Powering cities with tap water

With the preparation, distribution, and treatment of drinking water consuming 2-3% of the world’s energy, policy-makers are recognizing the importance of improving the efficiency of urban water systems. Turbines mounted into the water pipes could help offset some of this power consumption. Now, researchers have developed a method to find the optimum positions of the turbines to maximize the production of electricity without sacrificing the reliability and pressure of the water network. They tested their approach on the water network of the Swiss city of Lausanne and published their findings in the journal Water Resources Management.

Water networks work under pressure. It is the only way to ensure that water reaches the top floor of a building and still provides a satisfying shower. Even in flat terrain, the water pressure varies across the network and can exceed required values, especially in hilly cities. Rather than letting this excess energy go to waste, researchers working with Anton Schleiss propose harvesting it by fitting the water pipes with small turbines.

Determining exactly where to locate the turbines to maximize the amount of power produced is far from simple. “Water networks form a grid, and when you make a change in one place, it can have repercussions throughout the network,” explains Irene Samora, a PhD student who carried out the study under the supervision of Anton Schleiss and his colleague Helena Ramos in Lisbon. On top of that, flow-rates across the network change constantly throughout the day, and a minimal pressure has to be maintained at all times.

To see through all of this complexity and find the optimal setup for the turbines, Samora and her coauthors resorted to an iterative computational optimization approach called simulated annealing, which, in an abstract way, mimics the way that heated metals are cooled down slowly to avoid the formation of flaws. “We start by testing an initial set of positions for the turbines. Then we make small random changes in this set of locations. If these changes improve the performance of the system, we keep them, and if they don’t, we reject them most of the time, but not always. Then we repeat the process until we find the best solution,” explains Samora. By retaining some of the poorly performing solutions, she can ensure that the algorithm explores as much of the water grid as possible.

She then tested her approach using five years worth of hourly flow measurements from the city of Lausanne. “Our goal was to determine the optimal locations for up to three turbines in a 17-kilometer sub-grid of the city’s water supply network,“ she says. The best-performing setup using three turbines – out of over six million possibilities – would have allowed her to recover approximately 5% of the energy required to pump the flow through the part of the network she studied.

While this part of the project focused on the development of the optimization algorithm, Samora expects to follow up her study using more accurate turbine data, measured in the lab. “Our next goal is to determine the actual economic impact of micro-hydropower production in urban water supply networks,” she concludes.