US R&D establishment Battelle has customized its ‘EluciData’ analytical engine for use in the oil and gas vertical. EluciData’s analytical and machine learning methods were originally developed for data-heavy applications such as national security and healthcare. The development combines geoscience with advanced statistics to develop predictive models for optimizing well design. One approach is to identify statistically the factors that separate good wells from poor performers and to relate production rates to geologic formations.
The big data toolset includes data mining, clustering, machine learning and experimental design. Inputs can be structured and unstructured data. Battelle has also adapted its computer solutions for oil and gas equipment monitoring applications, replacing manual inspections with a drill site camera. Natural language processing is also in the mix.
Under the hood there is a Hadoop/NoSQL data environment with extensions for business intelligence tools such as Spotfire. Battelle is also offering training and consulting in data science to oil and gas clients. If you want to learn more about decision trees, random forests, Bayesian and Markov chain models then talk to Battelle.
© Oil IT Journal - all rights reserved.