, rather than related to soil loading associated with a geohazard.

This paper seeks to demonstrate the value in assessing IMU bending strain signatures as part of a
geohazard management program and introduce pipeline operators to the basic ground movement
mechanisms that induce bending strain. The paper provides an overview of common bending strain
signatures induced by geohazard loading scenarios and provides an explanation of the ground
movement mechanisms and loading that create the patterns. The paper reviews not only what the
signatures look like in vertical and horizontal bending strain profiles, but also how they appear in
position, pitch and heading plots. Common construction-related bending strain signature examples
are also provided. The examples provided are simplistic to demonstrate the key signatures and
loading mechanism; however, in most cases, the loading induced by landslides is complex and the
bending strain signatures are often much more complicated. Additionally, bending strain is only one
component of the total longitudinal strain demand induced by landslide loading, and IMU bending
strain assessments need to consider the proportion of axial loading anticipated based on the length
and orientation of ground movement in relation to the pipeline. In practice, assessing IMU bending
strain data for evidence of ground movement impact requires a deep understanding of landslide
processes and mechanics and should be completed by qualified subject matter experts (SMEs). This
paper seeks to enable pipeline operators to understand how they can best leverage the value of IMU
bending strain reports in geohazard assessment and prepare them to understand how loading
induced by ground movement produces identifiable patterns within bending strain data.

Background and Previous Work

Inertial measurement units have been used for pipeline geometric survey since the late 1980’s to
detect pipeline deformation related to ground movement. As described in Chyz and Adams (1994),
the pitch and heading is measured and combined with the odometer distance to provide the
northing, easting and elevation, based on surveyed tie-in locations along the pipeline route. This
orientation information can be differentiated with the odometer distance to calculate pipeline
curvature, which is proportional to bending strain. Bending strain reports are now a common add-
on to ILI tool runs. Within these reports, vendors provide a list and plots of bending strain “features”
that were either induced during pipeline construction or from post-construction ground movement
or third-party impacts.

In a typical bending strain report, ILI vendors distinguish between intentional bends (i.e., cold field
bends or manufactured elbows) and unformed bends by considering a combination of amplitude
and length of a bending strain feature. Intentional bend signatures have a single, relatively high strain
bending lobe (i.e. a curved/rounded shape similar to half of a sine wave) contained within a single
joint of pipe (typically 40 feet in length) in between two girth welds and are typically provided in

https://doi.org/10.52202/072781-0075 1342
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

operator pipe books, while unformed bending signatures typically extend over much longer pipeline
lengths and multiple pipe joints and have relatively lower bending strain amplitudes. Unformed
bending strain features are typically those identified in a vendor bending strain report.

A typical vendor bending strain report may have anywhere from tens to hundreds of identified
bending strain features over a given pipeline segment. However, as found by others, (Scheevel et al.,
2022; Theriault et al., 2019; and Hart et al., 2019), the majority of vendor-reported bending strain
features (typically over 90%) are not related to ground movement but are likely the result of
construction. When IMU bending strain features are assessed in conjunction with geohazard
inventories and high-resolution topographic data (i.e., lidar data), bending strain features can be
further screened to reduce the number of IMU features that may be related to ground movement.
Theriault et al. (2019) assessed whether bending strain magnitude and orientation (horizontal vs
vertical) correlated with geohazard presence, finding that higher magnitude bending strains greater
than 0.35% and horizontal bending strain greater than 0.15% did tend to correlate with geohazards.
However, just because a bending strain feature may coincide with a documented geohazard, it does
not mean the feature is related to ground movement. Also, relatively small magnitude bending strain
features that correlate with a geohazard may be indicative of ground movement impact. Scheevel et
al. (2022) advocates not only assessing whether bending strain features correlate with mapped
geohazards but also reviewing the bending strain signatures with the anticipated ground movement
mechanism and site history to evaluate whether the bending strain does indeed indicate ground
movement impact. In their assessment, Scheevel et al. (2022) found that by reviewing IMU bending
strain signatures with this additional insight, a prioritized subset of previously identified geohazard
sites was able to be refined to a prioritized list approximately 10% the size of the original inventory.
This method of combining an understanding of the slope movement mechanism, site history, and
bending strain signatures can be used to effectively identify which sites are impacted by ground
movement, providing the clearest picture of whether geohazards are impacting the pipeline. Being
able to distinguish IMU features consistent with ground movement from construction-related ones
is valuable as using magnitude or bending strain orientation alone can lead to unnecessary integrity
digs or missing key geohazard sites.

