Upscaling of Borehole Resistivity Image Logs in a Complex Fractured Carbonate Reservoir: Improving Drilling Location Selection

Manfred Frass¹, Efrain Mendez², and Sean Boerner^3
1 Veritas VHR, Villahermosa, Mexico
² Pemex E&P, Villahermosa, Mexico
³ Veritas VES, Houston, TX

Fracture intensity curves, obtained from borehole resistivity image logs are guided by 3D seismic inversion data (post stack and pre-stack) and other seismic attributes, to generate a fracture intensity volume, in a complex carbonate reservoir. Three different types of features are interpreted from the borehole image logs: Conductive fractures, Resistive fractures and Bed Boundaries. The aperture of the conductive fractures, the spacing and the resulting fracture intensity, is calculated for each fracture. Seismic attributes include P-wave velocity (Vp), Shear wave velocity (Vs), density, and mu-rho models from stochastic simulations, a number of other geometric attributes calculated from the Rp seismic volume and fracture density and envelope, are calculated from an azimuthal AVO (VAZ) process. All these attributes are loaded into a neural network system. Three wells were used to generate models using stepwise multi-linear regression. During the stepwise multiple regression, the system attempts to use the data from two wells to predict the third. The production behavior was integrated into the model looking for correlations of the fracture intensity and the production log data, of the same well. A fracture intensity 3D model is generated using the selected geometric and stochastic inversion attributes. This model was used to guide well trajectories for new infield and exploration drilling.