Numerical Studies On Model-Independent Retrieval Algorithm In Mie Scattering Particle Sizer

Particle sizers, which make use of the Mie scattering theory as their theoretical basis instead of the conventional Fraunhoferal diffraction theory, are gaining more and more popularity, because of the high numerical precision and wide particle size span feasibility.Model-dependent algorithm is extensively employed as the default inverse method for retrieving particle diameter distribution of a particulate system. The inverse computation has since remained to be the key and challenge. The practical application is confined, despite of its simple and quick characters, to specifying some distribution function sets in advance. On the contrary, the model-independent algorithm can work out any sort of particle size distributions in theory without any pre-assumptions for the particle system. The model-independent means is, therefore, anticipated to be the default or trend of future generation particle sizers for data processing.Determining particle size distribution by measuring scattered light intensities in different directions is a typical inverse problem. Using the model-independent method will make the solution unstable because it is sensitive to small disturbance in the input data, which we call the ill-posed problem.In this dissertation work, we studied the model-independent algorithm in particle size distribution retrieving. The Phillips-Twomey regularization method and the Projection onto Convex Sets method (POCS) are investigated for acquiring the solutions. The thesis also discussed the parameter choice, criterions and applicability, and presented our simulation conclusion as well: the Phillips-Twomey method out-performs other methods in a much better way.