Adrian Banica,
Synodon, Canada,
presents support
for a new,
technology-enabled
method for detecting
small pinhole
hydrocarbon
pipeline leaks.
L
eak detection is a key part of a liquid
pipeline risk management system and
several different approaches are used to
ensure early detection of releases. These
approaches include SCADA based leak detection,
acoustic and fibre sensors, aerial patrols and
sometimes ground patrols. Historically, operators
have shown a strong preference for leak detection
that utilises field equipment that is already in place.
This accounts for the dominance of leak detection
by pressure/flow monitoring and CPM on pipelines,
since the monitoring is already provided by
the SCADA system, and the CPM is a relatively
inexpensive addition to an existing metering
infrastructure.
Needs assessment
PHMSA published a report entitled Leak Detection
Study DTPH56-11-D-000001.
1
Notionally, this
report highlights an opportunity for additional
technologies to be applied to enhance pipeline
leak detection. This report aggregates data over a
large number of leak sizes and as such, does not
accurately reflect the real-life performance of CPM
systems.
Leak detection by CPM is limited by the
accuracy of the metering and uncertainties in the
line fill, both of which can have errors in the 1 - 5%
of the pipeline flowrate (certain ideal pipeline
configurations can achieve slightly better results
over shorter distances but this looks at an industry
wide view). Therefore, conventional SCADA leak
detection thresholds are typically at best 1% of
throughput.
To get a clear understanding of the
effectiveness of CPM at detecting various leak
sizes, the raw leak data catalogued by PHMSA was
analysed. Figure 1 breaks down the performance
of the CPM system per leak size category and, as
25
