Constraining 3D Static Models To Seismic And Sedimentological Data: A Further Step Towards Reduction Of Uncertainties

Abstract

Abstract The prediction of the spatial distribution of petrophysical properties within a heterogenous reservoir is affected by significant uncertainties when based only on well information. However, integrating additional constraints such as 3D seismic data and sedimentary concepts can significantly improve the accuracy of reservoir models and help reduce uncertainties on predictions away from wells. This paper addresses these issues with an approach for 3D modelling of static reservoir properties that involves the combined use of sedimentary modelling and seismic inversion techniques. The final results are multiple 3D petrophysical models with associated uncertainties. These models are coherent with the sedimentary model and consistent with seismic amplitude data. These results may then be used as an input to uncertainty analysis on static properties such as distribution of pore volume and hydrocarbon in place. An example from a North Sea shallow-water siliciclastic reservoir is presented. This example demonstrates how integration of seismic and sedimentological can effectively influence the reservoir model both in terms of spatial distribution of reservoir properties and volumetric calculations. Introduction Integrating sedimentological and seismic contraints in the process of modeling petrophysical properties is a key step to reducing uncertainties on static properties of a reservoir. Indeed, in the absence of production data, sedimentological concepts and seismic attributes provide the only 3D framework to extrapolate petrophysical properties away from wells. Furthermore, sedimentological and seismic data bring complementary information on reservoir properties. Seismic data are indirect measurements of porosity, fluid type, mineralogy and rock compressibility. They can therefore be used to guide extrapolation of pore volume, Net to Gross, fluid contacts, and extent of cementation. On the other hand, the type of depositional environment controls the degree of sorting, granulometry, and transitions between facies. This information may be used to characterize flow units within the reservoir and improve permeability prediction away from wells. The recent emergence of geostatistical techniques has greatly improved the ability to incorporate sedimentological and seismic data in reservoir models. It is now common practice to condition petrophysical properties in reservoir models to paleogeographic maps or 3D seismic attributes using kriging with drift, collocated cokriging or cosimulation (see Dubrule (1) for review). In this paper we propose to combine 3D sedimentary modelling and seismic inversion techniques to produce a 3D seismic-sedimentary framework to distribute petrophysical properties. We first describe an interpolation technique for constructing realistic 3D models of depositional environments. The proposed method constrains the 3D sedimentary model with well data (environments, accomodation cycles), a priori geological knowledge contained in paleogeographic maps, correlation panels, and stratigraphic gridding and seismic impedance results when available. This technique makes full use of the sedimentological interpretation at the well and, when necessary, captures fine scale heterogeneities observed at the well. We then present a methodology for introducing seismic information in reservoir models using geostatistical inversion, a powerful technique to constrain seismic inversion with well data and geological information. This method generates, within the framework of a stratigraphic grid, a family of 3D impedance realisations that are all compatible with seismic amplitude data and honor well data.