An Open Digital Twin Platform for Co-Simulation and Optimization of Geothermal Plant Operations

Abstract

Digital twins are increasingly transforming the operation and management of geothermal energy systems by enabling real-time simulation, monitoring, and optimization. This paper presents an open-architecture digital twin framework developed for low-enthalpy deep hydrothermal geothermal plants, designed to support co-simulation and optimization with real-time data streams. The framework integrates physical models, machine learning models, different data streams, and control algorithms to improve operational efficiency and reliability. Through case studies, we demonstrate how the digital twin can be applied to optimize the performance of electrical submersible pumps (ESPs), enhance plant-level operation, and predict potential failures before they occur. In a second case study, we show how large language models (LLMs) can be integrated into the digital twin environment to provide fast and intelligent access to operation and maintenance documentation, reducing response time for troubleshooting and decision support. For the second application, details of the retrieval-augmented generation (RAG) workflow for processing text data will be presented. The open-source nature of the tool lowers adoption barriers and enables broader collaboration across the geothermal sector.