On the delineation of the flow separation zone in open-channel confluences

Abstract

In the context of a T-shaped open-channel confluence, where the tributary flow separates at the downstream junction edge and creates a separation zone in the post-confluence channel, this paper provides a critical reflection upon the terminology and the delineation methods for the separation zone and its proxies: the recirculation zone and the reverse flow zone. Following the fluid-mechanical literature on flow separation, the separation zone is delineated by a dividing surface formed by the three-dimensional (3D) streamlines separating from the downstream junction edge. Based on a dense 3D time-averaged velocity field obtained from a Large-Eddy Simulation of a lab-style confluence flow, the separation zone turns out to be open. Near the bed, the dividing surface exhibits a gap, through which mainly tributary fluid enters the separation zone, leading to a limited discharge flowing through that zone. Downstream of the section where the merging flows are the most contracted, the dividing surface does not reattach to the inner bank, prohibiting the determination of a maximum length of the separation zone. Downstream of that section, the dividing surface gradually loses its role as a skeleton for the separation zone shear layer. Due to mixing, the separation zone and the adjacent maximum velocity zone gradually dissolve and transform into the flow recovery zone, as in a wake. By means of flow visualization with 3D streamlines, revealing different types of recirculating motions, the maximum recirculation zone extent is determined. Then three simplified methods commonly used in the literature to quantify the dimensions of the separation zone or its proxies are evaluated by comparison with the 3D streamline method. By adopting various simplifying assumptions about the flow, the pseudo-streamline method, the zero-unit-discharge method and the zero-velocity method yield definite reasonable maximum lengths for the recirculation zone and the reverse flow zone. With the exception of the zero-velocity method, all simplified methods yield reasonable estimates for the maximum widths and the corresponding contraction coefficients of the separation zone proxies. Still, contrary to the simplified methods, only the 3D streamline method can provide physical insight into the flow features occurring in the separation zone or its proxies.