Vortex merging and evolution mechanisms have been characterized across the 180° sharp bend region of matrix-cooled serpentine channel with the help of stereo Particle Image Velocimetry (PIV) system. Complete evolution of flow phenomenon is captured through measurements in multiple vertical and horizontal planes, and the mechanism responsible for evolution of vortices at a typical Reynolds number (Re) of 6500 has been explored. It was observed that the matrix subchannel typically produces a kind of swirling pattern downstream of the matrix structure, which is similar to the swirling flow occurring in various other application areas. Two co-rotating vortices emanate out of matrix subchannels (observed in secondary plane), which immediately merge into a single vortex at the onset of the bend and propagate further downstream of the bend section, where a pair of side-by-side counter-rotating vortices are observed later. Numerical simulations have also been performed to get the pressure distribution and flow characteristics within oblique/inclined planes along the bend, which are difficult to obtain experimentally through PIV. Results demonstrate that the vortical structures lose their strength during merging and turning, significantly reducing turbulence while passing through the bend. Based on the combination of experimental and numerical observations, a conceptual diagram depicting flow dynamics has been proposed, which provides the overall perspective of vortices evolution and merging across the bend.

Many thermal applications involve fluid flow across the sharp bend. The fluid flow across the sharp bend causes entire modification due to back pressure and flow separation, etc. Quantifying pressure across the sharp turn due to the modification of flow behaviour is always a complex challenge. One of the difficulties associated with the intrusive techniques of pressure measurements is to disturb the flow and the lack of pressure field estimation in the entire domain. This article presents a complete methodology for pressure estimation before and at the sharp bend using stereo-particle image velocimetry (PIV) data. The method relies on knowledge of the PIV velocity field, which is used to calculate the pressure field analytically and numerically. The outcomes of the study show that how the sharp bend in the fluid path results in a sharp pressure drop. Pressure at the entry of the bend is found to be ~5 Pa (absolute), and at the exit side, it drops to ~3 Pa for a turbulent in-flow condition (Reynolds number = 5500). The time-averaged streamlines are also superimposed over velocity magnitude and pressure data for a better understanding of flow and pressure simultaneously.

This paper presents the flow patterns and heat transfer distribution across the bend region in a two-pass rectangular duct for an inlet turbulent flow regime (Re = 6500). The results are presented at various vertical and horizontal planes located at different positions across the bend along the flow progression. Two divider wall configurations (in-between two-pass of the duct) are studied, i.e., (1) sharp-corner turn and (2) smooth curved turn. It has been noticed that flow gets completely modified because of divider wall shape. The numerical results reveal that primary and secondary flow in turn regime displays combined features of a bend-induced, Dean-type circulation. The flow dynamics and local heat transfer vary significantly with different divider wall configurations. For the duct with a sharp divider wall, a pair of counter-rotating Dean vortices induce after 90° near the end-wall of bend region. In contrast, for the duct with a smooth curved divider wall, vortices pair is induced close to the divider wall and stretched across the entire width. The results show that Dean vortices play an important role in enhancing localized heat transfer across the bend regime. Interestingly, the duct with a sharp divider wall exhibits higher localized heat transfer, whereas the duct with a smooth, curved divider wall exhibits lower localized heat transfer but more uniform distribution.