HYDROLEAKS2.0

Descripció del Projecte

The HYDROLEAKS2.0 project is an experimental development activity based on the need of water distribution network managers to improve the control of their installations. It stems from the data and results obtained in the previous HYDROELAKS project and is based on collaboration between technology-based companies offering different solutions in the field of digitisation, water distribution network management bodies and clusters. HYDROELAKS2.0 proposes the development of new artificial intelligence tools to calculate water operator leakage, using high temporal and spatial resolution models that take into account both real and apparent leakage. As a final result, the project proposes to generate a Digital Clone or Digital Twin of a supply network and to undertake new solutions that, through the use of AI, allow the detection of leaks. The results are managed through a digital platform, which already represents an important operational and economic improvement for companies operating in the water cycle, and a reduction in network management efforts. All this will be a first step that will open the door to further developments in future phases of the project.

In later phases, the project will focus on a solution oriented towards preventive maintenance, new algorithms for the operational management of the water networks and the water flow or the economic forecasting of investments, again based on AI solutions. The main objective of the present project is, therefore, to calculate with high resolution and frequency the real, baseline and incremental flows from the combination of different digital tools. In this sense, the objectives derived from this broader general objective are:

• To have reliable observed data, which implies having processes that detect and correct anonymous data (including the absence of data), at the sensor level (edge validation) and at the database level (database validation).
• Have processes that detect the presence of leaks in the network and the most accurate and automatic localisation possible. Certain processes may or may not require a previous hydraulic model (data-driven vs. model based), so that a subset of these processes can be applied to other supply networks that do not yet have a calibrated model.

This project will lead to an improvement in the management of water distribution networks for the operators participating in the project, with the possibility of being replicated and scaled according to the needs of other managing bodies. This improvement will result in a better cost of service for subscribers, in a reduction of water losses, with the consequent improvement in terms of sustainability and finally in the improvement of the operation and maintenance of the infrastructures of the network management companies.