Semester of Graduation

Summer 2025

Degree

Master of Science (MS)

Department

The Craft & Hawkins Department of Petroleum Engineering

Document Type

Thesis

Abstract

Abstract

This study presents a data-driven framework for modeling, interpreting, and forecasting heat recovery behavior in hydrothermal and fractured geothermal reservoirs. By combining scientific computing with unsupervised and supervised machine learning, this research offers a robust methodology for extracting dominant thermal patterns to optimize performance in various geothermal systems.

The approach begins with dimensionless modeling, identifying fifteen groups via inspectional analysis to represent key geophysical and operational dynamics of hydrothermal reservoirs. Self-Organizing Maps (SOM) cluster normalized production temperature profiles into four temporal regimes (early, early-intermediate, late-intermediate, and late). Non-negative Matrix Factorization with K-means (NMFK) extracts five latent temperature signatures tied to these regimes. Supervised models (XGBoost, Random Forest, and Deep Neural Networks) are trained on each regime, optimized with Optuna, and interpreted using SHAP analysis to identify the influence of key features such as temperature ratio, thermal Peclet number, fluid expansion, and reservoir geometry on the thermal regimes. These models achieve high accuracy (R² > 0.95).

To scale the workflow, 250 geothermal simulations are developed from five base cases encompassing varied fracture configurations representing enhanced geothermal systems (EGS). Each model incorporates deterministic and stochastic fractures, dynamic well placements, and consistent thermal-fluid-rock properties. Simulation outputs including temperature and pressure data from fracture and matrix zones are processed with NMFK to reveal temporal thermal signatures. These patterns differentiate between fracture-dominated regimes, with early thermal breakthrough, and matrix-buffered regimes characterized by gradual heat conduction.

This integrated framework combining dimensionless physics, machine learning clustering, and predictive modeling offers a replicable toolset for geothermal reservoir characterization and supports scalable, interpretable solutions for sustainable energy development.

Date

6-30-2025

Committee Chair

Dr. Mayank Tyagi

Available for download on Monday, June 28, 2032

Share

COinS