Key components: Integrated drilling simulator with virtual well bore and visualization; Real time automated interpretation, diagnostics and decision support; Utilization of increased bandwidth from borehole; New work processes, trainingsimulator and experience transfer.
Reservoir management
Results 2007
Norne field study :
Startup and definition of project
Prototype of a new reservoir simulator developed and tested
New history-matching method tested on the Norne field
Workshop.
Hiring and startup of 9 postdoc researchers and PhD candidates. This includes people from the petroleum engineering, systems and control, and work process domain.
Submitted scientific papers (~5), talks/presentations in scientific meetings and Master theses (~3)
Objectives/deliverables - Program 2.1
Embed model-based decision making methods/tools into new work processes for reservoir management. Try to challenge the existing mind-set. This includes improved methods for history matching and closed-loop reservoir management.
Specific for Case study 1 - Norne field
1. Develop methods for updating the reservoir model for a field at Halten Nordland using 4D seismic and production data.
2. Develop prototype simulator for use in closed-loop reservoir management
3. Identify the key challenges regarding interpretation of these new tools into the work process as a whole.
Application of results
Testing of new prototype and methods on case studies. Case study no.1 is the Norne field
Perform comparative studies of available and new metods using benchmark examples
Dissemination of results through interaction with partners. We start with Norne Petek.
Dissemination of results through academic channels - publications in scientific journals, at scientific conferences and IO center workshops
Technology/methods
Challenge the industry with new, much faster, and all together different simulator technology
Mimetic finite-difference discretization (MPFA/mixed FEM type method)
Fast transport simulation by reordering of unknowns
Generalized travel-time inversion (streamline-based history matching)
Tuning or history-matching (or data assimilation)
Many approaches: optimization based, Kalman-filter type methods, deterministic or stochastic formulations
Faster simulation technology allows us to use improve tuning. One example is the use of more accurate uncertainty models. We want to study alternative history matching/data assimilation methods and perform comparative studies of alternative approaches.
We will work with data assimilation from different data sources (production data, 4D seismic data)
Closed-loop decision making
Faster simulation technology allows us to use a receeding-horizon optimization approach as a means to approach closed-loop reservoir management. This approach has been a game changer in the process industries
Designing work processes and tools so as to make the use of model-based decision technology a natural part of the work flow is challenging. We will have a continued activity on this as an inherent part of the pilot studies.
Potential effect for improved field operation
Increased value of an asset by increased recovery
Faster and more reliable reservoir simulations result in better production scheduling and improved oil recovery
Improved understanding of human interpretation of the reservoir may lead to more optimal reservoir management
Today’s technology frontier/technology gaps
Today’s technology hinders the use of model-based decision making in new work processes for reservoir management.
Reasons for this are:
Direct simulation on geocellular models generally not possible
Lack of robust methods to help automate integration of available data
Further, there does not exist an established set of methods or commercial available software for fast updating of reservoir models based on available data as 4D seismic and production data. The same goes for closed-loop management. There is also a lack of understanding of how tools should be embedded into work processes for reservoir management.
The research challenges are huge. Hence, it is important to interact with relevant research environments in Norway and abroad. We have established close contact with research activities at Statoil, Hydro, and IRIS, and with research centers at TU Delft and Stanford university.
Activities,Deliverables, and Time
Activities
Deliverables
Time
1. Develop simulator technology for the Norne field based on new technology
Case study results Norne
2007 and 2008
2. History-matching, testing of methods at Norne both using new simulator technology and an Eclipse model
Workshop Norne field
2007 and 2008
3. Comparative study of methods for history matching
Prototype algorithms and methods
2009-2011
4. Closed loop reservoir management at Norne - preliminary study
Scientific publications (there will be many - including level 2 publications)
Master theses
Talks, invited talks, etc.
2007-2011
5. Study work processes at Norne as a basis for specifying model-based decision tools
Startup of new field study
2008
6. Define research tasks for 4 PhD candidates
4 PhD theses
2011
7. Identify new case study
Relations to other sub projects
Program 1. Updating geo-models from drilling data
Program 2. Subproject 2.2. Production optimization a decentralised approach
Program 3. Generic framework for predictive maintenance control
Program 4. Work processes and visualization
Resources
Sub project manager: Jørg Aarnes, SINTEF Applied Mathematics
Project team:
Professors (Jon Kleppe, Martin Landrø, Knut Andreas Lie (University of Oslo), Bjarne Foss) and 9 postdoc researchers and PhD candidates from:
Petroleum engineering
Applied mathematics
Systems and control
Work processes
Industrial participation:
Statoil Halten Nordland
Statoil Research Center
Deliverables 2007: These reports and other deliverables will available on the sub-project page (2.1), upon completion.
Field description (report)
History-matching, field testing of method (report)
Closed loop reservoir management field testing (report)
Production optimization
Results 2007
Troll oil field study. Development of a new optimization approach and initial results.
FMC study. Alternative formulations of proxy models for wells.
Startup of Shell production optimization study (in cooperation with Program 4)
Hiring and startup of 5 postdoc researchers and PhD candidates. This includes people from the petroleum engineering, and systems and control domain.
Submitted scientific papers (~2), talks/invited presentations at scientific meetings and Master theses (~3)
Objectives/deliverables - Program 2.2 Propose, develop, analyze and assess a new methodology for model-based production optimization on a tactical planning horizon (hours to weeks).
Propose, develop, analyze and assess methodology for production chain optimization. This includes merging new tools for system integration and applying ideas from supply chain management; ultimately covering the production chain from source to market.
