Standard production forecast techniques for unconventional asset development rely mostly on field data, which could suffer from limitation in both data quality and quantity. It is also a challenge to interpret subsurface dynamics directly from field observations. In this immersive session, we explore a workflow that can enable decision making for unconventional development through organic combination of field and simulation data. It provides physics based well performance forecasting, interpretation, and uncertainty estimation—all without running additional reservoir simulation during inference.
This session will include discussion about use of physics based simulation models, exploratory data analysis, as well as aspects of training and interpretation for machine learning models.
Associate Professor in Harold Vance Department of Petroleum Engineering
Texas A&M University