Oil and gas production systems stretch from reservoirs and wells via gathering systems to downstream facilities and beyond. They are large systems with many interconnected components. A production system is developed to drain an asset in some optimal way while obeying safety and environmental regulations. This requires planning and operational decisions on time scales from seconds to years, and such decisions may involve parts of or the whole production chain.
Program 2 places its emphasis on the following planning and operational decisions
The choice of reservoir drainage strategies and rates, and well locations
The choice of well rates and routing, and critical downstream parameters
The first set of decisions is made on time horizons from months to the lifetime of an asset, and requires intimate knowledge of the reservoir and wells. The second set of decisions is made on a day to week time horizon, and needs to consider the gathering and downstream system in addition to the reservoir and wells.
The tools and information available for planning and operational decisions is steadily increasing, and the trend is to try to integrate planning and operations along many axes:
Information from heterogeneous information sources
Along the production chain by linking reservoir and wells with gathering and downstream systems
Across time scales by combining short term and long term optimization
Across disciplines through teams with a variety of competences
Between functions by closely linking operations and maintenance
Between operating companies and vendors
It may be tempting to integrate “everything with everything”. This does, however, violate the “divide-and-conquer” principle which has been used quite successfully since engineers started to design and operate complex systems. All the same, we firmly believe that significant progress can be made by pushing integration on both the work science and the technology side. Program 2 will emphasize integration efforts along axes 1-4.
The vision is to develop technology and work processes with a significant economic potential. A guiding principle is to maintain a sound balance between integration and the “divide-and-conquer” principle. Means to this end will be the following:
Gain intimate knowledge on how work flows and technologies are actually used.
Develop much faster computational engines for assimilating heterogeneous data sources into models to produce predictions and manage uncertainty.
Computationally integrate the oil and gas production system.
Establish an understanding on the trade-off between integration and decentralized solutions.
Test new technologies and work processes in controlled environments.
Establish a strong on-going competence development program.
Seek answers to some fundamental problems – the “why” questions.
Program 2 objectives
1.Explore how novel model-based decision-support tools can be embedded into new work processes, and how work processes should be designed to fully exploit the potential of model-based decision-support tools.
2.Develop methods and prototype technology for model-based decision-support tools for real-time production optimization (RTPO) along the value chain from the reservoir to some downstream boundary , and for real-time reservoir management (RTRM). The planning horizon reaches from hours and days to years. Herein:
Methods for predicting the flow response of alternative well trajectories instantly – as basis for a draw-and-evaluate tool for well trajectory planning.
Methods for rapid updates of the reservoir model or geological model for the Norne field based on 4D seismic, production data, and other available data. The goal is to facilitate integration of new data, e.g. obtained during drilling, and to reduce risk and uncertainty due to mismatch between model and measurements.
Fast simulation technology with seamless model-reduction capabilities (model complexity selected semi-automatically at run-time to fit time-constraints and available computer resources). The goal is to enable increased use of flow information for decision support in operations with short time horizons.
Methods for closed loop reservoir management. The goal is to provide suggestions for future reservoir management which may include rates as well as well locations.
Methods to support real-time production optimization. The goal is to improve efficiency and predictability of the optimization methodology itself, and improve flexibility in terms of the types of proxy models that can be used.
Optimization technology that allows for decomposition of large problems into parts. The goal is to support a divide-and-conquer strategy which lends itself to common workflow patterns in engineering.
Develop proxy models from high-fidelity models. The goal is to create fit-for-purpose models along the production chain as a basis for integrated optimization.
3.Enhance the understanding of the link between production and field optimization, i.e. the interaction between long-range decisions and shorter-range decisions through the life-time of an asset. Draw experience from the process industries, through the way they structure the control and optimization hierarchy.
4.Communicate results with industrial partners through the use of established system integration tools, standards, and realistic and semi-realistic case studies.
5.Publish in high-quality journals and give talks at international conferences.
6.Educate PhDs and master students.
7.Give courses on IO technology to the petroleum industry.
8.Establish strong collaboration with TU Delft, Stanford University, Carnegie Mellon University and International Research Institute of Stavanger (IRIS) (exchange of students, establish common benchmarks, share data and resources, etc.).
9.Develop a benchmark data base for IO center research and trial activities. The database should use a field case and in particular promote comparative studies of alternative methods for history matching and ultimately closed loop reservoir management.
Industrial needs
Unified strategy for production and field optimization. More understanding of the interrelations between short horizon business decisions and long term recovery.
Develop, deploy and sustain real time model-based optimization tools. Real time means the frequency at which meaningful decisions can be made.
Faster integration and utilization of data in geomodeling workflows.
Technology and tools that facilitate use of new data (e.g. 4D seismic data) for decision support.
Uncertainty quantification related to short and long horizon business decisions.
Program structure
The program is divided into three main parts.
Program 2.1 Real-time reservoir management addresses technology and work processes for decision problems on a planning horizon where reservoir simulators like Eclipse traditionally is used. This means a decision horizon between some months and a couple of years.
Program 2.2 Real-time production optimization addresses technology and work processes for decision problems on a shorter planning horizon than program 2.1, typically between a day and a week. The whole value chain from the reservoir through wells, collection system and downstream processing to export is important in program 2.2 while as focus is on the reservoir itself in program 2.1. A background for the program 2 structure is shown below.
