and mapping services. Additionally,
Chevron, a major oil and gas operator,
will evaluate the technology’s future
commercialization.
Thousands of oil spills occur in US
waters each year for a variety of reasons.
While most are small, spilled crude
oil can still cause damage to sensitive
areas such as beaches, mangroves, and
wetlands. When larger spills happen,
pipelines are often the culprit. From 1964
through 2015, a total of 514 offshore
pipeline–related oil spills were recorded,
20 of which incurred spill volumes of
more than 1000 barrels, according to the
Bureau of Ocean Energy Management.
The timely inspection of subsea
infrastructure, especially pipelines and
offshore wells, is the key to preventing
such disasters. However, current inspec-
tion techniques often require a well-
trained human diver and substantial time
and money. The challenges are exac-
erbated if the inspection target is deep
underwater.
The SmartTouch technology now in
development at UH consists of remote
operated vehicles (ROVs) equipped with
multiple stress wave–based smart touch
sensors, video cameras, and scanning
sonars that can swim along a subsea
pipeline to inspect flange bolts. Bolted
connections have accelerated the rate of
pipeline accidents that result in leakage,
according to the Bureau of Safety and
Environmental Enforcement (BSEE).
BSEE is funding the project with a
US$960 493 grant to UH researchers
Zheng Chen, Bill D. Cook (Assistant
Professor of Mechanical Engineering)
and Gangbing Song (John and Rebecca
Moores Professor of Mechanical
Engineering), who are working in collab-
oration with Oceaneering International
and Chevron.
“By automating the inspection process
with this state-of-the art robotic tech-
nology, we can dramatically reduce the
cost and risk of these important subsea
inspections, which will lead to safer oper-
ations of offshore oil and gas pipelines
as less intervention from human divers
will be needed,” said Chen, noting that
a prototype of the ROV has been tested
both in his lab and in Galveston Bay. The
experiments demonstrated the feasibility
of the proposed approach for inspecting
the looseness of subsea bolted connec-
tions. Preliminary studies were funded by
UH’s Subsea Systems Institute.
“Corrosion is responsible for most
small leaks, but the impacts can still
be devastating to the environment.
Therefore, our technology will be highly
accurate in monitoring corrosion and will
also help mitigate the chances of pipe-
line failure from other factors,” said Song,
who has conducted significant research
in piezoelectric-based structural health
monitoring. His prior research efforts
include numerous damage detection
applications, such as crack detection,
hydration monitoring, debonding, and
other structural anomalies.
The SmartTouch sensing solution will
open the doors for inspection of other
kinds of subsea structures, according
to the researchers, by forming a design
template for future robotic technologies.
“Ultimately, the project will push the
boundaries of what can be accomplished
by integrating robotics and structural
health monitoring technologies. With
proper implementation, the rate of
subsea pipeline failure and related acci-
dents will decrease, and subsea opera-
tions will be free to expand at [a] faster
rate than before,” added Chen.
USING DEEP LEARNING
TO CLASSIFY
STEEL MATERIALS
OBJECTIVELY
The material typically used for rolling
bearings is surface-hardened steels
enriched with carbon. Surface hardening
is used to make the material durable so
as to prevent failure, fatigue, and crit-
ical crack growth in the components
due to cyclic loading, thereby averting
serious accidents. Critical microstruc-
tural characteristics in steel alloys are
typically nonmetallic inclusions and
larger-than-average grains. These crystal-
lites are formed during the steel produc-
tion process and are subject to constant
changes throughout the value chain. The
mechanical properties of steel are largely
determined by the grain size. Therefore,
for quality control purposes, it is essential
that they are reliably assessed.
Researchers at Fraunhofer IWM in
Freiburg, Germany, in collaboration
with Schaeffler Technologies AG &Co.
KG, have developed a deep learning
model for determining the grain size
of martensitic and bainitic steels (steels
with a hardened microstructure caused
by abrupt cooling). This model is
designed to supplement or replace the
time-consuming visual inspection carried
out by trained metallographers, who,
until now, have determined and classi-
fied the defects visually. They primarily
look for larger grains and other flaws, as
this is where the likelihood of failure is at
its greatest. However, as an interlabora-
tory round robin test revealed, the grain
size evaluations carried out by experts
differ from one another. Aside from the
grain size classification being inherently
subjective, the experts’ ratings were
shown to be inaccurate at times and
thus not sufficiently reliable—especially
for safety-relevant applications. The stan-
dard inspection procedure is also prone
to errors because it is based on small
samples and, due to the large amount
of work involved, the inspection of the
entire component is not feasible.
The deep learning model for grain
size determination, on the other hand,
can be used to assess arbitrarily large
component areas and exhibits high
accuracy and ideal reproducibility. To
achieve this, the model was supplied
with image data that previously had
been classified by experts. These
images from the industry partner were
used to train the model to recognize
and classify microstructures in steel.
The innovation here is that the grain size
can be assessed in a consistently objec-
tive and automated manner. The data
are subject to annotation noise due to
differences in the way metallographers
assess grain sizes. However, when opti-
mizing the model, the noise, that is, the
erroneous assessment, is filtered out.
By continuously receiving images anno-
tated with over- and underestimation of
SCANNER
|
INDUSTRYNEWS
12
M A T E R I A L S E V A L U A T I O N • J A N U A R Y 2 0 2 4
2401 ME January.indd 12 12/20/23 8:01 AM
Chevron, a major oil and gas operator,
will evaluate the technology’s future
commercialization.
