Sensitivity Analysis In The Study Of Combustion Mechanism Reduction And Uncertainty Propagation

Combustion is a very complex reaction process.It is widely used in exploring the high efficiency and low pollution combustion strategy.So it is necessary to analyze the combustion mechanism model.Through the analysis of combustion mechanism model,characteristic reaction of different fuels can be obtained,which is helpful to understand the combustion characteristics of fuel;Important reactions in different combustion mechanism models can also be obtained,which is helpful to reduce the mechanism scale;The uncertainty effect of different reaction model parameters can also be obtained,which is helpful to explore the source of the prediction uncertainty of combustion mechanism and reduce the prediction uncertainty.In this paper,the mechanisms of 1-hexane,1-hexane and 1-hexanol are analyzed by global sensitivity analysis.To investigate the oxidation and combustion performance of practical fuels,surrogate fuels including various types of fuels are usually introduced.The unique functional groups of different fuels dominate the fuel oxidation behaviors of different fuels,thus it is crucial to take account of the impact of fuel function groups for the development of the skeletal chemical mechanisms of surrogate fuels.In this work,by integrating the reaction class-based global sensitivity analysis and the decoupling methodology,a skeletal chemical mechanism of fuels is built,and the influence of the functional group was specially considered in the construction of the chemical mechanisms.First,the reaction class-based global sensitivity and path sensitivity analyses were employed to recognize the important reaction classes in the fuel-related sub-mechanism,and the reaction classes relevant to the fuel function group were identified.Second,a representative reaction was selected from each important reaction class by the rate of production analysis,and the skeletal fuel-specific sub-mechanism was obtained.Third,the initial skeletal chemical mechanism of fuels was formed by assembling the skeletal fuel-specific sub-mechanism with a detailed C₀–C₁ sub-mechanism and a reduced C₂–C₃ sub-mechanism based on the decoupling methodology.Finally,the optimization aiming at the ignition delay times and the concentrations of fuel,H₂O,CO,and CO₂was conducted based on the genetic algorithm by tuning the reaction rate coefficients in the fuel-specific sub-mechanism within their uncertainties to enhance the performance of the skeletal mechanism.Using the above method,a skeletal chemical mechanism for 1-hexane,1-hexene,and 1-hexanol was established containing 72 species and 243 reactions.The validation results indicated that decent consistency between the simulated and experimental data in premixed and opposed flames,jet-stirred reactors,and shock tubes was achieved for the three fuels over wide operating conditions.Moreover,the unique oxidation behavior of 1-hexane,1-hexene,and 1-hexanol was captured by the present skeletal mechanism due to the identification of the functional group reactions.Secondly,a series of uncertainty analysis methods based on sensitivity are used to analyze the mechanism of dimethyl ether,and then parameters that have a great influence on the laminar flame and the prediction of flame structure can be identified.In order to comprehensively explore the specific influence of flame speed and species concentration caused by the uncertainty of different parameters,the mechanism of dimethyl ether is analyzed from both macroscopic and microscopic levels under different operating conditions.The first-order screening method and the first-order uncertainty analysis method（Truanyi method）are employed in the macro-analysis.Then the reaction,thermodynamic and transport species that have a greater influence on the laminar flame speed and species concentration can be obtained.Based on the statistics of result of the Truanyi method,the uncertainty of predict laminar flame speed is mainly from the kinetic parameters.And the phenomenon of uncertainty propagation is more obvious under the high equivalent ratio,high dilution ratio and high pressure conditions.Through Monte Carlo analysis,compared with species concentration,the uncertainty propagation for flame speed is more obvious.In the micro-analysis,ANN-HDMR analysis was carried out for limited operating conditions and limited parameters.The results show that the reactions of H+O₂=O+OH shows different performance in macro-and micro-sensitivity analysis,which is caused by the discretization of uncertain space.Besides,the phenomenon of uncertainty propagation differs under different operation condition.Such as,the reaction CH₃+H+M=CH₄+M is of higher importance under high pressure conditions.This is because as the pressure increases,the rate constant of the reaction increases,so the uncertainty caused by the reaction increases.