Less leakage in drinking water systems

In Stockholm and surrounding areas, it is estimated that about 17 percent of the drinking water produced leaks out before it reaches consumers.

As the city's pipeline network consists of over 35,000 drinking water pipes, there are many possible sources of error. To minimize waste, AI models are used to help prioritize which pipes need to be repaired or replaced. The AI model takes into account both the age and material of the pipeline, as well as geological conditions to point out which pipeline stretches are at the greatest risk of leakage.

In real time, flows in the grid's pumping stations are also compared against an AI-based prediction of how large the flow should be, taking into account the time of day, season, temperature and precipitation. If the current flow differs from the prediction, this may indicate leaks or other disturbances in the pipe network, which can be quickly investigated and remedied.

Contact

E-mail: david.rehn@svoa.se

Updated