Typical data required for Geotechnical IMU Bending Strain Review

To describe how IMU signatures are evaluated in conjunction with an understanding of geohazard
mechanisms, a high-level overview of the necessary data is provided. The IMU data requirements are
based largely on the work outlined in Hart et al. (2019).

1343 https://doi.org/10.52202/072781-0075
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

Inertial Measurement Unit Data and Plots

To assess bending strain cause, bending strain reports should provide a series of plots for each flagged
bending strain feature. Hart et al. (2019), provides a useful outline of plots and pertinent details to
include. The recommended “seven panel” plots include:

x Vertical out of straightness or elevation profile

x Horizontal out of straightness profile

x Pitch

x Heading

x Vertical bending strain

x Horizontal bending Strain

x Total Bending strain

In some cases, vendors may provide individual plots for each field listed above or may combine
relevant data, such as pitch and vertical bending strain, in single plots. The plots should include the
location of girth welds, as well as the location of formed bends (derived either from as-built records
or based on the shape and length of the bending strain profile from the IMU data). Plots should be
scaled to observe detail along the bending strain features. This is often done automatically by the
vendor based on the bending strain feature length and magnitude. Due to the relatively lower
magnitude of unformed bending strains compared to formed bends, bending strain along formed
bends tends to be cut-off on bending strain plots appropriately scaled to assess unformed bending
strain. However, loading from slope movement, particularly axial-oriented movement, can deform
formed bends. Because of this, it is often beneficial to include additional vertical, horizontal, and
total bending strain plots scaled to show the bending strain magnitude of the formed bends.

As described in Hart et al. (2019), there are unique cases where vendor-provided plots may not suffice
to assess potential geohazards. These cases can include large landslides, where bending strain features
may only cover a portion of the pipeline length likely impacted by ground movement and the pure
axial component of strain (the uniform strain along the cross section of the pipe) may be much larger
than bending strain. In these situations, it can be useful to have the vendor provide processed digital
data in tabular format so that SMEs can create their own plots at the appropriate length and scale.

Single Run versus Comparison Run IMU Data

Single Run IMU data can provide an indication if the pipeline has been impacted by slope movement
in the past; however, there is typically a greater amount of uncertainty of in the assessment. As noted
by Hart et al. (2019), it can be difficult to distinguish positive ground movement impact from

https://doi.org/10.52202/072781-0075 1344
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

construction, using single-run IMU data alone. Deformation of formed bends (indicative of axial
loading) is particularly difficult to assess using single run IMU.

Multiple run (also referred to as run-to-run) IMU data can show whether continued ground
movement is occurring by showing changes in bending strain between the timing if the IMU runs.
SMEs can evaluate whether the change in bending strain is consistent with the ground movement
mechanism or due to repairs/construction activities. Additionally, multiple run IMU can indicate if
formed bends are changing over time, which is useful to identify axial loading. For multiple run IMU,
additional plots showing the difference between IMU runs are recommended to assess whether the
pattern of change is consistent with the anticipated ground movement mechanism.

In order to produce multiple run IMU plots, the IMU data from multiple runs must be carefully
aligned through the process of “odometer matching”, typically done using a subset of pipeline girth
welds or known bends (Clouston et al., 1999; Hart et al., 2020). Vendors can often do this with
tabular data from past IMU runs completed either by themselves or other vendors. However, due to
differences in tools and processes, there is typically some amount of noise inherent in multiple run
IMU data. Hence, many vendors have a bending strain change threshold of 0.04% to define bending
strain change features and typically do not call out bending strain change that may appear along
identified formed bends.

Multiple run IMU enables IMU data to be used as one method of monitoring geohazard sites, if
combined with other instrumentation (slope inclinometers or strain gauges), regular field
inspections, and/or remote sensing techniques (lidar change detection and/or InSAR).

The frequency at which operators should acquire IMU data for ground movement assessment is
dependent on that rate of ground movement typical within the region. From the authors’ experience,
a frequency less than 1 year is often too short to positively identify strain change in multiple run IMU
data, even in areas with relatively active ground movement, such as the Appalachian Plateau.
However, the maximum reassessment frequency required for interstate transmission pipeline (5 years
for liquid and 7 years for natural gas transmission pipelines) (49 CFR 195 and 49 CFR 192) is often
too long to effectively utilize IMU to monitor ground movement impact along a pipeline. Operators
are encouraged to engage with a geohazard SME to evaluate the optimum frequency to complete
IMU runs to detect bending strain change and enable operators to act should an integrity threat
become known.

