ANY industry that mislays 25-30% of its product in the process of delivering it might reasonably be thought to have a problem. Yet that, according to the World Bank, is the case for the world’s water companies. Though water is cheap, it is not free. According to a report published by the Bank in 2006, leaks even then were costing $14 billion a year. But to plug a leak you have to find it. Water mains are hard to inspect, particularly if they are underground. Many are old and thus decrepit. And outright theft is not unheard of, as the poor seek to fill their drinking vessels and the rich their swimming pools. An effective way of detecting leaks, both accidental and deliberate, would therefore be welcome.
TaKaDu, a firm based near Tel Aviv, thinks it has one. The problem, in the view of its founder, Amir Peleg, is not a lack of data per se, but a lack of analysis. If anything, water companies—at least, those in the rich world—have too much information. A typical firm’s network may have hundreds, or even thousands, of sensors. The actual difficulty faced by water companies, Dr Peleg believes, is interpreting the signals those sensors are sending. It is impossible for people to handle all the incoming signals, and surprisingly hard for a computer, too.
TaKaDu’s engineers have therefore developed a monitoring system called a statistical anomaly detection engine that is intended to identify clues in the data which might otherwise be missed. It applies a range of statistical tests (linear-regression analysis is one of the more familiar) to the data stream, and thus works out when the incoming signals are deviating significiantly from normal behaviour. Sometimes such deviations are caused by faulty meters. Sometimes they are caused by leaks. Either way, that is valuable knowledge.
To know what is deviant you have, of course, to know what is normal. Even a 1% change in flow rate can sometimes be significant, if it is persistent, but that is not always the case. Existing leak-detection systems therefore have thresholds built into them, to avoid false alarms. The price of this is that small leaks may go undetected and thus unrepaired, which often leads to larger leaks later. The detection engine attempts to work out what is important by using a process of continuous modelling to define normality. This identifies both obvious patterns—such as daily, weekly and annual flow-rates—and subtle ones, such as correlations between the behaviours of widely separated parts of the system that are brought about by things like similarities in network layout or in the behaviour of local customers.
If the system does detect an anomaly, it asks itself if there is likely to be an innocent explanation (a change in the weather, for example, or a public holiday). If not, it attempts to narrow down exactly where the puative leak is and alerts a human operator to go and have a look.
And it seems to work. Early last year Thames Water, which supplies London, tested the system on 3,000km (2,000 miles) of mains. The detection engine proved able to identify minor leaks up to nine days earlier than Thames’s existing systems could manage, and even picked up major bursts as much as 3½ hours more quickly. In light of that success, Thames has now extended the detection engine’s reach over the whole 10,400km length of its urban mains network.
Other firms are following suit. TaKaDu’s detection engine is now looking after the water supply in a dozen places in Australia, Europe and Latin America, as well as in Israel itself. One of its early victories was in Jerusalem, where it found that a fire hydrant was indeed being used to fill a private swimming pool—pulling the plug, as it were, on a very surprised water thief.
Back to TaKaDu News
Partner Login
