Accelerating CFD simulations of reactive flows via SPARC

In the intricate world of Computational Fluid Dynamics (CFD), simulating reactive flows with complex kinetic mechanisms has always been a computational challenge, due to the hundreds of species and thousands of reactions involved. However, the innovative Sample-Partitioning Adaptive Reduced Chemistry (SPARC) technique is changing the game. This blog post delves into the SPARC methodology, its benefits, and why it’s worth considering for your next CFD simulation project.

The Challenge

Traditional numerical simulations of reactive flows are notoriously CPU-intensive and time-consuming. The detailed kinetic mechanisms involved in such simulations lead to substantial computational costs, making high-fidelity simulations of realistic combustion systems a challenge.

What is SPARC?

SPARC stands for Sample-Partitioning Adaptive Reduced Chemistry. It’s a novel approach designed to mitigate the computational cost of detailed numerical simulations without sacrificing accuracy. The technique involves partitioning the thermochemical space for the generation of locally reduced mechanisms, significantly reducing the number of species and reactions at any given point in the simulation while maintaining the essential physics.

How Does SPARC Work?

• Dataset Generation: SPARC begins with generating a training dataset from multi-dimensional simulations of the system.
• Partitioning Thermochemical Space: The high-dimensional space spanned by the simulations is partitioned using clustering algorithms, identifying groups of similar points (clusters).
• Generation of Reduced Mechanisms: For each point in the dataset, a reduced mechanism is generated. These mechanisms are then combined for each cluster, creating a library of reduced mechanisms.
• Adaptive Simulation: During the actual CFD simulation, the most appropriate reduced mechanism is selected on-the-fly for each grid point based on its current state.

Benefits of SPARC

• Computational Efficiency: SPARC significantly reduces the CPU time and computational resources needed, enabling faster simulations.
• Accuracy: Despite the reduction in complexity, SPARC maintains a high level of accuracy in the simulations, as demonstrated in multiple studies comparing it with detailed mechanisms.
• Flexibility: The technique has been validated for both steady and unsteady laminar flames, showing its adaptability to different types of reactive flow simulations.

Why Consider SPARC?

In the pursuit of more efficient and accurate reactive flow simulations, SPARC presents a compelling solution. Its ability to maintain the delicate balance between computational efficiency and simulation accuracy makes it an invaluable tool for researchers and engineers alike. SPARC represents a significant stride in the field of CFD simulations. As we continue to push the boundaries of what’s possible in reactive flow modeling, techniques like SPARC will play a pivotal role in making these complex simulations more accessible and actionable for a wider range of applications.

References

• D’Alessio, G., Parente, A., Stagni, A. and Cuoci, A., 2020. Adaptive chemistry via pre-partitioning of composition space and mechanism reduction. Combustion and Flame, 211, pp.68-82, DOI: 10.1016/j.combustflame.2019.09.010
• D’Alessio, G., Cuoci, A., Aversano, G., Bracconi, M., Stagni, A. and Parente, A., 2020. Impact of the partitioning method on multidimensional adaptive-chemistry simulations. Energies, 13(10), p.2567, DOI: 10.3390/en13102567