Automatic Reduction Procedures for Chemical Mechanisms in Reactive Systems

Abstract

To deal with complex physical and chemical processes in reactive systems, such as combustion processes, it is necessary to find methods that simplify modelling in such a way that it becomes both more comprehensible and practically useful. In the work reported here a method for automatically reducing chemical reaction mechanisms has been developed for use in computer simulations of various combustion systems. The present method opens the way here for the application of adaptive kinetics, in which mechanisms are reduced at each operating point which allows a careful and efficient reduction to be carried out. The reduction in question is based on lifetime analysis and an assumption of quasi-steady state, abbreviated as QSSA. Species with short chemical lifetimes and/or with only a minor influence on the chemical system are selected to be in steady state, their concentrations being dealt with by means of simple algebraic equations. The concentrations of these species are calculated by being automatically implemented in FORTRAN subroutines. These routines calculate both the steady state concentrations and the remaining source terms for the non-steady state species by use of an internal iteration loop. Different procedures for selecting the steady state species are investigated for a range of different combustion systems. The simplest selection parameter is to use the chemical lifetime of the species only. By also introducing a maximum element mass fraction and the maximum enthalpy that a steady state species can occupy, violations of the required mass and energy conservations when applying QSSA could be avoided. This method was tested for a number of different physical setups or conditions, such as a chain-reactor sequence, use of a plug flow reactor (PFR) followed by a perfectly stirred reactor (PSR), a spark ignition engine (SI engine), a homogenous charge compression ignition engine (HCCI engine), and various flame configurations. For improved ranking a combined lifetime- and species sensitivity measure, termed here the level of importance (LOI), was developed as a selection parameter for automatically capturing species that although having short lifetimes, are of importance for the chemical system. It was shown to successfully identify the steady state species present in a mechanism. In order to enhance the efficiency of the reduction procedure, a method for the application of adaptive kinetics is developed and is applied to a case of ignition. At each operating point, a lifetime analysis is performed and species are moved in and out of steady state. In this manner, only true steady state species are considered throughout the computation. During ignition, the number of species treated in detail is larger than during afterburning, where most species are in steady state or contain small mass fractions.



Emphasis has been placed on increasing the efficiency of the reduction method and presenting a thorough validation. A method is developed for validating the selection parameters and testing for self-consistency. The method's performance is also compared with that of another, well established reduction procedure, the computational singular perturbation method, CSP.

Abstract (Swedish)

Popular Abstract in Swedish

En av dagens största utmaningar är de stadigt ökande utsläppen av förorenande avgaser och drivhusgaser. Den största delen av dessa föroreningar kommer från förbränningsanläggningar och fordon. För att kunna minska utsläppen måste man förstå förbränningsprocesserna, och hur man kan kontrollera förbränningen. Utmaningen är att förbränningprocesser är mycket komplexa. Det handlar om att modellera turbulenta gaser under höga tryck och höga temperaturer. Bara för att beskriva de kemiska processerna i detalj krävs att man inkluderer kanske tusentals olika ämnen och tiotusentals olika reaktioner. Ingen dator idag kan klara av att lösa så komplexa system. Man måste därför hitta metoder med vilka man kan förenkla systemet så att man inte tappar viktig information i beräkningen. Det gäller att hitta parametrar, med vilka man kan rangordna de kemiske ämnen, så att man automatiskt kan välja bort de som inte är viktiga för beräkningen. Dermed kan ämnen och reaktioner som är av mindre betydelse för att beskriva kemin tas bort eller beskrivas på ett mindre nogrannt sätt.

• author: Lövås, Terese
• supervisor: [unknown] [unknown]
• opponent: Dr Baritaud, Thierry
• organization: Combustion Physics
• publishing date: 2002
• type: Thesis
• publication status: published
• subject: Atom and Molecular Physics and Optics
• keywords: Mechanism Reduction, Combustion, Physics, Emissions, Fysik, Fysicumarkivet A:2002:Lövås
• pages: 278 pages
• defense location: Sal B, Fysicum, Lund
• defense date: 2002-09-06 10:15:00
• ISBN: 91-628-5307-4
• language: English
• LU publication?: yes
• additional info: Article: Nilsson, D., Løvås, T., Amneus, P. and Mauss, F., Reduction of Complex Fuel Chemistry for Simulation of Combustion in an HCCI Engine, VDI-Berichte 1492, pp. 511-516, VDI Verlag, Germany, 1999.Løvås, T., Nilsson D. and Mauss, F, Automatic Reduction Procedure for Chemical Mechanisms Applied to Premixed Methane-Air Flame, 28th Symposium (Int.) on Combustion}, pp. 1809-1815, The Combustion Institute, 2000Soyhan, H.S., Løvås, T. and Mauss, F., A Stochastic Simulation of an HCCI Engine using an Automatically Reduced Mechanism, 2001-ICE-416, ICE-Vol. 37-2, 2001 Fall Technical Conference, ASME 2001.Løvås, T., Mauss, F., Hasse, C. and Peters, N., Development of Adaptive Kinetics for Application in Combustion Systems, 29th Symposium (Int.) on Combustion}, In Press, The Combustion Institute, 2002.Løvås, T., Mauss, F. and Mastorakos, N., Comparison of Automatic Reduction Procedures for Ignition Chemistry, 29th Symposium (Int.) on Combustion}, In Press, The Combustion Institute, 2002.Løvås, T. and Mauss, F., Selection Parameters for Steady State Species for Automatic Reduction of Chemical Mechanism, Submitted to Combustion Theory and Modelling, The Combustion Institute, 2002.
• id: 5539f0cd-8d61-421b-8f9e-cbe88f1157e3 (old id 464819)
• date added to LUP: 2016-04-01 17:10:40
• date last changed: 2018-11-21 20:47:14