[URTeC 2021 Paper] Are unconventional well performance gains exhausted?

Are unconventional well performance gains exhausted? Investigating the drivers of year-over-year production improvements across the major US unconventional plays using machine learning

Talk Details::
- Tuesday, July 27th at 9:45 AM | Exhibit Hall – Station A
- Theme 7: Data-Driven Forecasting & Combining Physics and Machine-Learning Methods
- AUTHORS:: T. Cross, J. Chaplin, K. Sathaye, A. Cui (Novi Labs)
Changing production drivers through time in the Eagle Ford, Bakken, Midland, and Delaware. Completions and geologic sweet spotting have stopped making positive contributions since 2017, though operators still gain performance improvements through longer laterals

Abstract::

Objectives/Scope: Unconventional operators produced tremendous performance improvements over the last decade, with per-well oil production more than doubling in the Bakken and tripling in the Permian. This dramatic productivity improvement has been the result of some combination of longer laterals, more intense completions, and focus on higher-quality acreage and landing zones. Understanding the relative contribution of each of these factors will help answer whether productivity will continue to improve, stagnate, or even regress. To investigate this topic, we built machine learning models in the Williston, Eagle Ford, Midland, and Delaware Basins. We then use these models to analyze well performance and the contributions from subsurface rock quality and well designs across each basin.

Methods/Procedures/Process: We aggregated production for horizontal wells with known completions designs, surface-hole locations, and bottom-hole locations across the basins listed above. We then trained a series of models to forecast a three-year stream of production using interpreted subsurface, publicly available completions parameters, and well spacing. We evaluated model accuracy by withholding a random subset of 20% of the pads from each basin. After the initial model build, we trained a surrogate model to produce explanations for the model forecasts, specifically using SHAP values (SHapley Additive exPlanation), which isolate the contribution of each training variable on the model forecast.

Results/Observations/Conclusions: We examined average 1-year cumulative production values, analyzing both per-foot and absolute values, for all wells and also a subset of wells representing full development (“bound” wells). We found per-well production for bound wells peaked in 2017 in the Delaware and Midland, 2018 in the Bakken, and 2018 in the Eagle Ford. Longer laterals have had the largest impact on production, up to ~60% in the Delaware Basin, though the impact is muted in the Bakken where operators converged on 2 mile laterals much earlier due to state drilling unit regulations and favorable geology. Larger completions contributed a ~30-40% increase, though their impact plateaued beginning in 2017-2018. Geological high-grading shows a relatively limited impact of 5-10%, except in the Midland Basin, where activity shifted to the higher-quality northern portion of the basin from 2012-2018.

Applications/Significance/Novelty: The results of this study are consistent with the hypothesis that future performance gains will be relatively minor because the key innovations of longer laterals and upsized completions have been exhausted. Though our models have not yet seen a significant dip in average rock quality drilled across a basin, that will eventually become a headwind due to inventory depletion. Machine learning models combined with explainability datasets provide a powerful tool to quantify the impact of operator design choices, analyze basin-wide trends, and inform expectations for future developments.

Interdisciplinary Components: This work incorporates contributions from data science, reservoir engineering, and completions engineering.

Complete the form below and we will email the paper straight to your inbox.

First name(Required)

INTRODUCING CAUSAL MODELS

Accurate forecast on parent-child developments

In this live webinar, you will learn how Novi’s new algorithm improves model sensitivity for spacing and parent-child scenarios, providing powerful results for previously difficult-to-analyze problems.

Ted Cross, our VP of Product Management, will show you how this update improves spacing and infill scenario analysis without sacrificing model accuracy.