Investigation of Petrophysical Property Distribution of Reservoirs at Exploration Stage with Integration of Fuzzy Methods into Geostatistics

Abstract

Theories of fluid flow in heterogeneous porous media show that their transport properties are determined by the structure of spatial correlation in the permeability distribution. Geo-statistical methods have been used to provide equally probable reservoir descriptions that honor available data of a given reservoir for the past decades. During the early, exploration stage, the hard data (coring, well testing and well log) often come only from several wells with large well spacing. Though soft data (such as geophysical and geological interpretation, outcrop and seismic data) can be taken into consideration, it is still problematic to characterize a formation by deterministic Kriging or probabilistic simulation, because the formation property spatial variability in different directions is hard to capture by sparse hard data and soft data, especially in lateral direction.

In this paper, fuzzy methods are used to quantify the anisotropy and spatial correlation structure of formation petrophysical properties in a reservoir with several wells and some soft data, and fuzzy methods can be integrated into Kriging, simulation and fractal interpolation. As a result, some new approaches are suggested and applied to the prediction of the formation property distribution in Shanlan reservoir, West China. Compared to the same methods without fuzziness involved, the techniques show some more robustness and reliability in characterizing spatial variability and structure of the properties and in reservoir characterization.