## An autonomic reservoir framework for the stochastic optimization of well placement (2004)

### Cached

### Download Links

- [www.math.tamu.edu]
- [www.jhuapl.edu]
- [www.caip.rutgers.edu]
- [nsfcac.rutgers.edu]
- [www.caip.rutgers.edu]
- [nsfcac.rutgers.edu]
- DBLP

### Other Repositories/Bibliography

Venue: | Cluster Computing: The Journal of Networks, Software Tools, and Applications |

Citations: | 20 - 12 self |

### BibTeX

@ARTICLE{Bangerth04anautonomic,

author = {Wolfgang Bangerth and Hector Klie and Mary F. Wheeler},

title = {An autonomic reservoir framework for the stochastic optimization of well placement},

journal = {Cluster Computing: The Journal of Networks, Software Tools, and Applications},

year = {2004},

volume = {8},

pages = {2005}

}

### OpenURL

### Abstract

Abstract. The adequate location of wells in oil and environmental applications has a significant economic impact on reservoir management. However, the determination of optimal well locations is both challenging and computationally expensive. The overall goal of this research is to use the emerging Grid infrastructure to realize an autonomic self-optimizing reservoir framework. In this paper, we present a policy-driven peer-to-peer Grid middleware substrate to enable the use of the Simultaneous Perturbation Stochastic Approximation (SPSA) optimization algorithm, coupled with the Integrated Parallel Accurate Reservoir Simulator (IPARS) and an economic model to find the optimal solution for the well placement problem.

### Citations

445 |
The Grid 2: Blueprint for a New Computing Infrastructure
- Foster, Kesselman
- 2003
(Show Context)
Citation Context ...iguration and invocation of these algorithms, and the dynamic optimization of the reservoir remains a challenging problem. The overall goal of this research is to use the emerging Grid infrastructure =-=[7]-=- and its support for seamless aggregations, compositions, and interactions to realize an autonomic self-optimizing reservoir application. The application consists of: (1) sophisticated reservoir simul... |

224 | Multivariate stochastic approximation using a simultaneous perturbation gradient approximation
- Spall
- 1992
(Show Context)
Citation Context ...n our earlier work [15], we studied the use of the Very Fast Simulated Annealing (VFSA) [24] optimization technique. In this paper we use the Simultaneous Perturbation Stochastic Approximation (SPSA) =-=[27, 25]-=- algorithm for optimizing well placement. The reservoir framework consists of (i) instances of distributed multi-model, multi-block reservoir simulation components provided by the IPARS reservoir simu... |

123 |
Fundamentals of Numerical Reservoir Simulation
- Peaceman
- 1977
(Show Context)
Citation Context ...d from the reservoir by the given set of wells. Or, one may wish to minimize the amount of produced water. This last case is somewhat akin to preventing the water coning and water fingering phenomena =-=[20, 5]-=-. Note, however, that the (negative) cost of water production already appears as one term in the objective function defined above. 2.4 Optimization As mentioned above, viable methods for finding the m... |

118 |
Introduction to Stochastic Search and Optimization: Estimation, Simulation, and Control
- Spall
(Show Context)
Citation Context ...n our earlier work [15], we studied the use of the Very Fast Simulated Annealing (VFSA) [24] optimization technique. In this paper we use the Simultaneous Perturbation Stochastic Approximation (SPSA) =-=[27, 25]-=- algorithm for optimizing well placement. The reservoir framework consists of (i) instances of distributed multi-model, multi-block reservoir simulation components provided by the IPARS reservoir simu... |

116 |
Mathematical Models and Finite Elements for Reservoir Simulation
- CHAVENT, JAFFRÉ
- 1986
(Show Context)
Citation Context ...d from the reservoir by the given set of wells. Or, one may wish to minimize the amount of produced water. This last case is somewhat akin to preventing the water coning and water fingering phenomena =-=[20, 5]-=-. Note, however, that the (negative) cost of water production already appears as one term in the objective function defined above. 2.4 Optimization As mentioned above, viable methods for finding the m... |

