Abstract
One of the challenges of the petroleum industry is achieving maximum recovery from oil reservoirs. The natural energy of the reservoir, primary recoveries in most cases do not exceed 20%. To improve recovery, secondary recovery techniques are employed. With secondary recovery techniques such as waterflooding, an incremental recovery ranging from 15 to 25% can be achieved. Several theories and methods have been developed for predicting waterflood performance.
The Dykstra-Parson technique stands as the most widely used of these methods. The authors developed a discrete, analytical solution from which the vertical coverage, water-oil ratio, cumulative oil produced, cumulative water produced and injected, and the time required for injection was determined. Reznik et al extended the work of Dykstra and Parson to include exact, analytical, continuous solutions, with explicit solutions for time, constant injection pressure, and constant overall injection rate conditions, property time, real or process time, with the assumption of piston-like displacement.
This work presents a computer implementation to compare the results of the Dykstra and Parson method, and the Reznik et al extension. A user-friendly graphical user interface executable application has been developed for both methods using Python 3. The application provides an interactive GUI output for graphs and tables with the python matplotlib module, and Pandastable. The GUI was built with Tkinter and converted to an executable desktop application using Pyinstaller and the Nullsoft Scriptable Install System, to serve as a hands-on tool for petroleum engineers and the industry.
The results of the program for both methods gave a close match with that obtained from the simulation performed with Flow (Open Porous Media). The results provided more insight into the underlying principles and applications of the methods.
Efficient Algorithms for Identifying Loop Formation and Computing θ Value for Solving Minimum Cost Flow Network Problems
