New research into electric-vehicle (EV) charging systems could provide technology for electricity networks to use car batteries for storing excess power.
Engineers at Cardiff University are developing computer programs that will allow distribution-network operators (DNOs) to manage the charging and discharging of large numbers of EVs by treating them as a single unit of consumption or generation.
This could help prevent the grid from overloading when many EV owners charge their cars simultaneously and offers a way to store electricity in times of low demand, which some say is needed to make large-scale renewable energy generation viable.
It would be difficult to get manufacturers to agree on standard technology
Dr Liana Cipcigan
‘Charging an EV is a big load compared with a residential house,’ principal investigator Dr Liana Cipcigan told The Engineer.
‘If we are looking at an average for a national level it is not a big impact. The problem is the clusters that will be created in local parts of the network.’
Technology referred to as an EV aggregator will manage the relationship between individual charging points and the electricity network by forecasting the power needed by the cars and the market prices for different times of the day.
Aggregator software will either control a group of charging points in people’s homes or manage a single connection to the grid that provides power to multiple chargers in locations such as office or supermarket car parks.
The initial stage of the research will involve studying data from EV trials to build up a better knowledge of customer charging patterns, before developing several algorithms for running the aggregators and then trialling them using simulation programs.
‘Considering residential vehicles, we don’t know when these vehicles will charge and this is completely new for the DNOs,’ said Cipcigan.
‘We are expecting random patterns — for example, if they go to a supermarket it is random how long they will stay and charge their car.’
The research project will explore three models of EV management: one controlled by a central program, one where some decisions are taken by a hierarchy of local systems and one where EVs make their own choices based on shared information.
Cipcigan said she expected the centralised or hierarchical models to be more practical because the third option would require all EVs to have compatible intelligent software but it would be difficult to get manufacturers to agree on standard technology.
Cardiff University will work with several commercial partners, including power company E.ON, utility infrastructure firm UTL and engineering consultancy Mott MacDonald on the two-year project, which is supported by a £93,000 grant from the EPSRC.