Development of artificial neural network models to calculate the areal sweep efficiency for direct line, staggered line drive, five-spot, and nine-spot injection patterns

Abstract

The overall efficiency of the oil displacement processes, and the amount of oil produced after the primary recovery depend on areal (EA), vertical (EV), and displacement (ED) sweep efficiency. There are many correlations and graphical charts that are used to calculate EA, EV, and ED with variable degrees of deviation. This research focuses on calculating the areal sweep efficiency (EA) at and after water breakthrough for direct and staggered line drive patterns, also for five and nine-spot injection patterns by using artificial neural network models (ANN) and comparing them with the previous techniques. In addition to establishing a new ANN model for calculating EA for the nine–spot injection pattern, as there is no current model to calculate it. The results showed that the four proposed models achieve high coefficients of determination (R2) of (0.9995, 0.9996, 0.9996, and 0.997) and low root mean square error (RMSE) of (0.0038, 0.0023, 0.0026, and 0.0034) for a direct line drive, staggered line drive, five-spot, and nine-spot injection patterns respectively. These new models are presented to provide a more accurate calculation of the amount of oil produced by waterflooding.