Design of Flexible Reaction Models

Rutgers University New Brunswick, New Brunswick NJ

Investigators

Abstract

Research: This research project addresses the problem of modeling simulation and optimization of realistic reaction systems involving a large number of reaction steps and different species. The specific objectives are the description of complex kinetic networks through reduced reaction systems, the determination of the range of validity in terms of the initial conditions of the reduced models and the problem of uncertainty propagation through the reaction network. The methodology that is developed will be applied to fossil fuels combustion but the techniques developed are applicable to a diverse range of chemical processes. Nitrogen oxides (NOx) in the atmosphere cause photochemical smog, the destruction of the ozone layer and global warming. Atmospheric NOx increases have been attributed to the combustion of biomass and fossil fuels, systems that are mathematically represented by very complex kinetic reaction networks. Accurate kinetic models could help identify the parameters responsible for the production of the pollutant NOx in the atmosphere and be used to reduce their formation. The PI will develop methods of kinetic model reduction to reduce the computational complexity associated with the consideration of the detailed kinetic mechanisms of such systems. The approach is to develop a comprehensive framework for the analysis of uncertainty effects of complex kinetic models, quantify the range of validity of a reduced model, and determine more flexible reduced kinetic models with respect to kinetic parameters and initial conditions. This will be achieved through the following steps: - The development of an efficient approach to quantify the range of validity in terms of initial conditions of a reduced kinetic model based on the ideas of uncertainty analysis - The investigation of various uncertainty propagation techniques to evaluate the effects of uncertainty in kinetic model parameters on systems outputs - The integration of the uncertainty analysis procedure whtin the framework for kinetic model reduction that will result in the determination of reduced kinetic model with the required range of validity and pre-specified flexibility. Impact: Besides the training of undergraduate and graduate students at Rutgers, the PI is involved in SUPER (Science for Undergraduates a Program for Excellence in Research) at Douglas College for Women. It is expected that one student affiliated with SUPER will work on this project during each summer.