%0 Journal Article
%T A High Resolution Method for Fluid Prediction Based on Geostatistical Inversion
%A Zhen Yu
%A Jing He
%J International Journal of Geophysics
%D 2013
%I Hindawi Publishing Corporation
%R 10.1155/2013/845646
%X In order to predict the fluid in thin layer precisely, this paper proposed a high-resolution method for fluid prediction. The method used geostatistical inversion with lithology masks to calculate water saturation. We applied this method to theoretical model and real data. The result was compared with that of prestack AVA simultaneous inversion for fluid prediction. It showed that this method had high resolution both in vertical and lateral directions for fluid prediction and could also predict the fluid in thin layer efficiently. 1. Introduction Nowadays the main target of geological exploration has changed from a simple constructed oil and gas reservoir to complex structure reservoir and lithology reservoir. The need of precise exploration is increasing. The precise prediction for thin-layer fluid becomes an urgent problem for oil and gas exploration. Conventional fluid prediction is mainly based on AVO technology, including AVO analysis [1, 2], prestack AVA simultaneous inversion [3, 4], elastic impedance inversion [5, 6], and other methods. However, these methods do analysis mainly based on seismic data, and their vertical resolution is limited. So it cannot precisely predict the fluid in thin layer and cannot meet the needs of precise exploration. In this paper, in order to solve this issue, we presented a high-resolution method for fluid prediction. The method used geostatistical inversion with lithology masks to calculate water saturation. The result was compared with that of prestack AVA simultaneous inversion. 2. Method 2.1. Geostatistical Inversion with Lithology Masks Geostatistical inversion is based on geological information (including seismic, drilling, and logging). It applies random function theory and geostatistics (histogram analysis, variogram analysis, etc.) and combines with traditional seismic inversion technique. It can generate multiple optional inversion results with the same probability [7¨C9]. It includes stochastic simulation and stochastic inversion. Stochastic simulation includes sequential Gaussian simulation, sequential indicator simulation, and other simulation methods. The geostatistical inversion with lithology masks mentioned in this paper contained two parts: sequential Gaussian simulation with lithology masks and stochastic inversion. 2.1.1. Sequential Gaussian Simulation with Lithology Masks The basic principle of sequential Gaussian simulation is the application of Kriging algorithm for locally estimating the geological variables, which is using a set of values at grid points where it is known (it is generally the
%U http://www.hindawi.com/journals/ijge/2013/845646/