COMPUTATIONAL ANALYSIS FOR MIXING OF NEWTONIAN AND NON-NEWTONIAN FLUIDS FLOWING THROUGH

Abstract

In microfluidic systems, fluid mixing is critical for applications in chemical synthesis, drug delivery, diagnostics, and lab-on-a-chip technologies. However, achieving efficient mixing in microscale environments is challenging due to the predominance of laminar flow and low Reynolds numbers, which limit turbulence. This thesis addresses these challenges by computationally analyzing the mixing of Newtonian and non-Newtonian fluids in 3D micromixers using advanced numerical simulation techniques. The study systematically investigates the performance of various micromixer geometries, such as hybrid T-junction designs, focusing on their ability to enhance mixing through passive mechanisms. The analysis also examines how fluid rheology (the flow behaviour of fluids, particularly Newtonian vs. non-Newtonian) influences mixing performance. Non-Newtonian fluids, such as polymer solutions or biological fluids, often exhibit complex behaviours like shear-thinning or shear-thickening, significantly impacting flow patterns and mixing efficiency. A crucial part of this research involves studying operational parameters such as inlet velocity, Reynolds number, and fluid viscosity. These parameters affect the flow regime and mixing dynamics, and understanding their influence is key to optimizing micromixer design. Another novel aspect of the study is the incorporation of nanoparticles, which are increasingly used in microfluidic systems for enhanced mixing, heat transfer, and targeted delivery. The thesis explores how nanoparticle size, concentration, and distribution influence the mixing index—a quantitative measure of mixing quality. By comparing Newtonian and non-Newtonian fluid mixing in the presence and absence of nanoparticles, the study provides valuable insights for designing efficient micromixers tailored to specific applications. These findings advance the fundamental understanding of microscale fluid dynamics and offer practical guidelines for developing next-generation microfluidic devices. v The initial design is motivated by a simple-T micromixer using different twists and bends. Ansys fluent with finite volume analysis, conducts a simulation for this parameter. We investigate the impact of twist and bend on mixing performance and pressure drops. For data analysis, we consider the twist, bend angle, and Reynold number as input variables and mixing index and pressure drop as outputs variables for training using a neural network. With the moth flame optimization algorithm, the micromixer model was optimized for the maximum mixing performance and the minimum pressure drop. The optimal value for the twist and bend angle is 4 and 70°, respectively, at Reynold number (Re) 10. This technique has a significant benefit for microfluidic channel optimization. Using metaheuristic algorithms for a novel passive micromixer featuring bends and twists with offsets. The twists and bends cause constant changes in the direction of liquid flow, leading to chaotic advection that enhances species mixing while minimizing pressure loss. Although increasing channel length generally improves mixing performance, the proposed design's performance was higher than the reference channel for Reynolds numbers (Re) greater than 100. Key findings include achieving an excellent mixing performance of 87.76% at a Reynolds number of 400, with a bend angle of 60° and a twist factor of 4. The Harris Hawk Optimization (HHO) algorithm was most effective for optimizing microfluidic channel designs.