Optimal automatic reaction and species elimination in kinetic mechanisms

Journal Article · Wed Oct 15 00:00:00 EDT 2008 · Combustion and Flame

DOI:https://doi.org/10.1016/J.COMBUSTFLAME.2008.03.004· OSTI ID:21116101

Mitsos, A ^[1]; Oxberry, G M; Barton, P I; Green, W H ^[2]

---

1. RES Group/Numerica Technology, Cambridge, MA 02142 (United States)
2. Massachusetts Institute of Technology/Department of Chemical Engineering, Cambridge, MA 02139 (United States)

A known challenge for the simulation of reacting flow systems is that detailed chemical mechanisms contain hundreds to thousands of species and thousands of reactions, leading to high CPU requirements despite the use of state-of-the-art solvers. For specific conditions of interest (temperature, pressure, and composition), smaller mechanisms can predict the chemistry relatively accurately. One possibility for obtaining such mechanisms is species elimination from a detailed mechanism. Here, an automatic method for kinetic model reduction by simultaneous reaction and species elimination is proposed, based on an integer linear program (ILP) formulation. The solution of the ILP is an optimally reduced kinetic mechanism that reproduces the predictions of a reference mechanism within prespecified tolerances for finitely many reference points in the state space. The method is applied to generate optimally reduced models for isobaric, adiabatic homogeneous combustion. Case studies are presented for the combustion of n-heptane. Comparisons between the full and reduced models are shown and the tradeoff between species and reaction elimination is discussed. Tolerances in the ILP formulation control the error introduced by the model reduction. For increasing acceptable error, more species and/or reactions are eliminated. A method of quantifying this tradeoff between approximation error and reduction achieved is proposed, based on multiobjective optimization, and demonstrated in a case study. The effect of variable initial conditions is investigated. The mechanisms generated achieve significant reduction in the CPU requirement and can accurately predict the trajectories of the state variables (species mass fractions and temperature), as well as other metrics of interest, such as ignition time delay. (author)

• OSTI ID: 21116101
• Journal Information: Combustion and Flame, Vol. 155, Issue 1-2; Other Information: Elsevier Ltd. All rights reserved; ISSN 0010-2180
• Country of Publication: United States
• Language: English

Related Subjects

37 INORGANIC
ORGANIC
PHYSICAL AND ANALYTICAL CHEMISTRY
COMBUSTION
HEPTANE
COMPARATIVE EVALUATIONS
APPROXIMATIONS
COMBUSTION KINETICS
OPTIMIZATION
TOLERANCE
MATHEMATICAL SOLUTIONS
TIME DELAY
IGNITION
MASS
COMPUTERIZED SIMULATION
MATHEMATICAL SPACE
TRAJECTORIES
MATHEMATICAL MODELS
CALCULATION METHODS
ADIABATIC PROCESSES
Model reduction
Species elimination
Reaction elimination