Novosibirsk Researchers Bring Artificial Intelligence into Subsurface Exploration
Researchers at Novosibirsk State University (NSU) have developed LithoQB, or Litokub, a platform that combines neural network algorithms with traditional seismic data analysis methods to accelerate geological exploration and the development of digital deposit models. One of the most widely discussed potential applications is the search for lithium in underground brines.
For seismic surveys conducted at the surface or in a borehole, geologists generate a small explosion or mechanical vibration to send sound waves into the ground. By measuring how quickly the waves return and how they have changed, geologists can determine where rock layers meet, where faults and voids lie, and where dense rock formations begin. The resulting data are complex and can take specialists a long time to interpret. The new IT platform is designed to make that work easier.
One Platform, Multiple Tools
The platform brings together several intelligent tools that work in a single interface and automate some of the geologist's most time-consuming tasks. Its computer vision algorithms identify geological structures in seismic images, including faults, layer boundaries and distinctive rock bodies. But the software does more than recognize outlines: it attempts to isolate individual geological features from the full dataset, such as lenses or layers that could serve as reservoirs – rock formations capable of holding water, oil or natural gas.
The neural network can also perform seismic inversion. By analyzing how seismic waves are reflected, it reconstructs the physical properties of the rock itself, including its density and porosity, and what it may contain. Litokub also performs spectral analysis. Different geological features produce distinct responses at different frequencies: some stand out at lower frequencies, others at higher ones. Breaking a signal down into its frequency components can reveal subtle details of the subsurface structure.
The platform helps identify geological structures that may contain prospective brines, including faults, layer boundaries and other features needed to build a geological model. However, the presence of lithium still has to be confirmed through additional research, sample analysis and drilling. The system was originally designed primarily for oil and gas exploration, with lithium viewed as one potential area for expanding its applications. Some of its algorithms have already been tested on real-world data. The team has carried out service work on a site in southern Orenburg Region under a contract with a company belonging to the Gazprom Neft group.
Layers of Data
Alongside Litokub, several similar software products have been developed in Russia for geological exploration. They combine different data analysis methods and are intended to replace Western tools that are no longer available on the Russian market. NSU, working with oilfield services company BurServis, has launched Arkaim, software for interpreting seismic and well data. It is already being used to address operational challenges. Rosneft uses its RN-SEISM system to search for and identify rock formations within geological layers. A data interpretation process that once took geologists weeks of computing time now takes eight times less.
The Institute of Petroleum Geology and Geophysics of the Siberian Branch of the Russian Academy of Sciences has developed W-SEIS, a software suite for interpreting 2D and 3D seismic survey and drilling data. W-SEIS and other modules were later integrated into KRATON, an open platform being developed by a consortium that includes NSU. The aim is to let different software products communicate through an interoperability layer, while also allowing third-party developers to connect their own modules.
A Promising Project
Russia's geological exploration industry has long depended on foreign software suites. Last year, Oleg Kazanov, head of Rosnedra (the Federal Agency for Subsoil Use), identified software as one of the most difficult areas of import substitution, singling out systems for processing 2D and 3D seismic data and building three-dimensional geological models. Litokub is being developed to provide an intelligent tool for interpreting geodata. Over approximately five years, young NSU researchers have worked on seismic interpretation algorithms and published more than ten research papers. The project won the Student Startup competition and is now participating in the INDUSTRIX accelerator program run by Gazprom Neft.
Lithium is a strategic raw material for the battery industry and electric transportation. According to Russia's Ministry of Natural Resources and Environment, the country's balance reserves are estimated at approximately 3.5 million metric tons of lithium oxide. Rosnedra reports that 26 licenses have been issued for the study, exploration and production of lithium-bearing groundwater. Eastern Siberia, the North Caucasus and several oil- and gas-producing regions in the Volga and Ural areas are considered promising. Digital exploration tools could help companies assess these areas more accurately.
