Effective optimization of irrigation networks with pressure-driven outflows at randomly selected installation nodes.

This paper presents an innovative methodology for the design of pressurized irrigation networks. Compared to other methodologies proposed in the scientific literature, it features three novel aspects: (i) construction of peak demand scenarios based on the random selection of installation nodes for hydrant heads available in each sector of irrigated properties; (ii) realistic hydraulic modelling of outflows from hydrant heads by means of the pressure driven approach; and (iii) adoption of linear constraints to enforce the telescopic property in the distribution of diameters from the source towards the external areas of the network in the optimized design. The applications of the methodology to the real network serving an irrigated area of 750 ha in Northern Italy proved that the aspects (i) and (ii) contribute to the accurate modelling of the current network while highlighting its hydraulic deficiencies. The adoption of the linear constraints described in (iii) in the context of the bi-objective genetic optimization of network diameters resulted in the speeding up of the algorithm convergence. The results show how decision makers can choose the ultimate configuration based on budget considerations from the trade-off solutions obtained between installation costs and hydraulic performance, considering network layouts with different level of topological redundancy.