Formation evaluation techniques are the key to understand subsurface rocks properties from well-logs, especially those drilled in hydrocarbon exploration wells. Knowledge of the parameters related to different types of rocks is traditionally used in forward determination of lithology, porosity and water saturation, which can be refined by calibrating the input models. In this work, we perform well-log interval linear inversion with respect to formation density to investigate mineral and fluid properties in a real dataset. The method is based on an overdetermined problem, which supposes a homogeneous distribution of petrophysical parameters through stratigraphic layers and is applied in conventional reservoir rocks from the Norne Field (offshore Norway). Bulk density, gamma-ray and neutron porosity logs are employed in a workflow that relies on layer-by-layer least-squares regressions to estimate matrix, shale and fluid apparent densities. In this process, shale volume and the total density porosity are calculated depth by depth from empirical equations feed by the input logs. Therewith, the introduced inversion scheme stands as an alternative approach for well data interpretation that focuses on computing the densities of the rock constituents instead of fixing these parameters to invert fractional volumes. Furthermore, the application in two wellbores resulted in geological consistent individual densities in most intervals, except for a gas-bearing zone observed in one of the boreholes, where porosity uncertainty caused anomalous variation in grain and fluid densities.
Read full abstract