The Shared Earth Model Challenge

1
The Shared Earth Model Challenge
-Toward a new generation of integrated systems
David Archer, POSC Dan Schenck, POSC
2
Presentation Overview

• Background
• Shared Earth Model
• Todays Problems
• Causes
• Why SEM Helps
• POSC SEM Project

3
What its all about!

• The purpose of computing is insight not numbers
  
• Richard Hamming, Bell Labs

4
Analysis Synthesis

• Analysis (Machine Age - Newtonian thinking)
• Understand things by taking them apart
• Predicated on one-way causality
• If you understand all the parts, then you
  understand the whole
• Synthesis (Information Age - Systems thinking)
• Understand things by putting them together
• Mutual influences vs. cause effect
• Look for the big picture

5
Complexity

• Simple systems give rise to complex behaviors.
  Complex systems give rise to simple behavior. And
  most important, the laws of complexity hold
  universally, caring not at all for the details of
  the systems constituent atoms. (James Gleick,
  Chaos)
• Very little of importance is ever just the sum
  of its parts, except money. (Tyler Volk, NYU)

6
Learning

• Occurs through action, feedback, synthesis cycles

Virtuous learning cycle
7
Learning ...

• Occurs best in groups through
• Collaboration
• Sharing knowledge practices
• Building on the experience of others

• Integration
• Interoperability
• Information sharing
• Collaboration
• Knowledge Management
• Learning Organization

8
Current Earth Models
Seismic interpretation
Geologic modeling
Upscaling
Property modeling
Simulation
9
Integration via ...
Building integrated systems vs. integrating built
systems

• Integrating built systems
• Newtonian approach
• Activity of past decade
• Building integrated systems
• Systems approach
• Leading to the next generation

10
Shared Earth Models
If Boeing can build the 777 without making
drawings, why are we still making maps? - Tim
Warren (SIEP, AAPG98)
a model-centric approach to applications and
work flows to solve business needs. - Tim White
(LGC, OG J)
11
Shared Earth Model

• Puts traditional earth model information into a
  framework that supports collaboration improved
  decision making
• Shifts focus from data to datas context
• Allows building dynamic models that can be easily
• Shared
• Visualized
• Extended

The unifying concept is the earth model itself
not the many components from which it is
constructed.
12
SEM Characteristics

• Incorporates a comprehensive range of asset based
  information knowledge
• Consistent across EP domains
• Supports fusion of EP information
• Readily accessible to all team members
• Supports collaborative work
• Enables a shared understanding of the asset

13
SEM Vision
Geologic modeling understands depositional
environments, sequence stratigraphy, tectonics,
lithology, ...
Image consistency understands acoustic wave
propagation structural feature identification
SEM
Property modeling understands rock and fluid
properties, petrophysics, and property
distribution within and between fault blocks
Flow simulation understands conservation
of mass and energy, fluid flow pressure response
14
Identified Problems
Interpretations are inevitably sub-optimal
because they are based on models derived from
isolated subsets of the totality of available
empirical data.
15
Identified Problems
Rigorous testing of interpretations is often
impractical because current modeling
functionality does not support flexible
experimentation.
16
Identified Problems
Change is difficult and time consuming to
propagate through models because they are
developed in a linear sequence (output from one
provides input to another).
17
Identified Problems
There is no collective understanding of the
strengths and weaknesses of models, because the
rationale behind individual interpretations is
not captured.
18
Bottom Line Impact
Earth Models often fail to support technical or
commercial decision making in an effective,
efficient or timely manner. Resulting in...
EP investment and operational decisions that can
be sub-optimal or just plain wrong.
19
Causes..