Efficient development of proxy models targeted for production optimization
Application of results
Testing of methods on case studies. Case study no.1 is the Troll field
Perform comparative studies of available and modified/new metods using benchmark examples
Dissemination of results through interaction with partners - workshops
Dissemination of results through academic channels - publications in scientific journals/conferences and at IO center workshops
Technology/methods
The foundation for this research rests on the following three parts:
Optimization based on Lagrangian decomposition, dynamic piece linearization.
Model reduction techiques for developing targeted proxy models
Methods and tools for system integration
Further, model-based optimization methods and technology from the process industries will be studied as a basis for our work.
Potential value for improved field operation
Improved utilization of production facilities on a tactical planning horizon.
Faster and more reliable computations.
More transparent technology through new system integration tools. This helps to embed the technology into the work flow and build trust in the technology.
Today’s technology frontier/technology gaps
Technology for tactical planning needs to take into account all relevant constraints and couplings in the production systems. Current technology solves relatively large nonlinear optimization problems. This induces high computational loads and convergence problems are often encountered. These challenges increase when model complexity increases, for instances when using nonlinear well models instead of linear well models.
Computational IO (system integration) with model-based optimization is an exciting, new, and challenging approach to integrated operations. The approach is not currently used by the industry on a full scale, for many obvious reasons, but the industry has given a clear indication that this is an ultimate goal within IO. The concept is to (electronically) pipe together operations from far upstream to far downstream, including reservoir, production, flow assurance, processing, market, and full-scale economics. Local (operation) and global optimization is a pre-requisite.
There are major research challenges within production optimization. Hence, it is important to interact with relevant research in Norway and abroad. We have established close contact with the Center for Advanced Decision Making at Carnegie Mellon University, and also other groups with long-standing experience in model-based process control and optimization, and supply chain management.
Activities,Deliverables, and Time
Activities
Deliverables
Time
1. Further develop and study optimization method based on a decentralized approach
Case study results Troll on production optimization
2007 and 2008
2. Develop methods for model reduction for proxy models
Network modeling using new system integration tool
2007 and 2008
3. System integration - network modeling
Prototype algorithms and methods - discussion with commercial actors
2009-2011
4. Study work processes at Troll and Norne as a basis for specifying production optimizations tools
Scientific publications (there will be many - including level 2 publications)
Master theses
Talks, invited talks, etc.
2007-2011
5. Define research tasks for 4 PhD candidates
4 PhD theses
2011
Relations to other sub projects
Program 2. Subproject 2.1 - Reservoir management
Program 3. Generic framework for predictive maintenance control
Program 4. Work processes and visualization
Resources
Sub project manager: Professor Bjarne Foss - NTNU (for the time being)
Project team:
Professors (Curtis Whitson, Erik Ydstie (CMU)) and 5 postdoc researchers and PhD candidates from:
Petroleum engineering
Systems and control
Work processes
Industrial participation:
Hydro Troll Oil and Hydro Research
FMC
Shell Draugen
Deliverables:
A new method for production optimization analyzed and implemented for pilot testing.
Software code - suitable for research purposes
Comparative study with state-of-the-art technology (e.g. GAP software from Petroleum Experts and MaxPro from FMC) on at least one field example
2 PhD theses, 2 MSc theses and 6 scientific publications
Application of results The method shall be applied for tactical planning from hours to weeks. The 1st testing will be on a part of Hydro¡¦s Troll oil field. Based on the above results, the methodology will be promoted towards other partners.
During 2007 the activity will consider the production chain from the near well region of a reservoir to the 1st stage inlet separator. Downstream extension, ie. including the process equipment typically found on an offshore platform is planned for 2008 and beyond in co-operation with Program 3 Operations and maintenance
Potential effect for improved field operation
Improved utilization of production facilities on a tactical planning horizon.
Faster and more reliable computations.
More transparent technology because of a decentralized nature and, hence, probably easier integration into the work flow and build trust in the technology.
Deliverables 2007: These reports and other deliverables will be on the sub-project page (2.2) , upon completion.
M.Sc. theses, scientific publication
Scientific publication
Meetings, assessment report
PhD theses project, scientific publications
Reports, scientific publications
Course: Control & monitoring of oil & gas processes, New pilot studies
Objectives / Deliverables - Program 2.3
Supply a course on control, monitoring and supervision with two simulator-based exercises based on oil and gas systems; one being an offshore process and the other a downstream gas/condensate plant.
Provide 4th year students with hands-on experience and realistic problems related to control, monitoring and supervision of oil and gas processing.
Application of results To be applied in a new, regular course “Advanced control in industrial processes” for 4th year students (7.5 study points). The course will be offered during the Spring semester of 2008 and thereafter every Spring semester. Responible teacher: Professor Morten Hovd, Department of Engineering Cybernetics, NTNU.
Estimated number of students: 50
The course is applicable to students from several disciplines.
Potential effect for improved field operation Improved quality in education
Deliverables 2007: Additional information will be published on the sub-project page (2.3), as it becomes available.
Two simulators adapted for course use
Simulator integrated into a new, regular course at NTNU
New Pilots studies (2.4)
Objectives / Deliverables - Program 2.4
Expanding activity 3.1 and 3.2 to new test cases in other partner companies than those mentioned above, in order to develop further and generalize the methods for reservoir management and production optimization.
Application of results
Production planning and optimization
Potential effect for improved field operation
IOR and accelerated production rate
Deliverables 2007:
These reports and other deliverables will be made available on the sub-project page (2.4), upon completion.
Field testing reports
Seminars for consortium participants, exchanging experience from case projects in different companies
