An Analysis of Multiplicative Schwarz Procedure in Coupling of Reservoir and Networks in Next-Generation Reservoir Simulators
Cenk Temizel,Dike Putra,Ihsan M. Gok,Ming Zhang,Tanabordee Duangprasert,Sinem Aktas,Suleyman Tek,Karthik Balaji,Rahul Ranjith
DOI: https://doi.org/10.2523/iptc-18585-ms
2016-01-01
Abstract:AbstractWith the increasing use of reservoir simulation, not only in conventional reservoir modeling, but also in complex and large reservoirs, the need for faster and more robust modeling of reservoirs, along with the facilities and surface networks, has become crucial to meet the requirements of modeling either large or multiple sub-reservoirs with thousands of wells and surface networks, including pipelines and gathering centers, connected all the way to the separators. Next-generation reservoir simulators have begun replacing conventional reservoir simulators over the last decade in conjunction with the addition of more features that enable them to model different phenomena of reservoir fluid flow and recovery techniques, putting them at the forefront of reservoir simulation. While next-generation simulators offer more convenient solutions through surface-subsurface coupling and faster simulations that save time in decision-making processes in reservoir management, there are still some difficulties for the user, as well as in the development of these simulators from a computational point of view. The objective of this study is to outline the strengths and weaknesses along with the challenges in this process to serve as a guide for more efficient use.This study presents an analysis of an important component in the reservoir-network-coupling process in which the reservoir unknowns and network equations are solved iteratively. In the Schwarz procedure, the overlap region consists of reservoir cells included in the network set. An optimum solution for the number of cells extended into the network is important, because as the number of reservoir cells included in the network increases, the cost of the network to solve increases, which can become quite expensive, since bigger and more complex networks are needed to find the solution.This study highlights the effect of using different options for a better understanding of the important components of the coupling process; the results serve as a guide for a more efficient use of this procedure.Reservoir simulation involves several complex calculations to model the full-physics of the subsurface flow, as well as the well and surface facilities. It is not straightforward and easy to model and computationally evaluate the modeled phenomena. Discretizations and approximations contribute to error. Run time is another component that must be managed to find a balance between accuracy, precision, and time. Literature lacks practical examples of the Schwarz procedure, which is another important component in coupling reservoir and surface networks that should be used wisely and efficiently to find a balance between accuracy and computational time. A solid, practical example with the theoretical background that this study outlines will benefit the users of complex and large reservoir simulation models.