30 / 56
I
n recent years, the marine and offshore industries have
benefited from a rapid expansion in the use of remote
inspection technologies (RITs). Their advantages, including
safer, more efficient and lower cost procedures, have seen
RITs such as unmanned aerial vehicles, remotely operated
underwater vehicles (ROVs) and robotic crawlers widely used
for inspection of offshore risers, mooring chains, cargo tanks
and confined spaces.
One of the major challenges during remote inspection is
performing a coating condition assessment. Consequently,
the development of more effective RITs has enabled safer and
more efficient visual inspections of the condition of protective
coatings in historically hard to access and dangerous locations
on marine and offshore assets.
The application of protective coatings to steel surfaces to
prevent structures from corroding is well-understood, and the
benefits clear in terms of reducing risks during operation and
potential damage to marine and offshore steel structures.
Proper maintenance of the coating, leading to a longer
service life for the assets, further benefits the owners with a
decrease in asset lifecycle cost, but the scale of the task in
keeping coatings in good condition is considerable.
A study fromUS National Association of Corrosion Engineers
(NACE) estimated the annual cost of coatingmaintenance for
the shipping industry in the US alone at US$2.7 billion.
Visual inspections by properly trained and highly
experienced surveyors continue tomake up themajority of
maintenance surveys, but the growing use of RITs provides an
opportunity to augment human skills with computing power.
And while RITs can providemore convenient access for
inspectors evaluating the condition of coatings onmore flexible
schedules, there are twomajor issues with the current scope
of RITs.
Inspectionchallenges
RITs are potentially transformational in their approach to
onboard inspection but can be challenging for inspectors when
identifying potential coating failures because they generate a
significant amount of data.
For example, in early 2019 during a drone inspection
onboard a barge carrier in the Great Lakes area, approximately
40 GB of images and videos were generated for a coating
assessment of one single cargo tank. All the data generated
needs to be visually examined by inspectors within a certain
amount of time, either onsite or back in the office. The task is
both time-consuming and often tedious, creating the potential
for flaws to go undetected.
The second issue concerns the objectivity of the inspector
(sometimes called inter-inspector variability) during the
decision-making process. Generally, for coating condition
assessments, inspectors need to estimate the size and severity
of each coating failure area, some of which are separated
sparsely or with complex shapes that are difficult to evaluate
accurately through just visual inspection.
Gu Hai, ABS, Singapore,
explores how the application of machine learning
can be used for coating condition assessments in offshore structures.
COATING
CONUNDRUMS
28 |