Primary tab navigation

Managing Ireland's water supply with predictive analytics
Helping reduce loss and save resources

As urban populations increase, so will the strain on established and aging city water systems. The increasing demand for water will have to be supplied through water treatment and distribution networks that have inherent capacity constraints. These networks will therefore become more expensive to operate and maintain, which could lead to water shortages and lower water quality. Critically, the uncertainty in factors such as rainfall predictions, energy prices, and sensor data affects the decisions required to operate urban water networks.

At IBM Research - Ireland, researchers are dealing with these points of uncertainty through state-of-the-art optimization, data mining and data assimilation techniques in order to reduce water leakage and save money for citizens and water authorities in Ireland and around the world.

Our vision is an intelligent, integrated, and instrumented water system for our cities of the future, where both the quantity and quality of our water resources is constantly being monitored, predicted, and optimized, from rain to drain and back again.

Susara van den Heever, Research Manager, IBM Research - Ireland

Reducing leaks

In Dublin, latest published estimates illustrate that 29 percent of the water produced never reaches the consumer due to water leaking from the city’s aging distribution system1. Finding leaks is not an obvious task, as they’re often in pipes buried deep under the ground. Once leaks are found, replacing the infrastructure is a costly and time-consuming effort which can be disruptive when streets need to be dug up.

To ease this effort, Dublin City Council is collaborating with IBM Research - Ireland to better manage the water pressure in the network, allowing municipalities and utilities to reduce and prevent leaks without disrupting city life. IBM Research is combining predictive analytics, sensor data, and optimization models to recommend the best places in the network to install pressure reducing valves. Once the valves are installed, the models recommend the best settings to minimize leakage while meeting household, industrial and fire-fighting pressure requirements. This way, water can flow through pipes to the consumer at a pressure that is adequate for them, while also benefiting the utility by forcing less water out of leaking pipes.

A study in Dublin's Chapelizod neighborhood showed potential to reduce leakage up to 44 percent, allowing water providers to serve 15 percent more households without investing in additional production capacity or adversely affecting consumers.

We’re using the very latest sensor technology, providing dynamic data, to allow us to reduce our water consumption and find leaks … it’s replicable anywhere else in the world.

Tom Leahy, Executive Manager, Dublin City Council

Renewable energy saves water production costs

Treating and distributing water also has tremendous energy usage implications. In County Cork, Ireland, 95 percent of the water network’s energy cost is associated with pumping water into the treatment plant. By optimizing the pumping operations and making use of dynamic pricing schemes that incorporate renewable energy the plant can achieve significant savings.

Some energy providers now offer dynamic pricing schemes where prices change dynamically throughout the day, depending on the availability of renewables. For example, where energy is generated by wind turbines the price will drop when the wind blows. These pricing schemes typically yield lower overall prices, but introduce uncertainty into the decision process.

To this end, IBM researchers developed a robust pump scheduling prototype that takes the uncertainty in energy prices into account. It recommends pump schedules which minimize energy costs, while effectively hedging against unexpected price changes in real-time, similar to how a good investment portfolio hedges against market fluctuations.

An initial study indicates that cost savings of 10-20 percent are possible through a combination of switching to a dynamic pricing scheme using renewable energy and optimizing the pump operating schedules.

The future of water management

IBM Research - Ireland is continuing to expand their smarter water analytics to better understand how citizens use water, and how this affects a city’s water network. This improved understanding will guide the development of tools for water providers to better optimize their operations, and for citizens to contribute to a sustainable future where clean water is always available.

Share this story

More on predictive analytics and water management

Meet the researchers

  • Susara van den Heever thumbnail

    Susara van den Heever

    Research Manager,
    Constrained Resources and
    Environmental Analytics,
    IBM Research - Ireland

  • Bradley Eck thumbnail

    Bradley Eck

    Research Staff Member,
    Constrained Resources and
    Environmental Analytics,
    IBM Research - Ireland

  • Niall Brady thumbnail

    Niall Brady

    Lead Research Engineer,
    Intelligent Buildings/Energy Analytics,
    IBM Research - Ireland

  • Martin Mevissen thumbnail

    Martin Mevissen

    Research Staff Member
    Smarter Cities Technology Centre,
    IBM Research - Ireland

Share this story