, J. Fuller 
, and B. Granier 
by
T.A. Lyon , J. Fuller , and B. Granier
ABSTRACT
This paper presents a case study drawn from an integrated reservoir study on a faulted and fractured carbonate reservoir offshore Abu Dhabi. Particular emphasis is put on demonstrating how the results of 3D seismic interpretation impacted the structural understanding. The paper also explains how the seismic interpretation, well-bore observations and production data were integrated to provide an improved reservoir model.
INTRODUCTION
The objective of the work was to build a 3D model of the reservoir that could be input to a reservoir simulator for use in predicting future reservoir performance, and allow quantitative evaluation of future production wells. The model was required to include faults identified from seismic and to make predictions of the fracture permeability consistent with log, core, and production observations as far as practical.
The hydrocarbons are hosted in a faulted and fractured structure capable of producing from 3 main zones. These zones here in referred to as zones W, X, Y and Z are seen in the left-hand column of figure 1. This Lower Cretaceous reservoir has been produced off and on since the 1960's. The reservoir is also the subject of a gas injection pilot program.
Figure 1 - Reservoir Layering and Dominant Rock Type
The inputs to the study include 3D seismic data, various types of bore hole data, the production history and monitoring data from the ongoing gas injection pilot.
The 3D seismic survey acquired in 1994-1995 has revealed a large number of faults, most of which were not previously mapped. These faults have a consistent, roughly NW to SE, strike direction.
An analysis of the fracture data has been performed using a combination of fracture descriptions from core and borehole images. The analysis has been used to estimate fracture spacing and fracture aperture; these two characteristics have been used to generate an estimate of the fracture permeability and porosity.
Review of the sedimentology and stratigraphy has, in combination with a rock type approach, assisted in establishing the flow characteristics of the matrix and a reservoir-layering model.
The model building was designed to provide a dual porosity/dual permeability model.
MAIN TASKS
Listed below are the generic issues in the model building process discussed in this paper:
1. Fault mapping and building a structural framework with non-vertical faults based on the mapping of 3D seismic calibrated to well bore data.
2. Determination of a layering system for the matrix based on sequence stratigraphy and incorporating a rock type approach to assist in populating those layers with porosity and permeability values.
3. Fracture data analysis aimed at modelling fracture controls, fracture porosity and fracture permeability.
4. Building a flow simulation grid that honours the fault geometry, integrates the geological model, and respects the assumptions implicit in the flow simulation algorithms.
5. Populating the geological and flow simulation models with the appropriate levels of heterogeneity.
Throughout the study integration of the data from different disciplines has been emphasised so that the critical issues are addressed and a workflow developed which will allow revised data interpretations or additional information to be incorporated in an efficient manner.
DATABASE
Core Data -: Approximately 4,000 feet of core recovered from 16 wells, 5,000 thin sections and a total of 65 Mercury injection samples were available.
Fracture data -: Fracture description of cores in eleven wells (cored entirely or partly) from the reservoir. In addition the fracture interpretation from bore-hole image data in 16 wells some of which were included in the eleven cored wells.
The core fracture data included measurements of fracture strike, dip, azimuth, frequency (density), intensity, length, width, effective width, morphology and origin. The FMS/FMI fracture data includes fracture strike, dip, azimuth, frequency and morphology in addition to computed values of fracture porosity and fracture aperture in some of the wells. The geographic spread of these data across the field is good.
Well log data -: The wireline log data from 38 wells has been collated in and around the reservoir. During this process, where necessary, the data was validated and interpretations revised to give a consistent data set, particularly in relation to porosity logs and Sw curves.
Production data -: The study has utilised data from PLTs, TDTs, the production history, pressure transient tests and PVT data.
Seismic data -: The seismic data used in the study is from a 3D Ocean Bottom Cable survey acquired over the reservoir during 1994-1995.
FAULTS AND STRUCTURAL MAPPING
The output from the 3D seismic interpretation is one of the fundamental inputs to the new model building process. It is sufficient to emphasise that the 3D seismic (see Figure 2) has revealed the existence and orientation of many more faults than could be identified using the well data alone. See (Figure 3) for a comparison of maps generated with well data only and with 3D seismic.
Figure 2 - 3D seismic data and TWT surface
Figure 3 - Top reservoir depth maps - (left: Wells only; right: with Seismic)
The key outputs from the 3D seismic structural interpretation can be summarised as follows-:
· Top reservoir depth map with fault polygons.
· Near base reservoir depth map with fault polygons.
· Fault plane surfaces in depth for the 38 seismically detectable faults.
The interpretation of the seismic data, validated by the well data, suggests that a minimum fault throw of 12 to 15 feet is detectable.
Figure 4 - Time dip map on Top Reservoir
Time dip maps, such as figure 4, where used to highlight lineaments and assist in mapping the faults in cross-section. Figure 5 is a typical seismic cross-section through the field.
Figure 5 - Seismic Cross-section
