Abstract (download available at SPE ONEPETRO WEBSITE)
This paper reviews several Big Data analytical initiatives in the Marcellus Shale. We describe how application of Big Data technology evolved, share challenges and benefits derived from Big Data analytical processes, and discuss lessons learned. We present an overview of Big Data methods employed, show how we integrated results with economic analyses to guide field development, and summarize the significant impact on development economics.
This paper will help operators, analysts, and investors “de-mystify” Big Data technology, and provide insights and guidance to those embarking on Big Data initiatives. We discuss an ongoing initiative that employs cognitive analytics to generate production type curves via machine learning and couples the results with integrated economic analyses to guide field development. Challenges associated with data management, such as automated data QA/QC, sparse datasets, interpolation/extrapolation, model training and evaluation are discussed. Benefits derived from integrating Big Data-generated type curves with economics analyses to guide well/field optimization are also presented.
Our past big data experiences have taught us several important lessons. First, Big Data initiatives are journeys, not destinations, so expect to constantly feel like there is more to learn and do. Nonetheless, implementation of Big Data processes along the journey can add significant value to an asset, as demonstrated in this paper. Second, it is critical to clearly define the problem to be solved; without a crystal-clear mission statement, scope creep is inevitable, because Big Data technology is capable of so much. Finally, partnering with someone that has experience solving similar problems can significantly accelerate the process and add value.
Using Machine Learning to generate forecasts allows the engineers to focus their efforts on increasing business value, rather than managing and manipulating data. In the end, we will demonstrate how a process that once took multiple man-weeks of effort was solved within a single man-day of time. Finally, we present an example of an optimization opportunity identified with the potential to yield approximately 15 Bcfe in additional cumulative production, while maximizing future drilling inventory in the Marcellus Shale. (Note – this is presented as a “theoretical” example in the body of the paper.)