Neural network helps track earthquakes

The technology is targeted at areas of intense underground resource extraction.

Petroleum and gas extraction is one of the areas of industrial activity capable of provoking seismic activity. Micro-earthquakes can be caused by various processes, such as changes in porous pressure, geochemical reactions, temperature processes, or reactivation of preexisting fissures.

Nowadays, detection of seismic events is greatly helped by specialized algorithms. Vladislav Sudakov, Deputy Director of the Institute of Geology and Petroleum Technologies, comments, “Our current objective is to work out techniques of detection of natural seismic waves and program algorithms of automatic detection of seismic events near petroleum and gas deposits.”

There are some hindrances in the process. For instance, many important signals have very low amplitudes and cannot be filtered out of interference; moreover, not all signals mean real danger because they are too minor. However, small-scale fluctuations can give information about local seismic activity.

The algorithm assists in filtering signals, clearing them from technological noise, and in detecting signals with low signal-to-noise ratio.

Source text: Adelya Shemelova, Media Center

Translation: Yury Nurmeev