Integration of Geology, Geophysics, and Numerical Simulation in the Interpretation of a Well Test in a Fluvial Reservoir
- R. Raghavan (Phillips Petroleum Co.) | T.N. Dixon (Phillips Petroleum Co.) | V.Q. Phan (Stanford U.) | S.W. Robinson (Phillips Petroleum Co.)
- Document ID
- Society of Petroleum Engineers
- SPE Reservoir Evaluation & Engineering
- Publication Date
- June 2001
- Document Type
- Journal Paper
- 201 - 208
- 2001. Society of Petroleum Engineers
- 5.8.8 Gas-condensate reservoirs, 5.1.5 Geologic Modeling, 1.2.3 Rock properties, 1.6.9 Coring, Fishing, 5.6.1 Open hole/cased hole log analysis, 5.1 Reservoir Characterisation, 5.5 Reservoir Simulation, 5.2.1 Phase Behavior and PVT Measurements, 4.6 Natural Gas, 5.1.7 Seismic Processing and Interpretation, 2.4.3 Sand/Solids Control, 4.3.4 Scale, 5.5.8 History Matching, 5.1.2 Faults and Fracture Characterisation, 5.1.1 Exploration, Development, Structural Geology, 1.6 Drilling Operations, 5.5.3 Scaling Methods, 5.6.4 Drillstem/Well Testing, 4.1.5 Processing Equipment, 5.1.3 Sedimentology, 5.1.8 Seismic Modelling
- 0 in the last 30 days
- 701 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 12.00|
|SPE Non-Member Price:||USD 35.00|
The focus of this paper is the incorporation of geologic and geophysical data in the analysis of pressure tests in a fluvial reservoir. Using a 3,105,851-cell "porosity cube" derived by seismic and well-log data, this work outlines the steps involved in developing a 3D model with 7,128 cells that was used to estimate reservoir properties (such as permeability, skin, permeability/porosity relationship, reservoir connectivity, and fault factor). The principal steps involved in developing the reservoir model include:
identifying all points in the porosity cube that were connected to the wellbore, with porosities above a specified cutoff to identify high-porosity channels and low-porosity overbanks,
upscaling the porosity cube to a reasonable size for well-test simulation while preserving pay thickness, pore volume, and connectivity between high- and low-porosity materials; and
matching the observed pressure and pressure-derivative responses with a numerical simulator and a regression program to automatically adjust the reservoir permeability description.
A well test in a gas-condensate reservoir is used to demonstrate the ideas outlined in this work. This approach allowed us to identify not only horizontal permeability but also vertical permeability. The issues that need to be considered and the conclusions that are derived on the basis of this work differ significantly from conventional interpretations.
The reservoir considered here is a gas-condensate discovery. Four wells or sidetracks have penetrated the reservoir. In this paper, the test from one of these wells is considered. The purpose of this work is to demonstrate the value of integrating seismic and well-log data into the interpretation of a well test. We believe that this work shows the value using all available data in well-test interpretation.
A numerical well-test-simulation model was upscaled from a porosity cube generated by seismic data. Then the well test was matched using automatic history-matching procedures.1 In this paper, we first discuss the development of procedures for generating a 3D well-test model from the porosity cube. Then, we present the process used to obtain a satisfactory history match of the well test. Earlier attempts to match well-test data either by conventional well-test models or by assuming variations in properties derived from amplitude maps were unsatisfactory. The conventional approach suggested that the well produced a two-layer, noncommunicating reservoir, with the high-permeability layer bounded on all sides and the low-permeability layer acting as if it were infinite in extent.
This paper is organized into six sections. We begin with a brief review of numerical well-test procedures that are directly pertinent to this work. The second section addresses geological considerations, as this work presumes that a geologic setting is established, and a geological model based on the depositional environment, petrophysics, and seismic interpretations is derived. The end product of the geologic description is a 3D model that quantitatively describes both reservoir properties and reservoir architecture. In the third section, we focus on determining volumes connected to the well and the process of upscaling. Many techniques for upscaling are available and will not be discussed here. In our opinion, for analyzing well behavior, it is important to concentrate on upscaling in the vertical direction. The upscaling procedure we have followed in this study is conceptual and is based on the constraints imposed by the data available to us. The fourth section addresses issues concerning the preparations to be made for the analysis of the specific test under consideration. The preparations are principally governed by operational considerations and the duration of the test. In Sections 2 through 4, we outline our perspective on steps that need to be taken and issues that need to be addressed before pressure data may be analyzed by the outlined scheme. The fifth section analyzes the pressure response both qualitatively and quantitatively. The focus is on reservoir description. Three specific models are considered (all equally reliable). In this section, we show how the production of a well supplements the geologic description. We conclude the paper by noting observations based on this experience. That reservoir description may change as a result of additional measurements is recognized. We should also note, parenthetically, that the sequence of this paper essentially follows the methodology we have used in evaluating pressure tests described in this work. In general, the work-flow process will be governed by the information available to analyze the tests and the objectives of the analysis.
Numerical Well-Test Review
The numerical simulation of well tests has been conducted for several years.2-4 These early papers were concerned more with numerical than with geological considerations. But in recent years, there has been an increase in the use of numerical simulators to incorporate seismic and geological data into the interpretation of well tests. In general, this work falls into three types: the use of the numerical model to predict the geological model's effect upon the well-test response; the conditioning of geostatistically generated geologic models to the well-test response; and the history matching of the numerical model to observed well-test data. Massonnat and Bandiziol5 review the interdependence between geology and well-test interpretation. Several field examples of how the geology affects well tests or how the well test is used to confirm a geological model are presented. Holden et al.6 used a numerical model of a channel sand to interpret well tests and condition several geostatistical descriptions. Zheng et al.7 used a numerical model to simulate the pressure-test response of wells in meandering channels. The effects of well location within the channel, channel shape, and completion ratios were studied. Corbett et al.8 used a numerical model of braided fluvial reservoirs to calculate well-test responses. A geoskin concept was developed from these data, to be used in a full-field model. Savioli et al.9 used a 1D radial model and regression to analyze well tests. Core permeability/porosity data were used to reduce the number of regression parameters. Matthai et al.10 numerically simulated well-test signatures caused by geologically realistic faults in a sandstone reservoir.
|File Size||858 KB||Number of Pages||8|