Petrophysical interpretation of logging-while-drilling borehole measurements in the presence of electrical anisotropy, mud-filtrate invasion, noise, and well deviation effects

Abstract

Petrophysical interpretation of logging-wile-drilling (LWD) borehole measurements in the presence of electrical anisotropy, mud-filtrate invasion, measurement noise, and well-deviation effects remains a challenge in formation evaluation. In high-angle and horizontal wells penetrating electrically anisotropic sandstones, phase and attenuation apparent resistivity logs exhibit different values that are larger than resistivity parallel to bedding planes. Consequently, traditional interpretation methods often overestimate hydrocarbon saturation when using the deep-sensing apparent resistivity logs in petrophysical calculations. We implement two approaches to estimate horizontal and vertical resistivities from apparent resistivity measurements. The first approach is based on the numerical simulation of LWD resistivity measurements under a wide range of conditions of electrical anisotropy and well inclination to derive analytical models. The relative error of the analytical approximations is less than 6%. In the presence of measurement noise and shoulder-bed effects, we implement a Bayesian method for well-log inversion and uncertainty quantification. The relative error of Bayesian inversion is less than 2%. Estimates of horizontal resistivity are then used for the assessment of fluid saturation. A challenging set of synthetic and field examples of high-angle and horizontal (HAHZ) wells penetrating electrically anisotropic formations is selected for examination and verification of the new methods. For radial lengths of invasion shorter than 20 cm (8 in), sensitivity analysis shows that the effects of conductive mud-filtrate invasion on deep-sensing resistivity logs are negligible. The deep-sensing phase (P40H) and attenuation (A40H) apparent resistivities are used to estimate formation horizontal and vertical resistivities. Estimates of water saturation using horizontal resistivity are in good agreement with saturation-height models. Compared to conventional interpretation methods, the new approach improved the estimation of hydrocarbon saturation by 10%.