Signal Processing Methods And Simulation Of Microwave Solid Flow Meter

In the producing and controlling process of modern industry, measuring the solidmass flow accurately and real-timely is always being a challenging problem in industryand science. As one kind of these measuring devices, Microwave Solid Flow Meter hasunmatched advantages, such as low cost, loose environment requirement, and strongreal-time ability. However, there are some problems in the solid flow measuring withmicrowave, such as an inadequate theory and outdated signal processing methods,which result in the low accuracy of products in the market.Aiming to solve some of the problems existing in Microwave Solid Flow Meter,this paper will give a deep analysis and research on the signal processing methodsinvolved, with a background of measuring the mass flow of granular solid inpneumatically conveying pipes and on the basis of analyzing the theory and techniqueof traditional measuring methods. The main work can be divided into four parts:1. The research results domestic and overseas are synthesized and the physicalbasis of measuring solid flow with microwave is studied theoretically. The linearrelation between echo power and square of solid mass is proved according to theory ofelectromagnetic scattering. Also, mass flow formula of velocity-density method isderived and the parameters determining ways are researched.2. Some reasonable assumptions are proposed with the specific applicationscenario as well as their space geometry relations. The stochastic characteristics areanalyzed, based on which the Doppler echo model is established. Then a simulativesignal sample is generated on computer.3. Considering the fact that the Signal to Noise Ratio of echo is low and noise isusually ignored in traditional methods, echo de-noising process is studied. The waveletde-noising method is introduced into the signal processing. By means of sufficientsimulations, it is found that wavelet de-noising has a better performance than thede-nosing methods based on frequency-domain filtering. Moreover, the results ofde-noising with different wavelets in different decomposing levels are comparedthrough simulations, and the best choices of these factors are selected according to the standard of maximum Signal to Noise Ratio.4. The traditional parameter extraction approach is improved by means of spectrumestimation, due to which the steps of parameter extraction are simplified. During thespectrum estimation, classical methods and AR parameters model methods aresimulated. At last, the Doppler bandwidth is also estimated to enhance the precision ofparameter extraction.The simulation results illustrate that the signal processing methods studied in thispaper have a better performance and may be applied in real occasion to increase theaccuracy of Microwave solid flow meter.