Spiraling Toward Cleaner Combustion: A CFD Analysis of Swiss-Roll Combustors

Junji Ito’s horror classic, Uzumaki, emphasizes how few shapes are as hypnotic—or as unsettling—as the spiral. But beyond the eerie allure, this simple shape appears everywhere in nature, from snail shells and whirlpools to galaxies and hurricanes. In engineering, the spiral offers an interesting advantage. It can enhance combustion efficiency, helping to minimize unburned methane emissions in burners, engines, and chemical reactors. 

Specifically, Swiss-roll combustors are used in the oil and gas industry, where natural gas (primarily methane) co-generated during production, storage, and distribution is burned off in a process known as flaring. Flaring helps depressurize extraction equipment, manage the excess gas release, and dispose of natural gas that is impractical to use. This matters because methane has a much higher global warming potential than carbon dioxide when released into the atmosphere. Thus, mitigating its release is a major step in combating anthropogenic climate change.

Classic flaring involving tall towers with flames on top is a great first step. However, these systems are susceptible to many issues (irregular gas flow, high cross-winds, or even complete loss of flame) that significantly reduce the combustion efficiency, leaving considerable amounts of methane still unburned.

Now, your next question might be: How does the Swiss-roll combustor minimize the issue of unburned methane emissions? To start, the flame is fully enclosed inside the device—but that’s not all. As shown in Figure 1, the combustor is composed of two channels arranged in a double-spiral pattern with a shared core in the center. Air enters the inlet channel and mixes with the methane, which is injected into the inlet channel upstream of the core, where the mixture is combusted. In this design, the inlet and outlet gases flow in channels adjacent to one another; thus, the hot exhaust gases preheat the incoming air, resulting in a higher reactant temperature. This preheated inlet air leads to higher burning rates and improved flame stability, which allows for much leaner air-fuel mixtures (Φ = 0.3) to be combusted and results in lower methane and NO_x emissions. This is in contrast to traditional combustor designs that require air-fuel mixtures of at least Φ = 0.5 to sustain combustion. All of these factors result in a more stable flame that burns cleaner and ensures significantly higher methane destruction, unaffected by external factors.

The influence of several design factors must be studied to make informed modifications to Swiss-roll combustors. These include identifying flame anchoring locations, recirculation regions, and areas of incomplete combustion, as well as studying the temperature distributions. However, experimentally analyzing these complex interactions is challenging due to a lack of visual access and intrusive measurement methods. As a result, obtaining spatially resolved information on the internal processes of the combustors is impractical through experiments alone.

In this context, computational fluid dynamics (CFD) simulations play an important role in analyzing such systems, enabling engineers to visualize phenomena that are difficult to measure experimentally, shed light on crucial information, and predict the formation of harmful emissions. In this blog, we will briefly discuss how our flagship product, CONVERGE CFD software, is used to simulate a Swiss-roll combustor test bench created by Advanced Cooling Technologies Inc., who collaborated with us by sharing their design details.

Experimental Setup

The experimental setup, as shown in Figure 2, consists of a 9 inch diameter Swiss-roll combustor made from siliconized silicon carbide, a unique ceramo-metallic material that can operate up to surface temperatures of 1350°C. An array of thermocouples and pressure transducers is placed at key locations to measure temperatures and pressure drop in the combustor. A gas analyzer at the exhaust records the concentrations of O₂, CO, and NO_x, with gas chromatography sampling used to measure CH₄ at steady state. Dedicated air and fuel mass flow controllers are installed to control the flow of reactants.

Simulation Setup

Figure 3A shows the model simulated in CONVERGE. Two operating conditions were considered: an ultra-lean mixture (Φ = 0.25) and a higher equivalence ratio mixture (Φ = 0.30)—still significantly below regular flammability limits—as a baseline. Figure 3B shows the CH₄ injector and the slits along its length. The size of the slits reduces over the length of the injector, similar to the experimental setup.

