[URTeC 2024] Machine Learning vs. Type Curves in the Appalachian Basin: A Comparative Study

Machine Learning vs. Type Curves in the Appalachian Basin: A Comparative Study

Technical Paper Details::
Technical Presentation: Tuesday, June 18th, 2024
Theme 9: A Future of Production Forecasting: Data-Driven Models and Physics-Based Solutions II
Topic: Machine Learning vs. Type Curves in the Appalachian Basin: A Comparative Study

Authors: A. Cui1, A. Yanke2, T. Dao2, P. Ye2, T. Cross1, B. Davis*1, 1. Novi Labs, 2. Equinor.

Abstract:

This study aims to compare the outcomes of machine learning (ML) models with the traditional method of type curves (TC) for forecasting new well production in the Appalachian Basin. The primary objectives are to evaluate predictive accuracy, computational efficiency, and enhancements to decision-making when utilizing ML versus type curves.

We gathered production data, completions data, and subsurface grids for the Lower Marcellus formation in the Appalachian Basin. We limited the production dataset to records up to 1/1/2022, which is the date of internal TC creation. Approximately 30 TC Areas were defined based on existing well performance and geological parameters. For each Area, Type Curves profiles were constructed by fitting a multi-segment modified hyperbolic curve to the median profile of hand-selected analogous wells. ML models were developed using the same dataset, targeting the initial 3 years of production. We generated forecasts for all 2022+ wells using both TC and ML approaches for comparison.

ML models maintain or improve predictive accuracy of production forecasts over the TC methodology. This study also analyzes the differences in standard TC Area definition versus a ML-generated Rock Quality Index. Furthermore, the ML models were used for optimizing well performance and different what-if scenarios, such as increasing completions intensity.

ML production forecasts can enhance accuracy and optimize resource extraction strategies while saving substantial man-hours during this process. Implementing these improvements has empowered engineers and asset managers to influence operational choices.

Please fill out the form below, and we’ll email you the paper as soon as it becomes available.

First name(Required)
This field is for validation purposes and should be left unchanged.

Meet Novi Labs' knowledgeable experts.

Invite to Speak

Our team of experts is passionate about sharing valuable insights on US energy. Please provide us some information about your event, and we will get back to you as soon as we can

Date
This field is for validation purposes and should be left unchanged.