Nevada Geothermal Machine Learning project

GBCGE project personnel: Dr. James Faulds (PI); Dr. Bridget Ayling; Elijah Mlawsky; Dr. Cary Lindsey; Connor Smith; Dr. Mark Coolbaugh

Project collaborators: United States Geological Survey; Hi-Q Geophysical; Massachusetts Institute of Technology

Project duration: 24 months: 1 August 2019 – 28 February 2022.

Total project funding: $500,000

Funding agency: U.S. Department of Energy Geothermal Technologies Office (award number DE-EE0008762).

Project goal: Apply machine learning (ML) techniques to develop an algorithmic approach to identify new geothermal systems in the Great Basin region and build on the successes of the Nevada geothermal play fairway project. The reason for this is that an algorithmic approach that empirically learns to estimate weights of influence for diverse parameters may scale and perform better than the original workflow developed for play fairway analysis. Project activities include augmenting the number of training sites (positive and negative) that are needed to train the ML algorithms, transforming the data into formats suitable for ML, and development and testing of the ML techniques and outputs.

Three maps of different geological variables used for the NV machine learning project
Examples of feature types encountered in preparing the original NV play fairway datasets for ML analysis. (a) The map of favorable structural settings as an example of the categorical features. (b) The map of fault traces as an example of a dense binary feature. (c) The map of horizontal gravity gradient as an example of a continuous numerical feature. Figure from Brown et al., 2020.

Publications to date:

  • Faulds, J.E., Brown, S., Coolbaugh, M., DeAngelo, J., Queen, J.H., Treitel, S., Fehler, M., Mlawsky, E., Glen, J.M., Lindsey, C., Burns, E., Smith, C.M., Gu, C., Ayling, B.F., (2020). Preliminary Report on Applications of Machine Learning Techniques to Nevada Geothermal Play Fairway Analysis. Proceedings of the 45th Workshop on Geothermal Reservoir Engineering, Stanford University, February 10-12th, 2019. SGP-TR-215.
  • Brown, S., Coolbaugh, M., DeAngelo, J., Faulds, J., Fehler, M., Gu, C., Queen, J., Treitel, S., Smith, C., and Mlawsky, E., (2020). Machine learning for natural resource assessment: An application to the blind geothermal systems of Nevada: Geothermal Resources Council Transactions, v. 44, 14 p. https://www.geothermal-library.org/index.php?mode=pubs&action=view&record=1034262
  • Smith, C.M., Faulds, J.E., Coolbaugh, M., Brown, S., (2020). Initial Results of Machine Learning Techniques Applied to the Nevada Geothermal Play Fairway Analysis. Transactions, Geothermal Resources Council annual meeting, October 19-23, 2020, Vol. 44. https://www.geothermal-library.org/index.php?mode=pubs&action=view&record=1034351

• View More Current Projects