Geohazard Information

In addition to IMU bending strain plots, it is important to understand whether they intersect
documented geohazards or potential geohazard morphology as well as understand what the likely
mechanism and movement direction of the hazard would be. To do this, the extents of the bending
strain features should be plotted within a Geographical Information System (GIS) workspace that
includes aerial or satellite photographic imagery, lidar data, and the outline and extents of known

1345 https://doi.org/10.52202/072781-0075
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

geohazards. With this data, a geohazard SME can evaluate whether the IMU strain feature intersects
a known geohazard and can also consider the anticipated ground movement direction and loading
patterns. Remote sensing data such as lidar change detection and InSAR can provide helpful insight
in evaluating the landslide movement direction and differing rates of movement within larger
landslide complexes. Understanding the mechanics of the geohazard and then comparing these to
the signature within the bending strain plots is what enables a geohazard SME to assess whether an
IMU feature is the result of ground movement.

Site History

Any construction history or past repair records at a site should be compiled into the previously
mentioned geospatial database. Information on past integrity digs, past slope repairs and mitigations,
pipeline cutouts and replacements, and adjacent pipeline construction is useful in evaluating IMU
signatures and in evaluating whether a bending strain feature may be related to a one-time event
(such as lateral deflection due to a downslope pipeline installation) or an ongoing hazard that will
continue deforming and inducing additional strain on the pipeline. Records of integrity digs are
useful to confirm whether vertical bending strain signatures are related to settlement signatures from
past digs.

Bending Strain Classification

As outlined in Scheevel et al. (2022), assessed bending strain features should be categorized as either
ground movement related, construction related, or unknown. The classified sites should be reviewed
with a pipeline operator’s integrity team to rate and prioritize for future action. A system outlined in
Scheevel et al. (2022) has been utilized as an effective means for classifying bending strain features
(Table 1). This classification scheme not only differentiates strain based on cause, but also indicates
the level of possible severity of a bending strain feature, both for ground movement and construction-
related strains.

https://doi.org/10.52202/072781-0075 1346
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

Table 1. Bending Strain Features tier definitions and typical follow-up actions
(modified from Scheevel et al., 2022).
Strain Due to Geohazard-Related Ground Movement?

Likely Unclear Unlikely

Tier 1 2 3 4 5
Site Description Critical site Site with Site with Strains not Strains not
(Combination of ground possible ground consistent with consistent with
slide activity, site movement movement expected ground expected ground
condition and related related strains. movement; max movement; max
strain level) with strains. girth weld girth weld bending
ground movement bending strain strain <0.2%.
related strains. •0.2%.
Typical Detailed Detailed Baseline
Recommended inspection as soon inspection inspection as Determined by geomorphic factors, no
Baseline SME as practical within 6 to part of normal influence from strain presence.
Field Assessment (< 6 months). 12 months. managed field
program.
(>12 months).
Typical On-going As per SME/Operator review: Routine monitoring. Re-evaluate for
Management x Mitigation strain change when new IMU data
Actions available.
x Site-specific monitoring
x Routine monitoring

Common Characteristics of Ground Movement Induced Bending Strain Features

A bending strain signature consistent with ground movement impact (Tier 1 or Tier 2 as provided in
Table 1) tends to have all or most of the following attributes:

x The bending strain feature is located at the crossing of a known landslide or subsidence prone
area.
x The direction of the bending is consistent with what would be expected, considering the
direction of ground movement.
x The bending feature may be located at a point of differential ground movement such as the
margins of a landslide or boundaries between landslide blocks or movement zones. Differential
ground movement transverse to the pipeline axis is typically accommodated by pipeline bending.
x The bending feature may be located at an area of suspected longitudinal tension or compression.
x There is a “sinusoid” bend pattern with adjacent bends and reactionary “side lobe” bends which
are distinct from typical formed and roped bends.
x The bends do not follow the typical construction signature of formed or roped bends in terms
of bend length, location, uniformity, and curvature (see the later section on construction-related
bending strain features).

1347 https://doi.org/10.52202/072781-0075
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

The last two bullets are discussed in more detail in Hart et al. (2019), which provides a thorough
description of bending strain signatures from ground movement but does not discuss evaluating the
pattern in the context of potential landslide loading.

The following attributes are also common and strengthen the assessment that the bending strain
signature is consistent with ground movement impact:

x Multiple runs of IMU are available and reveal a pattern of bending strain change which indicates
ongoing pipeline deformation (e.g., bending strain growth)
x The landslide is known to be active and spatially intersecting the pipeline (e.g., through
geotechnical instruments, lidar change detection, axial strain ILI tools or other independent
measurements)
x The magnitude of bending strains interpreted to be caused by ground movement impact are
different from typical construction bending strain magnitudes (particularly roped bends).

