An Improved Empirical Approach for Prediction of Formation Water Saturation and Free Water Level for Uni-modal Pore Systems

Abstract

Abstract Minimizing the uncertainty associated with the prediction of formation water saturation can be achieved through the integration of formation evaluation (log analysis), capillary pressure data, and pressure gradient information. To reduce the uncertainty, it is often necessary to determine porosity, permeability and water saturation using integrated methods. The magnitude of uncertainty reduction associated with predicted petrophysical parameters is a function of the inherent quality of the various data acquisition devices. Thus, reduced uncertainty in hydrocarbon pore volume can often be achieved by combining explicit probabilistic formation evaluation and a quantitative capillary pressure model. An empirical quantitative capillary pressure-water saturation model has been developed that relates the capillary pressure and height above the free water level, FWL to the wetting phase saturation. This relationship generalizes the correlation for all existing rock types that exhibit uni-modal pore geometries. The model constants are expressed as continuous functions of the interval speed that is defined as the square root of core measured permeability divided by porosity (√k/ϕ) data. The water saturation is calculated continuously as a function of the height above the FWL using the log predicted permeability-porosity ratios. In the absence of a clear fluid contact, the new model provides an excellent tool for describing the fluid contacts by iterating on saturation. The FWL is adjusted until a good agreement exists between a base water saturation and the predicted water saturation from the empirical capillary pressure relationship. This paper describes the new saturation model and its application for a siliclastic layered reservoir. The model parameters are tuned based on a statistically significant number of measured high-pressure mercury injection capillary pressure curves. Furthermore, the model parameters are calibrated to data from a key well. The results of the model are then used to predict formation water saturation in neighboring wells.