Rheological identification of jetted fluid using machine learning

The understanding of flowing properties fluids and the knowledge related rheological are crucial from both a research industrial point view. To determine complex fluids, many devices have thus been developed, so-called rheometers. main objective present paper is to identify fluid jetted using continuous inkjet (CIJ) printing process by comparing morphology aforementioned dataset known (rheologically speaking) jet morphologies viscosity, surface tension, density large datasets CIJ process. When ejecting fluid, ejection competes among several forces: inertial, viscous, elasticity, which affect resulting jet. Also, under certain conditions, unique directly fluid. We want use this uniqueness be compared with, obtain its properties. Using numerically generated Newtonian jets, we show in article that identification viscosity neural network not only feasible but also proves very accurate with an average error less than 1% for range viscosities.