2026-09-19T06:30:43Zhttps://dspace.mit.edu/server/oai/request

oai:dspace.mit.edu:1721.1/1629742025-10-07T04:14:45Zcom_1721.1_7582com_1721.1_7581col_1721.1_131023

Edelman, Alan Ouko, Edwin O. Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science 2025-10-06T17:37:40Z 2025-10-06T17:37:40Z 2025-05 2025-06-23T14:03:11.469Z https://hdl.handle.net/1721.1/162974 Geothermal well arrays, which organize multiple geothermal wells into carefully planned geometric configurations, provide an opportunity to enhance energy production capacity and increase fault tolerance of geothermal systems. Closed-loop geothermal systems (CLGS), a type of geothermal well design, promises to allow harnessing of geothermal energy in any location with minimal adverse environmental impact. I demonstrate how the development of these emerging geothermal technologies could be accelerated by recent advances in large language models (LLMs) in conjunction with high-level high-performance programming languages like Julia. In particular, I focus on how LLMs could be used in design brainstorming and to increase efficiency in numerical modeling. I assess the potential of state-of-the-art LLMs such as ChatGPT, Gemini, Claude, Grok, and a domain-specific model, AskGDR, as expert assistants in geothermal research. Owing to the unpredictable reliability of LLMs, there is a constant need for objective evaluation benchmarks in various domains. I propose a novel approach, leveraging Google’s recently introduced AI tool, NotebookLM, to accelerate the generation of quantitative geothermal benchmarks with only new unpublished questions. In addition, I propose the use of blackbox optimization as a computationally less costly alternative to approximate the optimal configuration of CLGS wells in a geothermal array to minimize thermal interference and improve heat energy production. I evaluate several optimization strategies such as Bayesian optimization, particle swarm optimization, natural evolution strategies, differential evolution optimization, Nelder-Mead, and simulated annealing on various performance characteristics such as convergence speed and highest production capacity attained. M.Eng. Massachusetts Institute of Technology In Copyright - Educational Use Permitted Copyright retained by author(s) https://rightsstatements.org/page/InC-EDU/1.0/ Efficient Modeling, Optimization, and LLM-Assisted Decision Support for Geothermal Well Arrays Thesis application/pdf Master Master of Engineering in Electrical Engineering and Computer Science Publication unknown unknown <Publication xmlns="https://www.openaire.eu/cerif-profile/1.1/" id="65984016-7349-49d9-8b00-b4dfe7081293"> <Type xmlns="https://www.openaire.eu/cerif-profile/vocab/COAR_Publication_Types">http://purl.org/coar/resource_type/c_1843</Type> <Title>Efficient Modeling, Optimization, and LLM-Assisted Decision Support for Geothermal Well Arrays</Title> <PublishedIn> <Publication> </Publication> </PublishedIn> <PublicationDate>2025-05</PublicationDate> <Authors> <Author> <DisplayName>Ouko, Edwin O.</DisplayName> <Affiliation> <OrgUnit> </OrgUnit> </Affiliation> </Author> </Authors> <Editors> </Editors> <Publishers> <Publisher> <DisplayName>Massachusetts Institute of Technology</DisplayName> <OrgUnit /> </Publisher> </Publishers> <License>https://rightsstatements.org/page/InC-EDU/1.0/</License> <Abstract>Geothermal well arrays, which organize multiple geothermal wells into carefully planned geometric configurations, provide an opportunity to enhance energy production capacity and increase fault tolerance of geothermal systems. Closed-loop geothermal systems (CLGS), a type of geothermal well design, promises to allow harnessing of geothermal energy in any location with minimal adverse environmental impact. I demonstrate how the development of these emerging geothermal technologies could be accelerated by recent advances in large language models (LLMs) in conjunction with high-level high-performance programming languages like Julia. In particular, I focus on how LLMs could be used in design brainstorming and to increase efficiency in numerical modeling. I assess the potential of state-of-the-art LLMs such as ChatGPT, Gemini, Claude, Grok, and a domain-specific model, AskGDR, as expert assistants in geothermal research. Owing to the unpredictable reliability of LLMs, there is a constant need for objective evaluation benchmarks in various domains. I propose a novel approach, leveraging Google’s recently introduced AI tool, NotebookLM, to accelerate the generation of quantitative geothermal benchmarks with only new unpublished questions. In addition, I propose the use of blackbox optimization as a computationally less costly alternative to approximate the optimal configuration of CLGS wells in a geothermal array to minimize thermal interference and improve heat energy production. I evaluate several optimization strategies such as Bayesian optimization, particle swarm optimization, natural evolution strategies, differential evolution optimization, Nelder-Mead, and simulated annealing on various performance characteristics such as convergence speed and highest production capacity attained.</Abstract> <Access xmlns="http://purl.org/coar/access_right" > </Access> </Publication>