Cognitive methodology for forecasting oil and gas industry using pattern-based neural information technologies

2nd International Conference "Cognitive Robotics"

工业技术;无线电电子学;计算机科学

Gafurov, O.^1 ; Gafurov, D.^2 ; Syryamkin, V.^1

National Research Tomsk State University, Tomsk, Russia^1

Krasnoyarsk Research Institute of Geology and Raw Materials, Krasnoyarsk, Russia^2

关键词: Classical methods;    Data base theory;    Hydrocarbon fields;    International practice;    Mathematical analysis;    Oil and Gas Industry;    Production wells;    Russian federation;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/363/1/012008/pdf
DOI  :  10.1088/1757-899X/363/1/012008

来源: IOP

【 摘 要 】

The paper analyses a field of computer science formed at the intersection of such areas of natural science as artificial intelligence, mathematical statistics, and database theory, which is referred to as »Data Mining» (discovery of knowledge in data). The theory of neural networks is applied along with classical methods of mathematical analysis and numerical simulation. The paper describes the technique protected by the patent of the Russian Federation for the invention "A Method for Determining Location of Production Wells during the Development of Hydrocarbon Fields" [1-3] and implemented using the geoinformation system NeuroInformGeo. There are no analogues in domestic and international practice. The paper gives an example of comparing the forecast of the oil reservoir quality made by the geophysicist interpreter using standard methods and the forecast of the oil reservoir quality made using this technology. The technical result achieved shows the increase of efficiency, effectiveness, and ecological compatibility of development of mineral deposits and discovery of a new oil deposit.