A Fully Coupled Network Model: Comparison With Other Integrated Models on Field Cases

Abstract

This article, written by Special Publications Editor Adam Wilson, contains highlights of paper SPE 173251, “A Fully Coupled Network Model—Practical Issues and Comprehensive Comparison With Other Integrated Models on Field Cases,” by H. Cao, SPE, P. Samier, SPE, H.M. Kalunga, E. Detige, and E. Obi, SPE, Total, prepared for the 2015 SPE Reservoir Simulation Symposium, Houston, 23–25 February. The paper has not been peer reviewed. There has been increasing interest in integrated simulation of reservoirs, wells, and surface facilities, which is particularly important for companies with major assets in deep offshore fields. The basic approach for integration can be split into decoupled or separated (between a reservoir simulator and a facility simulator), iteratively coupled, and fully coupled networks, with increasing stability and efforts in implementation. This paper covers the development of a state-of-the- art fully coupled network inside a next-generation commercial reservoir simulator and compares different coupling approaches on real field cases. Introduction Two basic and diverse strategies exist for integrated simulation. The separated strategy focuses on loosely coupling separate software together, typically between an existing reservoir simulator and an existing facility simulator. The combined strategy focuses on tight integration within a single piece of software, typically a reservoir simulator that also models surface facilities. Both of these strategies have their advantages and disadvantages, and both are actively used in the industry. In the separated strategy, a reservoir simulator is used for reservoir and wells modeling and a facility simulator is used for surface-facility modeling. The information exchange [well inflow performance relationship (IPR) from reservoir to facility and well pressure/rate constraint] happens per timestep or periodically between the two simulators through a custom-made controller. The main advantages of this strategy are its low cost in development (mostly associated with the controller) and its flexibility in allowing users to use the reservoir and surface simulators that fit their needs best. On the other hand, as a loosely coupled (here referred to as decoupled) method, it often suffers from numerical instability because the IPR calculated at the beginning of the coupling timestep is not representative of the IPR at the end of the coupling timestep, which is particularly true for the kind of fine and full-featured reservoir models used in reservoir-simulation studies, where well-perforation cells are much smaller than the well-drainage radius. There are also other issues for the strategy. For instance, the information exchanged is based on IPR of individual (and independent) wells without proper accounting of well interference anywhere in the coupling scheme and in the IPR generation and format. However, well interference can become important when opening and shutting nearby wells and when modeling smart wells with well laterals and segments controlled individually. The separated strategy typically requires a combination of software across multiple platforms. In the combined strategy, a reservoir simulator typically is extended to model the facility, thus everything is contained within a single simulator. Contrary to the facility model in a standalone advanced facility simulator, the facility model in a reservoir simulator is a more simplified one in both topology and modeling. The iteratively coupled methods can differ further according to the frequency of coupling and the definition of timestep convergence. Recently, recognizing the fact that both wells and surface facilities model flow in pipes, researchers started to examine the idea of using a general pipe-network module to model both. This network module can replace the traditional well module inside a reservoir simulator in all aspects and, hence, achieve fully implicit coupling among reservoir, wells, and surface facilities. The combined strategy takes considerably more time and effort to develop, but, in turn, it ensures much-improved stability and speed, yielding generous payouts well worth the initial investment. Because of its ease of use (using only a single piece of software) and much-improved stability and speed, especially for fine and detailed reservoir models, this combined strategy has been recommended to and is often preferred by reservoir engineers for reservoir studies of deep offshore fields. The fully implicitly coupled network is the most advanced next-generation approach available. It is unconditionally stable and offers the best speed and performance possible. This approach has been featured in most next-generation industrial reservoir simulators.