Bearing the responsibility of ensuring
the continuous flow of energy resources through vast networks of pipelines, the
custodians of pipeline integrity face an ever-present challenge: geohazards.
From seismic shifts to landslides, these
relentless forces pose an intriguing and formidable threat that demands our
unwavering commitment to safeguarding the arteries of energy transportation.
In response, ROSEN and Teren have
leveraged modern technologies and methods to devise a solution that supplements
standard geohazard management approaches with unparalleled efficiency.
Join the front lines of the unseen
battles beneath our feet, where the fusion of engineering precision and an
innovative strategy is redefining the way in which we manage and mitigate the likelihood
of geohazards impacting our pipeline systems.
LiDar
and AI
Using LiDAR data to detect landslides
and landslide potential is not a novel concept. LiDAR has been widely valued by
pipeline operators and integrity consultants for geohazard identification, and
it is now a key tool for geotechnical experts assessing pipeline rights-of-way.
But how do we leverage modern
technologies to ensure we are extracting maximum benefit from this data? The
growing prominence of artificial intelligence (AI) has allowed for the development
of a data-driven approach to significantly increase the efficiency of geohazard
assessments in order to reduce timelines and scale the assessment process to a
size that effectively meets operators’ needs.
Computer vision algorithms can detect
geohazards within the pipeline right-of-way automatically based on the following
variables:
Additional algorithms can then be
implemented to characterize the structural components of the hazard, and a
manual review is performed by an expert analyst to finalize their extents and
remove false calls.
Using multiple LiDAR datasets,
subsequent algorithms can be used to automatically identify changes to existing
geohazards over a given time interval (e.g., land movement). Using a set of
change detection analytics, identified ground movement can be due to:
An example of this change detection process (Figure 1). Material added is signified by the green-shaded areas and material removed by the red-shaded areas.
While the proposed approach for
identifying bending strain anomalies and coincident landslides boasts
significant benefits, it also contains inherent limitations. For hazards
identified as severely affecting the pipeline, additional assessments are
required.
Inertial
Measurement
Using inertial measurement units (IMU)
to detect regions of bending strain and pipeline movement has been widely
adopted by the pipeline industry. The popularity of this approach continues to
grow, primarily due to its efficiency and effectiveness. Because almost any in-line
inspection (ILI) tool can be equipped with IMU, it is easy to integrate and
perform post-ILI functions.
The IMU does not measure bending strains
directly. Rather, it measures differentials in the unit’s orientation relative
to a calibrated starting position. The orthogonally mounted accelerometers and
gyroscopes measure linear and rotational accelerations across six degrees of
freedom, allowing the position (pitch, roll and yaw) of the tool to be determined
in three-dimensional space.
This method assumes that the pipe begins
as a straight segment and that any change to the pipeline trajectory is due to
cold-bending of the pipe. It does not account for nonhomogeneous materials or changes
in the shape of the pipe cross-section.
Once these calculations are performed, the entire pipeline can be screened for bending strain anomalies using a predetermined set of criteria. This typically includes a required length, maximum strain and pattern characteristics that indicate unintended bending of the pipe.
When two or more datasets are available,
it is possible to compare strains across a defined length of time and determine
whether pipeline movement has occurred across that interval. When trajectory
information is accessible, it is possible to tie two IMU datasets together at a
localized interval to determine the magnitude of movement that has taken place
in an area of concern.
Advantages,
Constraints
By combining these independent
assessment methods, a powerful and highly efficient framework emerges that can
effectively support geohazard management programs everywhere.
Pipeline operators are able to determine
– with lower resource expenditure than ever before – not only whether a
landslide exists within the right-of-way but also whether it has affected the
pipeline. As with any engineering process, advantages and constraints both
exist.
An important distinction between the
proposed method of geohazard detection and the conventional right-of-way
geohazard review is the time, level of detail and final product provided to the
operator.
Leveraging modern data processing speeds
and computer vision AI shortens the timelines and reduces expert availability
requirements. LiDAR can be acquired, geohazards identified and characterized
automatically, and quality checks performed. This means the geohazard
assessment does not constrict the proposed assessment process.
Because errors within an AI program are
systematic, they can be identified and corrected as this service approach
matures. A single geotechnical expert can review the results of multiple lines
in a fraction of the time required for a standard manual geohazard assessment,
reducing both the time required and the likelihood of human error.
