AusNet Services Demonstrates Power Of Data Analytics

AusNet Services has developed multiple analytics applications to extract consumer and network value from smart meter data.
Published: Wed 22 Apr 2015

With a mandated rollout of smart meters to most of its approximately 700,000 customers, including all residential and some commercial and industrial customers, Victoria, Australia utility AusNet Services has also taken the opportunity to deploy a range of data analytics solutions.

“With 30-minute integrated interval meter reads, as well as leveraging the distributed sensing capabilities of the network to provide power quality data at five minute intervals across all meters in the fleet, our aim is to provide customer benefits and to improve network performance and safety,” Jethro Kairys, network intelligence engineer in AusNet Service’s Network Intelligence & Analytics division, told Engerati in an exclusive interview.

With many utilities still uncertain how to move beyond billing and extract consumer and business value from the wealth of smart metering and smart grid data, AusNet Services offers some interesting lessons. Kairys will be presenting these, along with detailed insights into one particular application, ‘Loss of neutral’, at Asian Utility Week 2015.

Kairys explains that as an early adopter of AMI, the company developed much of its analytics capabilities in-house. “Our analytics system is centred around a highly optimized database, combining data from our asset management, SCADA, AMI and billing systems. The Network Intelligence & Analytics then develops analytics applications using this database along with a combination of ETL and analytics platforms from a number of vendors including open source.”

AusNet Services analytics applications

Kairys cites several examples of analytics capabilities that have been developed. These include identifying the phase connection of LV customers, detection of HV fuse operation, detection of a fallen live HV conductor on the ground, and monitoring of asset loading information. Visualization tools for both data discovery and network operations have also been developed.

Of these development the ‘Loss of neutral’ application, which detects failures on the LV network, is the first to be formally handed over to the company as a ‘business as usual’ process.

Detection of LV network failures

Kairys explains: “Each customer connection point is connected to the LV distribution network via an LV service conductor, which varies across the network in terms of age, condition and type. If the neutral conductor in this service cable degrades or fails, dangerous voltages will appear at earthed points around the customer premise, usually on metallic plumbing and taps. 

“The loss of neutral application detects the degradation and failure of these service cables and connections. It detects a potential fault, rates the confidence in the fault being present, and automatically passes the alarm to the network operations team, at which point a truck is rolled to confirm and repair the faulty connection at the customer premise.” 

The ‘Loss of neutral’ programme has been recognized with an Engineering Excellence award from Engineers Australia in 2014. Among outcomes from the programme is a 50% reduction in the number of electric shocks reported within the network.

Lessons from AusNet Services

Kairys says that his group is actively pursuing capabilities in many areas to better manage the network and improve the safety of the network. And whereas to date, most of the analytics developments have been used to determine the state of network components in the recent past and then reacting, efforts in the future will move towards determining these insights as they happen.

“We’ll then start integrating the outputs of these processes more into the way we manage our network in a more direct and automated manner, moving towards predictive capabilities in the coming years,” he says.

Based on the company’s experiences, Kairys offers several lessons for other utilities:

• Use known faults to educate and train analytics systems

• Trial and error is important and don’t expect everything to work first time 

• Look for “quick wins” and use your success to justify further work

• Measure the outputs of your analytics processes as these demonstrate the merits of your work and help justify further work. 

Looking ahead to Asian Utility Week, he says there are several topics that merit further discussion, including analytics workflow and interaction between the business and the ICT department, and real-time and predictive analytics and systems.

Jethro Kairys’s presentation at Asian Utility Week 2015 ‘Using AMI data to detect failed neutral service conductors and avoid shocks!’ is on June 9 at 11.30 am.