Multivariate Analysis to Identify Variables Controlling Production and Quantify Their Economic Impact - Midland Basin, Texas
Objective/Scope: The Wolfcamp formation in the Midland Basin of Texas has long been a hotbed of activity, with a recent dramatic increase in horizontal well development. Operators are now focused on optimizing their development efforts, but the number of variables that control production can make the problem seem intractable. This study demonstrates the use of multivariate non-linear regression techniques to identify and rank the variables controlling production and presents the economic impact of design changes. Methods/Procedures/Process: It is widely acknowledged that multiple variables control production in unconventional plays. In this multivariate analysis, we include both geological variables such as porosity, resistivity, thickness, and source rock maturity, as well as engineering variables, such as lateral length, proppant, fluid, wellbore orientation, and well spacing. Further, the model also allows for quantification of each variable’s relative contribution to this production. We then use this predictive model to estimate the future productivity and economic returns of various optimized drilling and completion scenarios. Results/Observations/Conclusions: There are many seemingly logical ways to increase unconventional production from horizontal wells. Theoretically, longer laterals, increased proppant loading, and increased proppant concentration should lead to improved performance. Similarly, wells in reservoir rock with higher porosity or pressure should outperform. However, in the study area, our model indicates that while some of these factors have the expected impact, others have surprisingly little. Further, the combinations of variables that do optimize production may or may not result in improved economic returns. Indeed, in some cases, we find that the best returns can be realized from wells with moderate – but not the best – production, because of their low cost. Complicating matters more, depending on the metric to be maximized – rate of return, return on investment, net present value, or even estimated ultimate reserves – the optimal development design varies. Novel/Additive Information: This work is notable in that it moves us from using educated guesses to optimize development, to the use of a data-driven mathematical model that allows for quantification of our design change impacts. However, its most valuable conclusion may be even simpler: the days of bivariate analysis are over. The multivariate nature of unconventional (or complex conventional) reservoir development demands multivariate tools to optimize.
AAPG Datapages/Search and Discovery Article #90335 © 2018 AAPG 47th Annual AAPG-SPE Eastern Section Joint Meeting, Pittsburgh, Pennsylvania, October 7-11, 2018