Teaching old data new tricksPeter Keller, General Manager at AGGS. Photo: Halfdan Carstens

Teaching old data new tricks

Old, publicly available seismic data from the NCS is given new life via the use of artificial intelligence and machine learning. A research project has shown that new insights in the petroleum system can be found with already existing data.

As the oil and gas industry moves away from the old, “high risk, high reward” mindset towards a new reality with generally lower investment appetite and tighter budgets, geoscientists are increasingly taking advantage of the power of computers to increase the exploration success rate while keeping costs low.

On the Norwegian Continental Shelf (NCS), vast amounts of public domain vintage seismic 3D data represent a genuine data treasure trove that can be used in big-data analysis to bring forward new understandings.

At the upcoming digital DIGEX conference, Peter Keller, General Manager at AGGS, will present about a consortium project (AGGS, PSS-Geo and Cama Geosciences, sponsored by Lime Petroleum) that is developing methods to predict lithology and fluids from seismic and well data using artificial intelligence and machine learning.

The ongoing research has so far proved successful, and at the conference, Keller will present examples for all investigated cases, including first results from 2D seismic sections and a small 3D test-cube.