It is often burdensome to handle large amounts of wells files (*.LAS), coming from different exploration campains, locations and operators, and possibly containing different subsets of log curves.
In this project we built a tool to facilitate the screening of such well files and to assist operational geoscientists in rapidly selecting relevant well data thanks to natural language queries.
The solution we designed is a web-based application, showing an interactive map of the well locations, and the possibility to highlight well data matching given conditions ( e.g. wells with gamma ray values between 1200 and 2500 meters )
Feedbacks from users underlined the time saved in retrieving subsets of LAS files with wanted characteristics. Consequently, geological studies are accelerated, and the risk of overlooking key log data is minimized.
Future work could focus on visualizing production data along with the well geological properties, and on integrating other data formats such as PDF reports and geoscience documents. The Natural Language Processing query system could also be extended, to further increase interactivity and flexibility in the queries. Finally, advanced statistical algorithms could be run directly in the tool to exploit the data gathered, such as the ones developed for production profile prediction.