Non-Newtonian Fluid Flow Modeling in the Inertial Viscometer with a Computer Vision System

Purpose of research . Development theoretical premises for the new inertial viscometer, as well development an approximate model viscosity fluid flow using convolutional neural networks and laser speckle contrast imaging data. Methods The study consists two parts. first is devoted to a in toroidal channel еру viscometer. mathematical includes dimensionless equations Navier-Stokes convective heat conduction, analysis which made it possible estimate conditions uniformity pressure temperature fields. numerical solution simplified equation was obtained by control volume method. computational experiment identify additional operating second part aimed at solving problem predicting values shear strain rate on tour surface rate. based ensemble trained data from speckle-contrast visualization transparent tube. Results recommendations parameters viscometer studied types liquids given range are obtained. An has been developed form deep networks, makes determine volumetric images. Conclusion result viscous can be used determination kinematic viscosity. So, necessary characteristics, such surface, found pretrained network speck test any non-Newtonian capable reflecting coherent radiation. In particular, physiological fluids, including blood.