An outline of the Ant Colon Reductionmethodology and its applications
Tailoring of mechanisms
The ACR method is not the only way to reduce mechanisms, a range of methods have been developed during the last few decades and many of them are available in the open litterature. Our ACR method has two strong advantages: It can produce smaller mechanisms than the other methods, and it can be used to tailor mechanisms of different size and accuracy. The tailoring process makes it much more releveant for industiral applications: you get exactly what you need, no more , no less!
A brief description
The Tailored Chemistry team has developed the Ant Colony Reduction (ACR) algorithm for reduction of the number of reactions in extensive chemical mechanisms. ACR is an adaptation of the family of algorithms called Ant-Colony Optimisation (ACO), which are used to find an optimal path from a start node to an end node through a graph of edges and nodes. The problem to solve is “how to minimise the number of species and reactions, while maximising the agreement with chosen targets?”. In this case, the nodes are species and the reactions are edges.
Which species and reactions that are important for a given target might not always be known and comprehensive mechanisms can contain tens of thousands of reactions and thousands of species. The ACR approach uses ants in a similar fashion as any ACO algorithm. The start node is an initial species and the end node is a product species. Ants walk along the reactions with probability governed by a static value related to rate of production of the comprehensive mechanism (instead of inverse distance) and a dynamic pheromone value. When an ant has found the end node, the reactions and species it passed forms a reduced mechanism.
The ACR method was developed for reduction of complex mechanism describing combustion chemistry of biofuels. It was implemented to create small mechanisms that accurately reproduced the results of complex mechanisms with respect to ignition, flame propagation and extinction. First academic projects concerned combustion of small alcohol biofuels, where ACR produced the smallest available mechanisms, as described in scientific publications 1 and 5 below.
It has since then been further developed for alkanes that are surrogates for petroleum fuels, like n-heptan presented in publication 4, and biodiesel model fuels like methyl decenoate described in publication 6. The figure below is a summary of publication 6 where the ACR mechanism has a size of 0.7% of the complex mechanism. The size of the reduced mechanisms vary dependent on the demands for accuracy and the range of conditions that the mechanism should be able to describe, but commonly the number of species are in the range 15 to 50 and number of reactions 30 to 200. The lower part of these ranges are for small hydrocarbon and alcohol fuels, while the larger mechanisms are for long chain fuels like the biodiesel surrogates.
The ACR method can be applied to any chemical system and it has, in an academic context, been applied to models of air pollution chemistry in street environments, see publications 1 and 3 below. The used of reduced mechanisms for air pollution chemistry enable CFD simulations of street canyon environments, including both turbulence and chemical reactions. The reduced mechanisms are used in a scientific project at Lund University, where researchers from the division of Combustion Physics and the department of Energy Sciences collaborate to bring the best chemistry models and the state of the art CFD to the air quality modeling community. A long term goal is to make CFD of street environments become efficient enough to aid in city planning.
Recently projects have been initiated to use ACR to reduce chemical mechanisms for modeling of catalysis process and for battery fires. These are new and challenging fields that we hope to add to our commercialcapabilities within short.
Future challenges are simulations of biomedical systems and industrial chemical reactors. In short, we can
reduce the size of any chemical reaction system.
Tailored chemical mechanisms for simulation of urban air pollution
Joelsson, Lars Magnus T., Christoffer Pichler, and Elna JK Nilsson. WIT Transactions on Ecology and the Environment 230 (2018): 165-176.
Reduced kinetic mechanism for methanol combustion in Spark-Ignition engines
Pichler, Christoffer, and Elna JK Nilsson. Energy & fuels 32.12 (2018): 12805-12813.
Tailored reduced kinetic mechanisms for atmospheric chemistry modeling
Joelsson, L. M. T., C. Pichler, and E. J. K. Nilsson. Atmospheric Environment 213 (2019): 675-685
Analysis of important chemical pathways of n-heptane combustion in small skeletal mechanisms
Pichler, C., Nilsson, E. J. K, Energy & Fuels, 34 (2019): 758-768
Pathway analysis of skeletal kinetic mechanisms for small alcohol fuels at engine conditions
Pichler, C., Nilsson, E. J. K., Fuel, 275 (2020): 117956.
Composition of Reduced Mechanisms for Ignition of Biodiesel Surrogates
Pichler, C., Nilsson, E. J. K., Fuels, 1 (2020):15-29.