Record Details

Title Deploying Digital Twins for Geothermal Operations with the GOOML Framework
Authors Iraklis KONSTANTOPOULOS, Paul SIRATOVICH, Grant BUSTER, Nicole TAVERNA, Jon WEERS, Andrea BLAIR, Jay HUGGINS, Christine SIEGA, Warren MANNINGTON, Alex URGEL, Johnathan CEN, Jaime QUINAO, Robbie WATT, John AKERLEY
Year 2023
Conference Stanford Geothermal Workshop
Keywords field optimization, digital twins, plant operations, algorithms, machine learning, Wairakei, strategy, modelling
Abstract We present GOOML, a Geothermal operations optimisation framework based on machine learning and components modelling under constraints informed by physics. The framework was developed in the real world, using data from steam fields of various types (e.g., brine and dry steam) and configurations (e.g., single plant, binary, direct heat) to develop digital twins. GOOML aims to increase the output of steam fields all over the world by allowing operators to run countless scenarios by simulating changes to their fields. In doing so we remove the need to experiment in the physical world, therefore removing the significant costs associated with developing new operational strategies. During the research phase GOOML has suggested strategies that can bring about a few percent increase in electricity generation. In the next phase we hope to deploy this system across the world.
Back to Results Download File