Session: 16-01-01: Government Agency Student Poster Competition

Paper Number: 148662

148662 - Data-Driven Solutions for Co2 Sequestration and Thermal Optimization in Energy Systems 

In response to the urgent need to address climate change and curb CO₂ emissions, our research involves innovative and interdisciplinary modeling, simulation, and machine learning (ML) solutions for empowering energy systems and carbon mitigation. This work highlights the vital need for cutting-edge ML techniques in two synergistic projects that address crucial approaches to carbon emission reduction: addressing the critical issue of carbon dioxide (CO₂) sequestration to mitigate climate change during the energy transition and understanding the thermal behavior of porous media for optimizing heat transfer applications. Both projects employ physics-based ML to address gaps that cannot be solved using conventional modeling approaches.

CO₂ sequestration, a critical method for CO₂ storage, is an important technology for mitigating the acceleration of climate change during the energy transition and involves pressurizing and injecting CO₂ into underground formations, commonly saline aquifers. Optimizing this process requires managing interfacial tension (IFT) between the CO₂ and a liquid (underground salt water), a highly nonlinear relationship that affects storage capacity and the risk of CO₂ leakage. There is no universal closed-form theory-driven model for IFT, and measuring IFT experimentally can be cumbersome, time-consuming, and costly. Data-driven ML approaches have demonstrated exceptional predictive capabilities, offering a cost-effective and efficient alternative to traditional experimental methods. We have developed ML-based IFT prediction models for a novel set of parameters. Our investigation has also shown improved performance compared to the existing work. Optimizing CO₂ sequestration design will further advance carbon capture, utilization, and storage (CCUS) technologies.

Multiphase flow and heat transfer phenomena in porous media are critical in various scientific and engineering applications. Examples include using phase change materials (PCMs) as the heat transfer medium in thermal energy storage (TES) systems for concentrated solar power (CSP) plants, flow and condensation in the regenerators of thermoacoustic coolers using air as the working fluid, and energy harvesting from porous media heat flow in geothermal applications. Recent advances in data-driven approaches have demonstrated that machine learning can be used to study the underlying physics of thermal fluid problems, leading to optimized designs for better thermal performance. Our second project focuses on utilizing advanced AI techniques to understand the characteristics of porous media, such as closed channels, structural gaps, and thermal performance for heat transfer applications.

Our work, through these projects, highlights the significant impact of ML in advancing energy technologies and sustainability.

(This material is based upon work partially supported by the National Science Foundation under Grant No. CBET-2223078.)

Presenting Author: Kashif Liaqat Rice University

Presenting Author Biography: Kashif is a 3rd year Ph.D. student in the mechanical engineering program at Rice University. His areas of research interest are AI for Sustainability, Concentrated Solar Power, Energy Systems, and Machine Learning.

Authors:

Kashif Liaqat Rice University
Laura Schaefer Rice University