Study On Pattern Recognition's Application In Average Paticle Size Measurement By Acoustic Emissions In A Reactor

At present, granular materials holding an increasingly important position in modern industry, the building of national defense and high-tech fields,are widely used in sectors like medicine, chemical engineering, metallurgy, electronics, machinery, light industry, construction and environmental protection.Many important characteristics of granular materials are determined by parameters including the average particle size and particle size distribution, for example, The particle size distribution of white granulated sugar affects the quality of crystal group; The particle size of a cement influences its setting time; The pigment's particle size can affects its color, hide and settling characteristics; Particle size of phosphor plays an important role in deciding brightness and clarity of televisions, monitors and other display screen; Catalysts' particle size also partly determines their catalytic activity. In addition, the particle size also has a great influence on taste of food, effects of medicines, sintering of powder metallurgy and explosion property parameters of explosives. Thus, along with the development of science and technology, the technology of particle size measurement by universal attention, has been gradually developed into an important branch of modern surveying.There are many ways to size particles. More than 200 devices have been developed and produced based on different work principles. And new particle size measurement methods and measuring equipment are emerging continuously. Sieving analysis, microscopic analysis, sedimentation analysis and electric induction are traditional particle size measurement technologies, while light scattering, mass spectrometry and measurement based on Brownian movement developed in the recent years are novel ones for particle size measurement.Although the particle size measurement methods and measuring equipment have made a great progress. The on-line particle size measurement technologies in chemical industry remained up in the air. Because the major problem of particle sizing is sampling. But chemical equipments are usually of big size. Particles are in complex movement. Due to the existence of reactions, measuring devices often have to face the harsh temperature and pressure conditions as well as the dense dust. The monitor of particle size in a chemical equipment shows its outstanding difficulties: bad environment surrounding, the real-time requirements, online sampling and proper sampling.Particle size measurement by passive acoustic emissions is a new particle sizing technology based on acoustic emissions. With sensitive detection, environmental security, non-contact, non-invasive and real-time online features, it is particularly suitable for chemical equipments.Based on two kinds of most commonly used reactors (gas-solid fluidized bed reactors, stirred tank reactors), application of pattern classification in average particle size measuring technology by passive acoustic emissions are studied. Passive acoustic emissions are collected and converted to energy feature vectors with wavelet packet decomposition. By using the correlative component analysis, the most informative classification characteristics are extracted which are then used to determine average particle size by least squares support vector classifiers. The result shows that the system made of CCA-LSSVM(HVH) method can measure average particle size well. In the gas-solid fluidized bed reactor and stirred tank reactor applications, the method is proved of a good applicability.