1.
© THE FINANCIAL TIMES LIMITED 2013
WEDNESDAY NOVEMBER 20 2013
A
utomation and
digitalisation have long
been a factor driving
productivity in oil and gas. Now,
large corporations need to take
automation to the next level to
address industry challenges such
as increasing operational risk
and complexity with decreasing
front line experience. To fully
reap the benefits, oil and gas
companies will have to adopt
some of the practices of big, data
intensive industries like
aerospace to become more agile
and able to match the speed of
the information and automatic
changes that digital technologies
will generate.
Four challenges
The challenges driving the call
for increased automation and
digitalisation are often relevant
to other asset-intensive
industries as well but the focus
of this article is on oil and gas.
The first major challenge is the
increasingly complex operations
that oil and gas companies are
undertaking. Most oil majors
have a long and thin tail of
mature assets that require an
efficient maintenance schedule
within tight operating expenses
to be profitable. At the same
time, increasingly remote
locations, tough geologies, high
pressure/high temperature
resources, and deepwater wells
require companies to run reliable
remote operations and improve
their supply chain management.
In addition unconventional oil
and gas such as oil sands and
shale oil shale necessitate even
more efficient, manufacturing-
like operations.
The second challenge is the
zero tolerance for health, safety
and environmental incidents.
Addressing this challenge
requires real time monitoring of
assets and real time or near real
time assessment of component
health to predict and avoid costly
failures. Increasingly, safety will
mean moving people out of
harm’s way, using automation to
replace people in the highest risk
roles.
The third challenge is the
increasing talent and experience
gap. Experienced personnel,
often hired in the late 1970s and
early 1980s, are currently retiring
and this requires documenting
and industrialising processes by
using digital technology to avoid
the loss of valuable knowledge
and experience. In addition,
human beings are not consistent
decision makers; presented with
the same facts, different people
will make different decisions.
This can be highly costly in
critical situations where only one
choice is the right choice – data
driven decision-support is
essential for safety, efficiency,
and yield.
The fourth major challenge is
the growing data overflow
already influencing businesses
today. Given that any single
asset may have in excess of
50,000 electronic tags, and that
the data from these might be
structured in different formats as
well as located in different
systems and subsystems, the
ability to connect the dots by
developing algorithms to
interpret the data are crucial.
In addition the complexity of
asset portfolios, and thus the
typical supply chain, have
increased. Moreover, growing
fragmentation of the supply
chain as specialisation increases
will require aggregation of data
across operators such as oilfield
services and equipment and
other vendors.
Defining and applying
automation
There are many attempts in the
literature to define exactly what
is meant by automation in a
technology context. We define it
broadly as the use of technology
(eg, sensors, data, storage,
connectivity, the Internet,
computing power, and control
systems) for performing tasks
that, for instance, are dangerous,
high-cost, error-prone, or of low
value, and making or aiding
decisions with reduced human
intervention. This definition is
wide enough to include
everything from self-driving cars
to big data visualisation of (eg a
3D image of the geology of an
oilfield) but it allows us to
address specific elements, such
as sensor connectivity, within a
framework.
We believe the future of
automation is being shaped by
the emergence of the Internet of
Things, which refers to the
massive increase in devices
equipped with the ability to
connect to the internet, usually
through wireless networks; thus
making what used to be isolated
entities, such as artificial lift
pumps, part of a large and
growing network of networks
able to communicate with each
other.
An automation system connects
observations about the physical
world to changes in that world
electronically. Traditionally
these types of systems were
closed and proprietary. However,
during the past few years there
has been a significant change in
this system architecture – a
change that will continue to
shape the use of automation in
the next decade. We are
beginning to see networked
sensors that can be identified
and authenticated via directory
technology, with identity
management that is similar to
what’s seen on the Internet for
people. Systems today are
connected to networked data
sources in a much more open
architecture that enables
companies to assemble the
building blocks to meet their
specific demands.
