DATA-DRIVEN SMART

MAINTENANCE:
THE NEXT GENERATION
OF EQUIPMENT MONITORING
AND MAINTENANCE OPERATIONS
© COGNITE
CONTENT

3 Introduction

5 Cognite Data Fusion (CDF)

7 Asset Data Insight (ADI)

10 Smart Maintenance in Action: 5 Use Cases

18 How Cognite Supports Your Pursuit

of Smart Maintenance

2
© COGNITE
DATA-DRIVEN
SMART Introduction
MAINTENANCE:
THE NEXT GENERATION Smart Maintenance is a set of data-
driven opportunities that increase
OF EQUIPMENT uptime, boost performance, reduce
MONITORING over-maintenance, and improve
AND MAINTENANCE planning.
OPERATIONS
But for too many companies in heavy-
asset industries, Smart Maintenance is
an opportunity that remains unseized.

There are many reasons why industrial

companies haven’t been able to scale
Smart Maintenance:

1 Choosing the wrong use cases, 4 Forgetting to change as an

often because of a lack of in-house organization, shaking up work
analytics competence. processes and management
structures to support a data-driven
way of working.

2 Development issues, sometimes 5 Finally, not focusing on the

connected to the complexity and scalability of solutions from the
time required to build the beginning. Case studies, pilot
necessary data models. projects, and proofs of concepts
must be scalable to maximize their
value.
3 The lack of an appropriate hosting
environment to deploy the models.

This paper will explain how Cognite’s Smart Maintenance Service Line helps customers solve these
issues and capture value. Through a combination of subject-matter expertise and technological
solutions such as Cognite Data Fusion (CDF), a data package that supplies contextualized data to 3
drive industrial applications that increase efficiency and drive revenue; and Asset Data Insight, a
dynamic application for operations and maintenance enabling real-time health-overviews and
analysis, Cognite lays the foundation for companies to extract the massive value potential in Smart
© COGNITE

Maintenance.
 From Reactive to Proactive

Smart Maintenance will substantially impact Smart Maintenance means a move to a

how companies in heavy-asset industries smaller, more centralized maintenance
operate. Today, many companies run large operation with end-to-end ownership. Data is
maintenance operations with distributed easy to access, and can be shared with
ownership. Information such as equipment technology partners and OEMs. Maintenance
health history, sensor data, and is performed proactively as a result of the
documentation are locked away in different actual condition of each component and data-
systems, meaning that alarms, calendars, and driven analyses that flag potential problems
personal experience drive maintenance before they happen. Communication flows
analyses and decisions – not data. seamlessly across the organization, making it
simple to weigh costs and benefits to optimize
operational decisions.
In addition to a lack of data, there is also a
lack of communication; collaboration is
hindered by too many middlemen and This transformation won’t happen on its own.
functional silos, coordination between To unlock the insights that can transform their
maintenance and production optimization maintenance operations, companies in heavy-
4
decisions is all but nonexistent, and Original asset industries first need to collect, clean, and
Equipment Manufacturers (OEMs) are contextualize the data that their installations
are already producing.
© COGNITE

completely disconnected from maintenance

work.
 Cognite Data Fusion (CDF)

CREATE VALUE ACROSS AN

ECOSYSTEM OF PARTNERS
Liberate data by making it possible
for the contextualized data to be
used to create value across an
ecosystem of partners
DROWNING IN DATA
STARVING FOR INSIGHT
Liberate data by turning it into
contextualized knowledge, thus
unlocking data from its raw state
LIBERATE DATA limitations
OUT OF SILOS

Few companies have successfully scaled Smart

Maintenance. Of the many obstacles they
face, one stands out: a lack of liberated,
usable, and shareable industrial data.

Cognite Data Fusion (CDF) was developed

specifically to tackle that challenge. With CDF,
contextualized data is available as a service,
easily shareable through a secure Application
Programming Interface (API).

