Main fields of the Republic of Tatarstan in Russia entered late stage of development. In recent years more than 60 methods on enhanced oil recovery and stimulation of producing wells have undergone comprehensive tests in a wide range of geologic conditions. Based on geologic information and experience of fields development there have been created geologic and hydrodynamic models for analysis and forecasting of technological indices of oil production, as well as data bases on application of methods for enhanced oil recovery and stimulation of wells.
Based on artificial intelligence, procedure is developed for selection of enhanced oil recovery method using model of residual reserves for different structures of development. Models apply neural network, capable to synthesize recommendations based on training by data base for realization of more than 60 methods on enhanced oil recovery and wells stimulation at similar structures of Tatarstan fields. These technologies are characterized by a set of criteria, including data on reservoir, formation fluids, operating conditions of wells, information on peculiarities of the technologies application. Models are trained on more than 2200 examples, reflecting correct conditions of realization of EOR and stimulation technologies, using considered technologies. During testing more than 70% of data sets are confidently recognized, which correspond to criteria of application of enhanced oil recovery methods, realized through injection wells, and more than 90% of methods on wells stimulation.
The developed procedure allows to forecast and rank the application of different methods for enhanced oil recovery and stimulation of wells at specific producing and injection wells in real-time operation mode. This procedure is used in conditions of fields development in the Republic of Tatarstan.
The conducted work allows to combine two modern approaches - modeling of field development and application of artificial intelligence for selection of most effective technologies on enhanced oil recovery and stimulation of wells. Data base has been developed for methods of enhanced oil recovery and stimulation of wells for conditions of Tatarstan fields.
Introduction
While development of oil fields in Tatarstan various improved oil recovery technologies involving most of scientific achievements known to date are tested and introduced. Choice of enhanced oil recovery technologies depends on formation structure, reservoir characteristics, saturating fluids, capabilities of water flooding system, and other technological requirements.
It is reported about some studies on application of analytical technologies based on artificial intelligence intended for proper choice and optimization of the process of oil field development, carried out by AMEC Process & Energy Ltd and University of Aberdeen [1], Escuela de Minas de Madrid [2], Kappa Engineering [3], Ecopetrol-ICP and I. D. A. Ltd [4], etc [5–7]. The authors' investigations [8–12], NIPER/BDM-Oklahoma, Inc. and Rhonda DOE-Bartlesville Project Office [13], New Mexico Petroleum Recovery Research Center [14], SWORD (PETEC Software & Services AS) are aimed at the problem of application of the optimum EOR method involving steam flooding, injection of CO2, polymers, acid, and other agents. However, direct application of the investigations' results in conditions of Tatarstan is impossible due to the reasons stated below.