Imagine what is possible with
sensors connected to a network
map, with the mapping data and
sensor data being aggregated
back into a data store able to
blend heterogeneous feeds from
many sources. Having such a
repository makes analysis
possible, which in turn is driving
algorithmic innovation to extract
the required information from a
vast tsunami of data.
Finally, with these huge
amounts of data, visualisation
and presentation of data then
becomes a critical operating
element connecting the ability to
assess that data to human
cognitive capacities. The second
main part is close loop actuation
that gives feedback and potential
corrective actions to eg the
control systems, which can really
drive changes in operational
performance. It can be as simple
as opening and closing a valve or
it could be as complicated as
steering an autonomous
(unmanned) vehicle through a
complicated course. Clearly, the
structures of automation systems
are changing fundamentally and
becoming more open, which in
turn will allow for much more
open innovation in the next
generation of automation.
FT.COM
Building the digital oil and gas enterprise
By Hans Henrik Knudsen, Tor Jakob Ramsøy and Roger Roberts

2.
When applying, Internet of
Things-driven, innovation to real
world business problems in oil
and gas, we see six important
categories of value.
1. Tracking behaviour. This
means monitoring people and
things through time and space.
2. Enhancing situational
awareness. This is about
achieving real time awareness of
the physical environment, which
could be at the bottom of the sea
or on a remotely operated
platform.
3. Central-driven (eg in a global
operating model) decision
analytics. This is about
connecting people to global
information and best practice
insights and allowing them to do
deep analysis and use data
visualisation to see the answer
quickly from very complicated
information flows.
4. Automated control of closed-
loop subsystems in a more fine
grained and dynamic way. One
example is a refinery process
control system optimising
production through improved
assets utilisation
5. Optimising resource
consumption across a distributed
network. Here you may think
about a large scale heavy oilfield
that is using algorithms to
balance steam injections into the
network based on historical data
and live feed of current steam
requirements.
6. Control of complex
autonomous systems. This could
be anything from automated
warehousing to Google’s
autonomous cars to the
application of these technologies
to airborne vehicles that are
autonomous in terms of their
ability to execute tasks,
relatively independently in a
complicated and rapidly
changing environment. For an
oil company this could be robots
executing preventive
maintenance.
All of these applications will be
the subject of significant
investment going forward, in the
energy industry and elsewhere.
Winning by automating
The automation winners of the
future will be data driven and
agile corporations that are able
to change as they receive new
information, with the tools in
place so that their assets
physically change themselves
based on the new data without
human intervention. Data-driven
decision-making will mean that
from the board room to the field
engineer, decisions will be based
on a solid fact base and an
understanding of the data. The
starting point will be an
automated recommendation
created by big data analytics.
Companies will also need to
adopt a more agile approach,
incorporating key elements from
the Lean toolbox. In an
automated world, the agile
method of software development,
as well as the Lean continuous
improvement mindset, become
more critical. Leveraging the
many small ideas of employees
on how to tweak key algorithms,
and their suggestions on how to
better use data, will be very
important.
Becoming more agile, for
example, when dispatching work,
will be enabled by the
deployment of mobile devices
such as tablets to highlight, for
instance, a safety hazard spotted
by an engineer. A mobile device
will also allow this engineer to
submit a process improvement
suggestion. These changes can
then, via tablets, be rolled out to
the rest of the employees.
The commercial potential of
automation and digitalisation is
tremendous in the oil and gas
industry. The winners will be
those that use technology to
smartly tackle their biggest
challenges, and are able to shed
their old ways of working to
become truly agile and
datadriven corporations.
Hans Henrik Knudsen is a
consultant in McKinsey &
Company’s Copenhagen office.
Tor Jakob Ramsøy is a director
in their Oslo office and Roger
Roberts is a principal based in
their Palo Alto office
© THE FINANCIAL TIMES LIMITED 2013