Thousands of oil spills occur in US
waters each year for a variety of reasons.
While most are small, spilled crude
oil can still cause damage to sensitive
areas such as beaches, mangroves, and
wetlands. When larger spills happen,
pipelines are often the culprit. From 1964
through 2015, a total of 514 offshore
pipeline–related oil spills were recorded,
20 of which incurred spill volumes of
more than 1000 barrels, according to the
Bureau of Ocean Energy Management.
The timely inspection of subsea
infrastructure, especially pipelines and
offshore wells, is the key to preventing
such disasters. However, current inspec-
tion techniques often require a well-
trained human diver and substantial time
and money. The challenges are exac-
erbated if the inspection target is deep
underwater.
The SmartTouch technology now in
development at UH consists of remote
operated vehicles (ROVs) equipped with
multiple stress wave–based smart touch
sensors, video cameras, and scanning
sonars that can swim along a subsea
pipeline to inspect flange bolts. Bolted
connections have accelerated the rate of
pipeline accidents that result in leakage,
according to the Bureau of Safety and
Environmental Enforcement (BSEE).
BSEE is funding the project with a
US$960 493 grant to UH researchers
Zheng Chen, Bill D. Cook (Assistant
Professor of Mechanical Engineering)
and Gangbing Song (John and Rebecca
Moores Professor of Mechanical
Engineering), who are working in collab-
oration with Oceaneering International
and Chevron.
“By automating the inspection process
with this state-of-the art robotic tech-
nology, we can dramatically reduce the
cost and risk of these important subsea
inspections, which will lead to safer oper-
ations of offshore oil and gas pipelines
as less intervention from human divers
will be needed,” said Chen, noting that
a prototype of the ROV has been tested
both in his lab and in Galveston Bay. The
experiments demonstrated the feasibility
of the proposed approach for inspecting
the looseness of subsea bolted connec-
tions. Preliminary studies were funded by
UH’s Subsea Systems Institute.
“Corrosion is responsible for most
small leaks, but the impacts can still
be devastating to the environment.
Therefore, our technology will be highly
accurate in monitoring corrosion and will
also help mitigate the chances of pipe-
line failure from other factors,” said Song,
who has conducted significant research
in piezoelectric-based structural health
monitoring. His prior research efforts
include numerous damage detection
applications, such as crack detection,
hydration monitoring, debonding, and
other structural anomalies.
The SmartTouch sensing solution will
open the doors for inspection of other
kinds of subsea structures, according
to the researchers, by forming a design
template for future robotic technologies.
“Ultimately, the project will push the
boundaries of what can be accomplished
by integrating robotics and structural
health monitoring technologies. With
proper implementation, the rate of
subsea pipeline failure and related acci-
dents will decrease, and subsea opera-
tions will be free to expand at [a] faster
rate than before,” added Chen.
USING DEEP LEARNING
TO CLASSIFY
STEEL MATERIALS
OBJECTIVELY
The material typically used for rolling
bearings is surface-hardened steels
enriched with carbon. Surface hardening
is used to make the material durable so
as to prevent failure, fatigue, and crit-
ical crack growth in the components
due to cyclic loading, thereby averting
serious accidents. Critical microstruc-
tural characteristics in steel alloys are
typically nonmetallic inclusions and
larger-than-average grains. These crystal-
lites are formed during the steel produc-
tion process and are subject to constant
changes throughout the value chain. The
mechanical properties of steel are largely
determined by the grain size. Therefore,
for quality control purposes, it is essential
that they are reliably assessed.
Researchers at Fraunhofer IWM in
Freiburg, Germany, in collaboration
with Schaeffler Technologies AG &Co.
KG, have developed a deep learning
model for determining the grain size
of martensitic and bainitic steels (steels
with a hardened microstructure caused
by abrupt cooling). This model is
designed to supplement or replace the
time-consuming visual inspection carried
out by trained metallographers, who,
until now, have determined and classi-
fied the defects visually. They primarily
look for larger grains and other flaws, as
this is where the likelihood of failure is at
its greatest. However, as an interlabora-
tory round robin test revealed, the grain
size evaluations carried out by experts
differ from one another. Aside from the
grain size classification being inherently
subjective, the experts’ ratings were
shown to be inaccurate at times and
thus not sufficiently reliable—especially
for safety-relevant applications. The stan-
dard inspection procedure is also prone
to errors because it is based on small
samples and, due to the large amount
of work involved, the inspection of the
entire component is not feasible.
The deep learning model for grain
size determination, on the other hand,
can be used to assess arbitrarily large
component areas and exhibits high
accuracy and ideal reproducibility. To
achieve this, the model was supplied
with image data that previously had
been classified by experts. These
images from the industry partner were
used to train the model to recognize
and classify microstructures in steel.
The innovation here is that the grain size
can be assessed in a consistently objec-
tive and automated manner. The data
are subject to annotation noise due to
differences in the way metallographers
assess grain sizes. However, when opti-
mizing the model, the noise, that is, the
erroneous assessment, is filtered out.
By continuously receiving images anno-
tated with over- and underestimation of
SCANNER
|
INDUSTRYNEWS
12
M A T E R I A L S E V A L U A T I O N • J A N U A R Y 2 0 2 4
2401 ME January.indd 12 12/20/23 8:01 AM
