Automatic Determination of Well Placement Subject to Geostatistical and Economic Constraints

Automatic Determination of Well Placement Subject to Geostatistical and Economic Constraints Karl P. Norrena; Karl P. Norrena University of Alberta Search for other works by this author on: This Site Google Scholar Clayton V. Deutsch Clayton V. Deutsch University of Alberta Search for other works by this author on: This Site Google Scholar Paper presented at the SPE International Thermal Operations and Heavy Oil Symposium and International Horizontal Well Technology Conference, Calgary, Alberta, Canada, November 2002. Paper Number: SPE-78996-MS https://doi.org/10.2118/78996-MS Published: November 04 2002 Cite View This Citation Add to Citation Manager Share Icon Share Twitter LinkedIn Get Permissions Search Site Citation Norrena, Karl P., and Clayton V. Deutsch. "Automatic Determination of Well Placement Subject to Geostatistical and Economic Constraints." Paper presented at the SPE International Thermal Operations and Heavy Oil Symposium and International Horizontal Well Technology Conference, Calgary, Alberta, Canada, November 2002. doi: https://doi.org/10.2118/78996-MS Download citation file: Ris (Zotero) Reference Manager EasyBib Bookends Mendeley Papers EndNote RefWorks BibTex Search Dropdown Menu nav search search input Search input auto suggest search filter All ContentAll ProceedingsSociety of Petroleum Engineers (SPE)SPE International Thermal Operations and Heavy Oil Symposium Search Advanced Search AbstractOptimal well placement is a complex problem that requires detailed models of the reservoir structure/geometry and the petrophysical properties such as facies, porosity, permeability, and fluid saturation. The reservoir development team attempts to integrate all of these aspects when devising a well plan for optimal reservoir exploitation. Ideally the well locations would be selected with the assistance of a flow simulator; however, this is impractical due to time and CPU requirements. This paper presents a technique for selecting optimal well locations for fine-tuning with a flow simulator. The technique constructs the well placement problem as an optimization problem to be solved with simulated annealing. The global objective function consists of multiple component objective functions. Each component represents a desirable feature or constraint in the problem. Optimality is defined as the best balance among the component objectives. The format of the technique is flexible and can incorporate 3-D geostatistical models of uncertainty and multiple constraints. The proposed method iteratively refines initial well locations and trajectories until the global objective is maximized. Several examples are shown. Optimal well placement in a steam assisted gravity drainage context is illustrated.IntroductionOne task of a reservoir development team is to set a well plan that, given all available information, is reasonable. A well plan is set with the help of a reservoir model. A reservoir model highlights candidate regions for well placement. This initial well plan is static because it does not account for the dynamics of fluid flow. The static well plan is adjusted to a dynamic well plan with the aid of a flow simulator. The process is iterative. This process can be expensive in terms of professional and CPU time. Selecting good static well plans is important because it will reduce the iterations required and lead to better decisions.Assembling a good static well plan is difficult due to heterogeneities and uncertainties in the subsurface reservoir parameters. Accounting for this information in a decision-making framework is the subject of this paper. Various approaches have been proposed including optimization techniques such as mixed integer programming and neural networks1,2. Integer programming requires the objective function to be expressed as a linear function, and neural networks require training and a library of training images. One common technique for accounting for uncertainty involves selecting the P05, P50, P95 realizations and selecting a well plan that is jointly optimal on these using a flow simulator. This approach is impractical in most cases due to the number of possible well locations and the computational expense required for evaluating the realizations.This paper proposes a technique for selecting a good static well plan. The problem is posed as an optimization problem for simulated annealing. Simulated annealing is an optimization routine particularly well suited to optimizing highly combinatorial problems such as the problem of selecting well locations. Consider the placement of two wells on a 2-D reservoir model on a 50 x 50 grid. If the reservoir development team were to exhaustively evaluate every location on the grid there are many combinations:Selecting well locations with respect to uncertainty would require L·C22500 evaluations, where L is the number of realizations. In practice a reservoir model may have a grid size of millions of cells and hundreds of realizations. The combinatorial becomes incomprehensible in size. Keywords: objective function, reservoir model, alignment, component objective function, reservoir characterization, alberta, placement, well location, spe ps-cim choa 78996, well plan Subjects: Reservoir Characterization, Geologic modeling This content is only available via PDF. 2002. SPE/PS-CIM/CHOA International Thermal Operations and Heavy Oil Symposium and International Horizontal Well Technology Conference You can access this article if you purchase or spend a download.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.