CDF integrates seamlessly with existing IT

and OT infrastructures to liberate a wide
variety of industrial data from separate, siloed
source systems, collecting it all as
a comprehensive set in the cloud, securely
and without space limitations. It then
automatically structures the sensor data
in relation to other relevant data (e.g., process
diagrams, 3D models, event data, and more).
This contextualization process effectively
creates a comprehensive digital twin 5

of an asset or system, making data available in

a way that’s intuitive to the human user and to
© COGNITE

the algorithms at work behind the applications

powering Smart Maintenance models.
 PROBLEM
Liberate Your Data Model development is slow as feature
engineering on time series data is
computational heavy and often done on local
Where companies run into model
computers.
development problems, CDF offers a solution:

SOLUTION
Companies can perform computationally
PROBLEM
heavy processes in CDF, tapping into the
Data access is cumbersome, as information is
cutting-edge computational power and
stored across multiple data sources,
functionality of the cloud. Google Cloud
potentially with different data owners and
Platform.
system experts.

SOLUTION
Data has the potential to be industry’s most
With CDF, companies get immediate access to
valuable asset. That’s what Cognite Data
historical and real-time equipment data
Fusion reveals: A complete, virtual view of
through one standardized and well-
industrial reality – both past and present –
documented API.
made accessible and meaningful for humans
and machines.

PROBLEM
Relevant data is hard to identify, as there is
limited context available to allow rapid
identification of all data relevant for one
particular piece of equipment.

SOLUTION
With CDF, companies can identify all relevant
data quickly, using data typing to identify all
data relevant for the equipment.

PROBLEM
Labeled data is sparse, of low quality, or
nonexistent.

SOLUTION
With CDF, customers can access data labels
through the same interface as time series 6
data, and its toolbox provides efficient
labeling in cases of low-quality or nonexistent
© COGNITE

labels.
 Asset Data Insight

ASSET DATA INSIGHT IS YOUR ENTRY POINT

FOR SMART MAINTENANCE

Cognite Data Fusion (CDF) vastly improves By enabling users to create a model for one
companies’ data science workflow. Once they component and then roll it out to an entire
have a pipeline of liberated, usable, and installation, ADI minimizes the manual effort
shareable industrial data in place, companies required from users, freeing up time for
can then implement applications that higher-level analyses and effective decision-
accelerate solving Smart Maintenance use making.
cases.

Asset Data Insight (ADI) is one such For problems that require deeper
application. Developed by Cognite, ADI investigation, users can efficiently navigate
enables real-time health monitoring and from Asset Data Insight into specialized
investigation. Through in-depth data applications or dashboards such as Grafana
exploration, users are able to create detailed and Power BI, among others.
analyses and make informed, effective
decisions at every level of an organization.

Asset Data Insight transforms the cleaned and

contextualized data within Cognite Data 7
Fusion into actionable insights via asset-
centric dashboards that intuitively visualize
© COGNITE

links between different data sources.

 Meaningful Health Overviews, sensor time series data, stored in another
Effective Investigation system. She consults design information,
design curves, and Piping and Instrumentation
Today, much of the maintenance work that Diagrams (P&IDs), all located in separate PDF
heavy-asset companies conduct is governed files.
by calendars. At set intervals, be it once a year
or once a month, an information system will This siloed approach to data management
automatically generate a work order to check causes excessive downtime, as engineers
the health of a piece of equipment, and a field spend excessive amounts of time identifying
worker will complete the task. This causes the problem before they can figure out how to
over-maintenance, as workers spend time fix it. And since an alarm can sound at any
checking equipment that may be in fine time, it upends maintenance schedules
working condition. offshore and planning processes onshore. As a
result, there is little time for advanced
Then an alarm sounds on the installation. analyses or learning from previous cases of
A pump has stopped functioning as intended. equipment failure.