Typical Ground Movement Signatures

Landslides deform pipelines due to lateral, vertical and axial deformation. Lateral deformation and
bending strain have been recognized as key indicators of landslide impact in IMU data (Theriault et
al., 2019; Hart et al., 2019); however, ground movement also can be apparent in vertical bending
strain signatures, particularly at the margins of active ground movement. The following subsections
describe typical signatures in IMU data that may indicate ground movement impact and explain the
loading scenarios responsible for the signatures.

General Bending Strain Patterns from Ground Movement

The pattern/signature of a bending strain feature refers to the magnitude, direction, uniformity,
length and shape of the bend or series of bends, and how the bending strain corresponds to other
formed bends and girth welds.

The most common signature of landslide impact is a pattern of sinusoidal bends; adjacent bends
with opposing bend direction created by differential ground movement (e.g., at a scarp or toe
location). These bend signatures are sometimes referred to as “S-shapes”, “W-shapes”, or “sinusoid
patterns”. Typically, the adjacent bends will have opposite bend directions formed by
bend/reactionary-bend sequences, but this may be masked-by or overprinted by formed bends and
construction related bends. These signatures are characterized by a directional pattern which is
consistent with the landslide loading direction and differential movements at the landslide
boundaries or at points of differential movement within a landslide, discussed in the following
subsections.

https://doi.org/10.52202/072781-0075 1348
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

Bend signatures generated by landslide loading tend to have a longer total bend length (tens of
meters) relative to formed bends (a few meters), and more variable bend curvature geometry
compared to formed bends which tend to have nominally uniform curvature through the bend. The
longer bend signature length also means the bends tend to span girth welds, causing elevated girth
weld strains compared to other sections of the pipeline outside of the landslide extents. Pitch and
heading plots are also useful for identifying landslide-induced bends. Formed bends typically have a
consistent ramp in pitch and heading plots that levels off at girth welds, as the curvature is uniform
over the bend and is within one pipe joint. The long and variable bending caused by landslide loading
produces curved pitch and heading ramps that often cross girth welds (Figure 1).

The magnitude of bending strain induced by slope movement is reflective of the stiffness of the
trench material at the interface between stable and moving ground or transitions between zones of
differential movement in a slide mass. An abrupt transition between relatively stiff substrates (i.e.,
bedrock or stiff clay) can cause strain concentration leading to high reactionary bending strain over
a fairly short pipe length due to the differential stiffness of the bedrock walls along the in-place trench
and the softer trench material within the slide mass. High confining pressures due to depth often
induce acute bending strains where directionally drilled pipelines intersect landslide slip surfaces,
making it fairly easy to identify ground movement induced bending strain within a horizontal
direction drill bore.

Figure 1. IMU plots (pitch, heading, vertical bending strain, horizontal

bending strain and total bending strain) showing typical signatures of
unformed and formed bends. The unformed bending in this plot is induced by
landslide movement. Slope movement has produced both vertical and
horizontal bending strain that track – the over bend aligns with the dominant
right bend. Plot produced by Baker Hughes.

1349 https://doi.org/10.52202/072781-0075
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

Lateral Ground Movement Bending Strain Signatures

Often the clearest bending strain signature indicative of slope movement is that induced due to lateral
deflection of the pipeline in the direction of slope movement. Lateral deformation is typical on side
slope construction, as well as in cases where landslide movement is encroaching on a pipeline at a
slope toe or along a ridgeline. The lateral deflection produces a horizontal bending strain signature
with a broad horizontal bend across the maximum zone of deflection and opposite-direction
reactionary bends at the margin of slope movement (Figure 2). In an IMU bending strain plot, the
strain pattern commonly forms a W-shape. This W-shape is indicative of slope movement impact if
the horizontal bending strain pattern is consistent with the direction of slope movement (as assessed
from geomorphic data or in-place instrumentation). If related to ground movement, the horizontal
“W” shape will often occur across the full width of moving ground, hence it can be particularly useful
for characterizing the extent of active ground movement.

Consider a landslide moving across the pipe from the 9 o’clock to 3 o’clock position (Figure 2). In
this case, slope movement pushes the pipeline to the right (3 o’clock position) within the slide mass.
This results in a broad left bend along the section of maximum deflection. Reactionary right bends
are apparent where the pipeline enters and exits the slide mass. In the IMU plots, the horizontal
bending strain shows a W-shape going from zero to positive (right) to negative (left) to positive (right)
and returning to zero. The bending strain typically extends over multiple pipeline joints, causing
elevated strain on girth welds. As bends extend across girth welds, the heading will be “ramped”
rather than level, indicating the pipeline orientation is changing across the weld (see Figure 2). In the
horizontal out of straightness (OOS) plot, there is typically a clear deflection in the direction of the
lateral slope movement. The OOS estimate from this plot can be compared to OOS measurements
made in the field (e.g., based on pipe locator data). The width of the landslide will influence the
horizontal bending strain signature. A smaller landslide may produce a single clear right-left-right W-
shaped bending strain signature. However, if the landslide is wide, a long length of pipe may be
relatively evenly deflected through the middle portion of the slide mass. In this case, the pattern may
be right-left-left-right (corresponding to an S-shape at both “shoulder” locations of the landslide rather
than a W-shape).

