Sistema Inteligente de Dados de Água

In Portugal, the distribution of water for human consumption is managed at the municipal level, with different management models (direct, delegated, or concession). Sustainable management of these services requires the continuous collection and updating of data (pressure, flow, consumption) to generate useful information for the control, management, and operation of the systems, as well as for the planning and asset management of the infrastructure.

Depending on the maturity of the managing entity, data can be collected manually and stored in files or on paper, or continuously and in real-time through telemetry systems. This data includes records of pressure, flow, water volume, reservoir levels, energy consumption, and water quality parameters (chlorine, pH), as well as information about the infrastructure, consumers, and weather conditions.

This research project aims to develop new algorithms and models to extract relevant information from this data, helping managing entities (MEs) improve system management, reduce water losses, and ensure water supply in both quantity and quality. The multidisciplinary team combines technical knowledge of water supply systems with data science, artificial intelligence, and mathematics.

The project has five main tasks: exploratory data analysis; predictive consumption analysis; spatial location of leaks; identification of anomalous events; and development of a prototype and recommendations for decision support. The prototype will enable the processing of data series, prediction of consumption and quality parameters, detection of leaks, and identification of anomalous events (illegal consumption, meter wear).

This tool aims to reduce water losses, improve the efficiency of water resources, reduce energy consumption, and make systems more resilient to climate change, allowing MEs to prepare contingency plans for extreme events, such as droughts. The data analyzed will be provided by three managing entities: Câmara Municipal do Barreiro, EMAS de Beja, and Infraquinta, representing different management dimensions and scopes. The results could be applied to other 256 MEs in Portugal, mostly managed by municipalities.