Negatively buoyant jets are turbulent flows that are frequently employed by the desalination industry to disperse reject brines into oceanic environments. Although such brines are characterised by elevated concentrations of the same elemental components as the discharge environment contains, there is significant potential for marine ecosystem damage if this waste is not diluted properly. Numerous workers have analysed the dilution and spatial characteristics of negatively buoyant jets, but published data demonstrates notable inconsistencies. An important reason for these discrepancies is the variety of bottom-boundary conditions employed. This complicates comparison with predictions by integral models typically employed for discharge design, as these generally have not been developed with consideration to boundary interaction. In the present study, negatively buoyant jet experimental data is collected where bottom boundary distances are sufficiently large to avoid boundary influence at the point where the discharge returns to its source height (the return point). Near-field centreline dilution data is measured under still ambient conditions, for the source inclinations of 15–75°. Considerable attention is paid to experimental data quality, and all relevant issues are mitigated where possible. In order to ensure the boundary has no influence, source heights in this study range between 2.33 d F0 and 8.07 d F0. A variety of time-averaged and temporal statistics are calculated, and these statistics are compared with published experimental data and predictions by integral models. Normalised trajectory and dilution data from the source through to the return point collapses well at each inclination. The attention to signal quality and the self-consistency of derived experimental results in this study suggest a high level of accuracy, and large distances to the bottom boundary ensure that results are not confused by boundary interaction. Data for dilution rate at the return point supports the use of higher source inclinations (60° and 75°) to maximise dilution capability. A new ‘forced jet’ model is developed that incorporates the concept of a reducing buoyancy flux as the flow rises to maximum height. While this model is not applicable above source inclinations of 60°, predictions at other inclinations are reasonable. Dilution predictions are notably improved when compared to those from existing integral models. Finally, CFD simulations of negatively buoyant jets are conducted using the k-ε turbulence model. Despite the sophistication of this model, the quality of spatial and dilution bulk flow predictions at the centreline maximum height are no better than those obtained from the forced jet model or analytical solutions of Kikkert et al. (2007).

In urban areas, discharging wastewater into rivers is a common way to dispose of contaminants, and it is usually the most economical. Accurate information about how effluents are distributed in the receiving water body is desirable when designing industrial plants. Flow structures will be influenced by an effluent’s dilution processes during the mixing. Meanwhile, the cross-stream motions resulting from the streamline curvature can redistribute both the velocity and the shear stress, which favors the mixing behavior compared to a straight channel. However, the interactions between jet mixing behavior and the bend flow requires further investigation. In the present study, jets with different densities were discharged horizontally into a laboratory flume with a 135-degree open channel bend, and both the main and secondary flow behaviors in the bend were observed after the introduction of effluents. The acquired three- dimensional velocity data were used to validate numerical models of the effluent-bend flow. Numerical turbulence models such as the standard k-ε eddy viscosity model, non-linear k-ε model (Shih quadratic k-ε), and the k-ω SST (shear stress transport) model were employed to evaluate their accuracy. OpenFOAM was selected in the analysis for proposing better numerical models since it gives high-quality results to individualized complex fluid flows, and as an open source CFD software it can be beneficial to further develop and maintain. The first part of this study presents the implementation of the physical modelling of the proposed problems. Detailed descriptions of the experimental process were elaborated. Specifically, the three velocity components at four cross-sectional planes in the bend section were measured with and without saltwater jets by using the stereo Particle Image Velocimetry (PIV) ii technique in the laboratory flume. The experimental results show that the more pronounced effects with the jets were found at the beginning and exit of the bend. Although the jets had little effect on the maximum streamwise velocity, it was found that the occurrence of the negatively buoyant jets would affect the patterns and properties of the secondary flow in the bend. The second part of this study investigated the mechanisms underlying the two cells system, particularly when interacting with a discharged effluent jet. Detailed experimental data were used in interpreting the large center-region cell as well as small structures in the 135-degree open channel bend. A term-by-term analysis of the downstream vorticity equation was executed to investigate the various mechanisms underlying these cross-stream flow motions considering the influence of the negatively buoyant jets. The results indicated the generation and the dissipation of the streamwise vorticity with the effective terms of the vorticity equation. The third part of this study evaluated the performance of three different turbulence models with the experimental measurements. It can be concluded that fully 3D numerical models are capable of simulating the primary flow pattern in a strongly curved channel with the presence of a negatively buoyant jet. The comparison also shows that, although the outer bank cell was not predicted, the k-omega SST model can satisfactorily predict some of the smaller flow features in bend flow, such as the inner bank circulation cell and the overall form of the vorticity distributions. The results enable more reliable predictions for the characteristics and development of jets in a bend.