To climb down the cost curve, upstream companies need to fundamentally change how they operate – technology, people and processes. The industry has reached close to the maximum threshold on the number of individual point solution applications (and associated processes and siloed departments) that are in use today. To remain relevant and thrive, upstream companies must firstly buy time, then digitalize and lastly, position more effectively for the energy transition. This means taking a “systems thinking” approach that focuses on the way that a production system’s constituent parts interrelate, how they work over time and within the context of larger systems. This presentation will outline the role of field-wide models, which when operationalized with real-time data, result in a digital twin that is highly effective in achieving production system optimization. These models when run in the Cloud, then enable the remote optimization center and generate synthetic data able to train AI algorithms for machine learning with limitless potential.
Improved Upstream Production Efficiency with Remote Optimization Centers, Field-wide Models and AI/ML
1. Andy Howell
Chief Executive Officer
KBC (A Yokogawa Company)
November 12, 2020
Improved Upstream
Production Efficiency with
Remote Optimization Centers,
Field-wide Models and AI/Ml
2. Agenda
1. Why Production Efficiency?
2. The Approach
3. Remote Optimization Centers
4. Field-Wide Models
5. AI/ML in Upstream
6. Conclusion
7. Industry Challenge
4. Production Efficiency—Highest Stakes Ever!
Maximize what
we have
Climb down
the cost curve
Fundamentally change
how we operate
Beyond max. threshold
of point applications
Siloed departments,
inefficient processes
Remain Relevant and Thrive
Buy time Then digitalize And position more effectively
for the energy transition
5. Avis versus Hertz: A Tale of Two
Digital Transformations
Take a “people and systems thinking”
approach, focused on the way that
production system constituent parts
interrelate, and work over time, within
the context of larger systems…and
how they empower humans
The Approach
Source: Eric Kimberling, Jul 11, 2019, https://www.thirdstage-consulting.com/avis-vs-hertz-a-tale-of-two-digital-transformations
Focus on the right things in your digital
transformation
AVIS HERTZ
Focused on fundamentally
enhancing overall business mode
Tried to automate
existing processes
Focused on the
customer experience
Focused on “building
a web site”
Led with business processes Led with technology
Still operating Chapter 11 bankruptcy
Traditional Thinking Systems Thinking
6. The Approach
Efficiency
Safety and Security Availability and Reliability
Human Reliability
▪ Improve total energy efficiency
and reduce CO2 emissions
▪ Ensure profitable operation by optimizing asset
lifecycle and supply chain
▪ Realize flexible and lean production
▪ Achieve zero incident operations
▪ Improve overall HSSE management
▪ Comply with legislation, regulations
and standards
▪ Capture and transfer knowledge
▪ Build autonomous and intelligent expert system
▪ Create better workforce-development training
▪ Eliminate unplanned outages
▪ Maximize plant uptime while minimizing
lifecycle costs
▪ Realize predictive operation and maintenance
Production Priorities
Engineering Priorities
▪ Improve project economics
▪ Mitigate project risks
▪ Optimize delivery schedule
▪ Realize flawless engineering
▪ Flexibly manage changes
▪ Comply with industry standards
7. The Approach—Stochastic Meets Deterministic
Reservoir Team
They live in a stochastic world
of probabilities
Drilling Teams
Their focus is on the optimal
method of drilling and
well design
Facilities People
Their focus is on how best to
achieve the design point
through the facilities. Manage
conditions changes and allow
e.g., for future secondary
lift mechanisms
Commercial People
Focused on maximum revenue
generation as a function of
system availability
Well Teams
Live in a world of maximizing
well production and minimizing
well damage
Multiple Factors Need Consideration and Behave in a Non-linear Manner
▪ Reservoir production potential profile
▪ Reservoir fluid composition
▪ Changing power demands
▪ Flow regimes and pipe diameters
▪ Type, size and cost of associated production
and export facilities
▪ Secondary and tertiary lift mechanisms
▪ Well lift and type curves
▪ Well decline and composition changes
8. Remote Optimization Centers
The key steps
If NASA can remotely operate the
unmanned MSL on its mission to Mars,
which is approximately 140 million miles
away, shouldn’t the upstream industry
be able to remotely operate unmanned
(or minimally manned) oil and gas
exploration and production assets right
here on earth?
Source: https://www.arcweb.com/blog/remote-operations-upstream-oil-gas
9. ■ Align disciplines
■ Empower humans with visualization
■ Manage multiple assets remotely
■ Balance stochastic versus deterministic
■ Host field wide models
■ Experiment with AI/ML
■ Infer data gaps
■ Scale at speed
■ Needle in the haystack
■ But something needs to train the ML!
Remote Optimization Centers
Goals
10. Field-wide Models
Takes a field-wide holistic
approach to deliver the
required return on capital
Creates the major cultural
change underpinning
digitalization that will see a
step change in profitability
An integrated asset
model enables more
effective economic
evaluation and portfolio
management decisions
11. Field-wide Models—Talking 2+ Languages
Correlation-based analytics tools are
convenient, simple to set up, and quick
Process and Production
Facility Engineers
Understand the value of first principles simulation
models to design and operate the asset
Reservoir Engineers
Models have a heavy dependency on
data-driven, correlation-based analytics
Intensive and continuous exercise of using spreadsheets
in an attempt to update the correlations against reservoir
history, before they lose their predictive validity
Know that chemical and physical interactions
and dependencies must be respected
Need to draw safe and meaningful conclusions
12. Field-wide Models
■ Form basis of digitalization
■ Survive fluctuating
data quality
■ Matching well deliverability
to topside power generation
and compressor availability
for 20+ wells feeding a
“self-powered” FPSO that
has processing capacity of
90,000 b/d of oil, 10–20
MMscfd of fuel gas handling
and treated water injection
rates of up to 30,000 b/d
■ Oh and don’t spend
any Capex!
Challenge: Create a single
field-wide modelling
environment to boost FPSO
production by >10,000 b/d
and deliver incremental
profits of >$180 million per
annum
13. Field-wide Model
Well models and
sub-surface templates
FPSO processing trains
Gas Turbine power generation
Inlet Manifold set pressures from onboard compressors
14. AI/ML in Upstream
Well models and
sub-surface templates
MP/HP Compressors
Onboard
Power Turbines
■ Infer missing field data
■ Lift curve envelopes
■ Gas lift injection points
■ Ingest data at scale
and predict failure
AI/ML used to generate
15. Conclusion
■ A single modelling environment to boost FPSO
production by >10,000 b/d and deliver
incremental profits of $180 million per annum
■ No CAPEX investment, uses only
onboard equipment
■ Matches sub-surface to surface pressure,
flows in a field-wide model
■ AI/ML used to infer data and provide injection
starting points
■ First time power-production
balance implemented
■ Tested on physical asset
■ New production regime confirmed
■ Production rates and value attained
16. Where can your people deploy
field-wide models, which when
operationalized with real-time data,
result in a digital twin that is highly
effective in achieving production
system optimization?
Are you ready to run in the Cloud
and enable remote optimization
centers and generate synthetic
data able to train AI algorithms
for machine learning with
limitless potential?
Industry Challenge
1 2
17. The names of corporations, organizations, products and logos herein are either registered trademarks or
trademarks of Yokogawa Electric Corporation and their respective holders.
Thank You!