A Computational Fluid Dynamics MacroModel for the Design of Bed Adsorbers

This paper presents and validates a novel three-dimensional computational fluid dynamics macro-model that incorporates pore adsorption occupation rate to accurately simulate CO2 adsorption in packed beds, demonstrating that a new high-surface-area geometric design significantly enhances process productivity compared to traditional cylindrical configurations.

The Gist

The Big Picture: Catching Carbon with a Smart Sponge

Imagine you have a giant, busy kitchen (the atmosphere) that is getting filled with too much smoke (Carbon Dioxide, or CO2). You need a way to catch that smoke before it escapes. Scientists use adsorption, which is basically like using a super-smart sponge (called a Zeolite) to soak up the smoke.

This paper is about building a virtual simulation (a computer model) to figure out how to make these "sponges" work better, faster, and cooler.

---

1. The Problem: The Sponge Gets Hot and Slow

When you soak up smoke with a sponge, two things happen:

1. The Sponge gets full: It grabs the smoke particles.
2. The Sponge gets hot: Just like your hands get warm when you rub them together, the chemical reaction of grabbing the smoke releases heat.

In traditional designs, these sponges are packed into a single, thick cylinder (like a big soda can).

• The Issue: Because it's so thick, the heat gets trapped in the middle. It's like trying to cool down a thick stew in a pot without stirring; the center stays boiling hot while the edges cool down. This heat slows down the sponge, making it less efficient at catching more smoke. Also, once it's full, you have to wait a long time for it to cool down before you can use it again.

2. The New Tool: A "Super-Computer" for Fluids

The authors built a new 3D Computational Fluid Dynamics (CFD) model. Think of this not just as a calculator, but as a high-definition weather forecast for a tiny world.

• Old Models (1D/2D): These were like looking at a shadow of the sponge. They could guess the average temperature, but they couldn't see the hot spots or how the gas moved around inside the complex shape of the sponge.
• The New Model (3D): This is like putting on 3D glasses. It simulates the gas, the heat, and the sponge particles in full 3D detail. It can see exactly where the heat builds up and how the gas flows through every nook and cranny.

The Secret Sauce (PAOR):
The authors added a new "ingredient" to their math called PAOR (Pores Adsorption Occupation Rate).

• Analogy: Imagine a parking garage (the sponge). Old models assumed the cars (CO2 molecules) parked evenly everywhere. The new model realizes that as the garage fills up, the available spots change, and the rate at which cars park depends on how full the garage already is. This makes the simulation much more realistic.

3. The Test: Did the Virtual Model Work?

Before trusting the computer, they had to prove it was right. They simulated three different scenarios:

• Pure CO2 (100% smoke).
• A mix (50% smoke, 50% Helium).
• A light mix (15% smoke, 85% Helium).

They compared their computer results to real-world experiments done by other scientists.

• The Result: The computer model was spot on. It predicted exactly when the sponge would get full (the "breakthrough") and exactly how hot it would get, matching the real-life experiments perfectly.

4. The Innovation: From a "Soda Can" to a "Honeycomb"

Once they proved their computer model was accurate, they used it to design a brand new shape for the adsorber.

• The Old Design: A single, solid cylinder (like a thick log).
• The New Design: Instead of one big log, they designed seven smaller tubes packed together inside a larger shell (like a honeycomb or a bundle of straws).

Why is this better?

• Surface Area: The seven small tubes have much more "skin" (surface area) touching the outside air than one big tube does.
• Cooling: Because there is more surface area, the heat can escape much faster. It's like the difference between trying to cool a thick steak versus seven thin slices of steak. The slices cool down almost instantly.
• Efficiency: Because the new design cools down faster, you can run the "catch and release" cycle (adsorption and regeneration) much more quickly. This means you can capture more CO2 in the same amount of time.

5. The Conclusion: Why This Matters

This paper shows that by using advanced 3D computer modeling, we can stop guessing and start designing better machines.

• The Takeaway: We don't just need better sponges; we need better shapes for the containers holding the sponges.
• The Future: This new "seven-tube" design could make carbon capture plants smaller, cheaper, and much more efficient, helping us fight climate change by cleaning the air faster.

In a nutshell: The authors built a super-accurate 3D simulator, proved it works, and used it to redesign a carbon-capture machine from a "thick log" into a "cooling honeycomb," making the whole process faster and more efficient.

Technical Summary

1. Problem Statement

Carbon capture and storage (CCS) is critical for mitigating climate change, with Pressure Swing Adsorption (PSA) and Temperature Swing Adsorption (TSA) emerging as promising technologies. However, the design and optimization of fixed-bed adsorbers are hindered by a lack of understanding of the coupled phenomena governing adsorption: mass transfer, fluid dynamics, and heat effects (specifically the exothermic nature of adsorption).

Existing numerical models often rely on:

• 1D or 2D approximations: These fail to capture local gradients in temperature, pressure, and concentration, particularly in beds with non-conventional geometries or complex heat loss paths.
• Oversimplified source terms: Traditional macroscopic CFD models often assume uniform adsorption rates and neglect the Pores Adsorption Occupation Rate (PAOR) (gas loading) inside the porous particles. They typically assume that the entire particle volume participates equally in mass and heat exchange, ignoring the internal porosity and non-uniform adsorption distribution.

The authors aim to develop a robust 3D transient CFD model that accurately predicts breakthrough curves and thermal front propagation while introducing a more rigorous physical formulation for volumetric source terms. Furthermore, they seek to apply this model to design a novel adsorber geometry that improves thermal management without sacrificing adsorption capacity.

