POSSIBILITIES OF GEOLOGICAL-GEOPHYSICAL DATA INTERPRETATION BY TAUGHT NEURAL NETWORKS

D. O. Gafurov 1, O. M. Gafurov 2, V. A. Kontorovich 3

1 Krasnoyarsk Research Institute of Geology and Mineral Resources, Krasnoyarsk, Russia; 2 InformGeoServis, Tomsk, Russia; 3 Trofimuk Institute of Petroleum Geology and Geophysics SB RAS, Novosibirsk, Russia

The study provides procedures for data interpretation of well log survey and seismic evidence from the Talakanskoye field by the neural network analysis. The Osinsky reservoir properties and structure have been predicted. Geological-geophysical data were integrated by means of the mathematical apparatus of neural networks realised in the NeiroInformGeo intellectual geographic information system developed in the OOO InformGeoServis. The InformGeoServis developed methods of geological section prediction and the Osinsky reservoir quality assessment. The developed methods (the Talakanskoye field) were adapted by the OOO InformGeoServis together with ZAO Krasnoyarskgeofizika. During a period from 2004 to 2016, the technique was further developed as applied to various petroleum regions of the Krasnoyarsk Territory, and in 2011 the RF patent and certificates were received.

Keywords: neural networks, data integration, well log survey, Fletcher-Reeves conjugate-gradient method, dynamic attributes of seismic wavefield.

DOI 10.20403/2078-0575-2017-3-85-94

Download PDF