Predictive model uses pressure data to help reduce water leaks in pipes

Water is a resource essential for life, as valuable as it is limited. For this reason, and especially in contexts of water scarcity, preventing it from being wasted is a key objective for those who are responsible for transporting it from catchment sources—reservoirs, swamps, springs, etc.—to taps at homes. It is a journey that water makes through piping systems that are often obsolete or that do not meet the necessary performance and efficiency requirements.

With the aim of enhancing the management of drinking water, a team from the University of Cordoba has developed a predictive model whose implementation can help reduce one of the main problems suffered by these distribution networks: pipe leaks due to excessive pressure.

Published in Energy Nexus, the work was carried out by a team in the Department of Applied Physics, Radiology and Physical Medicine, composed of researchers David Muñoz Rodríguez, Alberto Jesús Perea Moreno and María Jesús Aguilera Ureña; together with Manuel J. González Ortega, from the University of Seville; and Andrés Ortega Ballesteros, a student from the University of Cordoba who recently completed his doctoral thesis in this field.

As the authors explain, the pressure at which the water circulates has a direct relationship with the appearance of leaks and the volume of water that is lost through them. Having data on this variable allows distribution network managers to detect anomalous patterns and adapt the pressure to the necessary levels and demands, which is done through valves. In this way, it is possible to prevent leaks and, thus, unnecessary water loss.

The system proposed by the UCO team was developed following the Box-Jenkins methodology, which is mainly used to analyze and make forecasts through time series, and was generated and validated through real data on the water distribution network for the Cantabrian town of Noja, a small municipality with significant population fluctuations during holiday periods, affecting water demand and pressure needs.

The resulting predictive model, scalable to other similar networks, allows companies operating water distribution networks to overcome a major limitation: the absence of real-time data to guide them when making dynamic pressure adjustments. The information provided by this system allows them to detect variations in advance and take preventive measures.

The benefits of this model, the researchers explain, go beyond reductions in leaks and consequent water loss. By adapting the pressure to demand, it improves the quality of the service users receive.

At the same time, the appearance of new breakdowns is minimized, thereby lowering maintenance costs and also extending the useful life of the pipeline network, an essential infrastructure whose proper functioning is relied on to supply drinking water, a truly precious resource.