DOI: 10.1615/ICHMT.2022.CONV22
ISSN Online: 2642-3499
ISBN Online: 978-1-56700-523-3
PREDICTION OF FLOW PATTERNS OF LIQUID-LIQUID FLOWS IN TSHAPED MICROCHANNELS USING MACHINE LEARNING APPROACHES
SINOPSIS
In the last decade, microfluidics has become one of the most important areas of science due to the rapid development of microchannel devices and technologies. The extremely high surface area to volume ratio makes it possible to use microchannel flows to remove high heat fluxes, conduct highly efficient reactions, and create laboratories on a chip and organs on a chip. In the modern scientific literature, there is already a wide range of works aimed at constructing flow pattern maps of gas-liquid flows and flows of immiscible liquids in microchannels. Depending on the flow rates, the physical properties of liquids, and the properties of the material of the microchannel and its geometry, different flow regimes are realized. The occurrence of a certain flow regime is caused by the forces prevailing in the system for given values of the control parameters. The complex interrelation between a large number of adjustable parameters makes the analytical approach unreasonably difficult and a combination of adjustable parameters for flow pattern prediction can't be obtained. Thus, flow pattern maps or semiempirical models are used.