Servicios de Txingudi, Txingudi, is a public utility company responsible for the water supply in the municipalities of Irun and Hondarribia (Spain). Txingudi believes that proper water management is defined by a commitment to the sustainability of the region and the quality of life of its inhabitants. The company has an innovative approach to technology and focus on end customers and the environment. When it comes to water supply network management, an emphasis is placed in meeting present and future needs of the region to ensure a safe and clean water supply. It was one of the first water utilities in Spain to install smart meters (currently 40.000 of its water meters are smart) and to develop a mobile phone app to communicate with its customers. 

• Water consumption in this area is dynamic, greatly impacted by tourism and weather conditions. The water utility set new objectives to improve the efficiency and visibility of water network management. 
• They were looking for a compatible tool that could meet its needs for growth, a system with dynamic thresholds to detect and repair leaks quickly while avoiding false positives and a system capable of recognizing seasonal patterns. 
• Finally, although the current network is divided into small DMAs, future plans existed to potentially create larger DMAs. Larger DMAs could help to avoid water quality problems and to improve network resiliency. An AI-based solution responded to all of these criteria. 

SIWA Leak Finder

• SIWA Leak Finder (standard version) was implemented across the whole network covering 285 km pipe length. Real-time mode was activated. 
• The implementation was straightforward and simple taking less than a month. Data from existing installed sensors (flow meters, pressure sensors, and chlorine sensors) was sent automatically in a secure way to the cloud server. The artificial intelligence algorithms were trained with a 3-week dataset. 
• The tool processes following inputs: Sensor data (flow, pressure, chlorine), Customer water meter data (daily readings), Average error data for each customer water meter, GIS shapes​
• The system is configured to detect following types of events: Leaks, Flow trend, Chlorine, Pressure increase and decrease, Sensor failure, Missing data, Flow increase and decrease 
• The following KPIs are calculated automatically by the system for each DMA: Performance, Infrastructure Leakage Index, Delivered Water, Registered consumption, Water losses, Real and apparent losses, Average error, etc.​
• Scope of supply: Standard and advanced software versions were installed in the cloud. Sensors were already in place prior to implementing the tool. Water meters had been installed before implementing the tool.