Study Looks at Well Construction as Engineering With Data

_ This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 217665, “Well Construction in the Era of Big Data: It’s Not Data Analytics, It’s Engineering With Data,” by Gregory S. Payette, SPE, ExxonMobil; Jeffrey R. Bailey, SPE, consultant; and Paul E. Pastusek, SPE, ExxonMobil, et al. The paper has not been peer reviewed. _ The authors write that data science captures value in well construction when data-analysis methods such as machine learning are underpinned by first principles derived from physics and engineering and supported by deep domain expertise. The authors provide a summary of important data pertaining to the well-construction process and discuss high-level areas where data science can add value to well construction through analysis of such data. Introduction The authors write that it is important for organizations to grow data science and physics-based modeling solutions together; well-construction data science should be what they term an “engineering with data” endeavor. A large portion of the complete paper is dedicated to discussing ways that organizations can improve their abilities to derive value from data-science efforts. Most discussion focuses on steps that data science teams can take today. However, the provided commentary on data management and governance is forward-looking. The paper concludes by examining several historical data-science case studies for well construction. Each example highlights the need to connect data and some physical or engineering process to deliver value through data science. Improving the Effect of Data Science in Well Construction Data Management and Governance. A step that can be taken to enable better holistic well-construction data management is for data collectors and creators to organize their data using data models that center around what the authors term “data entities” that are common and stable across data collectors. Common well-construction data entities include items such as Wells, Wellbores, Hole Sections, Surveys, Casing Strings, Drillstrings and Bottomhole Assemblies (BHAs), Tubing Strings, and Downhole Completion Equipment. These data entities are capitalized in this paper to denote their roles as data entities. The industry is accustomed to organizing data around entities. Two common types of databases particularly relevant to data-science efforts commonly used by operators include the following: - Well Information Management Databases (WIMD)—These databases are used to document information about the life cycle of wells. - Wellsite Information Transfer Standard Markup Language (WITSML) Databases: WITSML databases are primarily used to store time- and depth-series sensor data collected during drilling and completion activities. For the purposes of their paper, the authors do not include other data entities that can be captured through WITSML data repositories. Even though WIMD and WITSML databases share common data entities, such entities typically are not linked to one another through formal relationships at the original database levels. This creates a barrier for data-science projects that require data from both source systems.