Sensor Analytics in the Water System

“Water losses are becoming a huge problem,” says Andrew Whittle, the MIT Edmund K. Turner Professor of Civil and Environmental Engineering. “Cities in the developed world typically lose anywhere from 10 to 30 percent of water supplied through the underground pipe networks due to leaks and bursts, primarily caused by antiquated infrastructure.”

Cities are not only wasting water, but they’re also wasting the energy and money spent purifying the water in the first place. It also means that we tend to overlooked approach to stem the enormous volumes of clean water that are lost in leaking water distribution systems. While the leaks are bad enough, bursts have additional side effects, releasing massive amounts of water that can cause considerable damage. “A few years ago in Boston, we had a water burst in which the water got into the gas mains and affected heating systems,” Whittle says. “It caused no end of problems.”

Replacing infrastructure is an obvious solution, but it’s expensive and “massively disruptive,” Whittle says. The more common solution is to find and repair the leaks. Utilities typically send crews of technicians around the city with acoustic detection devices to locate leaks hidden underground. This is costly and inefficient, however, and is further limited by the fact that the work must be performed at night when it’s quieter.

Now, a research group that emerged from the Singapore-MIT Alliance for Research and Technology (SMART) called the Center for Environmental Sensing and Modeling (CENSAM) is trying a different approach: monitoring water systems with a network of wireless sensors. Whittle has joined CENSAM as a geotechnical engineering advisor, helping the team work with the Public Utilities Board of Singapore to set up a sensor network in Singapore called WaterWiSe. CENSAM goals are to develop and deploy new environmental sensor technologies and to incorporate data from these and other sources into representations of the natural and built environment that are linked across spatial scales, from an individual constructed facility to the meso-scale of the city-state, the regional ocean and atmosphere and the global climate.

Water networks are “a bit of a detour” for Whittle. For decades, he has studied large urban construction projects, analyzing soils and rock used as engineering materials. He’s worked around the world on projects such as the Big Dig in Boston, oil rigs in the Gulf of Mexico, and the Crossrail tunnels in London. All these projects share the same essential challenge facing CENSAM: identifying and understanding what lies beneath.

When CENSAM started looking into the water system monitoring problem, they were surprised at what little research had been done on the topic. “There had never been such a thing as a smart water system,” Whittle says. “Before our project, there was no proper instrumentation for water distribution systems.” Indeed, only recently has the cost of wireless sensors dropped to the point where sensor networks have become feasible. As prices drop further, one could conceivably place sensors just about everywhere. However, the cost and disruption of installing and maintaining them will continue to limit their numbers.

“Monitoring a water distribution system in a city is very difficult,” Whittle says. “There are tens of thousands of kilometers of pipes in this very complex network, so we’re looking for the best places to place the instruments for continuous monitoring to gain the maximum information. We’re trying to create a diagnostic of the entire system.” Compared to finding leaks, detecting bursts is relatively straightforward. “For the bursts, we’re able to detect pressure waves going through the pipes to figure out where they occurred,” Whittle says. “By the time people know about a burst, it’s usually because water is pumping out of the ground, flooding the whole area and causing massive disruption. If you can detect it when it happens, the damage is much less.” In many cases, WaterWiSe has detected bursts hours before the water reached the surface.

Leaks are a trickier problem, but the WaterWiSe project has made good progress here as well. The project has developed data-mining algorithms to analyze output from a network of pressure and acoustic sensors. “We’ve come up with analytics systems that are very good at spotting changes in state, detecting the little leaks that occur over time,” Whittle says. “We have an end-to-end system for making measurements and all the analytics supporting it.” The water management analytics are now being commercialized by a CENSAM spinoff company called Visenti, he adds.

Reducing energy consumption and checking for quality. WaterWiSe is also helping utilities predict usage to maintain consistent availability at the highest efficiency. In addition to the energy required to clean water, utilities consume massive amounts of energy to pump it around. This collectively costs utilities billions of dollars every year and adds to their carbon footprint.

“The water industry is catching up with the electrical industry in that they’re figuring out to make their system much smarter in how they pump and store it,” Whittle says. One challenge is that water networks are even harder to control than power networks, making it difficult to predict flow dynamics.

“In water pipes, the hydraulics really control everything,” Whittle says. “There are large pressure pulses in the pipes every day as people consume water. As we learn more about the dynamics of the system, we can better predict and plan for consumption. We look at the daily cycles to figure out when we need to pump more water for greater efficiency.”

In addition, WaterWiSe is using sensors to improve water quality measurements, thereby reducing the threat of dangerous water-borne diseases. “Deteriorating water quality is usually only discovered days afterward, after testing for bacteria, so we’re trying to come up with real-time warning systems,” Whittle says.

Whittle is currently working in London on the Crossrail project, a complex tunneling system for high-speed rail. Sensor Networks in Underground Construction. “One of the biggest challenges is to ensure safety and prevent damage when tunneling under buildings,” Whittle says. “We are using monitoring data from overlying buildings and attempting to relate this to the controls used in the tunneling machines.” Instrumentation is not enough, however. The data also has to be comprehensible. Years ago, Whittle was called upon as an expert witness in a lawsuit regarding an excavation collapse in Singapore that led to several fatalities. “There had been plenty of instrumentation, but the data wasn’t made available in an intelligible way to engineers,” Whittle says. “I became interested in how you can extract sensor data in a timely format that’s presentable and usable. CENSAM is now starting to work on the integration of monitoring data with models of soil-structure interactions for big underground construction projects.”

Urban Heat Pumps and Floating Storm Surge Barriers. Geotechnical engineering is increasingly important as crowded cities begin to build downward. Whittle is currently working on a project sponsored by the Chinese government that is considering the application of ground source heat pumps for seasonal heating and cooling buildings in large urban districts. “In the summer you store excess heat underground and recover it in the winter,” he explains.

Such a project is much more feasible in the few remaining countries like China that are still building new cities. It’s a lot easier when you’re starting from scratch. “The bigger challenge we face is renovating old cities,” Whittle says. “Underground urban heat exchange becomes a big challenge when you’re trying to share shallow underground resources, especially where you’ve got tunnels and subway systems, which give off a lot of heat. You also have to address who owns the land and the rights, and under what jurisdiction.” While projects like the Chinese heat pumps look to reduce the carbon emissions that cause climate change, other projects are helping cities survive climate change. “Sea level rise and big storm events need to be factored into all the coastal projects we’re building today,” says Whittle, who has been working with CVN, the consortium that is constructing mobile flood barriers to prevent increasingly frequent acqua alta events. “In Venice, the barrier must be invisible except when they need it, so it’s a minimally intrusive floating barrier. It’s a remarkable piece of engineering.”

According to Whittle, part of the attraction of geotechnical engineering is that it touches on so many different disciplines. The core focus is geology — analyzing the properties of the soils and rocks at a particular construction site. But that’s just the beginning. “We are at the intersection of many different disciplines,” he says. “We work closely with material scientists, and interact with the structural engineers. We have to deal with the uncertain loads of underground construction and look at the impact of adjacent facilities. There are no two cities where the geology is exactly the same, or where identical solutions get adopted.”

Watch Andrew Whittle Talking about
Smart Water Systems through Sensor Networks


MIT News