Session: Government Agency Student Posters

Paper Number: 172162

A Numerical Investigation of Oscillating Flow and Condensation in Natural Porous Structures Using the Lattice Boltzmann Method and Deep Learning 

Multiphase flow and heat transfer within porous media are essential phenomena across a diverse range of scientific and engineering disciplines, including oil recovery, subsurface contamination and remediation, and fuel cells. The introduction of oscillating flows in conjunction with porous structures can further influence thermal transport due to the increase of surface area and the improvement of enhanced convection, applications are prevalent such as Stirling engines, pulsating heat pipes (PHP), and oscillating water columns (OWC).

 

Recently, many advanced energy technologies have emerged where oscillating multiphase transport phenomena in porous media play a significant role in their performance. Notable examples include the use of phase change materials (PCMs) as the heat transfer medium in solar thermal energy storage (TES) systems for concentrated solar power (CSP) plants, flow and condensation in the regenerator of thermoacoustic coolers using air as the working fluid, and the operation of enhanced geothermal systems that extract energy from fractured rock formations.

 

Despite the prevalence of these applications, a comprehensive understanding of the central mechanisms of oscillating multiphase flow and heat transfer in porous media remains unclear due to the intricate porous structure. The existing literature presents conflicting findings, highlighting the complexity of these interactions. For instance, Fu et al. (2001) observed increased average Nusselt numbers for oscillating flow through aluminum foams. In contrast, Al-Sumaily and Thompson (2013) reported an initial enhancement followed by a reduction in the average Nusselt number for flow within a porous channel. Conversely, Forooghi et al. (2011) documented the opposite trend in a channel featuring two distinct porous layers. These discrepancies highlight the necessity for a more fundamental investigation into how porous structures influence oscillating multiphase flow and heat transfer, thereby enabling enhanced thermal management.

 

This study utilizes natural porous structures, specifically coral rock, to investigate these phenomena. The selection of coral is motivated by its natural environment, where it is subjected to oscillating ocean currents and thermal gradients. Micro-computed tomography (micro-CT) is employed to capture the complex, multiscale internal channel network of the coral, with features ranging from micrometers to millimeters. From the reconstructed three-dimensional images, three distinct porous structure samples will be extracted. The selection criteria are designed to isolate key geometric parameters: two samples will possess identical porosity, while another pair will share the same effective porosity. These 3D samples will serve as the computational domains for simulating condensation under oscillating flow conditions. The simulations will be performed using Palabos, an open-source solver based on the Lattice Boltzmann Method (LBM).

Considering the complexity of porous media, which poses significant computational challenges, recent advancements in artificial intelligence (AI)—particularly in efficiently processing large datasets with reduced computational resources—offer promising solutions. In this study, a deep learning model, YOLOv11, will be implemented to assist in detecting the presence of condensation, enabling the prediction of condensation phenomena within porous structures.

This research aims to investigate the influence of porosity, channel connectivity, and oscillation parameters on the condensation process by using the LBM simulation. And deep learning will be implemented to assist in condensation prediction. The findings are expected to provide a fundamental understanding and  a prediction tool of multiphase oscillating flow and heat transfer within complex porous media. Ultimately, this work seeks to facilitate the prediction and optimization of designs, inspired by natural porous structures, for enhancing multiphase heat transfer in the presence of oscillating flows.

Presenting Author: Lichang Zhu University of Houston

Presenting Author Biography: Lichang Zhu is a PhD student and graduate research assistant at UH, and he graduated with his BS from China University of Mining and Technology in 2020. His research interests include multiphase heat transfer, flow and heat transfer in porous media, and machine learning.

Authors:

Lichang Zhu University of Houston
Ben Xu University of Houston