Computerised scheduling for electric vehicle charging
Computerised scheduling of charging for electric vehicles (EV) could ease the strain on grids and still meet customer demands, according to researchers at Southampton University.
The team has devised a system where electric vehicle owners use computerised agents to bid for the power to charge their vehicles and also organise time slots when a vehicle is available for charging.
Increased uptake of electric vehicles is predicted to put considerable strain on local distribution networks, especially at peak times.
To address this, some distribution companies that are already experiencing significant EV use (such as the Pacific Gas and Electric Company in California) have introduced time-of-use pricing plans for EV charging that attempt to dissuade owners from charging their vehicles at peak times, when the local electricity distribution network is already close to capacity.
Researchers have also begun to investigate the automatic scheduling of EV charging — where individual vehicle owners indicate the times at which the car will be available for charging, allowing automatic scheduling while satisfying the constraints of the distribution network.
However, because these approaches separate the scheduling of the charging from the price paid for the electricity (typically assuming a fixed per-unit price plan), they are unable to preclude the incentive to misreport (for example, an owner may indicate an earlier departure time or further travel distances in order to receive preferential charging).
The researchers’ solution was to leave some available units of electricity unallocated, as Dr Alex Rogers of Southampton explained to The Engineer.
‘The clever part of the mechanism ensures that you don’t need to strategise over these bids — you can report them truthfully to the system in the knowledge that it is impossible to get a better deal by waiting later, or saying you need the car earlier, or you are going to driver further than is actually the case.
‘All other scheduling solutions assume that this information will be truthfully revealed by the user. In our case, we ensure that it is always in the user’s best interest to report this information truthfully, as by doing so they get the best possible price for the electricity.’
To test their system the researchers used data from a real-world trial of EVs in the UK. By retrospectively applying their mechanism to the data, the number of electric vehicles that could be charged overnight, within a neighbourhood of 200 homes, increased by as much as 40 per cent.