Using neural network to predict relative permeability andcapillary pressure diagrams

Determining the fundamental properties of oil reservoir rocks and the parameters governing fluid flowthrough porous media is a complex aspect of reservoir characterization. Among the most critical reservoirrock properties are relative permeability of gas-oil and water-oil systems, which serve as essential inputs forreservoir simulation models. Traditionally, laboratory core analysis is considered the gold standard forobtaining accurate information about reservoir rocks. However, the high costs and time-consuming nature ofcore analysis, coupled with the limited availability of core samples, especially in the early stages of fielddevelopment, necessitate the exploration of alternative techniques for predicting these essential properties.This project aims to develop a methodology to reduce uncertainty in relative permeability and capillarypressure, utilizing neural networks to predict the corresponding parameter curves

کلیدواژه ها:

Prediction ، neural network ، relative permeability ، capillary pressure