Landslide movement oblique to the orientation of the pipeline can produce an asymmetric horizontal
bending strain signature. This orientation of loading typically induces compressive strain at the
downslope side of the lateral deflection and tensile strain at the headward end of the deflections.
This results in more acute bending strain in the downslope section of pipe and broader strains in the
upslope section of the slide mass. Oblique loading often leads to elevated total longitudinal strain,
particularly at the location of maximum deflection and at the downslope reactionary bend.

Horizontal bending strain signatures related to slope movement can often be detected in single-run
IMU bending strain data if the strain signature overlaps landslide morphology and is consistent with
the direction of slope movement. In areas where landslide morphology is not as clear, such as colluvial
aprons at slope toes, multiple IMU runs may be needed to confirm whether the horizontal bending

https://doi.org/10.52202/072781-0075 1350
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

strain is from ground movement or from construction roping. In cases of side slope pipeline
construction on shared corridors, lateral deformation and strain may be induced by downslope
construction. Multiple run IMU data is important in these cases to confirm whether ground
movement ceased following construction or has continued deforming the pipeline.

Figure 2. Diagram of a lateral landslide loading scenario and the IMU signature
produced by the loading. Note the downslope deflection in the out-of-
straightness plot (top) and the right-left-right horizontal bending strain
signature (bottom). Plot produced by Baker Hughes.

Vertical Ground Movement Bending Strain Signatures

Vertical ground movement signatures can be observed in IMU bending strain plots where vertical
loading occurs on a pipeline. In a landslide, this typically occurs where a pipeline crosses a zone of

1351 https://doi.org/10.52202/072781-0075
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

downward vertical movement at a headscarp (the upper boundary of a landslide where ground
movement has offset the slide mass downward from the adjacent stable ground) or an internal scarp
(a vertical ground displacement feature within the slide mass) (Figure 3). Upward vertical
displacement typically occurs at the toe of a landslide, where the slide plane reaches the ground
surface (Figure 3). Vertical bending strain is also induced where the pipeline is pulled downward over
an area of subsidence such as a sinkhole related to karst processes or underground mine collapse.

At a landslide headscarp, the typical bending strain pattern, from upslope to downslope, is an over–
sag pattern (see Figure 2). This occurs because the predominant vertical movement at a headscarp is
downward. In oblique and axial landslide movement, this pattern is not always very high in
magnitude and may not appear at all due to tensile axial strain, typical in this upper portion of the
slide mass. This tensile strain can limit the magnitude of bending strain induced on the pipeline.

The most pronounced vertical bending signatures within a landslide can be at the toe, where the
slide plane(s) thrust up to the ground surface, creating a “toe bulge” (see Figure 3). The portion of
the pipeline that crosses this toe bulge is upthrust as well, producing a characteristic over bend strain
bounded by reactionary sag bends. If oriented oblique or axial to the direction of slope movement,
this pattern may be more pronounced due to compressive axial strain at the toe of the landslide.

In many cases, there is a high degree of uncertainty for vertical bending strain features, even if they
do intersect a known landslide feature. Vertical bending strain induced by ground movement is not
as apparent as horizontal slope movement. Vertical loading can be complex within a landslide mass,
which can produce multiple vertical strain patterns. Unlike lateral deflection, the pipeline lengths
that are impacted by vertical deformation in a landslide tend to be relatively short and may often
only be one to two pipeline joints in length. If vertical bending strain does extend multiple pipe
lengths, ramping will be observed across girth welds in pitch plots. Vertical bending strain patterns
may appear to mimic topography, which is a typical of construction-induced bending strain
signatures. If slope movement is oblique, the unformed vertical bending strain signature often will
follow or “track” with the horizontal bending strain signature (see Figure 1). This can be helpful in
identifying ground movement induced vertical bending strain. If ground movement is actively
occurring, the best way to verify the source of the strain is through multiple IMU runs to detect
changes in the vertical bending strain.