100 |
Petroleum reservoir simulation
- Aziz, Settari
- 1979
(Show Context)
Citation Context ...lity of a component, and Pm the pressure of a phase. Additional equations specifying volume, capillary, and state constraints are added, and boundary and initial conditions complement the system, see =-=[2, 9]. Finally, Nm = -=-Smρm with Sm denoting saturation of a phase. The resulting system (omitting gravity terms for simplicity) is ∂(φρmSm) ∂t − ∇ · (ρmKλm∇Pm) = qm. (2) In this paper we consider wells that... |

87 |
Finite Element and Finite Difference Methods for Continuous Flows
- Russel, Wheeler
(Show Context)
Citation Context ...his model is discretized in space using the expanded mixed finite element method which, in the case considered in this paper, is numerically equivalent to the cell-centered finite difference approach =-=[23, 1]-=-. Time discretization can be either fully implicit, semi-implicit or sequential; here we only consider the sequential method in which two linear systems of equations, the pressure equation and the con... |

65 | Mixed finite elements for elliptic problems with tensor coefficients as cell-centered finite differences
- Arbogast, Wheeler, et al.
- 1997
(Show Context)
Citation Context ...his model is discretized in space using the expanded mixed finite element method which, in the case considered in this paper, is numerically equivalent to the cell-centered finite difference approach =-=[23, 1]-=-. Time discretization can be either fully implicit, semi-implicit or sequential; here we only consider the sequential method in which two linear systems of equations, the pressure equation and the con... |

30 |
Global optimization methods in geophysical inversion
- Sen, Stoffa
- 1995
(Show Context)
Citation Context ...) the design of a prototype policy-driven framework for autonomic reservoir optimization in Grid environments. In our earlier work [15], we studied the use of the Very Fast Simulated Annealing (VFSA) =-=[24]-=- optimization technique. In this paper we use the Simultaneous Perturbation Stochastic Approximation (SPSA) [27, 25] algorithm for optimizing well placement. The reservoir framework consists of (i) in... |

27 |
Multiphase flow and transport processes in the subsurface
- Helmig
- 1997
(Show Context)
Citation Context ... ∇ · Um = qm. (1) Here, φ is the porosity of the porous medium, Nm the concentration of a component m, and qm the sources (production and injection rates). The fluxes Um are defined using Darcy’=-=s law [9] which, -=-with gravity ignored, reads as Um = −ρmKλm∇Pm, where ρm denotes the density of a component, K the permeability tensor, λm the mobility of a component, and Pm the pressure of a phase. Additiona... |

23 | DISCOVER: An environment for Web-based interaction and steering of high-performance scientific applications. Concurrency: Practice and Experience
- Mann, Matossian, et al.
- 2001
(Show Context)
Citation Context ...e provided by IPARS, define the structure and properties of the reservoir to be simulated, and list required parameters. Deployment and management of IPARS instances use services provided by Discover =-=[14]-=- and Globus [6], and build on the CORBACoG Kit [18]. 3.3 SPSA Optimization service The SPSA Optimization service runs on the Optimization peer and implements the SPSA algorithm presented in Section 2.... |

23 |
A new generation EOS compositional reservoir simulator
- Parashar, Wheeler, et al.
- 1997
(Show Context)
Citation Context ... time step. This discrete model is solved by the IPARS (Integrated Parallel Accurate Reservoir Simulator) software developed at the Center for Subsurface Modeling at The University of Texas at Austin =-=[10, 28, 19, 21, 30, 31, 13, 29]-=-. IPARS is a parallel reservoir simulation framework for modeling multiphase, multiphysics flow in porous media. It offers sophisticated simulation components that encapsulate complex mathematical mod... |

