cognitive reservoir analytics
TRANSCRIPT
Wake up with Watson
Cognitive Reservoir Analogues
Renato Cerqueira
November 15, 2016
IBM Research Big Plays in Oil & Gas
Capital Projects,
Operations, Optimization
Resource Characterization
and Management
Nanotechnology Applied to Enhanced Oil Recovery and
Mining
Cognitive Computing for Natural Resources (Oil & Gas)
HPC Industry Cloud
2
Selected Use Cases
3
UC2: Basin / Reservoir
Analogues Adviser
UC3: Seismic Interpretation
Adviser
UC1: Basin / Reservoir
Knowledge Base Builder
UC5: Machine-learnt Well Log
Analytics
UC4: Capital Project
Management Adviser
UC6: IoT-driven Reservoir
Digital Twin
Decision-
Making
Operation &
Monitoring
Data
Transformation
Living
ModelInsig
hts
Actio
ns,
stra
teg
ies
Big Data IoT
News and
social media
Specs and
requirement
s
Rules and
RegulationsContracts
ReportsLessons
Learned
Capital ProjectsManagement
Adviser
Image /
video feeds
Projects schedules
Weather
Risk factors
TE
XT
UA
L /
UN
ST
RU
CT
UR
ED
D
AT
A
Cognitive
Computing
Technologies
Optimization
Simulation
Forecasting
Look to the past to understand the
project’s current situation
Look to the future and identify main
points of attention
What can be done to avoid potential
impacts?
Define optimal course of action and
implement measures SeismicCoherence
Selected Use Cases
4
UC2: Basin / Reservoir
Analogues Adviser
UC3: Seismic Interpretation
Adviser
UC1: Basin / Reservoir
Knowledge Base Builder
UC5: Machine-learnt Well Log
Analytics
UC4: Capital Project
Management Adviser
UC6: IoT-driven Reservoir
Digital Twin
Decision-
Making
Operation &
Monitoring
Data
Transformation
Living
ModelInsig
hts
Actio
ns,
stra
teg
ies
Big Data IoT
News and
social media
Specs and
requirement
s
Rules and
RegulationsContracts
ReportsLessons
Learned
Capital ProjectsManagement
Adviser
Image /
video feeds
Projects schedules
Weather
Risk factors
TE
XT
UA
L /
UN
ST
RU
CT
UR
ED
D
AT
A
Cognitive
Computing
Technologies
Optimization
Simulation
Forecasting
Look to the past to understand the
project’s current situation
Look to the future and identify main
points of attention
What can be done to avoid potential
impacts?
Define optimal course of action and
implement measures SeismicCoherence
UC 2: Cognitive Reservoir Analogues Adviser
5
UC 2: Cognitive Reservoir Analogues Adviser
6
Basin/ Reservoir Similarity + DeepQA
Data Driven Analytics
(Machine learning
modeling/prediction)
Data Management
(Injection, Edition)
Unstructured Data Sources
(Technical Articles, Reports from OnePetro) RES.
ID
PERMEAB
ILITY
BASIN
CODE
… VISCOSIT
Y
POROSITY
TYPE CODE
R1 112.5 2-1-2 … ̶ 8
R2 63.0 2-3 … 6.5 24
R3 ̶ 1-2-4 … 12.0 18
… … … … … …
R1200 75.0 2-3 … 0.75 ̶
Structured Data Sources
(Known Basin/Reservoir Databases)
“What is the average
permeability in the F1
field?”
Answers both
directly based on
known facts from
data sources
as well as
extrapolated from
analogues
Questions about
reservoir
properties and
analogues in
natural language
“What is the reservoir most
similar to R7 in terms of
structural properties?”
“Average permeability
in the F1 field is 22mD
with 90% confidence.”