One of the installation’s maintenance workers

– be it an asset maintenance manager,
offshore planner, onshore equipment expert,
or any one of the many users involved in
maintenance – is on it. But in order to When I’m notified, it’s too late to solve
diagnose the issue, she has to consult several the problem. I wish we had the time
different systems that don’t communicate with and tools to catch problems before
one another. She checks the equipment they affect our control room operators.
history, contained in one system. She looks at
– Equipment Engineer

WITHOUT COGNITE DATA FUSION AND ASSET DATA INSIGHT

End user looks into

MANY different systems

Application
Grafana
dashboards (e.g.,
dashboards SparkCognition)

8
CMMS Document
Historian
(e.g., SAP, mgmt. (e.g.,
(e.g., Pi)
Workmate) Documentum)
Vibration Oil Reports from
© COGNITE

service
Control system (e.g., monitoring providers (e.g.
System 1) (e.g., Invicta) Karsten Moholt)

Condition monitoring
All of these frustrations revolve around one Through a combination of predetermined
issue: centralized and complete data access. performance criteria and their own expertise,
users can identify problems before they
Asset Data Insight solves that issue. The happen and arrange for them to be fixed,
application brings health monitoring insights avoiding costly production shutdowns.
from across an ecosystem of courses together
into a single view, so that users can find the Greater data availability lowers the threshold
data they need in one application. Related for using data in day-to-day operations.
industrial data in the same view provides the Instead of being something reserved for office
necessary context for users to analyze health workers, data becomes another resource in
monitoring insights and take quick actions. the field worker’s toolbox. This supports
bottom-up digital transformation, ensuring
These features help Asset Data Insight that organizations change along with their
transform the workday and daily effectiveness maintenance routines.
of maintenance workers. Instead of leaping to
action when an alarm sounds, maintenance
workers can at any time access the application
on a computer or handheld device to
proactively monitor equipment health.

ACCESS ANY RELEVANT DATA THROUGH ONE UNIFIED INTERFACE

End user can access any data

through one, unified interface

ASSET DATA INSIGHT

Application
Grafana dashboards and
dashboards and models (e.g.,
models SparkCognition)

COGNITE DATE FUSION

CMMS Document
Historian
(e.g., SAP, mgmt. (e.g.,
(e.g., Pi)
Workmate) Documentum)
Vibration Oil Reports from
service
Control system (e.g., monitoring providers (e.g.
System 1) (e.g., Invicta) Karsten Moholt)

Condition monitoring

9
© COGNITE
SMART
MAINTENANCE
IN ACTION

Smart Maintenance is not some far-off fantasy. It’s industrial reality, happening at installations
around the world, driven by digital frontrunners. The following collection of use cases
demonstrate how companies in heavy-asset industries, powered by Cognite Data Fusion, can
sharpen their competitive edge by maximizing their own capabilities with data.

How Cognite helped Aker BP

save time and costs by
automating maintenance

CHALLENGE Then, they had to use the system to pull up

Undertaking PSD (Process Shutdown) valve data on each individual valve, going one by
testing on Aker BP’s Valhall field used to be one in the slow system until every valve that
time-consuming work. First, engineers had to was scheduled for maintenance during that
obtain a work permit to access Aker BP’s session had been checked. This approach 10
offshore system. meant that engineers were spending part of
their shifts testing valves in good working
© COGNITE

condition instead of those in need of

maintenance.
 SOLUTION checking each valve and toward prioritizing
Aker BP used Cognite Data Fusion (CDF) to valves that were not meeting travel time
automate and optimize the PSD valve criteria. This change was estimated to have
maintenance process at the Valhall field. CDF reduced the number of hours that engineers
captured data from Aker BP’s control system, spent on testing PSD valves a year by about
logging when valves were used and 80 percent.
comparing the results to travel time criteria.
It also cut the number of annual maintenance
The data was then visualized in Asset Data checks by two-thirds, and the duration of an
Insight, an application for computers and average maintenance session in half.
handheld devices. The dashboard gave
operators and first-line onshore support staff By reducing the time it took to check safety-
an easy-to-use dashboard for monitoring valve critical equipment, Aker BP was able to
health. reinvest those resources toward maintaining
production-critical equipment.

