Future Workforce Education Through Big-Data Analysis for Drilling Optimization

Abstract

This article, written by Special Publications Editor Adam Wilson, contains highlights of paper SPE/IADC 184739, “Future Workforce Education Through Big-Data Analysis for Drilling Optimization,” by Y. Zhou, SPE, T. Baumgartner, G. Saini, SPE, P. Ashok, SPE, and E. van Oort, SPE, The University of Texas at Austin; M.R. Isbell, SPE, Hess Corporation; and D.K. Trichel, formerly of Hess Corporation, prepared for the 2017 SPE/IADC Drilling Conference and Exhibition, The Hague, The Netherlands, 14–16 March. The paper has not been peer reviewed. An operator partnered with the drilling-automation research group at The University of Texas at Austin to develop a work flow for big-data analysis and visualization. The objectives were to maximize the value derived from data, establish an analysis toolkit, and train students on data analytics. The operator provided data sets, business and technical objectives, and guidance for the project, while a multidisciplinary group of undergraduate and graduate students piloted an analysis work flow. Objectives The project stakeholders agreed on three main objectives. The foremost objective was to maximize the value from the tens of gigabytes of data gathered during drilling operations. Several work streams were selected to help identify key drilling-performance limiters and cost-saving opportunities. These work streams include assessment of the bottomhole-assembly (BHA) and directional-drilling performance by using measures such as wellbore tortuosity and time-based vibration data to create meaningful visualizations and implement standardized data structures. The second objective was to establish a standardized data-analysis toolkit. The steps toward such a toolkit were to identify, streamline, and document the working process to establish work flows and to build software tools that automate these work flows (e.g., perform analysis or visualization of the data). The third objective was to help educate undergraduate students and equip them with the skills necessary to tackle problem in a big-data world. Data Sets The data available for this project covered four pads in the Bakken Formation comprising 16 wells that were drilled from 2014 to 2015 for a total of 256 active rig days. The data were the product of a drilling-automation pilot project previously published by the operator. The data set for each well was comprehensive and included well-planning reports, geology information, surface-sensor data, directional surveys, daily drilling reports (DDRs), and extensive measurement-/logging-while-drilling and other downhole data. Deliverables Three main deliverables formed the basis for the data-analysis toolkit (Fig. 1). Data curation—This involved ensuring data quality and organizing data into a more-cohesive, -consistent, and -accessible format. Effective analytical approaches could be applied only after the data were classified and cleaned. Data visualization—This involved analyzing and visualizing key information in the form of interactive and informative graphics. Storyboard—This involved developing a work flow that guides the user through different visualizations, from an overview to a highly detailed level, in order to illustrate and quantify key elements of the drilling performance related to the drilling systems and well-delivery requirements.