Water flooding performance prediction by multi-layer capacitance-resistive models combined with the ensemble Kalman filter

In this study, a system of multi-layer capacitance resistive models (MLCRMs) is established for water flooding performance prediction in a layered reservoir. Three different types of observation data are considered for a layered reservoir: (1) both injection and production rates of each layer; (2) only production rates of each layer; and (3) only injection rates of each layer. The cross flow among layers and the bottom-hole pressure are considered in the models. A modified power law model is used to describe the ratio of water to oil production. The ensemble Kalman filter (EnKF) method is applied to estimate the connectivity coefficients for each layer and the well index numbers in the MLCRMs. Synthetic examples of a layered reservoir with different types of observation data are performed to validate the proposed models. The results show a good match of historical data and a reasonable prediction for future oil and liquid production.