Droplet Velocity and Film Thickness Studies of an Elongated Taylor Droplet in a Microchannel and Characterization Using Machine Learning

Droplet microfluidics is used in biological and chemical systems, but the behavior of non-Newtonian fluids in these systems is not well understood. The present study focuses on an elongated Newtonian Taylor droplet flow in a shear-thinning inelastic fluid inside a circular microchannel. Additionally, moderately viscous droplets are considered with a viscosity ratio in the range of 0.5 to 5. In these conditions, droplet dynamics are investigated using the isoAdvector Volume of Fluid in OpenFOAM. The shear-thinning fluid is characterized by the Carreau model. Droplet profile, droplet velocity, and film thickness are observed to be critically influenced by the shear-thinning variations and viscosity ratios. Accurate models to predict these complex droplet dynamics are critical in the development of microfluidic devices. Machine learning is utilized to develop a prediction model using Boosted Trees with LSBoost algorithm. The developed model is capable of accurately characterizing the droplet velocity and film thickness.