Abstract
Adsorbed natural gas (ANG) is a promising technology that can contribute to reducing the carbon footprint of the oil industry. This is because the ANG operates at a lower pressure relative to traditional methods. Topology optimization enables the generation of non-intuitive designs of adsorptive natural gas storage systems by inserting internal metallic structures to enhance thermal management. However, its application to adsorbent beds is often hindered by high computational cost and numerical instabilities when solving the fully coupled multiphysics problem. This work proposes a novel surrogate-based methodology that decouples the forward problem from the optimization stage. A steady-state heat conduction model, driven by transient heat-source profiles extracted from a single full-physics simulation, is used for topology optimization. In addition, to prevent non-physical material placement near boundaries that can arise from the simplified model, an Eikonal-based penalization function is introduced to define exclusion zones systematically. The approach is implemented in COMSOL Multiphysics\textsuperscript{\tiny\textregistered} with the full physics (Navier-Stokes and adsorption model) simulation validated using experimental data. Results show that optimized fin-like structures increase gas storage capacity by up to 7.6\% under specific volume fractions and filter sizes. Dimensionless analysis reveals that the Biot number serves as a threshold for the effectiveness of optimization, whereas the Fourier number governs the optimized loading cycle duration. The proposed method significantly reduces computational effort while preserving essential physics, providing a viable and generalizable framework for designing advanced adsorption vessels.

This publication has 22 references indexed in Scilit: