Water consumption analysis finds faults, leakages

IBM Haifa research lab and partners develop "Big Data" program.

By SHARON UDASIN
Alon Tal testing water supply 370 (photo credit: Courtesy Alon Tal)
Alon Tal testing water supply 370
(photo credit: Courtesy Alon Tal)
A new data analytics technology will enable water utilities to closely monitor water consumption patterns in order to pinpoint system glitches or leakages from afar.
In partnership with Israel’s Arad Metering Technologies – known for its Dialog3G and City-Mind smart water metering systems used around the world – Haifa-based IBM Research Lab scientists have developed “Big Data” algorithms that can harness the plethora of consumption data collected by utility meter readings and sensors.
Big Data is able to outline typical consumption patterns in individual households at various points of the day and from there, identify if an abnormality is affecting the system.
“When you have this data there are a lot of interesting things you can do with it, which is the beauty and the fun of the program,” Pnina Vortman, head of the Smarter Water Program at IBM Research Labs in Haifa, told The Jerusalem Post on Monday.
The partnership between Arad and IBM began four years ago, when IBM researchers began analyzing data collected by Arad smart meters from the north of Israel. From there, they were able to develop a series of algorithms based on machine learning that are able to analyze consumption data and reveal consumption patterns, Vortman explained.
“The more data you have, the better they are,” she said, noting that the machine “learns” and creates the algorithms based on the data.
“Fortunately enough with the Arad technology there is a lot of data.”
One problem that can be identified by analyzing the consumption patterns is a faulty metering system. If a meter reveals almost zero consumption, this could reveal that a family is on vacation, an apartment is empty, a meter has broken or system tampering has occurred, Vortman noted.
Today, water utilities send technicians in such cases to identify the source of the minimal usage.
However, by incorporating the new Big Data system, a more precise analysis of the family’s consumption patterns and historical information will allow the utility to reduce the need for an on-site technician, Vortman explained.
Field tests conducted by IBM and Arad have shown that the analytics will reduce the requirement of a worker visit by about 50 percent, according to the companies.
Likewise, Vortman said the technology can also predict if a leak is occurring in a home or an apartment building.
In addition to individual apartment water meters, an entire building has a common consumption meter. If a leak occurs in the overall system, the extra consumption will be divided equally among the histories of all of the households and will be easily identifiable. The analytics technologies can detect even the smallest of leaks, Vortman stressed.
“By analyzing the data that is produced during the night you can detect very small leaks,” she said. “If you see a very steady type of consumption, than it’s a leak.”
The new IBM consumption pattern analytics program will be integrated into the Arad smart metering system.
Vortman added that municipalities employing the system will have the opportunity to choose the service among their portfolio of options available from their Arad meters.