• integration/interoperability issues caused by
  lack of overall architecture spanning full SEM
  domain
• inconsistencies in semantics and incompatible
  data models between individual domains
• no support for versioning, e.g., inability to
  support multiple scenarios, no automatic
  propagation of updates across all alternative
  representations of the same object
• no accounting for uncertainties in data
  interpretations

20
Causes

• some objects not defined other than through their
  representations, and no standards in place
• inadequate and/or inconsistent representation of
  domain objects e.g. lack of support for special
  objects, e.g., complex faults
• lack of uniform co-ordinate systems
  transformation service that can handle full range
  of temporal and spatial scales and resolutions
  required
• i.e., basinÞfieldÞreservoirÞcompletion zoneÞpores
• eons Þfield lifecycleÞwell lifecycleÞwell
  testÞimpulse response

21
..Causes..

• data access cycle-time issues due to lack of
  seamless data pipeline between models
  (re-formatting required, etc.)
• inadequate support for concurrent access by
  multiple users - models can only be updated
  through a series of one-way streets
• poor workflow management (integration of best
  practices, collaboration support etc.)
• little support for storage retrieval of
  generated information and knowledge

22
..Causes

• and so on and so on

23
Why do we need a SEM?

• Sharing Business Objects
• Current Situation
• Vendor sharing through proprietary databases
• Export of data - Objects are exported from one
  application to the other through filters and
  files
• Limitations
• Objects are not shared, they are transmitted down
  the study flow
• Sharing objects is a key point for integration

24
Why do we need a SEM?

• Sharing Objects Why is it not enough?
• The business of EP is centered on the reservoir
  model
• When focusing only on business objects, a key
  part of the EP interpretation is left out the
  links between objects

25
Why do we need a SEM?
Are we doing the same interpretation 3 times?
26
Why do we need a SEM?

• Sharing Links Between Objects
• Saves the re-interpretation time at each step of
  the EP process
• Is safer because the interpretation is done in
  the context of the emerging model
• At each step of the interpretation we can go back
  to solve a consistency problem

An EP study consists of building a Shared Earth
Model to obtain, for example, a production
profile.
27
Why do we need a SEM?
28
Reservoir ModelRepresentation
The SIM model is the SEM model. - Bill
Quinlivan, GeoQuest
29
The Corresponding SEM
30
SEM Project Objective
To define specifications for open interfaces and
shared services that enable the implementation of
a new generation of interoperable earth modeling
products epiSEM
31
Project Strategy
Achieve the Objective (specifications for open
interfaces and shared services that enable the
implementation of a new generation of
interoperable earth modeling products) by ...

• Developing consensus on definitions, core
  requirements, conceptual model, and architectural
  principles through analysis, information sharing,
  and debate
• Promoting wide understanding of model centric
  computing through a high profile communications
  program
• Identifying opportunities to develop prototypes
  to demonstrate concepts and validate designs
• Leveraging off existing work inside and outside
  the EP Industry
• Focusing on specifications that deliver early
  benefits

32
Implementation?

• epiSEM Standard Interfaces to Standard Earth
  Model Services defined and described via UML and
  IDL.
• A typical vendor implementation of epiSEM might
  be based on CORBA/DCOM and might include such
  Components / Services as
• EP(I) Objects (data access, representation
  independence, domain relationships, e.g.,
  geological data model)
• Basic Model Services (definition, classification,
  identification, contents,..)
• Geometry Properties Calculations, e.g.,
  intersection, property evaluators

33
epiSEM Services?

• Examples of potential epiSEM Components /
  Services
• EP(I) Objects (data access, representation
  independence, domain relationships e.g.
  geological data model)
• Basic Model Services (definition, classification,
  identification, contents,..)
• Geometry Properties Calculations e.g.
  intersection, property evaluators
• Visualization Services
• Knowledge Engine (capturing/reusing the why?)
• Lifecycle Management (transactions/workflow/versi
  oning)
• Equations and Solver (system formulation and
  control)
• Interaction Models
• Others??

34
Illustration

• Assumes many object services are available
• Trader
• Generalized sample data representation objects
• etc..

