Study On Prediction Of Performance Of Chemical Reactor Network For Combustor With High Precision

With the increasingly prominent role of computational fluid dynamics,the numerical simulation technology of combustor has been developed rapidly.In recent decades,chemical reactor network(CRN)model method,an efficient combustion chamber design method,has been developed.This method not only reduces the time which is required to compute but also can use detailed chemical reaction mechanism to predict pollutant emissions.As a result,the prediction of pollutant emission is more accurate compared with CFD.In order to predict the combustor performance under different conditions with high precision as soon as possible,a CRN model method is proposed.Firstly,through studying the chemical reactors,chemical reactors with turbulent features and geometry characteristics were proposed.At the same time,a prediction simulation software for combustor with high precision was developed,using a limited number of reactors to simulate combustor as well as predict combustion emissions.Moreover,the pollution emission prediction program,using C# as development language,is designed to analyze chemical reaction mechanism,chemical reaction kinetics calculation and so forth.The combustion models used in this paper include PSR and PaSR.Furthermore,the characteristics of the combustion flow field inside the engine combustor were analyzed based on the numerical simulation method.According to the flow field and temperature field results of CFD,the combustor was divided into several regions,where the temperature,reaction volume,pressure and other characteristic parameters were set.There is a need to build networks,which would be simulated by appropriate reactors separately.The original CRN was simulated and the effects of back flow and PaSR reactor on the simulation results were considered.In this paper,according to the experimental data under different working conditions,the geometry characteristics of each reactor were optimized by using genetic algorithm based on the exit temperature and NOx emissions of combustion.Two optimization algorithms are proposed,one is fixed parameter optimization and the other is parameterized optimization.Fixed parameter optimization,which is referred to optimize characteristic sizes of each reactor according to several different experimental data,makes parameters of optimized reactor remained constant and the network topology applicable to the combustor under any working conditions.Based on the results,it is easier to determine which reactor should be used and whether there is a possibility to consider back flow.Parametric optimization means that every reactor under each working condition will be optimized based on the results of fixed parameter optimization in order to obtain the volume of each reactor under different working conditions.The relationship between input parameters and the volume of each reactor was established.At last,this method was applied to SPRF combustor and LPP combustor respectively.It can be seen that the exist temperature and nitrogen oxide emissions of the combustor,which are obtained through simulation of chemical reactor network after optimization of fixed parameters,are in good agreement with the experimental data,indicating that the chemical reactor network is reasonable.The fitting relations between the input parameters and the reactor parameters through parameterization optimization are established and verified.The results will be indicated that fitting relationships are feasible and are able to predict the combustor performance with high precision.