Vertical bending strain induced by ground settlement is often quite apparent and produces an over-
sag-over bending strain pattern, similar to what could be expected at a landslide headscarp. However,
the downward deflected area may be more extensive and thus, it can be easier to see, particularly if
there is evidence of subsidence in geospatial data (sinkholes and or documented historic mining
activity). To evaluate whether the strain is related to subsidence, additional information on karst
activity and active/historic mine activity are usually necessary, gathered either through desktop
literature reviews or field investigation programs.

https://doi.org/10.52202/072781-0075 1352
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

Figure 3. Diagram of an axial-oriented landslide (top), the vertical loading and deformation
produced in cross section (middle), and the IMU pitch and bending strain signature produced
in multiple run IMU data. Plot produced by Baker Hughes.

1353 https://doi.org/10.52202/072781-0075
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

Axial Slope Movement Signatures

IMU data cannot measure axial strain. However, axial loading can sometimes induce changes in
bending strain, particularly at formed vertical and horizontal bends (e.g., due to cable action P-ȟ
effects). In zones of tensile loading, formed bends will usually straighten/open, meaning the
magnitude of the bend curvature will decrease over time. Alternatively, in areas of compressive
loading, formed bends will close, or increase in bend curvature magnitude over time. For bends that
are closing, the bending strain can sometimes be observed extending into neighbouring pipe joints,
leading to strain on the girth welds bounding a formed bend.

While this deformation may be apparent in single-run IMU, typically multiple run IMU data can
greatly decrease uncertainty on whether the formed bends are deforming over time. To fully assess
formed bends, additional IMU plots may be required that show the full magnitude of formed bends.
The bending strain vertical axis should be scaled differentially than plots used to assess unformed
bending strains due to the higher magnitude bending strains found on formed bends. Often, large
axial loading occurs within larger landslide complexes, where individual vendor-provided bending
strain plots do not provide full coverage. In these cases, it is helpful to create plots using tabular IMU
data to allow for plot customization and to enable the bending strain profiles throughout the entire
landslide mass to be viewed.

While bending strain data does not provide axial strain, it can be used to estimate the axial strain
demand on a pipeline. If there is measurable OOS, a simplified axial strain calculation can be
completed based on the magnitude of OOS and the length of deflected pipeline. In some cases, it is
beneficial to evaluate the current strain demand on a pipeline by completing a finite element analysis
(FEA) that models the impacted pipeline including axial force effects. Bending strain data from IMU
can be used to develop the input ground displacement profile used in the FEA model so that the
model is calibrated to the soil loading being experienced by the pipeline. This enables the estimation
of a more accurate axial strain component and total longitudinal strain demand being imposed on
the pipeline. If the strain capacity of the pipeline and girth welds are known, an operator can evaluate
whether the strain demand estimated by the FEA is nearing any critical thresholds that may warrant
further action.

Typical Construction Induced Signatures

The majority of bending strain features identified in bending strain reports tend to be induced during
initial pipeline construction or from regular maintenance activities such as integrity digs. These
bending strain signatures are not expected to continue changing over time and once identified, do
not necessarily require additional monitoring unless they correlate with a known geohazard.

Tie-In Locations

https://doi.org/10.52202/072781-0075 1354
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

Construction related bending strain features are often observed at suspected tie-in locations
associated with road, rail, and watercourse crossings, which is supported by Theriault et al. (2019)
and Scheevel at al. (2022). These bending strain features may be oriented horizontally or vertically
and are induced when the line pipe is tied in with an adjacent pipeline segment. Typically, small
offsets between the two ends of the pipe to be joined require some amount of deflection to bring the
joint together, inducing unformed bending strain on both side of the joint. The bending strain
features tend to be short, typically limited to one or two pipe joints (see Figure 4). In vendor IMU
plots, these strain features can often be observed on pipeline joints adjacent to wall thickness changes
(often indicative of a crossing). These locations can typically be identified by assessing geospatial data
to see where road, rail and watercourse crossings are in relation to the pipeline. In some cases,
particularly at watercourse crossings at the base of slope, these construction strain features can appear
similar to vertical ground movement strains. Because of this, multiple run IMU data may be necessary
to confirm whether the bending strain is related to construction.

Construction Roping

An additional source of construction-related bending strain features is roping. Roping is common

where minor variations of topography do not require formed bends and where trench backfill creates
irregular loads on the pipeline. Roping often occurs after the pipeline is welded into a string, and
therefore bending strains may cross girth welds and be superimposed on formed bends.
In IMU data, roping results in irregular increases or decreases in the heading or pitch data and
irregular horizonal or vertical strains (see Figure 4). Common construction practices tend to produce
roping strains less than 0.2%, however this is dependent on the installation stresses and stiffness of
the pipeline. When a pipeline is routed through a landslide it is common to observe roping that
follows the benched and stepped topography typical of landslides, therefore it can be challenging to
differentiate landslide movement from roped bends that follow topography.

