Turbulent fluxes, integral model coefficients, and desalination discharge predictions

Abstract

Brine discharges from large-scale desalination plants are typically in the form of inclined negatively buoyant jets (INBJs). Existing simplified predictive (integral) models employed to predict the behavior of desalination discharges demonstrate significant errors when compared with physical measurements. The accuracy of these INBJ predictions is dependent on our knowledge of the mean and turbulent fluxes within the flow. In this study, the mean and turbulent fluxes of INBJs are extracted from a recently generated and validated Large Eddy Simulation (LES) dataset. The implications of this new information for development and implementation of integral models are then explored, with a particular focus on the ability of these models to predict characteristic parameters that are representative of physically measured flow profiles. Central to the development of integral models (as the name suggests) is the integration of mean and turbulent fluxes over cross-sections perpendicular to flow direction to determine model coefficients. These integral coefficients remain constant where the non-dimensional functional form of the perpendicular profiles remains unchanged, that is, independent of downstream distance (self-similarity). The assumption of self-similarity and associated constant integral coefficients has been shown to be valid for predicting jet and plume. Analysis of the new LES flux data enables the integral model coefficients to be determined directly from the mean and turbulent flux data at each cross-section for INBJs and to therefore determine the extent to which the assumption of self-similarity remains valid for these flows. Analysis of these fluxes also informs modifications to an existing INBJ integral model, which result in significantly improved model predictions. The latter confirms that a substantial source of previously noted discrepancies with measured data has been an oversimplified approach to determining integral model coefficients for INBJs. In addition, the new flux data provides a basis for cross-validating future experimental and simulation results, as well as informing further experimental studies of these flows.