Study Of Laser Granulometric Analysis Algorithm

The design and programming of the calculation algorithm is the core part of measuring particle size distribution by light scattering technology. A detailed analyse is gived to the light scattering particle size distribution and the inversion algorithm of tiny particles based on the granulometric principle of Mie scattering theory. Algorithms is improved in order to advance the computation speed and testing accuracy of laser sizers.The research on the following aspects is completed in the paper:Firstly, Mie scattering theory is studied after the light scattering theory and principle of laser sizers are mastered.The scattered light intensity distribution functions of individual and clusters particle are analysed mainly. Then mathematical model of particle scattering is established, which is based on the laser sizers'principle.Secondly, particle inversion algorithms are researched.The characteristics of dependent algorithms and independent algorithms are summarized,and the theoretical validation is performed. The range optimization of dependent algorithms is expanded and the speed optimization is improved by adding the genetic algorithm to dependent algorithms. The stability and computing speed of independent algorithms are improved by fuseing the advantages of both Non-negative Philips-Twomey(NNPT) and Least Mean Square (LMS).Thirdly, laser sizers'software design is completed by the VisualC++. Software structure is designed according to software features,which include several modules,such as parameter setting module, serial port control module, data acquisition module, data processing module, displaying results module and saving results module.Finally, the experimental tests of standard samples are completed after building experimental system.The experimental system and the design of entire software system of laser sizers is finished in the paper. The experimental results of standard samples testify that the measurement error of dependent algorithms is less than 6%, and the computing speed and accuracy of independent algorithms are improved, and the computing time is controled within one minute by the method which combinates independent algorithms with filter method.