Extraction of oil & gas is a complex process. Data are typically limited and uncertain. The processes are complex and non-linear. Physics models to predict the behavior of the reservoir are challenging to develop and take a long time to run.
PetroAI addresses these challenges.
PetroAI develops and applies science-informed ML models to predict oil & gas extraction based on site data and models.
PetroAI accounts for complex geomechanics processes.
PetroAI estimates uncertainties of model parameters and ML predictions.
PetroAI estimates sensitivities and provides decision support to extract actionable information.
Explore EnviCloud
Explore related demos
Try interactive public demos that bring these ideas to life.
- Public demo
Energy Production & Storage
Explore how science-informed models can compare energy production and storage strategies.
Open demo (opens in a new tab) - Public demo
3D Geo Modeling
View wells, terrain, and subsurface context in an interactive 3D scene.
Open demo (opens in a new tab) - Public demo
Model Diagnostics
Calibrate a model and identify influential, uncertain, or poorly constrained inputs.
Open demo (opens in a new tab)
Related publications
- Machine-Learning Predictions of the Shale Wells’ Performance (2021) — M. Mehana et al. — Journal of Natural Gas Science and Engineering
- Two-Stage Fracturing Wastewater Management in Shale Gas Development (2017) — X. Zhang et al. — Ind. Eng. Chem. Res.
Related presentations
- Predicting oil and gas production from unconventional tight-rock reservoirs using machine learning (2020) — V.V. Vesselinov — XXIII International Conference on Computational Methods in Water Resources (CMWR)
- Unsupervised and Physics-Informed Machine Learning Analyses for Characterization of Energy Production from Unconventional Reservoirs (2020) — V.V. Vesselinov — Machine Learning in Oil & Gas Conference
- Machine learning analyses for characterization of oil, gas and water production from unconventional tight-rock reservoirs (2019) — V.V. Vesselinov — AGU Fall Meeting