Back to search

PETROMAKS2-Stort program petroleum

Enhancing Reservoir Characterization by Applying Machine Learning

Alternative title: Forbedre reservoar-karakterisering ved å bruke maskinlæring

Awarded: NOK 9.0 mill.

Project Manager:

Project Number:


Project Period:

2020 - 2024


Oil and gas companies spend considerable time and resources to identify and evaluate prospects before they drill a well. They need to make a precise earth model which shows oil and gas presence, and whether it is economic to drill. A precise earth model is also important for safety and environmental consequences of the drilling campaign. In ReservoAIr we make a software that aid companies in creating a more precise earth model. The companies will use the software to predict sand and rock properties with a new machine learning method. The most important progress we have made to date is to predict facies bodies, like channels and injectites, and to predict rock properties. We have made a software prototype of an interactive workflow where the user can interpret complex geological formations, aided by machine learning. The results are promising and according to plan. The methods and software we develop to create a better earth model can also be used to evaluate and monitor CO2 reservoirs.

Funding scheme:

PETROMAKS2-Stort program petroleum