Inversion Algorithm Applying In The Particle Size Distribution Measured By Total Light Scattering

Recently, there are several popular methods to detect particle size, most of them canonly be used in a narrow range, and have a complex calculation process and they must workin the high requirements of the environment. And with the total light scattering particle sizingtechnique, the particle size distribution can be obtained by the measurements of lightextinction data at multiple wavelengths, this method has a high detect speed and can be usedin wide range, and the experiment’s operation is simple. And the most difficult of thistechnique is actually the solution of Fredholm integral equation of the first kind; it can besolved by inversion algorithm. This thesis studies the inversion algorithm applying in theparticle size distribution of particle system measured by the total light scattering.This thesis gives classification and development direction of particle measurementmethod firstly, and focuses on the analysis of the light scattering mechanism andmeasurement theory-Mie Theory, and then systematic analysis the feasibility of the Mietheory used in particle detecting. After discussing the inversion algorithm classificationmethod, it can be divided into two types of the inversion algorithm of grain size by whetherthe particle size distribution function is determined: algorithm of independent modelinversion algorithm and the dependent model inversion. Algorithm of dependent modelinversion algorithm has a great advantage in studying on those who do not know the particlesize distribution function, but it is more complex and slow than the independent model, sowe select the dependent model which is more suitable for on-line measurement as the mainresearch object. Dependent inversion algorithm including spherical particle size distributioninversion based on similar anomalous diffraction, matrix method, classification inversion,R-R unimodal distribution. This thesis mainly studies the R-R distribution. By using theoptimization algorithm, the particles size distribution is calculated based on the dependentmodel algorithm. Two results are agreed well, less than ten percents and proved it is feasible that R-R distribution inversion algorithm to measure the particle size distribution, and it has afast calculation speed and it is suitable for on-line measurement.