5 Comments

Summary:

South Bend, Ind., is using a sensor network and IBM software to prevent its sewer system from dumping sewage into rivers or backing up into citizens’ home, but it’s just a microcosm of global trend toward solving urban maladies using big data.

shutterstock_74603293

If you don’t spend a lot of time thinking about the design of your city’s sewer system, you’re not alone. If you’ve lived in South Bend, Ind., though, it might be a different story. There, heavy rain has sent sewage overflowing into the St. Joseph River and backing up into basements in the old part of town. Until data came to the rescue.

However, South Bend’s sewer situation is just a microcosm of the woes facing cities as they continue to grow — the planet is estimated to have 9 billion people by 2050, and much of that growth will happen in urban areas. Traffic congestion, air pollution, crime and just general inefficiency are among a host of problems that plague growing cities and that, data suggests, actually grow at a faster rate than cities themselves. As we’ve reported numerous times over the past few months, however, big data — from Xerox’s traffic-management systems in Los Angeles to disease-predicting models in New York — is increasingly looked at as part of the answer to many of these problems.

South Bend’s sewage problem

South Bend’s problems stem from its outdated combined sewer system that mixes rainwater and sewage in the same pipes. Heavy rains mean waste isn’t easily diverted away from the river and to the sewage treatment plant; backups occur when large streams of water hit bottlenecks in the system. (A rather detailed explanation of the system, straight from the city (and written before the IBM engagement), is available here.)

Fixing the problem meant either spending $120 million on rebuilding the sewer system to add capacity, or spending a fraction of that, $6 million, on several dozens sensors and a new software service to make sense of the constant data they’re putting off. South Bend chose the latter option, and it has proven to be a wise decision thus far.

According to South Bend Director of Public Works Gary Gilot, the project not only saved the city $114 million, but it also has reduced the number of sewage overflow incidents to 1 per year from 27 per year. The solution was a system that married 116 sensors (developed at the nearby University of Notre Dame), strategically dispersed throughout the 500-mile sewer system, with IBM’s cloud-based Intelligent Operations Center software. The reduction in overflows also saved the city another approximately $600,000 in government fines.

A screenshot of IBM’s IOC software.

With the sensor network and cloud software in place, the city can see in real time where problems are arising, as well as where there’s excess capacity. Armed with this information, city employees can intelligently divert sewage flows to ensure as much as possible makes its way to the treatment plant and as little as possible finds its way into people’s homes.

Gilot analogizes South Bend’s sewer evolution to the change from timed traffic signals to today’s signals that alter the timing of lights based on the flow of traffic: “We’re essentially taking something that’s been used for decades [in traffic management] and applying it to sewage instead of the flows of cars and trucks.”

Opening communication channels for data

But, like most cities, South Bend’s problems aren’t limited to sewers, and Gilot said he’d like to use data to solve some of the other ones. For example, he wants to connect data from the city’s water works department — which keeps tabs on which properties have water service turned on or off — with code- and law-enforcement data to cut down on vandalism and other crimes associated with abandoned properties. Knowing which areas have a disproportionate percentage of water-less properties will let those departments know where to focus their efforts, Gilot said.

Craig Hayman, general manager of IBM Industry Solutions, said his team is working with all sorts of cities to help solve their specific problems. In Rio de Janeiro, IBM is helping the city try to predict mudslides and prepare for both the 2016 Olympics and World Cup. In New York, IBM and the City University of New York are using the IOC software to analyze and monitor the city’s efforts around solar energy production. Elsewhere, IBM’s Smarter Cities team is working on everything from crime prevention to social services to, of course, predicting traffic.

Still, while big projects get the attention, citizens might feel the effects of municipal analytics efforts in much more mundane ways. Try going to city hall and getting a permit or doing anything that requires dealing with multiple departments, Hayman explained, and it can be a nightmare because most still operate in silos, disconnected both physically and digitally.

“This is very similar to the early days of the internet,” he said, referring to the pre-Web 2.0 and social web eras when data didn’t flow so freely between sites, service providers and applications. In five to 10 years, though, Hayman thinks cities will operate much like the web does today, so while cities are saving money on operations, you’ll be saving time bouncing between departments as you try to get that building permit.

Feature image courtesy of Shutterstock user Abramov Timur; abandoned house image from Wikipedia Commons.

  1. Very well written post…
    This is one of the excellent use cases of Big Data solutions helping to create smarter cities without huge investments…
    I just wish my city (Gurgaon) authorities could take a cue from such implementations…

    Share
  2. Maurice Walshe Friday, August 31, 2012

    How is only 500 miles of sewer 116 (woop do do) sensors big data?

    And trafic models have been used for decades before the internet existed – how do you think AT&T designes phone networks.

    Share
    1. Maurice: Yes. And yes. And yes!

      Traffic models were used for phone networks and roadway planning and yes, sewage networks, for DECADES prior to “Big Data” and the internet. I learned about this is my operations research classes back in the dark age of 1986. And it was nothing new! Dijkstra wrote algorithms that were implemented, with ease, using very “primitive” computers by comparison to what we have now.

      All of these models and systems have been in place for decades. I am puzzled as to why this is presented as anything new. I wouldn’t be surprised if South Bend, Indiana already had implemented something like this, 20 or 30 years ago, in fact!

      Share
      1. Well, perhaps it’s just that South Bend has finally caught up? Although, gadetry to workaround an inadequate sewer system wouldn’t be my preference as a rate payer. The sewer has obviously not been evolved by the city over time according to the demands and developments of modern times.

        I’d prefer living in a city with an old-fashioned but adequate sewer, rather than one with a smart but inadequate one. Just wonder what would happen if the power went out in a heavy rainstorm.

        Share
  3. very nice . the content is awesome

    Share

Comments have been disabled for this post