Automated construction of re duce d mechanisms and additive reaction modules

A bottom-up approach to assemble reduced combustion kinetics mechanisms is proposed as an alternative to conventional reduction techniques. Rather than relying on simulations using detailed mechanisms to identify the set of species and reactions to include in the reduced mechanism, the proposed building algorithm follows an add-as-needed approach, in which reduced mechanisms are progressively augmented with individual reactions carefully selected among a restricted list in order to properly capture combustion dynamics in increasingly varied operating conditions. The algorithm is first described in details, and its characteristics and performance are explored through several examples. In a first example, reduced mechanisms able to capture methane/air auto-ignition in a constant volume homogeneous reactor are built, and compared to those generated with a conventional graph-based reduction technique. In the second example, the selection behavior of the algorithm is explored at the medium (methane) and large (heptane) mechanism scale, showing some computational advantage in using a building, bottomup approach. Finally, the flexibility of the algorithm to add, onto a reduced mechanism, kinetic pathways that were initially not considered in the reduction is demonstrated, using the addition of a reduced representation of NO x pathways on a previously obtained methane oxidation reduced mechanism as example. The algorithm is found to yield similar results compared to reduction techniques informed by detailed mechanisms, while providing increased efficiency and flexibility to the end-user. (c) 2021 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

Automated summary

A bottom-up approach is proposed to assemble reduced combustion kinetics mechanisms, showing similar results to reduction techniques informed by detailed mechanisms but providing increased efficiency and flexibility. The algorithm augments reduced mechanisms with carefully selected individual reactions, demonstrating computational advantages in various operating conditions. Additionally, the flexibility of the algorithm to add kinetic pathways not initially considered in the reduction is highlighted.