Answering the Challenge of Up-Scaling a 900 Million-Cells Static Model to a Dynamic Model - G reater Burgan Field, Kuwait

Abstract

The Greater Burgan field is the world’s largest sandstone oil field. It has been producing since 1946 under primary depletion from natural water drive. Sub-surface modeling is an integral part of reservoir management and Kuwait Oil Company (KOC) has been investing significant amount of resources in this technology to support field development planning and depletion strategy. In 2001, the first comprehensive Greater Burgan full-field geological model was built with 65 million cells encompassing all the major reservoirs. Subsequently, a reservoir simulation study with a 1.6 million cells dynamic model was conducted in 2003 utilizing parallel simulation technology. In the last decade, active field development plans have resulted in major surface facility upgrades and more than 300 new wells drilled. The existing sub-surface models no longer sufficed to meet technical requirements and as a result, an unprecedented Greater Burgan sub-surface modeling project was commenced in 2009. This is a 4-year project consisting of structural, static and dynamic modeling. It started with Sequence Stratigraphy Study followed by Geo-modeling. The latter was completed in August 2011 and subsequently paved way to Dynamic Modeling phase of the study. This paper discusses up-scaling of the high resolution geological model and the specific problems that the study team had to overcome in the process. State-of-the-art technologies were applied to the construction of the biggest-ever geological model (900 million cells) of the Greater Burgan field. The high resolution of the static model was necessitated by not only the sheer size of the field, but also, by the complex depositional environment defining the internal architecture of the reservoir and the resultant heterogeneity in the system. Sedimentological and stratigraphic data were used extensively to describe the internal architecture of the reservoir, capturing the level of heterogeneity observed in the field. A primary use of this high resolution model was to create a basis for the flow simulation model used in reservoir management. Although computing technology has advanced significantly, conducting flow simulations on such a fine scale model demands prohibitive amount of computation and becomes impractical when a time constraint is imposed on the project. Therefore, model up-scaling is essential to conduct simulations in a reasonable run time. Preservation of volumetric quantities and flow features were the two key considerations for the successful up-scaling. While volumetric conservation can be achieved by following a strict procedure, preserving flow features across the various reservoirs imposes a great challenge. This paper addresses actual challenges encountered during the up-scaling process. The discussion focuses on the following topics: Choice of the model size considering both computational time and accuracy of simulation results. Need for multi-scale approach with three simulation models: Fine, Coarse, and Very Coarse - each to be used to answer specific questions of the study;Right balance between areal and vertical grid coarsening that ensures adequate model physics and preservation of geological features;Mechanistic modeling to support decisions made in the process of up-scaling;Preservation of flow features in various reservoirs, difference between massive and more heterogeneous reservoirs;Transferring water saturation between fine and coarse models: testing various approaches to find one that produces the best volumetric match.

Upscaling Polymer Flooding to Model Sub-gridblock Geological Heterogeneity and Compensate for Numerical Dispersion

“…For example, a 900 million grid block static model was subject to multiple upscaling methodologies, all of which are either averaging or flow-based single phase techniques. (Ma et al, 2013).…”

Reservoir Simulation to Support Pressure Maintenance Projects in the Greater Burgan Field, Kuwait

“…The final product was a 951 million cells geological model with various rock-types defined in each of the six major reservoirs. Details of geo-model construction and up-scaling were described comprehensively in two technical papers, AAPG#1235186 [7] and SPE paper#164187 [9] referred in the reference section of this paper. …”

Reservoir Simulation Modeling of the World

“…Details on the up-scaling process have been documented in SPE-164187 [1] and presented at the 2013 MEOS in Bahrain. To succeed, it required the use of an elaborate approach for grid coarsening and up-scaling/transfer of the properties from finescale static model to the coarse simulation grid.…”