Seismic exploration-focused
multi-attribute machine learning
Oil & Gas Exploration
ENRGeo has developed an exploration screening tool that uses multi-attribute seismic data combined with unsupervised and supervised machine learning approaches.
Our approach takes seismic data and converts it into high dimensional space to create statistically relevant clusters that can be visualised, detected and ranked. Our current ML tools are created in 26-dimensional space and the results can be used as training data to build another algorithm that can be applied anywhere to find the same features the user is interested in. The approach works in all geological environments and is easily scalable.
The results are similar to a seismic inversion but require no supervision and only use routine attributes. Our ML runs currently take several hours to output the machine learning features and visualisations, compared to up to several months for a seismic inversion. The tool is currently being turned into a SaaS product for exploration screening, examining massive datasets in short order e.g. as required in competitive licensing round application.
