Flow Properties Of Gas-Liquid And Gas-Liquid-Solid Multiphase Microreactors

Multi-phase flow pressure drop in microchannels is an important parameter to design and scaleup of microreactors. Flow patterns in microchannels greatly influence this multi-phase flow pressure drop. Although gas-liquid two-phase flow properties in microchannels have been extensively studied recently, it is not solved yet to classify the related flow patterns and to predict the pressure drop accurately. So far no report was found about the flow preoperties of three-phase flow in microchannels.In the thesis two-phase and three-phase flow properties were studied in microchannels with the diameter of 400μm. The liquid phase used was water, ethanol, CMC aqueous solution with two different concentrations, the gas phase was nitrogen, and the solid was silica nanoparticles.Flow patterns, such as bubbly, slug, slug- annular, chum and annular flow, were recorded for both two-phase and three-phase through high-speed photography. BP neural network was proved to predict the two-phase flow patterns in microchannels using WeLS,WeGS,ReLS,ReGS andρρP GSL as input parameters, its identified accuracy is 96.7%. While for three-phase flow in microchannels, BP neural network can also predict flow patterns with an identified accuracy 97.8% using WePSLS,WeGS,RePSLS,ReGS,ρρP GSL,WS as input parameters.Pressure drop prediction models were studied for different flow regions in gas-liquid microchannels according to indentified results of neural network. For the chum and bubbly flow, average kinetic energy model can be used to estimate flow pressure drops, the corresponding proportional coefficient was modified as: . For the slug flow, the pressure drop can be estimated by combining the equation of Kreutzer with the neural network model about the slug length. For slug-annular flow, Dukler model can provide good estimation of pressure drops in the transforming region between slug and slug-annular region, while average kinetic energy model is good to estimate the pressure drop in the transforming region between slug-annular and annular region. For annular flow, the model based on factor is good to estimate the pressure drop.In gas-liquid-solid microchannels, average kinetic energy model is good to estimate the pressure drop for the chum, bubbly, slug and slug-annular flow with a proportional coefficient of 0.015, which is independent of silica concentration. For annular flow, the model based onΦ_vo² factor is good to estimate the pressure drop.