Effect of the fluid rheology on low-density particle settling

Understanding how fluid rheology and particle geometry influence particle sedimentation is crucial for modeling transport processes in drilling and other industrial applications. The study presents an extensive experimental investigation of the sedimentation behavior of low-density particles—spheres and ellipsoids—in two fluids with similar densities but distinct rheological properties. A Newtonian glycerin–water solution and a shear-thinning Carbopol® solution were formulated and rheologically analyzed. A deep learning-based tracking system, combining dual-camera acquisition with Python object detection, was developed to reconstruct three-dimensional trajectories and extract kinematic parameters. The results reveal that shear-thinning behavior significantly modifies settling dynamics. In the Carbopol® fluid, smaller particles exhibited delayed or absent terminal velocity due to viscoelastic effects, whereas larger particles settled more stably. Drag coefficients were substantially higher in the non-Newtonian fluid, particularly for small spheres (CD > 50), and decreased with particle size (CD ≈ 0.8–1.6 for the largest particle cases). Particle geometry played a decisive role in the trajectory stability. Oblate ellipsoids followed the most stable and linear paths, while spheres exhibited greater lateral deviations, up to six times their diameter in the Newtonian fluid, particularly at higher Reynolds numbers. Prolate ellipsoids showed intermediate behavior. The findings highlight that neglecting fluid rheology and particle geometry can lead to significant underestimation of drag and misprediction of settling paths, with direct implications for drilling cuttings transport and wellbore stability. The study establishes the first experimental benchmark integrating deep-learning-based three-dimensional tracking with controlled rheology to describe particle settling in complex fluids. The resulting dataset provides a valuable reference for validating sedimentation and computational fluid dynamics models in non-Newtonian media, with direct implications for designing drilling fluids and optimizing solid transport in industrial processes.