“R11 is the reservoir most
similar to R7 in terms of
structural properties with
similarity .78”
DeepQA (enhanced to
consider similarity reasoning)
Knowledge
Base
Domain Ontology
ponents nents
Information
Extraction
Agents Reasoning
Engine
Knowledge
Graph
Query
Interface
Cognitive Reservoir Analogues – Technical Details
7
Structural Code = 8
Area (Km2) = 20
Porosity (%) = 12
Oil Viscosity (MPa.s) = 1.1
Permeability (mD) = 22
Lithology Code = 410
ID PERMEA
BILITY
BASIN
CODE … VISCOSITY
CR1 112.5 2-1-2 … ̶
CR2 63.0 2-3 … 6.5
… … … … …
CR1200 75.0 2-3 … 0.75
New Target Reservoir
Known Reservoirs
Databases
Resulting Analogues Rank ID Similarity
1 CR354 .82
2 CR180 .82
3 CR22 .81
… … …
Uncertainty Characterization
Target Reservoir with
predicted parameters Reservoir Analogues Toolset
Data Management (injection/edition)
+
Data Driven Analytics (machine
learning modeling/prediction)
Structural Code = 8
Area (Km2) = 20
Porosity (%) = 12
Oil Viscosity (MPa.s) = ?
Permeability (mD) = ?
Lithology Code = ?
Flow Diagram of the RA method
8
Reservoir Analogues Toolset
Data Management (injection/edition)
+
Data Driven Analytics (machine
learning modeling/prediction)
Application Example Computing ROI of a set of reservoir targets
9
Reservoir Visual Analytics (RVA)
EC
ON
OM
IC
EV
AL
UA
TIO
N
RESERVOIR SIMULATION
FA
CIL
ITIE
S
DE
SIG
N
RESERVOIR
MODELING
PRODUCTION
STRATEGY
Asset Uncertainty
Characterization
Business &
Market Information
Prototype Demo – Video (on site at event only)
10
Scenario
• How can a Cognitive Reservoir Analogues (CoRA) help Jonatan, a Petroleum Engineer,
make critical decision as to whether or not to go ahead with the well exploration in the
Sergipe-Alagoas Basin, Brazil?
You will see
• How CoRA helps Jonatan calculate the oil in place by identifying and estimating important
missing parameters
• How CoRA points out that Jonatan needs the value of Porosity of the rock to carry out the
task.
• How CoRA computes and shows evidences in support of its calculations.
As a result
• Jonatan was able to explore and inspect analogous reservoirs in his company’s DB as well
as external unstructured knowledge so as to estimate with high confidence unknown
parameters of the prospect reservoir.
Summing up
11
Natural Language
Understanding Analytics
Multimodal Dialog
Machine Learning
Visual Analytics
Uncertainty Quantification
Semantic-based
Integration
Multimedia Feature
Extraction
Automatic Reasoning
Knowledge Management
Predictive Models
Parallel Hypotheses Generation
Cognitive Computing is the use of computational learning systems to augment human cognitive capabilities and accelerate, enhance and scale human expertise to solve real world problems.
Cognitive technologies are ready to address the challenges of the Oil & Gas industry and to transform practices in the industry, in face of data overload, new frontiers, workforce and skills shortage.
Cognitive Systems can provide unprecedented gain in productivity in knowledge-intensive processes, such as exploration and production, giving geologists to tap quickly into information and draw relationships at a scale that is almost prohibitive today.
CoRA can support several different business processes, such ROI estimation of a target prospect, seismic interpretation, geological risk assessment, production planning and forecast, among others.
12
Thank you.
13
© IBM Corporation 2016
IBM, the IBM logo, ibm.com, and Watson are trademarks or registered trademarks of International Business
Machines Corporation in the United States, other countries, or both. If these and other IBM trademarked terms are
marked on their first occurrence in this information with the appropriate symbol (® or ™), these symbols indicate
U.S. registered or common law trademarks owned by IBM at the time this information was published. Such
trademarks may also be registered or common law trademarks in other countries. A current list of IBM trademarks
is available on the Web at “Copyright and trademark information.”
• SoftLayer is a trademark or registered trademark of SoftLayer, Inc., an IBM Company.
• Other company, product, and service names may be trademarks or service marks of others.
• References in this publication to IBM products or services do not imply that IBM intends to make them available in
all countries in which IBM operates.
Trademarks and notes