Stereo microscopic particle image velocimetry (PIV) was employed to investigate the complex three-dimensional flow structures within a micro hydrocyclone, with a focus on visualizing key flow features such as the locus of zero vertical velocity (LZVV). This measurement technique enabled detailed observation of the internal flow dynamics at microscale resolution, revealing the spatial distribution and evolution of vortical structures. The aim of this study was to assess how these flow characteristics influence the separation efficiency of circulating tumor cell (CTC) clusters.  

Developed a fully custom microscope system, built entirely from scratch to enable high-resolution scanning of microfluidic samples. Engineered and coded the complete control software, including automated stage traversal and image acquisition. Integrated hardware and software seamlessly to deliver precise, repeatable imaging across complex sample geometries. This end-to-end solution was designed, built, and programmed in-house, tailored specifically for advanced microfluidics research.

A micro-hydrocyclone is investigated as a high throughput particle/cell sorting microfluidic device. A complex flow structure has been reported by limited numerical works that increases the importance of undertaking a comprehensive experimental investigation. In addition to the flow, utilizing damageable bio cells in a micro-hydrocyclone, requires a deep understanding of the interaction between the cells entering the device, the flow structure and its instabilities forming in different operational phases. 

Hydrocyclones are widely used for separating particles from liquid suspensions in various industries. The physics inside a micro-scale hydrocyclone, however, is not fully understood due to the complex behavior of particles and fluids and difficulties in performing experiments at the micro-scale. This paper uses particle image velocimetry to investigate the flow behavior inside one of the smallest micro-hydrocyclones reported in the literature. 

In this project a convergent-divergent nozzle configuration is optimized in order to prevent the maximum T/W. 

For this purpose, Gambit, Fluent, and MATLAB linked together to operate the optimization cycle using the constrained genetic algorithm with 50 initial population.