Streamlining petrophysics workflows with machine learning

The oil and gas industry is not short of data, in the form of wells, seismic and other geophysical information.   However, the industry is notoriously poor at utilizing this information.  The complexity of workflows required to take raw information that is available in public or proprietary data stores and turn this into decision ready information on sub-surface geology and properties means that such workflows are time consuming, so that often only a fraction of available information is used.  Making better use of information, using modern data analytics techniques, and presenting this information in a way that is immediately useful to geologists and decision makers has the potential to dramatically reduce time to decision and the quality of the decision that is made. Abacus and EPCC are proposing to work on one aspect of this problem:  streamlining petrophysics and rock physics workflows.    A typical rock physics atlas developed by Abacus may contain 100-200 wells, and each well takes on average 7 days for an experienced petrophysicist to prepare.  However including more wells in the atlas will make them ever more valuable:  rather than searching 100 wells on a margin, if we can search 1000 or 10,000 the richness and diversity of information that is available is dramatically improved.  To do this we need to decrease the time taken to process a raw well, ideally by an order of magnitude.