Big Data for Small Towns, or Stopping the $300,000 Leak

The water burbling down the rural hillside, nurturing the thick foliage, certainly looked like a stream.

In reality, the water came from a broken water pipe leaking 280 gallons per minute, notes Carl Alexander, the GIS director at White House Utility District (WHUD), a municipal water utility in north of Nashville that serves roughly 33,000 homes and businesses.

Over a year, that meant 1.47 million gallons lost a year, or enough for 2,239 homes. Put another way, $300,000 a year was going down the drain.

"Our water treatment plant was having to run one hour a day just to feed that leak,” Alexander said at the 2017 OSIsoft Users Conference in San Francisco. “Without us know there was a problem in that area, we would have never been able to stumble upon that leak.”

Leakage, or non-water revenue loss, is a chronic, growing problem for water utilities. On average, a utility should expect to lose around 10% to 15% of its water on the voyage from the treatment plant to the faucet, notes Gary Wong, Industry Principal for water at OSIsoft. Unfortunately, the total is often much higher, thanks to aging infrastructure, small budgets, stretched staff and the challenge of patrolling hundreds, and often thousands, of miles of underground pipes spread across square kilometers of ever-changing terrain.

Chicago at one point was losing 60% percent of its water. Manila was likewise losing over 60% at one point.

Water losses mean more than just higher bills and water stress for drought-stricken areas. Water utilities are often the largest consumers of power in their area, so leakage directly leads to lost energy as well as chemical consumption.

This is where software and technology comes to the rescue. WHUD segmented its service territory into 39 sub-districts. It then monitored the inflow of water through 79 flow meters and the outflow with SCADA data from the wastewater treatment center. By analyzing the differentials in the PI System and mapping the hot spots on ArcGIS from ESRI, WHUD could pinpoint leaks. In a clever twist, WHUD didn't track daytime use. Instead, it compared inflow and outflow between 1:00 a.m. and 4:00 a.m. when any large water consumption would likely be leakage.

The red areas indicate a region that might have sprung a leak, handy info for the tech on duty that morning. 

whudd

The results? WHUD recovered $900,000 worth of water in 2015 and 2016. It also avoided $200,000 in hardware upgrades that would have otherwise been spent to stop the leaks and thousands more in employee time.

“It was taking our water analysts almost all day, six hours, to calculate all of the information needed for our DMA zones. Now it takes less than ten minutes. “We are seeing $30,000 in savings just by optimizing his work flow.”

Results started to be achieved in the first 3.5 days of the program. In the future, WHUD hopes to mine weather, sewer and other data to see if it can optimize their system.

Other cities have seen similar results Maynilad, the utility for Manila, has cut water losses by 640 million liters, or down to 32%. Halifax in Nova Scotia saves 38 million liters a day.

WHUD is one of my favorite case studies. Not only are the savings pretty spectacular, it shows that Big Data and IoT aren't just for large organizations. The power of these technologies will percolate everywhere.

Check out the talk and the presentation here.