United Energy using big data for EV charging patterns

Tobie de Villiers with the competition winners (United Energy)
Tobie de Villiers with the competition winners

Victorian electricity distributor United Energy will work to trial new techniques to better identify EV charging patterns on its network, following a first-of-its kind competition between students at Deakin University.

United Energy, the Centre for New Energy Technologies (C4NET) and Deakin University collaborated as part of the project, with C4NET acting as the data host and facilitated the distribution of data. Ten student teams at Deakin spent two months developing models using de-identified real-world data from smart meters across the United Energy network area, in Melbourne’s south-east.

Related article: New ‘solar soak’ trial to see if EVs can stabilise grid

Accurately detecting EV charging behaviours on its network will allow United Energy to better upgrade its infrastructure and improve planning for the expected mass take-up of electric vehicles in coming years.

Two thousand de-identified and anonymised voltage, current and power factor load profiles were supplied to the teams, including 35 hidden electricity use patterns, known to be the result of electric vehicle charging.

Teams were asked to look at the anonymised data to determine how an automated system could detect, with high probability, data profiles consistent with those of EV owners.

The winning team developed a model that was able to accurately identify most of the EV load profiles hidden within the data, with the team’s algorithm only throwing up a handful of false positives.

United Energy and the Centre for New Energy Technologies (C4NET) helped fund the Deakin University project, with prize money of $20,000 split between the top five teams and $9,000 allocated to the winners.

United Energy head of network intelligence Tobie de Villiers said the top teams delivered very impressive results, with elements of each likely to be used in some way across United Energy’s network of 702,000 customers across Melbourne’s south-east and the Mornington Peninsula—and potentially other networks in the future.

“While EV take-up is still low within our network area, we think it’s very important to plan as best we can now to ensure we’re ready when more people decide to purchase an EV as their next car,” de Villers said.

“We have been impressed by how these talented students have researched and tested potential solutions that will provide us the insight we need to accommodate more electric vehicles on our network.

Related article: Labor’s EV commitment to help pave electric future

“Knowing how customers charge their electric vehicles gives us the data we need to understand what investments we need to make in our network to support our customers. It was good to see this level of industry collaboration and how we can utilise the education sector to use real data to resolve industry problems.”

The winnings Deakin University teams were:

 First place teamSecond place teamThird place teamEqual fourth place teamEqual fourth place team
Student membersMr Yotam Barazani Mr Thuc Nguyen Ms. Thi Hoa (Hannah) Nguyen Mr Yang CaoMs Lusi Xiao 
 Mr Benjamin Archbold Mr Bao Duong Mr Islam Khalil Mr Rongxin XuMs Brigitta Febriani 
 Dr Fatemeh Ansarizadeh  Mr Nabeel Maqsood Ms Xiaoyan Wang 
Academic mentorsDr Adnan Anwar Associate Professor Lemai Nguyen Dr PrasadDr Quan Vu Dr Ali Tamaddoni, Associate Professor of Marketing 
 Dr Valeh Moghaddam Dr Thin Nguyen Sankar Bhattacharya Dr André Bonfrer, Professor of Marketing 
 Dr Sutharshan Rajasegarar     
Previous articleTesla opens US orders for “badass” semi electric truck
Next articleDeadlyScience founder wins Transgrid Indigenous Achievement Award