Abstract

One of the most substantial challenges currently faced in many unconventional plays is linking subsurface fundamentals and empirical results/measured data, e.g., type curves, to predictive production. One way to achieve this link between subsurface fundamentals and well production for optimal completion techniques is through multivariate analysis. Roth (2013) showed this method has been used for sweet spot mapping, but in field development of unconventionals today, the next best well that meets the economics will be drilled and completed regardless of whether it is in the sweet spot or not. The more powerful use of this approach, therefore, is marrying the rock to the completion to optimize well performance.

In unconventional reservoirs, production is driven by the complex interplay of engineering and geology. There are no bi-variate crossplots of engineering, geological, or geophysical variables that are capable of adequately describing the deliverability of the reservoir. Numerous engineering and geologic attributes must be considered simultaneously using multivariate techniques, such as multiple regression analysis, in order to properly model production. To understand key play driver relationships, well performance must be normalized with engineering data in order to isolate the impact of geologic attributes on well performance.

In this study, a well performance predictive model was created in the Utica from identified key play drivers that combined engineering, geologic, and geophysical parameters to test different completion designs across areas of varying geology. Using this predictive model, which was supported by independent data from current analytical modeling and competitor data, the completion design was changed for a well pad drilled and completed in early 2017. The model predicted the (inline equation) of this pad to be approximately 18, which was the upper limits of performance input into the model. Initial production and tubing pressure show this pad exceeded our expectations with an (inline equation) of 23, which is the highest in the field.

This content is only available via PDF.
You can access this article if you purchase or spend a download.