The applications of sensitivity analysis by gradient techniques to study detailed combustion chemistry will soon become more common as a result of the recent availability of increasingly efficient computer codes. This paper provides a compendium of the various uses and interpretive aspects of sensitivity gradient techniques applied to chemical kinetics. Illustrations are provided through analysis of the frequently studied elementary kinetic mechanism for the CO/H2O/O2 system. Elementary first and second order sensitivity coefficients, obtained by the Green's function method, are interpreted and discussed in detail for a fuel lean and dilute system reacting in N2 at 1100K and 1 atm. From the results of linear sensitivities, the effects of variations in initial conditions and uncertainties in rate constants and equilibrium constants on the model predictions are determined. Second order sensitivities are shown to yield valuable information on how these linear sensitivities vary when an initial condition or rate constant is varied. Elements of the Green's function matrix are shown to have a memory function interpretation, and results of small perturbations in each species concentration at various times during the kinetics on all other species concentrations at later times are presented. These latter sensitivities are unique in that they, in principle, are measurable in the laboratory. The results of feature sensitivity analysis (i.e., the effects of variations in input parameters on specific kinetic predictions, such as induction and kinetic periods) are presented for a wide range of equivalence ratios. Finally, elementary sensitivity analysis is shown to be extremely useful in experimental design as well as model development and evaluation.
All Science Journal Classification (ASJC) codes
- Chemical Engineering(all)
- Fuel Technology
- Energy Engineering and Power Technology
- Physics and Astronomy(all)