A CFD-based methodology for improved representation of complex hydraulic components in network-scale models: application to irrigation hydrants

ABSTRACT

Graphical abstract illustrating a multi-scale modelling framework for representing complex hydraulic components in network-scale simulations. The workflow starts with CFD simulation of an irrigation hydrant to quantify component-specific resistance coefficients (Ks) under different operating conditions. These coefficients are then incorporated into an EPANET hydraulic model to improve the representation of head losses and hydraulic behaviour, resulting in more accurate and reliable network-scale hydraulic simulations.

Hydraulic network models commonly rely on simplified representations of local energy losses, which can introduce inaccuracies when simulating complex hydraulic components. Multi-outlet hydrants are key control and distribution elements in pressurized irrigation systems, where localized hydraulic behaviour strongly influences pressure distribution and system performance. However, their complex internal configurations are rarely represented in detail within conventional network-scale modelling approaches. This study proposes and validates a computational methodology to improve the representation of complex hydraulic components by transferring CFD-derived hydraulic information to network-scale models developed in EPANET 2.0. The proposed methodology was applied to 12 representative multi-outlet hydrant configurations, with one representative manifold presented in detail as a case study. CFD simulations were used to derive configuration-specific resistance coefficients associated with the singular elements of the hydrant manifold, enabling the characterization of flow behaviour and head losses. The methodology was validated against laboratory measurements, showing high agreement, with root mean square errors below 0.12 m and relative errors generally below 1%. Compared to conventional approaches based on standard coefficients, the method reduces modelling uncertainty and improves head loss estimation. The proposed framework is transferable to other complex hydraulic components, enhancing the reliability of network-scale simulations for water distribution systems.