22 |
Multiphase Flow and Transport
- Helmig
- 1997
(Show Context)
Citation Context ...∇·Um = qm. (1) ∂t Here, φ is the porosity of the porous medium, Nm the concentration of a component m, and qm the sources (production and injection rates). The fluxes Um are defined using Darcy’s law =-=[9]-=- which, with gravity ignored, reads as Um =−ρmKλm∇ Pm, where ρm denotes the density of a component, K the permeability tensor, λm the mobility of a component, and Pm the pressure of a phase. Additiona... |

19 | A new generation EOS compositional reservoir simulator
- Wang, Yotov, et al.
- 1997
(Show Context)
Citation Context ... time step. This discrete model is solved by the IPARS (Integrated Parallel Accurate Reservoir Simulator) software developed at the Center for Subsurface Modeling at The University of Texas at Austin =-=[10, 28, 19, 21, 30, 31, 13, 29]-=-. IPARS is a parallel reservoir simulation framework for modeling multiphase, multiphysics flow in porous media. It offers sophisticated simulation components that encapsulate complex mathematical mod... |

18 | A distributed object infrastructure for interaction and steering. Concurrency and Computation: Practice and Experience 2003; 15(10):957–977. Copyright c○ 2005 - Muralidhar, Parashar |

17 | DIOS++: A Framework for Rule-Based Autonomic Management of Distributed Scientific Applications
- Liu, Parashar
(Show Context)
Citation Context ...ed resources, to run applications on these resources, and use Discover to connect to and collaboratively monitor, interact with, and steer the applications. • DIOS Interactive Object Framework (DIOS=-=) [12, 16]-=- that enables the runtime monitoring, interaction and computational steering of parallel and distributed applications on the Grid. DIOS enables application objects to be enhanced with sensors and actu... |

16 | Parashar M, “Discover Middleware Substrate for Integrating Services on the Grid
- Bhat
- 2003
(Show Context)
Citation Context ...ss, interrogate, interact with, and steer these applications using pervasive portals. Key components of the Discover collaboratory include: • Discover Interaction & Collaboration Middleware Substrat=-=e [3]-=- that enables global collaborative access to multiple, geographically distributed instances of the Discover computational collaboratory, and provides interoperability between Discover and external Gri... |

16 | A parallel multi-block black-oil model in multi-model implementation
- Lu, Peszynska, et al.
(Show Context)
Citation Context ... time step. This discrete model is solved by the IPARS (Integrated Parallel Accurate Reservoir Simulator) software developed at the Center for Subsurface Modeling at The University of Texas at Austin =-=[10, 28, 19, 21, 30, 31, 13, 29]-=-. IPARS is a parallel reservoir simulation framework for modeling multiphase, multiphysics flow in porous media. It offers sophisticated simulation components that encapsulate complex mathematical mod... |

13 | Iterative solvers of the implicit parallel accurate reservoir simulator (IPARS
- Lacroix, Vassilevski, et al.
(Show Context)
Citation Context ...on in the subsurface, and which execute on parallel and distributed systems. Solvers employ state-of-the-art techniques for nonlinear and linear problems including multigrid and other preconditioners =-=[11]-=-. It can handle an arbitrary number of wells each with one or more completion intervals. Although not used here, IPARS supports multiple physical models and their multiphysics couplings. 2.3 The econo... |

9 | A distributed computing portal for coupling multiphysics and multiple domains in porous media
- Wheeler, Wheeler, et al.
- 2000
(Show Context)
Citation Context |

8 |
Globus: A Toolkit Based Grid Architecture
- Foster, Kesselman
- 1999
(Show Context)
Citation Context ...on the Pawn P2P substrate, which provides JXTA-based [22] peer-to-peer messaging services, and the Discover computational collaboratory, which combines Grid infrastructure services provided by Globus =-=[6]-=- and interaction and collaboration services. The rest of this paper is organized as follows. Section 2 describes the well placement problem and introduces the underlying models and components. It also... |

8 | R.N.: Uncertainty assessment of well placement optimization
- Guyaguler, Horne
- 2001
(Show Context)
Citation Context ...ertainty and significantly reduce the risk in decisionmaking. Consequently, there is an increasing interest in the use of optimization algorithms for finding the optimum well location in oil industry =-=[4, 8, 17, 32]-=-. However, the selection of appropriate optimization algorithms, the runtime configuration and invocation of these algorithms, and the dynamic optimization of the reservoir remains a challenging probl... |

