--> Abstract: Capturing Credible Reservoir Architecture in 3-D Sub-Surface Models of Shoreface and Paleovalley Reservoirs: a Comparison of Some Geostatistical Versus Deterministic Approaches, by Paul Davies, Huw Williams, and Simon Pattison; #90078 (2008)

Datapages, Inc.Print this page

Capturing Credible Reservoir Architecture in 3-D Sub-Surface Models of Shoreface and Paleovalley Reservoirs: a Comparison of Some Geostatistical Versus Deterministic Approaches

Paul Davies1, Huw Williams1, and Simon Pattison2
1Reservoir Geology Consultants Limited, Rhandirmwyn, United Kingdom
2Department of Geology, Brandon University, Brandon, MB, Canada

Optimizing field development and design of surface facilities relies on accurate production forecasting. A key parameter in workflow efficiency is to minimize uncertainty in 3-D sub-surface models by only including credible sedimentary architecture.

The objective of this study was to define the best modeling strategy for the creation of realistic reservoir architecture in sub-surface models of wave-dominated shoreface/deltaic deposits. To achieve this goal, a fully-deterministic 3-D model was created which exactly matches a well-exposed series of outcrops which have adjacent sub-surface core and log data.

The exceptionally well-exposed near-horizontal Campanian strata of the Book Cliffs in eastern Utah were chosen as the basis for this study as they provide ideal outcrop analogs which have previously been used to develop, test and refine many sedimentological and stratigraphic models, including the principles and concepts of sequence stratigraphy.

Various probabilistic methods were used to create a series of different models using the same input well data; each result was then objectively compared against the answer provided by the fully deterministic model.

Results showed that the most successful approach to creating realistic architecture in models was to use deterministic methods where detailed geological knowledge of organisational patterns can be included.

Probabilistic models constructed using a minimum of geological input result in the poorest comparison to reality, with an unrealistic, disorganised scatter of sediment bodies. However, these models showed considerable improvement when (a) a high-resolution deterministic framework of timelines was introduced; (b) when geological trends were used to capture the broad-scale architecture of the reservoir; (c) when body shape and size were deterministically constrained.

 

AAPG Search and Discovery Article #90078©2008 AAPG Annual Convention, San Antonio, Texas