A novel parallel solver using decoupled chemical–physical integration for rapid multizone simulation of reactivity controlled combustion in engines
Lopullinen julkaistu versio - 4.09 MB
Salahi, M.M, Moradi, J., Golbaghi, K., Heidarabadi, S., Andwari, A., Hyvönen, J., & Mikulski, M. (2026). A novel parallel solver using decoupled chemical–physical integration for rapid multizone simulation of reactivity controlled combustion in engines. Energy conversion and management X 30. https://doi.org/10.1016/j.ecmx.2026.101865
© 2026 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/)
Lataukset7
Pysyvä osoite
Kuvaus
This study presents a novel acceleration strategy for the University of Vaasa Advanced Thermo-kinetic multi-Zone (UVATZ) RCCI combustion model. The method decouples stiff chemical kinetics from interzonal physical transport processes using an implicit–explicit integration strategy, where chemical reactions are solved implicitly while transport phenomena are handled explicitly. This approach reduces the size of the Jacobian matrix, enables zone-wise parallelization, and allows tailored numerical solvers for different equation classes. The solver is implemented in C++ using the Cantera framework and validated using three operating points from a Wärtsilä W31DF dual-fuel marine engine. For a 13-zone configuration using a 54-species/269-reaction kinetic mechanism, the accelerated solver achieves 14.7–19.3 × speed-up compared to the baseline implicit solver while maintaining excellent agreement in combustion predictions. Key engine indicators including peak pressure, IMEP, CA10, CA50, and net heat release remain within 1.5% deviation from the baseline results. With a larger mechanism containing 143 species and 746 reactions, the acceleration increases to approximately 24 × . Furthermore, the proposed method exhibits near-linear runtime scaling with the number of zones, in contrast to the cubic scaling behavior of the baseline solver. For a 40-zone configuration, the method achieves up to 249 × reduction in runtime, enabling high-resolution multizone simulations that would otherwise be computationally impractical. The proposed framework substantially improves the computational efficiency of detailed RCCI multizone simulations while preserving predictive accuracy. This capability enables large parametric studies, transient engine cycle simulations, and model-based design applications for advanced low-emission marine engines.
Emojulkaisu
ISBN
ISSN
2590-1745
Aihealue
Kausijulkaisu
Energy conversion and management X|30
OKM-julkaisutyyppi
A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä (vertaisarvioitu)