9 / 32
I N N O V AT I O N S • V O L . V I I I , N O. 1 • 2 0 1 6
7
Operators Driving Positive Change
Operators investing in modern ILI tools are seeing
more and more potential calls for corrosion of
or along the long seam, sometimes hundreds in
just a few dozen kilometers. The initial concern
with multiple indications of corrosion is how
to prioritize the threats. Which ones should the
company dig first to maintain the safest possible
pipelines – to get to the biggest threats quickly?
To overcome this issue, T.D. Williamson
(TDW), a pipeline services company, is developing
a classification model to help prioritize SSWC
corrosion threats.
To develop the model, the TDW team first
needed some base data points. Partnering with a
U.S.-based pipeline operator, TDW selected areas
to dig and analyze in the ditch, then compared
these data sets to existing ILI data produced by
the TDW proprietary Multiple Dataset (MDS)
platform, which provides various inspection
technologies on one tool.
Some high-level data analysis, regressions, and
many late nights later, data scientists were able to
produce a SSWC classifier model – a mathematical
model that uses the ILI data not only to establish
the probability that corrosion of or along the seam
is in fact SSWC, but also to estimate the depth and
length of the corrosion.
Better yet, this new model enables classifications,
at least in part, to be automated – eliminating many
of the man hours needed to manually prioritize a
run with hundreds of anomalies.
Fast Reaction to the Most
Imminent Threats
Results are promising.
In one preliminary test, more than 175
anomalies were identified, out of which 90
candidates were selected as potentials for SSWC.
With the new classifier model, just 14 were
categorized with a probability of 40 percent or
greater to be SSWC, and only 11 indicated a
probability of greater than 70 percent.
The research was presented to the PHMSA
Central Region in Kansas City, Kansas, in
December 2015. “They were keen on the work
and interested in the results as they develop,”
says Chuck Harris, manager of strategic
commercialization at TDW. “They even requested
to be in the ditch on some of the digs.”
Ideally, the ILI runs will one day generate a
complete prioritization of potential SSWC threats,
so operators can start their digs at the top of the
list and work their way down. The technology is
currently in field trials, and additional results will
be reported later in 2016.
New Classifier Model for Selective
Seam Weld Corrosion (SSWC)
as sampled in preliminary testing
175+
Anomolies identified
90
Candidates for SSWC
14
40% probable
11
70% probable
SSWC is of particular
concern for a few reasons:
• Traditionally difficult to detect with
conventional inline inspection (ILI) tools.
• SSWC can cause a pipeline to rupture
even under low stress conditions.
• Selective corrosion grows in service
faster than adjacent pitting corrosion.