Research On Inversion Algorithm For The Particle Size Distribution Of Non-Spherical Aerosol Particle

In the inversion of aerosol particle size distribution,We regard aerosol particles as spherical in generally.The Lorenz-Mie scattering theory is used to calculate the scattering characteristics of spherical particles,and inversion the aerosol size distribution.In reality circumstance,the composition of aerosol particles is complex and the morphology is different.Many shape of non-spherical particles is considered,it can make the results more suitable for the actual situation.Focusing on this problem,this paper proposes a Tikhonov regularization method improved by genetic algorithm to inversion the size distribution of non-spherical aerosol particles.Firstly,the T-matrix method is used to calculate the extinction,scattering,and absorption characteristics of non-spherical particles of different types,shapes and particle sizes.This paper compared the T-matrix result with the Lorenz-Mie scattering theory result,and verified the accuracy of the T-matrix method,which provides data support for the next step of inversion the size distribution.We determined the initial population,crossover probability,mutation probability and maximum evolution algebra of genetic algorithm by experiments.The optimal regularization parameter is obtained by calculating the fitness function.Under dust the conditions with particle size ranging from 0 to 3μm,this paper inversion size distribution of ellipsoidal particles and cylindrical particles with different shapes.Then,the non-spherical particles size distribution results inversion by Tikhonov regularization method improved by genetic algorithm,it compared and analyzed with the non-spherical particles size distribution results inversion by the GCV method.the feasibility of this algorithm is verified.It is found that no matter how the shape of the non-spherical particles changes,the inversion result of Tikhonov regularization method by genetic algorithm is better than GCV method.The results of the size distribution and volume distribution are more close to the ideal distribution.Through the analysis of anti-noise,the mean square error obtained by the Tikhonov regularization method improved by genetic algorithm is still less than the mean square error obtained by the GCV method,it is proved that the Tikhonov regularization method improved by the genetic algorithm is more stable than the GCV method.Finally,we can know from the experiment that the Tikhonov regularization method improved by genetic algorithm is an effective method to inversion particle size distribution,and the inversion result is better.