Date of Graduation

2023

Document Type

Dissertation

Degree Type

PhD

College

Statler College of Engineering and Mineral Resources

Department

Petroleum and Natural Gas Engineering

Committee Chair

Shahab Mohaghegh

Committee Member

Shahab D. Mohaghegh

Committee Member

Samuel Ameri

Committee Member

Kashy Aminian

Committee Member

Mohamed El Sgher

Committee Member

Qingqing Huang

Abstract

Geological carbon sequestration (GCS) offers a promising solution to effectively manage extra carbon, mitigating the impact of climate change. This doctoral research introduces a cutting-edge Smart Proxy Modeling-based framework, integrating artificial neural networks (ANNs) and domain expertise, to re-engineer and empower numerical reservoir simulation for efficient modeling of CO2 sequestration and demonstrate predictive conformance and replicative capabilities of smart proxy modeling.

Creating well-performing proxy models requires extensive human intervention and trial-and-error processes. Additionally, a large training database is essential to ANN model for complex tasks such as deep saline aquifer CO2 sequestration since it is used as the neural network's input and output data. One major limitation in CCS programs is the lack of real field data due to a lack of field applications and issues with confidentiality.

Considering these drawbacks, and due to high-dimensional nonlinearity, heterogeneity, and coupling of multiple physical processes associated with numerical reservoir simulation, novel research to handle these complexities as it allows for the creation of possible CO2 sequestration scenarios that may be used as a training set. This study addresses several types of static and dynamic realistic and practical field-base data augmentation techniques ranging from spatial complexity, spatio-temporal complexity, and heterogeneity of reservoir characteristics. By incorporating domain-expertise-based feature generation, this framework honors precise representation of reservoir overcoming computational challenges associated with numerical reservoir tools.

The developed ANN accurately replicated fluid flow behavior, resulting in significant computational savings compared to traditional numerical simulation models. The results showed that all the ML models achieved very good accuracies and high efficiency. The findings revealed that the quality of the path between the focal cell and injection wells emerged as the most crucial factor in both CO2 saturation and pressure estimation models. These insights significantly contribute to our understanding of CO2 plume monitoring, paving the way for breakthroughs in investigating reservoir behavior at a minimal computational cost.

The study's commitment to replicating numerical reservoir simulation results underscores the model's potential to contribute valuable insights into the behavior and performance of CO2 sequestration systems, as a complimentary tool to numerical reservoir simulation when there is no measured data available from the field. The transformative nature of this research has vast implications for advancing carbon storage modeling technologies. By addressing the computational limitations of traditional numerical reservoir models and harnessing the synergy between machine learning and domain expertise, this work provides a practical workflow for efficient decision-making in sequestration projects.

Download

Share

COinS