Abstract
Hydraulic fracturing advancements have improved shale well recoveries as evidenced in the Technically Recoverable Resources plus cumulative production (referred to in this paper as Estimated Ultimate Recovery (EUR) of type wells), which directly results in the "learning curve" concept in 2018 PRMS. However, post-pandemic trends show operators drilled shorter lateral length wells and applied less proppant for better profitability, which is contrary to previous aggressive stimulation strategy.
This study analyzed over 10,000 Permian Basin shale wells from 80+ anonymous operators during 2005 to 2021. The wells at study were selected unbiasedly. We then applied innovative data analytics to map fracturing trends in addition to identifying geographic sweet spots. This discovered shift marked a significant fracturing strategy change influenced by economic motivations.
This study showcased technological progress in stimulation in the Permian Basin. The study incorporated a multivariate least squares regression model, which applied various weighting factors based on the theoretical derivations and experience from reserves engineers. The new model was developed in Java and presents a handy auto-forecasting tool. This tool achieved reasonable forecasts in 15 minutes for the entire 10,000 well dataset for EUR for individual wells. This technique essentially enables real-time updates for massive production forecasts and greatly improves the evaluation efficiency that condenses weeks of labor into minutes. Additionally, a sophisticated, Tableau-based multi-layer analytical tool was employed for visualization analysis, along with a tornado chart for deterministic sensitivity analysis, probit plot in probabilistic study and type well development. Bubble map in time domain was also employed to track operator's performance and operation strategy across time and geographic area. This model is practically useful for basin analysis and due diligence for third-party evaluators on the buying side. This tool can also incorporate formation tops to generate structure maps, effortlessly identifying sweet spots for enhanced analysis.
Post-pandemic, there was a noticeable drop in EUR even though operators continuously increased fracturing intensity in the same sweet spot areas. This suggests strategic adjustment is needed for improved profitability. If evaluators develop type wells based on the "learning curve" concept, which includes the wells before pandemic, it can potentially overstate reserves. The EUR per lateral length per proppant curve suggests that operators prioritized drilling best locations in early development and the later wells might be in less "sweet" spots or experience frac hits. However, owing to the scarcity of publicly accessible well data post-2022, further conclusions are currently constrained. We will continue to monitor and incorporate the latest data for future analyses.
Based on the successfully developed new data analytics model, this study also suggests the necessity to review the "learning curve" concept in the next PRMS release.