GSOC 2022 November Technical Talk

 Amanda Trumbo

Unsupervised Classification Techniques

to Update Type Curves

BIO

Amanda Trumbo is a Senior Staff Geophysicist for Chesapeake Energy located in Oklahoma City, OK.  She earned a Bachelor’s in Mathematics and a Master’s in Geophysics, both from the University of Oklahoma.  She started her career as a geoscience technician for Devon Energy Corporation in Oklahoma City.  Upon completion of her master’s degree, she continued at Devon as a geophysicist working the Barnett Shale and Permian Basin.  After five years at Devon, she joined Transform Software and Services and a short while later, Chesapeake Energy.  During her current nine-year tenure at Chesapeake, she spent the first five years as a member of the reservoir characterization team in Geophysics Technology and currently works in the Marcellus geoscience team.  Amanda has experience spanning multiple, onshore US basins and has been published in many industry periodicals.

ABSTRACT

One key challenge of mature, unconventional plays is synthesizing large amounts of data at multiple scales (well stage data, logs, maps, and seismic).  Machine learning and multivariate techniques are becoming more commonly used to aid in interpretation and optimize development.  While there are numerous ways to utilize these techniques, this talk will focus on two unsupervised learning techniques, available in commercial softwares, and their application to identify and prioritize key subsurface play driver relationships.  The outputs will highlight how to streamline combining geologic and geophysical mapping with reservoir and completions analysis to inform and update type curves in the Marcellus Shale, NE Pennsylvania.