Roping patterns may occur in pipelines of any diameter but are more distinctive in small diameter
pipelines that are more flexible. Roping strain magnitudes in smaller diameter pipelines tend to be
higher as the pipeline is more flexible and can overlap into ranges typical of formed bends.
Understanding typical construction practices for a pipeline is important for identifying roped bends
and differentiating from ground movement bends and other causes. In these cases, multiple run IMU
data is often helpful to assess for changes in bending strain signatures over time to correctly categorize
the bending strain feature.

1355 https://doi.org/10.52202/072781-0075
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

Figure 4. A seven-panel plot showing examples of roped vertical bends

associated with straight sections of pipeline (circled in red), formed bends
(circled in yellow) and pup pipeline segments associated with a road crossing
(circled in green). Note the unform ramp in the pitch plots associated with
formed vertical bends and the irregular pitch signature associated with the
roped vertical bends.

Integrity Dig Signatures

Integrity digs produce readily apparent bending strain signatures that can easily be observed in
multiple run IMU data. These occur when a full 360-degree exposure of a pipeline is completed
either for an inspection or repair. Due to limitations with compacting backfill beneath and adjacent
to the pipeline, the underlying soil at the dig location tends to consolidate and settle during and after
backfilling. This creates a clear (over-sag-over) settlement signature along the pipeline in IMU vertical
bending strain plots (Figure 5). Often the length of this signature is limited to the pipe segment
exposed during the dig and the segments immediately adjacent. If a record of integrity dig locations
is available, these sites can be easily identified in both single and multiple run IMU data. In multiple
run bending strain reports, these locations are often identified as pipeline movement zones.

https://doi.org/10.52202/072781-0075 1356
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

Figure 5. Vertical bending strain and multiple run change plots for an integrity
dig site. Note the bending strain change occurs between 2019 and 2022, with
no change reported between 2022 and 2023. This is consistent with the date
of the integrity dig (2021). Plot produced by Baker Hughes.

Horizontal Directional Drill Installations

Horizontal directional drill (HDD) installations are made without formed bends; however, the
pipeline typically forms a broad roped bend as a result of being dragged into a curved drill path.
Typically, these bends are oriented vertically – with pipelines often having a broad sag bend (positive
or negative vertical bending strain, depending on the vendor) signature. Usually, the bending strain
is lower than 0.05% for the extent of the drilled portion but may be higher. Particularly with smaller
diameter pipelines which are more flexible, bending strains may be higher where a shallow bore
abruptly changed direction during installation. At the entry and exit of the drill a formed overbend
is often required to tie in the conventionally trenched section and because of the restricted position
of the HDD, roped bends are often also present as the less constrained conventionally trenched
pipeline is maneuvered to align with the HDD portion. If depth of cover or as-built information is
available, HDD installations can often be confirmed by installation depth.

Additional Considerations Regarding Bending Strain Assessment

Bending Strain is One Component of Total Longitudinal Strain

The bending strain derived from IMU data is only one component of the total longitudinal strain
state of the pipeline which also includes axial strain. IMU does not measure the axial strain
component resulting from elongation or compression of the pipeline, which in some cases can be
larger than bending strain. Axial strain is often assessed through a pipeline stress analysis (e.g., FEA)
if sufficient data on ground movement and pipeline impact are available or a simplified axial strain
calculation based on the OOS and length of OOS. Correctly positioned stain gauges can also be used
to provide information on axial strain at a specific location.

1357 https://doi.org/10.52202/072781-0075
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

A Bending Strain Assessment Cannot Rule Out Ground Movement Impact

A bending strain analysis cannot rule out ground movement impact because small amounts of ground
movement may result in pipeline bending that is indistinguishable from construction related bends
(particularly roped bends). Landslide loading may also be axial to the pipeline and can generate
strains which are dominantly longitudinal and do not result in significant (i.e., distinguishable)
bending. Even with multiple IMU runs, there is uncertainty due to the accuracy of the tool and data
noise, which may be indistinguishable from low magnitude changes in bending due to landslide
loading. Additionally, on vintage pipelines, bending strain induced by ground movement may have
occurred prior to the baseline IMU run, so a lack of change in bending strain from multiple IMU
runs does not rule out past landslide impact.

