In skeletal reduction of large chemical mechanisms, graph-based reduction methods such as directed relation graph with error propagation (DRGEP) are widely utilized. For the graph-based reduction methods, it is a common practice to pre-select a static set of target species for graph searching process. However, such practice faces some major challenges. The first one is that the reduction performance is strongly dependent on the selection of target species, a proper selection of which requires abundant experience and expertise. The second one is that some of the species in the static set can be consumed entirely and thus no longer important during some combustion period. In addition, a limited number of species in the pre-selected set of target species may not satisfy all sampled chemical states, which has the risk of missing some essential species as target species at certain sampled chemical states. In this study, an automatic dynamic target species selection (ADTSS) technique was proposed to replace the traditional pre-selection of target species and incorporated with DRGEP. After that, extensive validations against the detailed mechanisms and comparisons with utilizing static target species were conducted. It has been found that the ADTSS technique is reliable and significantly improves the reduction capability. Another great benefit of the ADTSS technique is that it removes the need of manually pre-selecting target species and their cut-off threshold on mass fraction, which makes it possible to obtain a good skeletal mechanism for researchers without much experience and expertise in mechanism reduction.
Read full abstract