Semantic Model for the Automated Assessment of Water Balance in Regional Water Supply Networks: the Case Study of Siciliacque S.P.A.

<p dir="ltr">Efficient water resource management represents a growing global challenge, particularly intensified by climate change and increasing demand. This paper presents a semantic data model based on NGSI-LD (Next Generation Service Interface - Linked Data) with two primary innovations: (1) processing non-equidistant time series data from manual operator readings of non-telemetered volumetric meters, enabling accurate water balance calculations despite irregular sampling; and (2) implementing a flexible framework that automatically adapts to changing data sources as smart meters replace traditional meters. These capabilities were validated through implementation in Siciliacque's regional water transmission network using adapted Delft FEWS technology for operational management meters positioned downstream from treatment plants and remotely read by the telecontrol system. The model was specifically designed to maintain compatibility with IoT smart meters that are gradually replacing traditional Woltman meters, ensuring future adaptability. A key technical challenge addressed by our NGSI-LD implementation is representing the network's dynamic nature, particularly the bidirectional interconnections between aqueducts that change flow direction based on operational requirements. The semantic model effectively captures these variable relationships while maintaining contextual integrity through the principles of linked data. The NGSI-LD model serves as the foundation for an automated water balance calculation system. This system was implemented using Delft FEWS, a platform developed by Deltares that was originally designed for flood early warning systems and adapted to work as water infrastructure management system.</p><p dir="ltr">This paper was presented at the 21st Computing and Control in the Water Industry Conference (CCWI 2025) at the University of Sheffield (1st - 3rd September 2025).</p>