Gazprom Neft develops machine learning technologies

Gazprom Neft, in conjunction with the engineering center of MIPT, is developing machine learning technologies: new algorithms that will improve the quality of operating data received from wells. Now the quality of this data sometimes raises questions, for example, there is no information for certain time intervals, some measurements do not correspond to the physical model, etc.
This can occur due to equipment malfunctions and because of the “human factor”. New technologies will avoid these errors.

Data Science methods make it possible to process huge data sets, to reveal new patterns and based on them to make forecasts. Already there are algorithms for searching for incorrect values ​​and restoring missing data, determining the processes of mutual influence of wells on each other, and others.

“Big Data-paradigm allows you to adapt business strategies for such” exploding “data growth rates. At the same time, modern methods of data analysis, such as machine learning, convolutional neural networks, allow to give a fundamentally new approach to solving actual problems, “the press service of the company quotes the Director General of the MIPT Engineering Center Timur Tavberidze.