Experts at RN-KrasnoyarskNIPIneft, part of Rosneft’s research and development division, have created a neural network based software algorithm. Its mission is to dramatically accelerate one of the early stages of seismic data processing by a factor of ten. Rosneft’s press service notes that the algorithm is packaged as a standalone PC application, designed for convenient use without the need for constant specialist supervision.
In a typical project where a specialist interprets seismic information, the area in question can span about 300 square kilometers and include roughly 30,000 seismograms. Traditional manual processing of this data volume can require as many as 80 working hours. By contrast, the new algorithm completes the same workload in about seven hours, performing the task with minimal direct human intervention.
Beyond the headline time savings, the solution brings another important advantage. It reduces labor costs while removing the potential for subjective judgement in interpretation that can arise from human visual perception. The program operates on solid mathematical models, delivering consistent results that are not influenced by individual perception or fatigue.
The program’s prototype has already undergone testing on actual seismic data gathered from two Rosneft licensed sites in Eastern Siberia, demonstrating its capability to work with real-world datasets and conditions.
Looking ahead, the technology holds promise for broader adoption in similar geophysical tasks, including operations in North American basins. By offering a scalable, computer-driven approach, this AI-based tool can support independent analysts and teams in delivering faster, more reproducible seismic interpretations while maintaining rigorous standards for accuracy and reliability.