MAPPING OF RESERVOIR PROPERTIES AND FACIES THROUGH INTEGRATION OF STATIC AND DYNAMIC DATA

Knowledge of the distribution of permeability and porosity in a reservoir is necessary for the prediction of future oil production, estimation of the location of bypassed oil, and optimization of reservoir management. But while the volume of data that can potentially provide information on reservoir architecture and fluid distributions has increased enormously in the past decade, it is not yet possible to make use of all the available data in an integrated fashion. While it is relatively easy to generate plausible reservoir models that honor static data such as core, log, and seismic data, it is far more difficult to generate plausible reservoir models that honor dynamic data such as transient pressures, saturations, and flow rates. As a result, the uncertainty in reservoir properties is higher than it could be and reservoir management can not be optimized. The goal of this project is to develop computationally efficient automatic history matching techniques for generating geologically plausible reservoir models which honor both static and dynamic data. Solution of this problem is necessary for the quantification of uncertainty in future reservoir performance predictions and for the optimization of reservoir management. Facies (defined here as regions of relatively uniform petrophysical properties) are common features of all reservoirs. Because the flow properties of the various facies can vary greatly, knowledge of the location of facies boundaries is of utmost importance for the prediction of reservoir performance and for the optimization of reservoir management. When the boundaries between facies are fairly well known, but flow properties are poorly known, the average properties for all facies can be determined using traditional techniques. Traditional history matching honors dynamic data by adjusting petrophysical properties in large areas, but in the process of adjusting the reservoir model ignores the static data and often results in implausible reservoir models. In general, boundary locations, average permeability and porosity, relative permeability curves, and local flow properties may all need to be adjusted to achieve a plausible reservoir model that honors all data. In this project, we will characterize the distribution of geologic facies as an indicator random field, making use of the tools of geostatistics as well as the tools of inverse and probability theory for data integration.