A conjugate heat transfer (CHT) setup was prepared to capture the heat transfer between the gases and the walls. Since conduction in solids occurs over much longer time scales than convection, directly coupled CHT simulations are typically computationally expensive. To address this, CONVERGE offers a feature called super-cycling, which accelerates CHT simulations by solving the fluid and solid domains at different, yet coupled, rates. CONVERGE’s SAGE detailed chemistry solver is used to capture the methane combustion and NO_x emissions. In this case, a reduced chemical kinetics mechanism proposed by Yuki et al [2], comprising 61 species and 509 reactions, is used.

To adequately resolve the flow, combustion, and heat transfer while minimizing computational resource usage, CONVERGE’s Adaptive Mesh Refinement (AMR) is applied to automatically refine the cell size during runtime, from 8 mm down to 0.5 mm, based on temperature and velocity curvature. This strategy maintains sufficient resolution at and around the flame fronts without inflating the total cell counts.

Simulation Results

Looking at the results for Φ = 0.25, we can see that the highest temperatures (Figure 4A), velocities (Figure 4B), and CO concentrations (Figure 5A) occur around the highest slit of the fuel inlet tube. This suggests the most intense combustion takes place in this area. In contrast, looking at the volumes near the lower slits, we can see lower temperatures and higher concentrations of O₂ (Figure 5B), CH₄ (Figure 5C), and NO₂ (Figure 5D)—a gas that forms primarily at lower flame temperatures [3]. 

Additionally, in Figure 6, an isosurface of OH mole fraction is plotted to visualize the flame front. Across both air-fuel ratios, we can see a decrease in flame penetration, as well as lower flame temperatures, along the length of the tube. This is likely due to less fuel being injected along the length of the tube, resulting from the reduced slit sizes, which makes the mixtures leaner in the channel near the bottom of the tube.

Finally, the concentrations of gases at the outlet of the combustor were compared to the experimental data to determine the level of accuracy of the simulations. Table 1 lists the CH₄, O₂, and CO concentrations from the experiments and the simulations for Φ = 0.25 and Φ = 0.30. The measured and computed results show strong agreement, confirming that CONVERGE reliably captures the complex interactions globally.

To dive deeper into the methodology and extended validation results, check out the full paper: 3D Numerical Simulations and Experimental Validation of Swiss-Roll Combustor Using Detailed Chemistry, Adaptive Mesh Refinement and Conjugate Heat Transfer, presented at the 2025 AFRC Industrial Combustion Symposium.

Have a combustion simulation of your own that you’d like to run? We’d love to discuss how CONVERGE can add value to your analysis and design process.

References

[1] Mistry, Z., Radyjowski, P., Avanessian, O., Pomraning, E., Liu, S., Wijeyakulasuriya, S., Carlson, D., Chen, C.-H., Jensen, D., Lieberknecht, E., Rao, P., Agarwal, P., “Numerical Simulations and Experimental Validation of Swiss-roll Combustor Using Coupled 3D CFD, CHT and Combustion Chemistry,” AFRC Industrial Combustion Symposium, San Antonio, TX, United States, Sep 15-17, 2025.

[2] Murakami, Y., Wang, Q.-D., Liu, S., Zhu, Y., Wang, P., Maffei, L.P., Langer, R., Faravelli, T., Pitsch, H., Klippenstein, S.J., Bergthorson, J., Bourque, G., Wagnon, S., Senecal, P.K., and Curran, H., “C3MechLite: An Integrated Component Library of Compact Kinetic Mechanisms for Low-Carbon, Carbon Neutral and Zero-Carbon Fuels,” Combustion and Flame, 282, 2025. DOI: 10.1016/j.combustflame.2025.114410

[3] Wang, X., Dai, G., Yablonsk, G.S., Vujanović, M., and Axelbaum, R.L., “A Kinetic Evaluation on NO₂ Formation in the Post-Flame Region of Pressurized Oxy-Combustion Process,” Thermal Science, 25(4), 2609-2620, 2021. DOI: 10.2298/TSCI200415236W