Tracking motion of thermal turbulent structures at the water surface in compound channel flows using block matching algorithm

Abstract

In this research a novel approach is used to track the motion of the coherent structures at the water surface of a compound channel flow with vegetated floodplain. Thermal sequences images in conjunction with a block matching algorithm (BM) approach is utilized to track the motion of turbulent structures by extracting a 2-D velocity field at the water surface. A laboratory flume was used and a compound channel of half cross-section was installed along the flume length. To mimic the vegetation elements, three diameters of wooden rigid, emergent cylindrical rods that located in staggered form were used. A thermal camera was vertically installed at section where a fully turbulent flow was achieved. Injected hot water was released from a water tank at the main channel-floodplain's edge at a small distance from upstream location of the thermal camera. A thorough assessment is performed to evaluate the performance of using the new approach. Where a Fast Fourier transform (FFT) is utilised to extract the energy of the eddies through arbitrary selected points between the cylinders wake and showed that the FFT successive peaks have different slopes. The exponent was deduced and showed different magnitudes from the Kolmogorov exponent (-5/3). The power law (-5/3) was found in the region where the fully turbulent flow is still developing, whereas the other exponents were found near to the cylinders where the velocity fluctuations are decreased in strength. Furthermore, the two-dimensional velocity field was extracted and analysed according to the BM approach and its coherent motion profiles are presented and discussed for all the experiments. The results showed a good estimation of tracking the motion of the coherent turbulent flows at the wake of the cylinders. The present results showed that the new proposed approach is primarily useful for providing visualisation of the coherent turbulent flows motion and can be extended for a wide range of turbulent flows as an easy and reliable technique.