Hydrodynamic characteristics of overland flow based on particle image velocimetry

Accurate measurement of overland flow velocity along flow depth is critical for hydraulic and soil erosion processes over hill-slopes, yet multipoint velocity along the flow depth has not realized a clear understanding of overland flow characteristics. Particle Image Velocimetry (PIV) breaks through the spatial simple point survey technology limit and not disturbing the flow due to optical measurement. This method could provide rich velocity information for overland flow. To match the shallow depth measurements, the resolution of PIV was improved up to 64 pixels/mm by adding the extension tubes and strengthen the light. Taking advantage of PIV, this study was to explore the hydrodynamics characteristics of overland flow, and the velocity in the streamwise and wall-normal direction were measured. The velocity was obtained by calculating the velocity of corresponding particles for two consecutive images. Experiments were carried out with seven overland flow conditions ensured by previous literature research, featured with flow depth changing from 0.55 to 1.1 cm, Reynolds number from 1 092 to 2 877, and Fraud number from 0.7 to 0.995, while an extra case of deep-water open channel flow was conducted as the control group. The statistical parameters of overland flow were studied, in terms of velocity profiles, correction coefficient, turbulence intensity, skewness, and kurtosis coefficient of instantaneous velocity. Results showed that 1) The velocities from flume bed to free surface were effectively measured using PIV. The correction coefficient equated the ratio of mean velocity to maximum velocity, which widely was used to dye and sault tracing methods, logarithmically increased with increasing Reynolds number when overland flow regimes belong to transition flow. However, the present mean correction coefficient equated to 0.77 in transition flow and was larger than 0.7 that acquired by dye and sault tracing methods. Because of the maximum velocity measured by dye and sault tracing methods were doubtable, the correction coefficient acquired by different methods were discrepant. The PIV had clear physical meanings, that could distinguish maximum and mean velocity. 2) The turbulent intensity was the second moment of instantaneous velocity and represented the pulse of fluid. Compared with deep-water open channel turbulent flow, the turbulence intensity and Reynold stress were not stable for overland flow. The streamwise turbulent intensity of overland flow was larger than that of deep-water open channel turbulent flow, while wall-normal turbulent intensity was smaller. As increasing flow depth and Reynold number, turbulent intensity became stable and closed to that of deep-water open channel turbulent flow. The parts of fluids affected by Reynolds stress was about 80% for deep-water open channel turbulent flow while that was less than 80% for overland flow. Moreover, the parts of fluids affected by Reynolds stress became larger with increasing Reynolds number for overland flow. 3) The skewness and kurtosis coefficient were the third and fourth moments of instantaneous velocity, which described the shape of the probability density function. The higher the order of the moment, the more accuracy of measurement was required. The skewness and kurtosis coefficient of control groups well agreed with the previous study, implying the reliability of the present experiments. The skewness coefficient and the kurtosis coefficient of overland flow were different from deep-water open channel flow and gradually closed to the curve of deep-water open channel flow as increasing flow depth. Based on the features of the skewness coefficient, more instantaneous velocity larger than the mean velocity occurred near the flume bed region, while more instantaneous velocity smaller than the mean velocity occurred near the free surface region. Most parts of the overland flow of the kurtosis coefficient were larger than 3, implying the probability that occurred with excessive velocity for overland flow was lower than that of deep-water open channel turbulent flow, due to the limitation of shallow depth and the coherent structure had not enough space to fully develop. Although PIV is not suitable for field tests and erodible flume bed due to block of camera view, PIV has a unique advantage, i.e. multiple point survey, contactless, and high-frequency measurement. Thus, this method could further apply in the research of soil and water conservation and could help study the water erosion mechanism.