8 | F.: “Multiphysics Coupling of Codes
- Peszynska, Lu, et al.
- 2000
(Show Context)
Citation Context |

8 |
Optimization of Nonconventional Well Type, Location and Trajectory
- Yeten, Durlofsky, et al.
- 2002
(Show Context)
Citation Context ...ertainty and significantly reduce the risk in decisionmaking. Consequently, there is an increasing interest in the use of optimization algorithms for finding the optimum well location in oil industry =-=[4, 8, 17, 32]-=-. However, the selection of appropriate optimization algorithms, the runtime configuration and invocation of these algorithms, and the dynamic optimization of the reservoir remains a challenging probl... |

7 | R.N.: Improved methods for multivariate optimization of field development scheduling and well placement design
- Pan, Horne
- 1998
(Show Context)
Citation Context ...ertainty and significantly reduce the risk in decisionmaking. Consequently, there is an increasing interest in the use of optimization algorithms for finding the optimum well location in oil industry =-=[4, 8, 17, 32]-=-. However, the selection of appropriate optimization algorithms, the runtime configuration and invocation of these algorithms, and the dynamic optimization of the reservoir remains a challenging probl... |

6 | Multiphysics coupling of codes, in: Computational Methods - Peszynska, Lu, et al. - 2000 |

5 | R.N.: Reservoir development and design optimization
- Bittencourt, Horne
(Show Context)
Citation Context |

5 |
M.: Computational engineering and science methodologies for modeling and simulation of subsurface applications
- Wheeler, Peszynska
(Show Context)
Citation Context |

5 |
Modeling subsurface flow on pc
- Wheeler, Peszynska, et al.
- 2000
(Show Context)
Citation Context |

4 |
Autonomic oil reservoir optimization on the grid. Concurrency and Computation: Practice and Experience
- Matossian, Bhat, et al.
- 2003
(Show Context)
Citation Context ...tions between instances of peer services are opportunistic, based on runtime discovery and specified policies, and are not predefined. In this paper we use our prototype autonomic reservoir framework =-=[15]-=- to investigate the policy-driven runtime selection and invocation of optimization services to determine optimal well placement and configuration. The specific objectives of this paper include: (1) ch... |

3 |
A corba commodity grid kit
- Parashar, Laszewski, et al.
- 2002
(Show Context)
Citation Context ...rties of the reservoir to be simulated, and list required parameters. Deployment and management of IPARS instances use services provided by Discover [14] and Globus [6], and build on the CORBACoG Kit =-=[18]-=-. 3.3 SPSA Optimization service The SPSA Optimization service runs on the Optimization peer and implements the SPSA algorithm presented in Section 2.4. It also offers interfaces and mechanisms for int... |

2 |
Adaptive stochastic approximation by the simulateous perturbation method
- Spall
(Show Context)
Citation Context ...e., ∆k in {−1, +1}. The gain parameters ck, ak are a decreasing sequence with respect to k. Although they may change according to the problem, we have found it suitable to define them as suggested=-= in [26]. F-=-or the present problem, we use c = 5 and a = 2 · 10 −5 . Step 2.3 and 2.4 are used to compute an 5sapproximation to the magnitude of the gradient given by gk. The reader may realize that the update... |

2 |
Fundamentals of Numerical Reservoir Simulation, 1st edition (Elsevier Scientfic
- Peaceman
- 1977
(Show Context)
Citation Context ...d from the reservoir by the given set of wells. Or, one may wish to minimize the amount of produced water. This last case is somewhat akin to preventing the water coning and water fingering phenomena =-=[5, 20]-=-. Note, however, that the (negative) cost of water production already appears as one term in the objective function defined above. 2.4. Optimization As mentioned above, viable methods for finding the ... |