USING WELL LOG DATA TO DETERMINE OIL RESERVOIR HETEROGENEITY

The description of reservoir characteristics can be either qualitative or quantitative. Reservoir characteristics include permeability and porosity, thickness, saturation, faults and fractures, rock facies distribution, etc. All reservoirs are characterized by a sum of matrix and fluids properties. They are evaluated by a complex investigation, consisting of core sample analysis, geological, geophysical and hydrodynamic investigation, as well as production data. These properties may be constant for the whole field when the reservoir is a homogeneous one, or they may be variable if the reservoir is heterogeneous. There are several statistical methods for determining reservoir heterogeneity: the coefficient of variation, the Dykstra-Parsons coefficient, the Lorenz coefficient, the Gelhar-Axness coefficient, etc. This work is focused on one of the statistical methods Lorenz coefficient method. Lorenz coefficient calculation involves the knowledge of porosity, permeability and thickness of the collectors. For an appropriate statistical calculation, the number of values for each parameter must be large enough and the parameter distribution has to be uniform across the field. It has to be emphasized that, in some cases, the available number of permeability data for a reservoir is not large enough for a statistical analysis. In such situations, the data obtained by well logging are very useful, being more representative and easy to get. This paper presents a comparative analysis of the Lorenz coefficient values obtained from well logging data and from rock samples data. For a more complete approach, the covariance and variogram have also been calculated. The technical goal of this paper is to demonstrate that reservoir properties obtained from well logging data are very reliable and useful for reservoir characterization.