This article presents applications of deterministic and conditional geostatistical reservoir characterization methods to the heterogeneous carbonates of the upper Shuaiba formation in Daleel field, Oman. High-resolution reservoir descriptions based on the integration of logs, core, pressure transient tests, geology, and seismic data are constructed; and upscaled for use in reservoir simulation models to history match field performance data. Generally, geostatistical techniques combined with geology and proper upscaling of permeability heterogeneity yield best results without artificial alterations in various fluid and rock properties. Although acceptable history matches can be obtained with compromised lessdetailed reservoir descriptions, these require modifications to reservoir data beyond reasonable ranges. Only detailed and concise reservoir descriptions result in history matches that are consistent with a variety of measured data sources. IntroductionReservoir characterization has gained a new momentum in the past decade, largely due to the introduction of geostatistical methods to the petroleum industry and rapid progress made in their advancement. 1 The keen interest in reservoir characterization arises because it is well recognized that reservoir heterogeneity has a profound affect on all phases of hydrocarbon recovery, ranging from oil in-place calculations to sweep and conformance efficiency determination of various injection processes. Thus, any improved understanding of a reservoir will aid in better management and better exploitation of its hydrocarbon recovery potential.The challenge in understanding and predicting reservoir performance is two-fold: first, to describe reservoir geologic heterogeneities realistically and quantitatively, and second to model reservoir flow behavior in the presence of all heterogeneities accurately and efficiently. 2 While large-scale reservoir features ͑such as main layers or major faults͒ can be described by deterministic techniques, less-correlated medium-scale and more-chaotic smallscale heterogeneities may be characterized by geostatistical methods or related interpolative techniques. This is especially true for estimating interwell reservoir properties based on a limited amount of information available at wells.The approaches to reservoir characterization fall into three categories: deterministic, stochastic, and combination of the two. The deterministic approach has been in use for several decades and ample success with it has been reported. The interwell properties are generally interpolated or extrapolated using algorithms based on the inverse-distance-square principle or variations of it. Usually, adjustments to the number of layers, gridblock properties, relative permeabilities, and even fluid properties are made in order to history match field performance. Some of these adjustments are warranted and some are solely knobs that are arbitrarily tuned in simulation models without physical bases. Thus, the resulting reservoir models may lack reliability and predic...

This paper presents applications of deterministic and geostatistical reservoir characterization methods to the heterogeneous carbonates of the upper Shuaiba formation in Daleel field, Oman. High resolution reservoir descriptions based on the integration of logs, core, pressure transient tests, geology and seismic data are constructed; and properly upscaled for use in reservoir simulation models in history matching of field performance data. Deterministic descriptions utilize three and six layer models to assess the impact of layering heterogeneity on reservoir performance. Conditional geostatistical methods of SGSIM and SISIM with and without explicit modeling of geology generate fine-scale realizations of porosity and hydraulic flow units, which are then used to construct permeability and initial water saturation distributions. Permeability heterogeneity is upscaled to coarse-scale simulation gridblocks by an improved upscaling method based on an upstream-weighted interblock transmissibility algorithm. The study shows that detailed and concise reservoir descriptions result in history matches that are more consistent with a variety of measured data. Generally, geostatistical techniques combined with geology and proper upscaling of permeability heterogeneity yield best results without artificial alterations in various fluid and rock properties. Although acceptable history matches could be obtained with less detailed reservoir descriptions, these required modifications to fluid and rock data beyond reasonable ranges, illustrating the need for detailed reservoir characterization.

No abstract

No abstract