Discovering Signatures of Hidden Geothermal Resources based on Unsupervised Learning
V.V. Vesselinov, M.K. Mudunuru, B. Ahmmed, S. Karra, R.S. Middleton
Stanford Geothermal Workshop2020
Summary
This work explores unsupervised machine-learning methods for discovering hidden geothermal resource signatures from large, heterogeneous datasets. It emphasizes unbiased feature extraction without heavy dependence on prior labeling and demonstrates how these methods can help characterize geothermal systems and guide exploration decisions in geothermal study areas such as New Mexico and The Geysers.