A Bending Strain Assessment Provides Insight into the Condition of the Pipeline

Despite not being able to rule out ground movement impact, a bending strain analysis which does
not identify evidence of impact still provides insight into the condition state of the pipeline by
limiting the severity or rate of bending strain change over an observation period. A bending strain
analysis which does not identify evidence of landslide impact implies that either the pipeline has not
been impacted by a landslide or it has been impacted but has only generated bending strains or
bending strain change at a low enough magnitude that the impact is not distinguishable from
bending strain from other causes. The insight this observation provides into the condition of the
pipeline needs to be considered on a site-by-site basis as scenarios with potential for axial pipeline
loading and longitudinal strain may not manifest in unusual bending.

Bending Strain Analysis Does Not Confirm a Pipeline’s Fitness for Service

Bending strain analysis may identify potential integrity threats, such as landslide impact, or provide
supporting information as part of a comprehensive integrity assessment but is not able to directly
confirm pipeline integrity or fitness for service.

Conclusions

Profiles of bending strain, pitch, heading and out-of-straightness from IMU data can be used for
geohazard impact identification and characterization, when the data is assessed with an
understanding of the geohazard mechanism. Loading scenarios induced by ground movement can
produce signatures that are readily apparent and different than those induced by construction. To
maximize the value of bending strain reports for geohazard management programs, pipeline operators
should not only assess the magnitude and location of bending strain features but should assess the
bending strain, pitch, heading and out-of-straightness plots in conjunction with an understanding of
ground movement and loading.

https://doi.org/10.52202/072781-0075 1358
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

Ground movement induced by geohazards, particularly landslides, is complex with different loading
patterns across the landslide that vary based on the orientation of the pipeline. Especially in large,
complex landslides, varying zones of ground movement can induce complex bending strain on a
pipeline. While the IMU signatures presented in this paper are from real sites, these cases are
considered exemplary in simplicity and clarity regarding the loading and bending strain signatures
produced. In most cases, bending strain induced by ground movement, particularly landslides,
should be assessed by SMEs with a thorough understanding of the mechanics of the geohazard and
the anticipated loading scenario.

IMU data is a powerful tool in that it can indicate whether the pipeline has been impacted by ground
movement, provide insight into the condition of the pipeline, and provides closely spaced (essentially
continuous) data along the entire length of the pipeline. This differentiates IMU data from discrete
point in-ground instrumentation such as slope inclinometers and location-specific instrumentation
such as strain gauges. However, IMU data is only one tool for assessing geohazard impact along
pipelines. It is important to assess IMU data in conjunction with other information such as lidar,
lidar change detection, satellite imagery, slope inclinometers and strain gauges. When this data is
fully integrated and assessed by knowledgeable SMEs, pipeline operators can leverage the maximum
value of the data within an effective geohazard management program.

Acknowledgments
The authors would like to thank Enbridge for their permission to use data from their geohazard
management program and the support and feedback they have provided on integrating IMU data
with geohazard assessment and characterization. The authors would also like to thank Baker Hughes
for providing permission to use their IMU plots and their collaboration on developing and producing
plots well-suited for geohazard assessment.

References

Clouston, S., Blair, G. and Hektner, D. (1999, November). Pipeline Out-of-Straightness Assessment
Using Pipeline Inertial Geometry Survey Technology. Alaska Pipeline Workshop. Anchorage,
Alaska, USA.

Czyz, J.A. & Adams, J.R. (1994, February). Computation of pipeline bending strains based on geopig
measurements. Pipeline Pigging and Integrity Monitoring Conference. Houston, Texas, USA.

Hart, J.D., Czyz, J.A., Zulfiqar, N. (2019, March). Review of pipeline inertial surveying for ground
movement-induced deformations. Proceedings of the Conference of Asset Integrity Management-
Pipeline Integrity Management Under Geohazard Conditions. Houston, Texas, USA.

Hart, J.D., Zulfiqar, N., and McClarty, E. (2020, September). Recommended Procedures for
Evaluation and Synthesis of Pipelines Subject to Multiple IMU Tool Surveys. IPC2020-9235,
Proceedings of the 13th International Pipeline Conference, Calgary, Alberta, Canada.

1359 https://doi.org/10.52202/072781-0075
 Pipeline Pigging and Integrity Management Conference, Houston, February 2024

Scheevel C., Dowling C., Hart J.D., & Cook, D. (2022, March). IMU bending strain: analysis as
geohazard screening tool. Pipeline Research Council International Research Exchange Meeting.
Orlando, Florida, USA.

Theriault, B., Hart, J.D., McKenzie-Johnson, A., & Paulsen, S. (2019, March). Correlation of single-
run ILI IMU bending strain features to geohazard locations. Proceedings of the Conference of Asset
Integrity Management-Pipeline Integrity Management Under Geohazard Conditions. Houston,
Texas, USA.

https://doi.org/10.52202/072781-0075 1360