Back to search

University of North Dakota

Exploring Geothermal Energy Development In The Deep Reservoirs Of The Williston Basin: An Integrated Analysis Using Empirical, Probabilistic, Numerical Simulation And Machine Learning Techniques

Abstract

dc:description.abstract

Geothermal energy, a renewable, clean, and abundant energy source, is gaining global attention as a key contributor to the transition towards a low-carbon economy. However, uncertainties in reservoir characterization and development risks hinder its widespread adoption, particularly in the Williston Basin. This study focuses on the Williston Basin, particularly the Red River and Deadwood formations, to explore the potential of deep geothermal reservoirs in North Dakota. The study integrates a multi-disciplinary technique in empirical, probabilistic, numerical simulation, and machine learning to reduce uncertainties and risks associated with geothermal energy exploration and development. Probabilistic methods (Parametric and Monte Carlo), incorporating geo-statistics, are employed to quantify geothermal resources and conduct uncertainty analysis. The probabilistic approach is further utilized to compare hydrothermal systems and CO2 Plume geothermal system within the deep reservoirs varying reservoir conditions. Permeability modelling, a critical factor in geothermal reservoir characterization, is evaluated utilizing multiple empirical correlations. Additionally, machine learning techniques, including supervised and unsupervised learning, are applied to abundant log data and limited core data to predict reservoir permeability, demonstrating their efficacy over traditional empirical methods. Thermal simulation of the 3D model is performed using KAPPA, a three-phase multi-component thermal and steam simulator, to assess the thermally recoverable energy in deep aquifers. Key results demonstrate significant geothermal potential in North Dakota’s deep reservoirs, with probabilistic analysis identifying critical factors influencing geothermal performance. Machine learning outperforms traditional methods in permeability prediction, reducing uncertainty. Numerical simulations confirm substantial thermally recoverable energy, highlighting the impact of geological heterogeneity on resource viability. This study highlights the untapped geothermal potential of North Dakota and the importance of an integrated approach for accurate geothermal resource assessment.

Degree

thesis:*
Name thesis:degree_name
Doctor of Philosophy (PhD)
Level thesis:degree_level
Dissertation
Discipline thesis:degree_discipline
Energy Engineering

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Gyimah, Emmanuel
Contributors dc:contributor
  • Stanley O. Tomomewo

Subjects

dc:subject × 1

Identifiers

dc:identifier.*
Repository record dc:identifier
https://commons.und.edu/theses/7111

Chain of custody

source
Harvested from
University of North Dakota
Base URL
commons.und.edu/do/oai/
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Gyimah, Emmanuel. Exploring Geothermal Energy Development In The Deep Reservoirs Of The Williston Basin: An Integrated Analysis Using Empirical, Probabilistic, Numerical Simulation And Machine Learning Techniques. Dissertation thesis, https://commons.und.edu/theses/7111