Modeling Incomplete Solute Mixing at Pipe Junctions in Water Distribution Networks: Parametric CFD Simulations and a Predictive Regression Approach

Abstract

Abstract High-quality drinking water is essential for modern settlements, and accurate modeling of solute mixing at pipe junctions is crucial in water distribution networks (WDNs). While complete mixing is often assumed, recent studies emphasize the need to account for incomplete mixing. This paper advances the understanding of incomplete mixing effects on water quality dynamics in WDNs. Computational fluid dynamics (CFD) simulations are employed to examine how pipe connection angles in cross junctions influence solute mixing. The results highlight the significance of the junction angle and the relative distribution of inflow and outflow rates, whereas the diffusion coefficient and the turbulent Schmidt number have a limited influence. A generic regression model is developed to capture the nonlinear relationships between the dominant factors, providing accurate predictions of mixing at pipe junctions with minimal computational cost. The model is validated against experimental data from the literature and can be integrated into traditional WDN solvers such as EPANET. The findings improve the simulation reliability of contaminant concentration profiles, supporting more robust water quality analysis and advancing the state of the art in WDN modeling.