2. Methodology

The study utilizes the open-source OpenFOAM (v2306) platform to develop a custom transient solver for gas adsorption in porous media.

A. Mathematical Modeling

The model solves volume-averaged conservation equations for mass, momentum, energy, and species transport under transient, compressible, multi-species flow conditions.

• Governing Equations:
  • Momentum: Navier-Stokes equations modified with an Ergun equation source term to account for pressure drop in the packed bed.
  • Mass & Species: Continuity and species transport equations including axial dispersion.
  • Energy: A single effective energy equation assuming local thermal equilibrium between the gas and solid phases.
• Adsorption Kinetics & Equilibrium:
  • Equilibrium: Modeled using the Dual-Site Langmuir (DSL) isotherm to account for heterogeneous adsorption sites on Zeolite-13X.
  • Kinetics: The Linear Driving Force (LDF) model is used to describe the rate of mass transfer.
• Novel Source Terms (Key Innovation):
  The authors introduce new volumetric source terms ($S_Y$ for mass and $S_T$ for energy) that explicitly account for:
  1. Particle Porosity ($\varepsilon_p$): Distinguishing between bed porosity and internal particle porosity.
  2. PAOR ($\Gamma$): A new term representing the Pores Adsorption Occupation Rate. Unlike previous models that assume $\Gamma = 1$, this model correlates $\Gamma$ with the feed-in gas concentration.
     • $\Gamma_Y$ and $\Gamma_T$ are functions of the CO2 feed-in percentage.
     • Below critical thresholds (0.25 for mass, 0.5 for heat), these terms are constant; above them, they increase, reflecting higher mass/heat exchange efficiency as gas loading increases.

B. Numerical Setup

• Geometry: A cylindrical fixed-bed (Reference Design) with an inner diameter of 2.82 cm and length of 6.4 cm, replicating experimental setups from the literature.
• Mesh: A grid convergence study was performed using three mesh densities (Coarse: 100k, Medium: 337.5k, Fine: 800k cells). The Grid Convergence Index (GCI) confirmed that the medium mesh provided sufficient accuracy.
• Boundary Conditions:
  • Inlet: Constant volumetric flow rates for different CO2/He mixtures (100%, 50%, 15% CO2).
  • Wall: Natural convection heat transfer modeled via a correlation dependent on local wall temperature and height.
  • Initial Conditions: Uniform temperature (294.6 K) and pressure (1.02 bar).

3. Key Contributions

1. Development of a 3D Macroscopic CFD Model: A fully coupled 3D solver capable of resolving spatio-temporal evolution of species concentration, temperature, and pressure in fixed-bed adsorbers.
2. Enhanced Physics-Based Source Terms: The introduction of PAOR-dependent source terms ($\Gamma_Y, \Gamma_T$) that explicitly account for particle porosity and gas loading effects. This corrects the assumption in previous models that adsorption occurs uniformly throughout the entire particle volume regardless of saturation levels.
3. Validation Against Experimental Data: Rigorous validation against experimental data (Wilkins & Rajendran, 2019) and 2D CFD results (Ramos et al., 2024) for three distinct CO2 feed-in concentrations (100%, 50%, 15%).
4. Geometric Optimization: Application of the model to a novel multi-tube bed design (seven inter-spaced cylinders) to demonstrate how 3D geometry can enhance thermal management.

4. Results

A. Model Validation

The new 3D CFD model demonstrated excellent agreement with experimental data across all test cases:

• Breakthrough Curves: Accurately predicted the timing and shape of the CO2 breakthrough curves for 100%, 50%, and 15% CO2 feeds.
• Thermal Fronts: Successfully captured the magnitude and timing of the temperature peaks caused by exothermic adsorption. The deviation in peak temperature was minimal (e.g., <0.8 K for the 50% case).
• Impact of New Terms: Simulations showed that using the new PAOR-dependent terms ($\Gamma < 1$) significantly improved the accuracy of temperature predictions compared to assuming $\Gamma = 1$, particularly at lower feed concentrations.

B. Application to New Bed Design

The model was applied to a new design consisting of seven smaller, inter-spaced cylindrical tubes (totaling the same adsorbent volume as the single reference cylinder).

• Adsorption Performance: The breakthrough time and total CO2 capture capacity remained virtually identical to the reference single-cylinder design, confirming that the change in geometry did not compromise adsorption kinetics.
• Thermal Performance: The new design showed a significant improvement in thermal management:
  • The peak temperature during adsorption was similar to the reference.
  • Cooling Phase: The post-adsorption cooling period was substantially reduced due to the increased external surface area and better heat dissipation to the surroundings.
  • Implication: Faster cooling allows for shorter regeneration cycles in PSA/TSA processes, potentially doubling the overall process productivity.

5. Significance and Conclusion

This research establishes that 3D CFD modeling is essential for the next generation of adsorber design, particularly when optimizing for thermal management in non-conventional geometries.

• Scientific Impact: The work advances the physics of macroscopic CFD modeling by integrating the concept of PAOR, providing a more realistic representation of the gas-solid interface dynamics than previous 1D/2D models.
• Engineering Impact: The study proves that geometric optimization (e.g., multi-tube configurations) can decouple thermal and mass transport limitations. By increasing surface area without increasing adsorbent mass, engineers can design adsorbers that cycle faster, leading to higher efficiency in industrial CO2 capture and gas separation processes.
• Future Work: The authors suggest moving toward meso-scale modeling where individual particles are resolved to capture anisotropy and non-uniform local properties, further refining the physics of adsorption in fixed beds.