Experts from RN-KrasnoyarskNIPIneft (part of Rosneft’s research and development unit) have developed a software algorithm based on a neural network that reduces the time of one of the first stages of seismic data processing by 10 times. The press service of Rosneft reported that the software algorithm is presented as a standalone application for a personal computer (PC).
The volume of a project in which an expert works to interpret seismic information is 300 km2, or about 30,000 seismograms. The “manual” processing of such a large volume of data can take up to 80 working hours, while the created algorithm allows you to perform a similar work in 7 hours without the direct participation of a specialist.
The main parameters for assessing the effectiveness of the developed solution are not only a significant reduction in labor costs, but also the exclusion of the factor of subjective assessment of the translator, since the work of the program is based on mathematical models. on visual perception of information.
The created prototype of the program was successfully tested on real seismic data obtained at two licensed sites of Rosneft in Eastern Siberia.