Program 2.3 Competence development program is included to address education and dissemination issues
Program 2 - Program Structure
.
Results 2007
Summary
The main tasks during 2007 have been to develop the program, define the first case studies, hire PhD students, develop collaboration with selected international research groups, and generate initial results.
The program is divided into three main parts. Program 2.1 Real-time reservoir management addresses technology and work processes for decision problems on a planning horizon where reservoir simulators like Eclipse traditionally is used. This means a decision horizon between some months and a couple of years. Program 2.2 Real-time production optimization focusses on technology and work processes for decision problems on a shorter planning horizon than program 2.1, typically between a day and a week. Program 2.3 Competence development program is included to address education and dissemination issues.
The main results from Program 2 are:
Two case studies (Norne field, Troll oil rim) have started and initial results have been developed.
Adaption of a streamline technology for the Norne field reservoir by using multiscale modelling concepts on the reservoir models. The goal is to use this for rapid reservoir model updating and hence increase the capability of (semi) closed-loop reservoir management. Included in the activity is the use of time-lapse seismic data for history matching and studies on work processes to embed model-based decision support tools into the normal work flow. Techniques for history matching include streamline-based history matching, ensemble Kalman filtering and optimization based history matching.
Development of a concept for decentralized design making systems which adapt to different operating conditions in the oil and gas field. The idea is to combine IO with a divide-and-conquer strategy which is commonly used in engineering. Initial results have been developed for a simplified case of the Troll oil rim production case. A joint project with the Center for Advanced Process Decision Making at Carnegie Mellon University (CMU) has been established.
A proxy model for gas coning in oil rim reservoirs has been developed in cooperation with FMC Technologies.
7 new PhD candidates have been recruited. Three other PhD students have been associated with and will provide valuable input to the program.
An advanced simulation tool for a new course on Advanced Control of Industrial Processes at NTNU has been developed.
Publication
2.1 Real-time reservoir management publications
Reports
Interdisciplinary operational decisions in the subsurface domain: An exploratory discussion of information requirements and support from integrated technologies – NTNU report
Petter G. Almklov
The purpose of this report is twofold. First, it outlines some topics that will be focussed in the “work and decision making processes” part of Program 2 at the IO center. The intention is both to develop independent research on integrated work processes, but also to provide input to other research and innovation, to help situate it within an operational context. This brings us to the second purpose of the report. Within the reservoir domain, StatoilHydro’s Norne field is our industry case partner, and this report seeks to outline some impressions from a an initial visit there. In particular operational challenges within, or maybe more importantly, on the boundaries of the reservoir domain may be possible areas of interest for the IO center’s new technologies and work processes.
Application of multiscale modelling concepts on the reservoir model for the Norne field – SINTEF report
Jørg E. Aarnes, Stein Krogstad, Knut-Andreas Lie
Multiscale modelling of flow and transport in porous media has become a hot research topic within the academic research community. Generally speaking, multiscale methods are numerical methods and strategies that aim to describe physical phenomena on coarse grids while accounting for the influence of fine-scale structures in the porous media. However, unlike traditional upscaling techniques, multiscale methods often provide a mechanism for also recovering an approximate fine-scale solutions. In this report we make an effort to demonstrate some applications where multiscale flow solvers, used in combination with various methods for fast computation of fluid transport, may spur new ways of using flow information as part of reservoir planning and management.
In recent years methods for streamline-based history matching have been developed. These methods have proved to be efficient for three reasons. First, streamlines delineate flow patterns and can therefore be used to define reduced inverse models. Second, streamline methods provide fast forward simulation. Third, streamline-based sensitivities can be evaluated directly based on one flow simulation for different reservoir responses. This report proved a literature review of streamline-based sensitivities and streamline methods for history matching.
This report presents the Ensemble Kalman filter (EnKF) which has become a popular tool for history matching. The report focuses on the link between EnKF and the use of Kalman filtering in traditional control applications. Further, properties of the EnKF is investigated on a couple of stylistic examples.
Scientific publications and presentations
Journal publications on reservoir seismics, including history matching, estimation of changes in reservoir parameters from 4D seismic and reflectivity from complicated reservoir model
Conference publications on reservoir seismics, including history matching, estimation of changes in reservoir parameters from 4D seismic and reflectivity from complicated reservoir model
This report describes the work performed in the project “Near well modelling”, under the NTNU IO centre. The work has been performed in close collaboration with FMC Technologies.
Near well models based on first principles and Darcy’s law have been developed in the project. The models are developed in order to support fast generation of dynamic GOR values for use in production optimization. The main focus is oil fields with a thin layer of oil and a gas cap. The model has been verified by simulation and gas coning has been illustrated. Moreover, a model describing wells lying in the water zone under the oil layer has been developed. Simulations showing water and gas coning have been presented.
Recommendations for further work have been presented. The main focus is model development for generation of dynamic GOR values, and possible incorporation in FMC Technologies software in the near future. This would also include further development of the near well models to be able to model up-coning of water, development of an estimation scheme for state estimation and a detailed study on the development of more simplified near well models by using black box techniques.
This report presents an alternative solution method for optimizing oil production. It is based on Lagrangian decomposition and dynamic piecewise linearization of nonlinear inflow and well models and pipeline models. The problem is cast into a mixed integer linear program (MILP) and solved using Xpress-MP. The method is tested on a part of the production system for Troll C with interesting results.