IMPACT
By using Asset Data Insight to inform their
maintenance routines, Aker BP’s engineers
were able to move away from manually

-50 % -80 % -66 %

Time reduction for an average Reduction in hours engineers Reduction in annual

maintenance sessions spend on testing maintenance checks

11
© COGNITE
 How Cognite liberated data to
provide fast, accurate condition
monitoring

CHALLENGE Cognite Data Fusion offered the capability to

Aker BP granted its control system provider collect, clean, and contextualize more various
access to all Ivar Aasen field data. The kinds of data automatically and without space
company pulled process data through their limitations. With Aker BP’s authorization, the
control system and most of the third-party company redirected their system to read the
equipment data bypassing the control system company’s data from the CDF, including data
and going straight to the information types that had previously been outside of the
management system (IMS). company’s scope.

However, there remained several other For Ivar Aasen, this meant 150,000 time
systems, usually inaccessible to the company, series, 20 billion data points, 400,000+
that held potentially valuable data, such as documents made available through a single,
work orders, work permits, document secure point of entry.
systems, HMS information, etc. Combining
these sources would vastly enrich their The control system provider estimates that
solution. attempts to run point-to-point integrations
with all the individual source systems and
SOLUTION ingest their data would have taken about
The control system provider turned to 1,500 hours. The process of integrating with
Cognite, whose cloud-native industrial data the CDF alone – pulling in all the same data –
platform was already in operation across Aker took only 200 hours. And a lot of that time
BP’s assets on the NCS. was actually spent looking at the data and
deciding how best to visualize it.

12
© COGNITE
 IMPACT
Artificial intelligence, machine learning, and
tested algorithms are what make the IMS
visualizations intuitive and effective. With
rapid access to live and historical data,
regardless of original source or format, via the
CDF’s single point of entry, the control system
provider was ready to present its powerful
processing, analysis, and display capabilities
to Aker BP.

ARMED WITH THE POWER

OF PREDICTION, THEY CAN ALSO

Boost safety by Order fewer Deploy the right Bring down the
limiting the replacement parts crew and tools cost of the
number of human and reduce the when consequences of
personnel on the “just in case” maintenance equipment failure
platform stock needs do arise (i.e., unplanned
downtime or
production
slowdown)

Especially in offshore environments, condition-

based maintenance (and eventually, predictive
maintenance) has the potential to
revolutionize business models and reduce
bottom lines.

13
© COGNITE
 How Cognite Data Fusion
integrated Power BI’s quick, off-
the-shelf visualization

CHALLENGE
Over the last few years, off-the-shelf
visualization from the likes of Power BI,
Tableau, and TIBCO Spotfire has gained
popularity. However, they are often not
readily available to everyone in an
organization.

OPERATIONAL STATUS DASHBOARD

SOLUTION IMPACT
Cognite Data Fusion can make this Today, large-scale data visualization is
technology available to all. Cognite available to all necessary users in the
developed a simple Power BI integration that organization. In a simple browser view. This
allows any authenticated user to leverage the enables instant access to easy use cases,
Power BI interface connected to Cognite fueling creativity and innovation.
Data Fusion via an authentication key. From
there, the user can utilize the primary time
series engine, as well as any other data they
are permitted to access on Cognite Data 14
Fusion.
© COGNITE
 How Cognite Data Fusion
enabled Framo and Aker BP
to transition to a new smart
service contract

CHALLENGE
Traditional industrial infrastructure has limited
supplier relationships to a transactional
format. The results include static reporting,
uninformed maintenance schedules, and a
waste of money and resources on both sides.

