United StatesEnergy and utilities2022 – ongoingSM Energy
Machine learning optimization of Midland Basin well completions
SM Energy tested around 25,000 completion designs with a machine learning model
Problem
SM Energy needed to choose well completion designs that improved returns on its Midland Basin development spending. Designs had to account for differences between wells.
Approach
SM Energy used a multivariate model to compare completion designs intended to maximize net present value per well. It continually updated the process with new information and evaluated it for nearly every well. Management attributed higher capital efficiency to the optimization effort.
Results
Takeaway
The model supported repeated design decisions as new well information arrived. The company did not isolate a numerical financial gain attributable to the model.
Categories
- Industry
- Energy and utilities
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and MLOptimization
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
Management described machine learning for Midland Basin completion optimization. Its example compared around 25,000 designs, with the process updated and evaluated for nearly every well.
Full details are in the original sources.