Machine learning can help protect urban water. Here’s how – The European Sting – Critical News & Insights on European Politics, Economy, Foreign Affairs, Business & Technology

(Credit: Unsplash)

This article is brought to you through The European Sting’s collaboration with the World Economic Forum.

Author: Shefali Rai, Project Specialist, C4IR India and Sanjiv Jha, Director, Smart Infrastructure, Amazon Web Services


  • Climate change has highlighted the importance of preserving fresh water as a resource.
  • Effective infrastructure management in cities is essential to conserve water.
  • Real-time sensor networks with machine learning can predict and troubleshoot water leaks more effectively.

Water is perhaps Earth’s most unique natural resource. As the most important natural resource, water covers about 71% of the Earth’s surface, but only 1.2% of fresh water is drinkable. This makes drinking water as rare and valuable as certain rare metals.

Fresh water from rivers and streams can be used for a variety of purposes: commercial, domestic, emergency and industrial. Given this versatility and a scarcity due to climate change, fresh water has become a point of contention in countries and cities with growing urban populations. To ensure a stable water supply, good urban governance and urban resilience, it is imperative that cities have strong water infrastructure.

A good water distribution network is essential

The water distribution network is an essential part of the urban hydraulic infrastructure. An efficient water distribution network has the capacity to meet the demand for potable water with minimal or no losses. Water leaks are one of the main causes of losses in the distribution network and, consequently, of the gap between supply and demand, especially in cities.

Water leaks can occur at any stage of the network – transport, treatment, distribution or storage. To some extent, the issue of water leakage has been tackled in various ways in cities, especially since decreasing the amount of unbilled water (water lost in transit) can lead to better socio-economic outcomes. -economic, environmental, health and safety for cities.

Cities typically focus on identifying the sources of water leaks and the extent of leaks through water management systems. Generally speaking, these management systems focus on quantifying the amount of water lost, detecting leakage hotspots, and effectively controlling current and future leakage levels.

Despite these measures, according to the World Bank, developing countries lose about 45 million cubic meters of water per day. This water, pumped but lost or unaccounted for, can cost up to US$3 billion per year. Being able to save even half of these losses can lead to an adequate water supply for at least 90 million people. This is a significant amount, especially for a country like India, which according to a report by NITI Aayog is expected to experience a double increase in demand for water supply by 2030.

Better management of water leaks is urgently needed

It is absolutely necessary to abandon traditional and existing methods of managing water leaks for three main reasons.

First, the existing methods are mainly corrective and not predictive. The predictive component of today’s water distribution network is predictive maintenance, which is pre-scheduled, done manually and takes time. Even to implement corrective measures in case of larger water leaks, large sections of the water supply systems are often closed, which negatively affects the daily lives of citizens, especially those who already lack water. uninterrupted water supply.

Second, while the magnitude of water leaks in cities varies in magnitude, most corrective actions are taken for larger water leaks, while smaller leaks often go unchecked for an extended period of time. . As a result, smaller, uncontrolled leaks result in water losses of significant magnitude, in volume and revenue. Civic organizations can lose between one hundred thousand and one million dollars in revenue over a five-year period.

Third, in the worst cases, a leak in the main water pipes has also led to high-current cable short circuits, which poses a deadly threat to the mass population. Such limitations raise the need for more efficient and effective water management systems that are both corrective and predictive.

A possible solution – real-time sensor networks with Machine Learning

A widely accepted solution for verifying and controlling water leaks in cities lies in the exploitation of real-time sensor networks with machine learning (ML). Real-time sensor networks mainly include components such as sensors, network monitoring, cloud storage, and supporting applications.

Image: Asian Development Bank

This real-time sensor network that monitors water flow in real time when interfaced with machine learning models will be able to predict an anomaly (water leak) and accurately detect it in the water distribution network. This system can have multiple benefits for the city.

First, cities can expect a more accurate prediction of the magnitude and complexity of water leaks. This is beneficial for cities that have an aging water distribution system that is more prone to water leaks. Accurate prediction of scale and complexity can lead to correct and quick identification of water leak location. This, in turn, will reduce water loss and increase city revenue.

Second, accurate and timely location of leaks accelerates the pace and efficiency of troubleshooting. Shorter turnaround time to resolve water leaks is critical when entire sections of the water distribution system are often shut down in search of the leak point. Additionally, it ensures a reduction in collateral costs that arise from activities such as digging/rebuilding roads to locate underground pipe leaks.

Third, the subsequent efficiency in dealing with water leaks will reduce the downtime of water supply systems, thereby improving water supply reliability and water quality. This is a high priority for cities that experience an acute water crisis during peak summer times and are highly dependent on water for their economic activities.

Finally, this system of real-time sensor networks and ML is relatively inexpensive and simple to implement. This system can also help local, municipal authorities and citizens to anticipate conditions of reduced water supply and adopt water conservation measures. The adoption of such systems can contribute to the desired urban transformation.

The socio-economic, environmental, health and safety benefits that come from using a real-time sensor network with ML are obvious. Furthermore, with rapid urbanization, rising water prices and drying up of water sources, the deployment of advanced technological solutions such as ML is urgently needed if cities are to avoid exacerbated urban water stress.

Sherry J. Basler