Modeling of Capillary Pressure Behavior Using Standard Open Hole Wireline Log Data: Demonstrated on Carbonates from the Middle East

Abstract

Abstract A new technique was developed to model capillary pressure behavior from wireline log data and applied to carbonate reservoir rock from a Saudi Aramco field. The method utilizes image analysis of petrographic thin sections, capillary pressure measurements, and neural network analysis of standard open hole wireline log data. Twenty capillary pressure curves and their associated pore type proportions (identified in thin section) arc the basis for the capillary pressure predictive model for the reservoir interval under study. Neural network analysis of the wireline log data was used to continuously predict pore type proportions downhole. The neural network-derived pore proportions were then applied in constructing wireline log-based capillary pressure curves using the capillary pressure predictive model. This method provides an accurate means of determining capillary pressure behavior from wireline log data and extends the applicability of the limited number of available capillary pressure curves. Once trained, the neural network may be applied to other wells in the field as long as the training set (both rock samples and wireline log types) is representative within the study area. The capillary pressure curves predicted from wireline log data can be used for the same purposes as capillary pressure curves measured on core samples, such as determining water saturation in intervals above and within the transition zone.