Evolutionary prediction of multiple vertical buoyant jets in stationary ambient water

The variables describing the merging and mixing properties of multiple vertical buoyant jets in stationary ambient water have not been generally formulated and reported. This work develops new evolutionary-based models to predict such variables, including for the non-dimensional vertical displacement of the merging point (ym/D), the non-dimensional centerline concentration (Cm/C0) where the jets start merging, the non-dimensional vertical displacement of the well-mixed location (yw/D), and the non-dimensional concentration where the jets are well mixed (Cw/C0). These parameters are crucial in the design of outfall systems and for environmental impact assessments. A fully three-dimensional (3D) computational fluid dynamics model that solves the Navier–Stokes equations and uses the re-normalisation group (RNG) k-ε turbulence closure is validated against the available experimental data. The validated numerical model is further utilized to carry out additional computations for 100 different cases to enrich the data sets. The data are employed to develop and evaluate the evolutionary-based models, and the model networks are optimized using multigene genetic programming. The proposed models can serve to compute the mentioned jet characteristic variables as a function of non-dimensional port spacing (Sp) and jet densimetric Froude number (Fr). Partial derivative sensitivity analysis is also conducted to assess how the varying input variables affect the jet characteristic parameters. An uncertainty analysis is also performed for the proposed models. The proposed explicit equations and codes facilitate easy estimation of the merging and mixing properties of vertical buoyant jets discharged from multiport diffusers.