Data driving real-time leak detection at Sydney Water

In times of drought water saving becomes vitally important – not just by customers but by utilities as well.

With dam levels across Sydney at about 50% capacity, Level 1 water restrictions are in place and the Sydney Desalination Plant is supplementing drinking water supplies.

Keeping leaks to a minimum is an important part of the water saving effort, and Sydney Water is taking steps to bring its non-revenue water down.

Sydney Water Customer Hub Manager Darren Cash said the utility is focusing on being proactive in its communications with customers and predictive in how it operates its assets.

“Much of the predictive work is done so that we can act to avoid or minimise customer impact, identifying problems before they affect customers,” Cash said.

“This approach has generated great results. Since September 2017 we have been able to avoid or minimise water interruptions to over 200,000 properties that would have otherwise been affected.”

This is in part due to software service TaKaDu, which helps the Sydney Water team effectively use vast amounts of data to identify anomalies in real time.

The cloud-based central event management service enables early detection of network events such as leaks, bursts, faulty assets, telemetry and data issues, changes in demand and operational failures.

It does this by aggregating data from different sources and learning from previous events to continuously improve its predictions.

For example, in the harbour-side suburb of McMahons Point, TaKaDu used statistical algorithms and machine learning to identify a deviation from the predicted flow pattern in the district metered area that supplies water to the suburb.

“There had been no reported leaks in the area, so in response to the unusual flow trend we initiated a leak investigation by our network technicians,” Cash said.

“They used acoustic listening equipment at selected points within the zone and easily located a hidden leak adjacent to Sydney Harbour. The leak was completely invisible to the public, with the water making its way under the road and straight into the harbour.”

The team found the leak was losing about 1.3 ML of water per day, which would have cost Sydney Water customers nearly $1 million a year if it had been allowed to continue.

Cash said it could also have developed into a main break, which would have had an even greater impact.

“This event showed how we can effectively use data and monitoring to identify issues before our customers do, serving as an introduction to a more predictive future with all of our teams working together to resolve the problem,” Cash said.

For Tony Kelly, former Managing Director of Yarra Valley Water and member of TaKaDu’s advisory board, the technology is about helping utilities meet the increasing expectations of their customers.

“The days of utilities waiting for their customers to call and tell them there is a burst water main, pressure problem or water quality problem are rapidly coming to an end,” Kelly said.

“Soon it will be utilities that are proactively calling their customers to assure them the problem is known, is being addressed and that the system will be back to normal within x minutes.”

Kelly said new technologies help organisations evolve into digital utilities, but that there are a number of challenges that come with this for leaders, including:

• Which technologies should you buy?
• How can you future-proof your business and not get locked into proprietary systems with short half-lives?
• The cost of communications and network sensors is coming down and the business case is getting better by the year, but should you jump in now or wait?
• More and more data is being created, but how can you turn this into useful actions and knowledge for your utility?

“While there is a lot of debate about exactly what a digital utility looks like, everyone is likely to agree it must include real-time, automatic monitoring of a utility’s network, combined with smart analytics,” Kelly said.

“With greater visibility, a utility can prioritise jobs more effectively and respond more quickly; know immediately if there is a change in the configuration of the system or a rapid change in demand; and monitor pressure.”

Related video:

https://www.youtube.com/watch?v=sgzZwOEJrxI&t=9s