Framo is a global OEM providing pumps to SOLUTION

the marine industry. Historically, they have Working closely with Aker BP, Framo utilized
been unable to access or leverage operational Cognite Data Fusion to access Aker BP’s
data for Oil & Gas. Data has only been made industrial data. With API access to Cognite
available for sporadic projects linked to Data Fusion, Framo was able to inform their
specific, isolated events, such as equipment product development.
failure. If Framo were to have access to their
clients’ industrial data, past and real-time, Integrating with Aker BP’s ERM system, they
they would gain the knowledge needed to could set work orders and enable a feedback
improve and innovate on their existing loop with design and engineering. From there,
products and services. Framo was able to develop and launch more
effective, service-based models informed by
real-time data.
IMPACT
Now that real-time data is instantly available,
users can experiment with advanced analytics
and machine learning to deliver an improved
performance and new products.

This new way of working lays a foundation for

redefined business models, allowing OEMs to
move from selling equipment to service-based
offerings.

-30 % -70 % +40 %

15
© COGNITE

Reduced maintenance Reduced shutdowns Increased pump availability

 How Cognite Data Fusion
enables quick maintenance
overviews of critical equipment

CHALLENGE SOLUTION
Aker BP’s offshore and onshore maintenance Cognite liberated and contextualized data
organization meets every Wednesday to from four different source systems, streaming
discuss how to prioritize the maintenance of them in real time to a Power BI dashboard.
critical equipment on all its installations over The dashboard uses color-coded labeling to
the next week. In order to run an effective help the offshore and onshore maintenance
meeting, the organization creates and organization decide where to focus its
distributes a spreadsheet ahead of time that maintenance resources for the next week and
gives attendees an overview of the discover opportunities for future maintenance
maintenance status of critical equipment. work.

Assembling the spreadsheet is a cumbersome

process that takes up to 15 hours a week, as
the engineers responsible for it have to access -15
a number of different systems to collect the hours
required information, including lube oil charts,
time series data, and maintenance work
orders.

Time saved from automating

data collection

16
© COGNITE
 IMPACT Beyond reducing time-consuming
The maintenance overview dashboard reduces administrative work, the dashboard helps Aker
the time that the offshore and onshore BP’s larger maintenance and operations
maintenance organization spends preparing organization get an overview of the status of
for its weekly meeting, enabling the critical equipment, which makes operations
organization to focus on setting its more efficient.
maintenance priorities. In addition to
providing an overview of critical equipment,
the dashboard also allows for deep dives into
individual work orders to find relevant
information such as estimated time to
completion.

17
© COGNITE
HOW COGNITE
SUPPORTS YOUR
PURSUIT OF SMART
MAINTENANCE

Smart Maintenance is the path to the next But technology alone won’t bring about this
generation of equipment monitoring and digital transformation. Our subject-matter
maintenance operations. experts will sit side-by-side with your own
experienced staff to figure out what will
Powered by future-proof technology deliver the most value for your organization.
designed to accelerate bottom-up digital Together, we will streamline the Smart
transformation, Smart Maintenance secures Maintenance use case process – from
both short- and long-term value generation identification through delivery – so you can
for companies in heavy-asset industries. create end-to-end digital value in your
operations as quickly as possible.
Companies are already collecting the data
needed to power this transformation. What We’ll tap into our ecosystem of partners to
they need are solutions that turn the data into find the ideal partners to build models and
liberated, usable, and shareable information. supply analytics, ensuring that solutions get
up and running quickly and continue to
generate value.

THAT’S WHERE COGNITE’S SMART

Simple use cases such as those detailed in this
MAINTENANCE SOLUTION CAN HELP.
white paper highlight the power of Smart
It begins with Cognite Data Fusion. CDF Maintenance. By liberating and
liberates a wide variety of industrial data from contextualizing data that previously was
separate, siloed source systems, then locked away in silos, companies in heavy-asset
automatically structures the sensor data in industries can slash the time it takes to
relation to the rest. This contextualization complete routine maintenance process by
step makes data available in a way that’s more than half, increasing uptime and
intuitive to the human user and to the avoiding multimillion-dollar production losses.
algorithms at work behind the applications
revolutionizing maintenance work today.