The application of IMU bending strain
analysis allows for hazards not currently impacting the pipeline to be
specified as lower priority and not needing immediate intervention while
identifying high-priority areas requiring a follow-up review. Lastly, the
efficiency and ability to scale this approach for entire pipeline systems is
unmatched by its manual counterpart.
Nevertheless, automation of geohazard
detection has constraints. Because geohazard identification is automated, it is
possible that some hazards may be overlooked by the program. The assessment
process is limited to characterization, severity and prioritization, so recommendations
for high-priority areas still require expert insight to determine whether
remediation or further assessment is required. Additionally, advanced
assessments need to be carried out for high-priority areas by a geotechnical
expert as well.
Nevertheless, the benefits of using this
approach to supplement any geohazards management program are apparent – as long
as it is applied appropriately. It is of paramount importance that the details
outlined above be examined during the planning stage and implemented during
execution and review.
It should be noted that IMU bending
strain analysis can also be applied with similar results to a more conventional
approach.
Case
Study: Products Pipeline
Using the methods and assumptions
outlined above, this assessment process has been applied successfully to
real-world data for areas of historical landslide activity. Let us take a look
at a 16-inch refined products pipeline located within the Kanawha section of
the Appalachian Plateau in Eastern Ohio. This pipeline was installed within the
banks of a historical mining impoundment dam.
The top panel of Figure 3
presents a comparison of the bending strains calculated for two in-line
inspections equipped with IMU. Strains are displayed as percentages. The first
inspection, indicated by the red line, took place in 2016. The second one,
indicated by the blue line, took place in 2022.
The green line is the calculated
difference in strain between the two inspections. The results in both vertical
and horizontal planes show that strain change (e.g., pipeline movement) has
occurred. The horizontal plane reflects a somewhat characteristic movement
response pattern. The vertical plane reflects a significant strain change as
well, although with less consistent quality. This is likely due to the orientation
of the landslide relative to this span and the 90-degree change in centerline
trajectory occurring within the landslide area.
Based on the observed changes in strain
and the specified sign convention, movement has occurred to the right and
downward when observing the pipeline from upstream to downstream according to
the direction of the ILI.
The bottom panel of Figure 3 shows the
vertical change in terrain (soil movement) for the landslide between spring
2020 and fall 2021. The landslide is outlined in purple; the warm tones are
areas of soil gain, and the cool tones are areas of soil loss.
Investigation of this site indicates
that the landslide is active with a significant amount of change at the toe
slope, and that it is moving downhill and to the right of the pipeline (red
line). The change detection of the toe slope (lower right of the landslide)
shows the largest amount of vertical soil gain.
Based on Teren’s ranking algorithm, this
feature was classified as a 6 on a scale from 1 to 10, with 10 being the
highest ranking. This ranking was assigned because the scarp slope intersects
the pipeline (end point), is moving away from the pipeline (downslope), is
relatively large (0.5 acres) and is on a gradual slope (17 degrees).
The results of the IMU-based pipeline
movement and LiDAR-based geohazard assessments are in strong agreement. Based
on the time intervals over which the two datasets were collected, it is
possible to attribute causation. In addition, it is possible to quantify the
amount of pipeline movement that has taken place as well as to determine the
characteristics of the landslide.
Although additional material limit calculations and a coincident anomaly review would be recommended during the prioritization stage for cases like this one, the presence of active movement of both the ground and the pipeline are key indicators that remedial interventions are likely required for this site.
Advancing
Forward
The presence and influence of geohazards
pose many pipeline integrity challenges because of the increased frequency of
soil-mobilizing extreme weather events, public awareness and regulatory
requirements. As concerns regarding ground movement and pipelines continue to
grow, finding efficient and scalable solutions will be demanding.
With powerful analytical tools at their
disposal, integrity personnel can be equipped with modern methods of identifying
and tracking geohazards, leveraging highly efficient processes and emerging
technology. Landslides can be detected using LiDAR digital terrain models and
AI analysis, while pipeline movement can be revealed using IMU-equipped in-line
inspection. In tandem with these measures, additional checks against material
limits and interactions with other pipeline anomalies allow for the efficient
and informed prioritization of areas where movement has been detected.
As the pipeline industry continues to
adapt and grow to meet modern challenges, such strategies will be required to
ensure the safety of operational assets. P&GJ