Top.Mail.Ru

An algorithm from the Gubkin Russian State University will speed up gas pipeline diagnostics tenfold

Scientists from the I.M. Gubkin Russian State University of Oil and Gas have developed an algorithm that reduces the processing time of in-line diagnostics data of gas pipelines from several hours to minutes—by tens of times. Pilot operation of the system at the facilities of the Unified Gas Supply System is planned until the end of 2026, the press service of the National Research University reported. The method automates the decoding of data from in-line flaw detectors. It is based on a mathematical model and software algorithm that independently identifies welded joints on magnetograms and analyzes the data even in the presence of interference. This eliminates manual processing errors (up to 10% of data was previously entered with errors) and reduces labor costs.

Associate Professor of the Department of Thermodynamics and Heat Engines Konstantin Zhuchkov noted that the technology is a step towards creating a risk forecasting system for the entire pipeline infrastructure of the country. At the next stage, they plan to supplement the algorithm with neural network technologies: a diagnostic specialist will become the operator of a robotic complex that transmits reports in near real time.

Accelerating data processing optimizes the costs of maintaining gas pipelines with a length of over 200 thousand kilometers. In the future, scientists intend to develop machine learning methods to predict damage to UGSS facilities and develop recommendations for maintaining equipment performance. The project is being implemented at the UGSS System Research Center, created in 2020 to develop import-independent software for the gas industry.