Influence of Geological, Completion, and Well-Spacing Parameters on First-Year Productivity in the Delaware Basin

The presented work focuses on the Permian Delaware Basin with the objective of determining how completion practices, geology, and well spacing influence wells’ first-year productivity (stb/1000ft). We demonstrate the results of a machine learning algorithm for predicting wells’ first-year productivity as a function of the completion, subsurface, and well spacing parameters within a pad. Finally, we provide the results of the model explainer - SHapley Additive exPlanations (SHAP). We extracted subsurface parameters from an in-house 3D geo-cellular model to obtain the values of those parameters at individual well locations. We combined these with individual well production and completion data obtained from S&P Global. The combined data set was used to conduct an analysis aimed at determining well spacing within a pad and its evolution over time. After conducting the analysis, we could explain 39% of the variance in the test dataset. Using SHAP, we can show that the top 5 variables that explain first-year productivity are lateral length, proppant use intensity, midpoint true vertical depth, average spacing within the pad in the vertical direction, and hydraulic fracturing water use intensity. Analyzing how first-year productivity is influenced by the combination of subsurface, completion, and distance parameters within pads is necessary to analyze the economic viability of production from pads apart from energy prices and other economic variables.