As electric cars grow in popularity, so does their energy impact. A Deakin researcher is leading an innovative project that will give energy utilities high-precision data on electric vehicles so they can manage demand on Australia’s power grid.
They are quieter and greener than most vehicles on the roads, but electric cars nevertheless leave an environmental thumbprint by way of the energy used to charge them. As more of us ‘go electric’, power grids will need to keep up with resulting spikes and congestion.
For Deakin Department of Information Systems and Business Analytics Professor Rens Scheepers, big data and predictive analytics offers energy providers a way to better forecast, and respond to, fluctuating demand patterns relating to electric car uptake. Demand insights into consumer behaviour around where, when and how often drivers charge their vehicles can contribute to a greener grid by improving capacity planning and integration of renewable sources.
Generating data on electric car usage
Professor Scheepers is leading an applied research project in partnership with United Energy and C4Net (Centre for New Energy Technologies) which aims to generate more accurate data on electric car usage, as they grow in prevalence and place pressure on power grids across the country.
'We are using predictive analytics to increase the sophistication of the intelligence drawn from energy use data. This granular data can help energy providers better forecast demand and consumption as electric vehicle use increases.'
So, where does the data come from?
'Energy meters relay back information about usage to energy providers constantly. Every 10 minutes there is a package of information that captures power consumption at that point. This is the basis of energy utility analytics,' explains Professor Scheepers.
The big data on electric cars
Professor Scheepers says the greater the accuracy of energy forecasts the less likely providers face disruptive, unmanageable spikes in demand.
Targeted education is key
'Utility companies then look at these packages in a time series and they can make high-probability inferences, for example that an electric vehicle is being charged daily at these locations.'
United Energy analysts are working alongside Professor Scheepers and his team to develop sophisticated but efficient tools [algorithms] that will enhance the reliability and depth of the inferences being drawn from the raw data. Due to the sheer scale of data, the tools need to be efficient enough not to over-burden server systems.
'If energy companies have data on where electric cars are in a city or suburb, and what the charging pattern is (for example, charged daily or short-term charging a few times a day) they can prepare and plan for that. It can inform critical decisions and support better planning for energy grid infrastructure.'
The team is looking at additional variables that can be added to the data to increase precision, such as the model or brand of car. 'Like refrigerators or washing machines, electric vehicles vary. Some charge differently depending on time of day, some are built for efficiency and some for performance, so finding out what kind of vehicle it is can boost data accuracy.'
Where to next?
Professor Scheepers said there are many 'unknowns' that need to be factored into the project. An example is the influence of 'early adopters' of technology versus late adopters.
'Early adopters are likely to be very energy conscious, to be passionate about doing it for sustainability. As we shift from these early adopters, we may see changes in charging behaviour as consumers seek to minimise costs over time. We also need to avoid ‘false alarms’ -- electric cars that are not actually being regularly charged, for example.'
Professor Scheepers hopes the outcomes of the project will lead to novel ways in which analytics can play a part in supporting sustainable energy use and emissions reduction.
This article was published by the Deakin Business School and features Professor Rens Scheepers. The original article can be read here.
The Centre for New Energy Technologies, C4NET, has contributed to the funding of this project. C4NET acknowledges the major funding contribution of its Core Participants and the Victorian Department of Environment, Land, Water and Planning. The views expressed herein are those of the author and not necessarily the views of C4NET