Rock Physics and Reservoir Inference Study from Cretaceous Sandstones from Espirito Santos Basin, Brazil
This work illustrates a methodology for sandstone inference in the Upper Cretaceous of the Espirito Santos Basin, Brazil. The goal of this methodology is to produce an indicator of reservoir in a 3D seismic volume. To create this indicator, we consider porosity, clay volume and saturation.
The degree of confidence that we have regarding these reservoir properties depends on the current state of information. The sources of information are prestack seismic data, petrophysical and geological information from well data and rock physics models (empirical and theoretical). The seismic data has been pre-processed in an amplitude-preserved fashion.
Following a Bayesian approach, these data sets are used to construct local probabilities representing each piece of information and associated degree of confidence, regarding the reservoir parameters. These probabilities are then combined by application of the Bayes theorem. As a result, we obtain a collection of posterior probability density functions distribution covering each cell of the reservoir. The desired parameters can be computed from the distributions (e.g., estimates and associated uncertainty) and presented as reservoir inference cubes for further analysis. In particular, it can be used to delineate reservoir conditions for a risk analysis workflow associated with time-lapse and reservoir development projects.