Abstract

Reservoir flow modeling involves two aspects: a functional model (flow equations and numerical methods) and a representation model (related to the reservoir description) - and this last aspect is usually the critical one. Reservoir representation is different from reservoir characterization. It is not a question of describing the reservoir in a more or less exhaustive or realistic way, but a question of incorporating relevant information into the flow simulation model, considering the syntax of the flow simulators and the relative impact of information on the simulation results. Reservoir representation makes the bridge between reservoir characterization and flow simulation.

This paper presents selected results from a cluster of applied research projects in reservoir representation for flow simulation. It includes:

  1. gridding issues in the context of integrated reservoir studies (discussion of terminology, grid specification, generic format reading, etc.);

  2. identification and representation of critical heterogeneities of turbidite reservoirs;

  3. assigning transmissibility multipliers across partially sealing faults in a field case;

  4. incorporation of production data in reservoir flow models with an example of application;

  5. hyperdocumention for reservoir simulation files (making use of HTML tags to get a much richer documentation of the flow simulation model).

Introduction

Reservoir flow modeling involves two complementary components: a functional model and a representation model (Fig. 1). The functional model consists of a set of differential equations and of numerical methods for solving these equations. The representation model mathematically describes a particular reservoir (rock and fluid) by spatial variable coefficients, external boundary conditions and initial conditions that complete the formulation.

From this perspective, the geometry of the reservoir, its porosities, permeabilities, etc. are not input data for the model; their representation of this knowledge is part of the flow model itself. This way of thinking highlights the intimate connection between the reservoir representation and the formulation of the flow problem of interest. Such a connection was evident in the old days of physical modeling: in fact, the functional and representation models were almost indistinguishable in this approach; there is an analogy with the classical relationship between form and content. However, in the context of numerical simulators and discipline specialization it became easy to forget how customized the reservoir representation should be in relation with the functional model and with the objectives of the study.

The critical component of a reservoir simulation study is the representation model: generally, if the study fails in quality or in timing, etc., the reason is that the reservoir was not properly represented. Reservoir characterization has received great attention and there were a lot of advancements but we still often find some lack of objectivity in the way the knowledge about the reservoir is Reservoir representation is different from reservoir characterization. It is not a question of describing the reservoir in a more or less exhaustive or realistic way, but a question of incorporating relevant information into the flow simulation model, considering the syntax of the flow simulators and the relative impact of information on the simulation results. Reservoir representation makes the bridge between reservoir characterization and flow simulation.

This content is only available via PDF.
You can access this article if you purchase or spend a download.