Hosted by IDTechEx
HomeWebinarsEventsReportsTVCareersAbout UsIDTechExTwitterFacebookLinkedInYoutubeRSSForward To Friend
Artificial Intelligence Research
Posted on May 10, 2019

$5.5 million to apply machine learning to geothermal exploration

The US Department of Energy has announced up to $5.5 million for 10 new projects to apply machine learning techniques to geothermal exploration and production. Machine learning - the use of advanced algorithms to identify patterns in and make inferences from data - could assist in finding and developing new geothermal resources. If applied successfully, machine learning could lead to higher success rates in exploratory drilling, greater efficiency in plant operations, and ultimately lower costs for geothermal energy. For more information see the IDTechEx report on Energy Harvesting Microwatt to Megawatt 2019-2029.
 
Geothermal energy is an important part of DOE's "all-of-the-above" strategy to advance toward energy dominance and ensure a secure, reliable, resilient, affordable, and enduring supply of American energy.
 
The projects selected by the Office of Energy Efficiency and Renewable Energy's Geothermal Technologies Office focus in two areas: machine learning for geothermal exploration and advanced analytics for efficiency and automation in geothermal operations.
 
The selected projects include:
  • Colorado School of Mines (Golden, CO): Applying new machine learning techniques to analyze remote-sensing images, with the goal of developing a process to identify the presence of blind geothermal resources based on surface characteristics. Colorado School of Mines will develop a methodology to automatically label data from hyperspectral images of Brady's Hot Springs, Desert Rock, and the Salton Sea.
  • Lawrence Livermore National Laboratory (Livermore, CA): Developing and applying new machine learning techniques to a multi-physics (magnetotelluric and seismic) dataset from the Raft River geothermal field. The overall goal of the project is to better identify and target fracture zones for drilling production wells.
  • Los Alamos National Laboratory (Los Alamos, NM): Developing an extendable, open source cloud-based machine learning framework called GTCloud (GeoThermal Cloud) that will incorporate local, regional, and continental scale geothermal data to estimate risk, cost, and thermal power production outputs for geothermal exploration.
  • National Renewable Energy Laboratory (Golden, CO): Improving geothermal reservoir management by using machine learning in conjunction with physics-based subsurface flow paths and interwell connectivity models. NREL will focus on two operational decisions: 1) where to drill and complete 'makeup' wells and 2) how to allocate injection into new and existing wells.
  • Pennsylvania State University (University Park, PA): Applying machine learning methodologies toward the study of microearthquakes (MEQs) and their linkages to probable zones of permeability, as well as the risks associated with induced seismicity in geothermal development. The project team has recently had notable success in predicting earthquakes at the laboratory scale, showing that passive seismic signals do contain information on the evolution of stress and fractures in the subsurface.
  • University of Arizona (Tucson, AZ): Building a single web-based platform to allow geothermal researchers and developers access to unique continuously growing scientific and exploration data. The project will programmatically parse the grammatical and visual relationships of words in the texts (e.g., noun, adjective) and use these relationships to build structured (e.g., spreadsheets) datasets for geothermal research. The project will address one of the most significant barriers to broader application of machine learning techniques in geothermal exploration, the unstructured (non-tabular) nature of most publicly available data.
  • University of Houston (Houston, TX): Developing a methodology to automatically detect subsurface fault/fracture zones from seismic images, and reliably characterize the fractures with the fault/fracture zones using the 'double-beam' method with machine learning. The investigators have already shown some success using the proposed techniques in an oil and gas setting (Marcellus Shale) and will now adapt these techniques to the more difficult geothermal environment.
  • University of Nevada (Reno, NV): Building on a prior GTO-funded project that was focused on defining geothermal 'play fairways' in Nevada; the previous project utilized several machine learning techniques to identify regions having high geothermal favorability but it relied to some degree on expert opinion where training data was lacking. This application addresses that shortcoming through the introduction of an additional 100 training sites, as well as the addition of an industry partner with extensive proprietary datasets.
  • University of Southern California (Los Angeles, CA): Developing novel data-driven predictive models for integration into real-time fault detection and diagnosis, and integrate those models by using predictive control algorithms to improve the efficiency of energy production operations in a geothermal power plant. The project will develop deep dynamic neural networks for fault prediction and predictive process control workflows to improve the efficiency of geothermal operations.
  • Upflow Limited (Taupo, New Zealand): Making available multiple decades of closely-guarded production data from one of the world's longest operating geothermal fields, and combining it with the archives from the largest geothermal company operating in the U.S. Models developed from this massive data store will enable the creation of a prediction/recommendation engine that will help operators improve plant availability.
 
/Source: US Department of Energy
Top image: National Renewable Energy Laboratory~