Vendor Product B Time-Depth Conversion
Vendor Product A Drilling Simulation

• BasicEM
• model name
• version
• last updated
• who updated
• properties defined

• Property Evaluator
• get_property( in BasicEM, id, in shape, in
  property_name, out measure)

• Geologic Model
• get_horizons
• get_faults

Intersector

• High value first steps...
• Well Planning
• GeoSteering

Horizon
Fault
Formation
Reservoir
epiSEM Interface
35
Conceptual Architecture View
EP Technical Applications Data Stores
System Management
Visualisation
Shared Earth Modelling Objects and Services
EP Domain Objects and Services
EP Common Objects and Services
Distributed Object Services
Base Computer Services
36
How? -Tactics

• Consensus
• Iteration and Review
• Build on current thinking and previous work
• Stepwise, Focused Effort

37
Who?

• Project Sponsors

- Agip - Chevron - IFP - Landmark - Mobil
- PDS - Schlumberger - Shell - Statoil
38
Phased Approach
SEM Domain
Focus
Focus
Phase ...
Phase 1
Project Initiation
Phase 2
1. epiSEM Functional Scope 2. epiSEM Principles
Concepts 3. Industry Activity Survey 4.
Preliminary epiSEM Specification 5. RFC Response
Project Plan Project Proposal
39
Phase 1 Work Packs
1. epiSEM Functional Scope 2. epiSEM Principles
Concepts
WP2 Develop epiSEM Requirements
4. Preliminary epiSEM Specification
WP5 ConductRFC Program
WP1Project Initiation
WP4 Prepare Preliminary epiSEM Specifications
WP3 Survey Industry Activity
Project Plan Project Proposal
5. RFC Response
3. Industry Activity Survey
40
Value Statement 1

• With minimal workflow constraints, EP project
  teams will be able to formulate and justify
  common cross-disciplinary interpretations of the
  composition, properties and behavior of the
  subsurface by using SEM products that conform to
  the epiSEM specification.

41
Value Statement 2

• Project and Asset teams will be able to test
  their theories against the totality of current
  empirical and derived information using epiSEM
  plug--play software components.

42
Value Statement 3

• Technical and commercial decision making
  information derived from epiSEM models will be
  more reliable and more timely than at present.

43
Schedule

• Finish specifications September 1
• Distribute RFC October 1
• Deliver RFC responses by year end 99

44
Activities

• Requirements Gathering
• Questionnaire developed and sent out
• Workshops held at Shell and IFP
• Workshops in May at Chevron Mobil
• Workshops in May at Landmark GeoQuest
• Industry Analysis
• Developing summary of current industry situation

45
Actitivies ...

• Design
• Team members identified
• Tools identified
• Reviewing requirements and implementation
  architectures

46
SEM Directions

• Many opportunities constrained by available
  energy
• Focus on model use and value, not model creation
  and maintenance
• Much will depend on RFC feedback
• Survey conducted problem statements derived
• Gives us a picture of issues/priorities

47
Problem Statement Issue Occurrence
Knowledge management
Inverse modeling
Administration/versions/data dependency
Sharable earth model
Unified time and distance scales
Data loading/other
Coverage/generality
Coherent spatial model
Uncertainty/quality
48
Summary

• SEM is a model centric approach to managing EP
  workflow
• It holds the promise of better asset decision
  making through more effective team collaboration
• POSCs epiSEM will be the EP industrys open
  specifications for SEM implementation technology

49
Earth Model-Centric Workflow Integration
True SEM is magnitudes larger than anything the
industry has done before. A stepwise progression
is the only way to approach a solution to this
challenge. - Tim White,LGC
Geophysics
Engineering
Geology
Source Landmark Graphics
50
POSC SEM Project

• Questions
• David Archer
• Phone [phone]
• Email archer_at_posc.org
• Project Contact
• Dan Schenck
• Phone [phone]
• Email schenck_at_posc.org
• Web Site
• http//www.posc.orgthen follow the link to
  Projects

51
Current Earth Models
Seismic interpretation
Geologic modeling
Upscaling
Property modeling
Simulation