Numerical Study Of Multi-wavelength Reconstruction Of Symmetric And Asymmetric Flame Parameters In Participating Medium

Environmental pollution is a major problem that all countries in the world need to face.The massive burning of fossil fuels will not only cause a high total amount of carbon emissions in the environment,but also cause serious air pollution.The combustion of particulate products is one of the main sources which causes of inhalable particulate.Accurate measurement of flame temperature field,particle concentration field and particle radiation parameters is an important part of achieving combustion diagnosis and accurately controlling the combustion process.However,due to the presence of dispersed particles(soot,carbon particles,etc.)of different particle sizes in the combustion products,these participatory medium can attenuate the radiant energy emitted by the flame,thereby affecting the temperature field,so that under the premise of the paricipatory medium,it is necessary to carry out high-precision flame parameter measurement research,which is of great significance for improving the combustion efficiency of energy-using equipment and reducing the emission of pollutants.The research on the radiation characteristics of burning particulate pollutants,especially the research on soot particles,is a hot spot in the field of international radiation heat transfer.The soot particles can emit radiant energy in the visible light band during combustion,and the tomographic reconstruction technology based on the radiation image can use the radiant energy as an input signal,and the accurate reconstruction algorithm can accurately measure the key parameters of the flame.It has the advantages of high precision,high dimensional,real-time measurement.However,as of now,the following problems remain in the technology: First,the absorption characteristics of the dispersion medium have an important influence on the flame radiation field.However,the current reconstruction algorithm based on self-absorption of soot lacks support for multi-wavelength information of flame radiation,and too many simplification conditions limits its measurement accuracy under high particle concentration conditions.Second,for pulverized coal flames,the effect of particle scattering which affected by both particle size and wavelength cannot be simply ignored.Absorption coefficient and scattering coefficient are two key parameters for characterizing the radiation characteristics of particles.The traditional solution is obtained by direct calculation of Mie theory on the basis of known optical constants of particles.However,the radiation parameters of these dispersion media are closely related to the flame temperature,soot particle size distribution,and particle number density.Among them,the solution of the particle size distribution and the particle number density itself is a problem that has not yet been solved.The accurate results can hardly obtain by directly applying the Mie scattering theory.Third,the asymmetric flame is a typical flame in actual production,and the traditional reconstruction algorithm has very limited reconstruction accuracy and reconstruction speed for the asymmetric flame under finite projection,which limits the development of measurement technology in practical applications.Based on the above problems,this paper first establishes an emission tomography reconstruction model including medium emission,absorption and no scattering,and constructs an iterative algorithm with optical thinness as the initial value,by combining non-iterative least squares Tikhonov regularization algorithm and the multi-wavelength decoupling algorithm achieves high-precision reconstruction of the axisymmetric simulated flame temperature and soot volume concentration.The results show that the combination of multi-spectral decoupling technology and soot self-absorption model can greatly improve the reconstruction accuracy.Under the condition of relative noise of 5%,the mean square error of temperature and soot volume concentration is 0.15% and 4.91% respectively,which higher than those of the traditional two-color method and optical thin conditions.Then,on the basis of the previous,this paper focuses on the inverse reconstruction method of scattering coefficient,constructs a reconstruction algorithm for reconstructing the non-uniform temperature and scattering coefficient of anisotropic scattering flame,and introduces the idea of multi-stream based on generalized source term.By establishing a functional correlation between the soot radiation coefficient and the physical property parameters,the reverse reconstruction of the scattering coefficient is finally completed.Thirdly,aiming at the defect of low reconstruction precision and slow reconstruction speed when the traditional algorithm reconstructs asymmetric multi-peak flame under finite angle,this paper establishes a priori ARTTV algorithm based on particle swarm optimization algorithm to optimize weight parameters,using asymmetric radiation source as a test goal.The regression performance of the proposed algorithm is tested.The results show that compared with the traditional reconstruction algorithm,the ARTTV-PSO algorithm has better inversion accuracy for multi-peak asymmetric flame than ART,SART,regularization,LSQR and other traditional algorithms.The accuracy between them with 1%-10% difference.On the other hand,in order to improve the slow reconstruction speed of the iterative algorithm,this paper uses the ELM neural network method to fit the performance of the ARTTV-PSO algorithm to improve the reconstruction speed of the algorithm.It can be seen from the simulation results that the ELM neural network based on "ARTTV-PSO core" has almost the same reconstruction performance as the original ARTTV-PSO algorithm,but its reconstruction speed is about 300 times higher than the ARTTV-PSO algorithm.The research of this paper starts from the radiation characteristics of diffuse medium,and deeply studies the reconstruction algorithm under optical thick conditions.At the same time,the neural network algorithm for both the reconstruction accuracy and the reconstruction speed is proposed for the asymmetric flame target.These technologies provide an effective solution for the on-line measurement of actual industrial flames.