Asset Data Insight is one example of what

companies can do with their contextualized
data. With the ability to see all their industrial
The real potential of Smart 18
data in one unified interface, companies can
Maintenance is in what we can
rapidly scale their Smart Maintenance build together. Get in touch with
capabilities, creating one model for one
© COGNITE

us today to start exploring the

component and then rolling it out to cover an possibilities.
entire installation.
ALL ABOUT Together, we have built Cognite Data Fusion
(CDF), a software package that empowers
COGNITE companies in industries like Oil & Gas, Energy,
Shipping, and Manufacturing to extract value
from their wealth of existing data by
Our Vision transforming it into useful information. CDF
supplies data contextualization as a service,
An industrial world powered by data and delivering reliable, meaningful data to
algorithms, freeing human creativity to shape industrial applications that increase safety and
a productive and sustainable future efficiency and drive revenue.

Curious about Cognite?

Our Team Call on us!

Cognite is a global software company These are just a few of our 250+ industry and
supporting the full-scale digital transformation tech specialists, all proud to be shaping the
of heavy-asset industries around the world, future of Oil & Gas.
from the U.S. to Japan to Austria to New
Zealand. Our impressive interdisciplinary team
includes more than 250 of the best software
developers, data scientists, designers, 3D
specialists, and industry professionals.

Sander van der Meer Director of Executive Sales and Business

Development
Sander spent 25 years in O&G and Power Generation with companies such
as General Electric, Bently Nevada, and Emerson, working in engineering,
services, and leadership roles. A former mechanical and control
and monitoring systems engineer, Sander specialized in large rotating
equipment such as turbines, compressors, and generators.
[email protected] | +4790117639

Paula Doyle SVP of Sales and Marketing

Dr. Paula Doyle holds a PhD in Industrial Automation from the University of
Limerick, Ireland. She has spent the last 15 years working in a variety of roles
within the Oil & Gas industry, for ABB and Siemens, both in Norway and the
Middle East.
[email protected] | +47 409 20 526

Jarle Skrebergene VP of Customer Success // EPC

19
Before joining Cognite, Jarle worked in Management Consulting in Arkwright
and Deloitte and Business Development in Kongsberg Gruppen. He also
© COGNITE

specialized in EPC projects in Kongsberg Oil & Gas. Jarle holds an M. Sc. in
Industrial Economics from NTNU.
[email protected] | +47 959 61 239
Marius Biermann Director of Customer Success
in Industry Solutions
Marius started his career in the O&G service industry, first with Aker Solutions
and later Schlumberger. Before joining Cognite, Marius spent three years as a
management consultant in Arkwright Consulting. Marius holds a M.Eng. in
Mechanical Engineering with Management from the University of Edinburgh.
[email protected] | +47 936 80 906

Jake Bouma Director of Condition Monitoring Services

in Industry Solutions
With a background in nuclear physics and experience building data science
teams across a number of industries, Jake has spent the last two years
focusing on industrial equipment and predictive maintenance, seeing
advanced analytics through from development to production.
[email protected] | + 47 948 52 005

Anna Olsson Director of Partners and Alliances

Anna has held a diverse set of solution sales, product marketing management,
business development, and leadership roles over her 20+ years in the software
industry. She has a degree in Information Technology and started out her
career as a developer. Today, Anna and her team focus on enabling a thriving
partner ecosystem.
[email protected] | +47 982 22 841

Tatiana Moguchaya Vice President of R&D

Before joining Cognite, Tatiana worked with Schlumberger, managing the
world’s leading E&P software development, and with Microsoft – managing
Norway's largest business areas to accelerate digital transformation with cloud,
analytics, and big data. Tatiana holds MSc in Computer Science, Mathematics
and Physics.
[email protected] | +47 468 50 624

20
© COGNITE
OUR LOCATIONS
Oslo, Norway
Houston, Texas, USA
Austin, Texas, USA
Tokyo, Japan www.cognite.com
Vienna, Austria [email protected]