Data-driven approaches are key for modeling and optimizing the operation of geothermal wells. However, data collected at different drilling sites are usually not standardized and contain missing or erroneous values which hinder the development of reliable models that can be used across a broad range of conditions and geological contexts. In this work, we present an end-to-end workflow for analyzing, modeling, and optimizing geothermal drilling based on probabilistic methods to account for data uncertainty and heterogeneity. The components of the workflow include: 1) self-organizing maps to visualize well operation data and process trajectories, 2) process mining to analyze drilling event logs and discover process models, 3) Bayesian models to predict relevant operation and performance metrics from incomplete and uncertain drilling data, and 4) algorithms to optimize well drilling under specific operational constraints. The workflow has been implemented as a web-based tool that facilitates geothermal drilling planning and operation tasks. Models used in the workflow have been trained and tested on the database of 113 geothermal wells representing various geological settings which was built in the framework of the EDGE project.

Press the Back button in your browser, or search again.

Copyright 2022, Stanford Geothermal Program: Readers who download papers from this site should honor the copyright of the original authors and may not copy or distribute the work further without the permission of the original publisher.