Graduate Seminar - 3/13, 12:00PM Duncan Hall 1064

Dr. Sebastien Matringe and Dr. David Castineira

QRI (Quantum Reservoir Impact)

"Big Opportunities for Data-Driven Models in the Oil and Gas Industry"

Hydrocarbon reservoirs typically have an abundance of both geological and production data that originates from very different sources (geological studies, seismic surveys, petrophysical analysis, production data, etc). The common approach in the industry over the last decades has been to integrate all this data into a 3D, full-physics reservoir simulation models that follow a set of prespecified governing equations. However, proper integration of all that field data into multimillion cell models is a complex and time-consuming mathematical process. In fact problems typically arise from algorithms quickly converging into local optima, which make many reservoir simulation models highly unreliable for deployment in real field operations.

Data-driven models is an emerging trend in the oil and gas industry where innovative technology such as signal processing, data mining and artificial intelligence offer alternatives to resolve the intrinsic problems that appear in managing a hydrocarbon reservoir. The idea behind data-driven models is to transform the abundant field data from a curse to a blessing. Furthermore data-driven approaches can provide valid quantitative solutions in a fraction of the time required to run classical reservoir simulations, allowing companies to quickly delineate optimum reservoir management strategies for the fields. Some of the mathematical and computational opportunities in building data-driven models will be discussed during this presentation.