Aqueducts are important water crossing structures that mainly consist of three parts: the inlet transition section, the tank body section, and the outlet transition section, and are widely used in water transmission projects. When the inlet transition section is too short, the cross-sectional area decreases sharply and the turbulence of the water flow increases, or the water flow changes from slow flow to rapid flow, which inhibits the flow capacity.
The problem of local head loss in the constriction section of the channel has been studied by many researchers. Nguyen et al. [
1] measured the local head loss coefficient for the transition from a rectangular channel to a pressurized pipe by means of physical model tests, and the local head loss coefficient was ζ = 0.63 (1 − downstream cross-sectional area/upstream cross-sectional area) when the constriction section was symmetrical. The hydraulic calculation manual [
2] and the design specifications of irrigation and drainage canal system buildings [
3] provide some empirical head loss coefficients without providing a clear correlation between them. Zhai Yuanjun [
4] designed and conducted increased flow tests on several typical channel transition section models, and performed regression analysis to obtain a one-dimensional quadratic function relationship between the local head loss coefficient and the contraction angle. Wu Yongyan et al. [
5] conducted a physical model test study on the flow characteristics of a transition section from a trapezoidal open channel to a horseshoe unpressurized tunnel and obtained the relationship between the local loss coefficient and the area ratio between the upper and lower reaches of the constriction section and the bottom elevation difference. Wang Songtao [
6] and Qu Zhigang [
7] revealed the mechanism of water surface fluctuation in the trough body of the aqueduct by numerical model simulation of a typical aqueduct in the South–North Water Transfer Central Line. Kazemipour and Apelt [
8] reported that, when the transition section changes dramatically, the local head loss can even account for more than 90% of the total head loss. Local head loss can be calculated by multiplying the velocity head rate by the local head loss coefficient, and the reasonable selection of the local head loss coefficient is key to the correct estimation of local head loss. The local head loss coefficient generally needs to be selected according to the form of the transition section, Henderson [
9] pointed out that, when the transition section forms from an abrupt change to a gradual change, the local head loss coefficient can be reduced by two-thirds; Chow [
10] suggested a local head loss coefficient of 0.1–0.2 for a twisted surface form of a transition section, 0.3–0.5 for a linear form, and 0.75 for an abrupt change. Yaziji [
11] measured a local head loss coefficient of 0.2–0.42 for linear and streamlined shrinkage transition sections. Zhang Zhiheng [
12] reported that, for twisted surface transition sections, the local head loss coefficient of the tunnel inlet is not a constant, but is related to the shrinkage angle of the water’s surface. Crispino Gaetano et al. [
13] conducted calibrated numerical simulations to assess the hydraulic features of supercritical bend manholes with variable deflection angles, curvature radii, and lengths of straight downstream extension elements. It was demonstrated that the hydraulic capacity of a bend manhole increased with increased curvature radii and straight extension lengths, whereas the effect of the deflection angle was less significant. Cheng Yong et al. [
14] studied the three-dimensional flow at the open channel bifurcation by numerical simulation using FLOW-3D software, analyzed the hydraulic characteristics of the recirculation zone and flow structure near the open-channel bifurcation, and obtained the equations for the inflow width of the surface and bottom layers in a trapezoidal channel. Tellez-Alvarez, J et al. [
15] studied the energy loss coefficients of three types of grate. The energy losses recorded at flow rates between 20 L/s and 50 L/s ranged between 0.15 and 3.41, showing a negative correlation. It is known that it is feasible to study the hydraulic characteristics of aqueducts using the three-dimensional numerical simulation method.
Based on the Flow-3D software platform, this paper conducts a numerical simulation of a typical aqueduct in the South–North Water Transfer Central Project to study the factors influencing the head loss coefficient of the inlet gradient section.
There are no clear formulas for calculating head loss in existing studies and codes, or the factors affecting head loss have been considered to be small. In order to highlight the influence of head loss factors, this paper, based on existing research, considers the influence of the water surface contraction angle, inlet and outlet overwater area, and water level difference of the gradient section and downstream water depth, and provides the calculation formula for the head loss of the gradient section to support the engineering design.