Horizon Blog

The real EV tipping point: artificial intelligence in domestic EV charging

As the 26th global Climate Change Conference (COP26) commenced this week and following recent petrol shortages and soaring increase in energy prices, the imperative shift towards more sustainable fuels and patterns of consumption is evermore glaring. However, the low uptake of available smart technologies is evident of distrust in the energy sector in the UK, as well as the difficulties of achieving technology adoption in domestic life.

Horizon’s Agile project ‘Domesticating Electric Vehicle Charging (DEVeC)’ will develop our understanding of the roles and implications of autonomy and smart technologies in residential electric vehicle charging. Over the next year, this project brings together a multi-disciplinary, multi-partner team to examine this, putting ordinary citizens central to our research through a methodology – design fictions with family interviews – that aims to get beyond the limitations of current consumer research.

Electric vehicles (EVs) offer to make both transport and the energy system more sustainable by switching to a potentially zero carbon transport fuel, foregoing dependence on imported oil, and for the same investments, have national networks of dispersed batteries (EVs) next to major sites of consumption (homes and workplaces). This fuel for powering EVs is also cheaper than petrol or diesel, both now and in forecasts, and can be especially cheap if generated by onsite production – such as via rooftop PV.  Using the EV’s battery to then feedback energy to the grid at times of high electricity demand and pricing, results in this fuel and flexible storage and provision becoming a source of income – about £4 per battery charge, and about £500 a year for feeding back to the grid. EVs are also already about 40% cheaper to maintain than combustion engines.

Realising the sustainability and affordability benefits anywhere near the magnitude possible however has two major hurdles. First, it requires a critical mass of EVs, and mass EV adoption will not happen until consumers commonly know they will benefit from sufficient cost savings to offset the greater initial cost.  Reducing battery prices and increasing economies of scale could see EVs and internal combustion prices equalise before 2025. However, at current prices, one is looking at eight years before the initial cost of an EV (approximately £7k over its fossil fuel equivalent) is compensated (@ 10k/pa miles and £200 & £150 saving on maintenance and road tax), without including income for charging from or discharging to the grid.

Secondly, after shifting us to EVs, the question of when we routinely charge needs to be considered, and dynamically! If all car drivers shifted to EVs overnight and were charged at the same time it would increase electricity demand by 10% in the UK – well within the range of manageable load fluctuation. However, if this 10% came on top of existing daily demand peak, e.g. between 6pm and 8pm, as it may well do, being the peak commuting and arrival home time, there would be limited transmission capacity to meet this demand.

The charging that will be required therefore is highly flexible, however our consumption of both energy and transport is notably inflexible. Efforts to address this fail by misunderstanding what drives human behaviour, especially for routines as emotive and culturally, politically and materially embedded as car, home and energy use.

A solution may well be one that does not entail considerable behavioural change. For this we need joined-up artificial intelligence (AI): smart EV charging in smart energy systems in smart homes – or at least homes fitted with more smart technology as standard than now.  For instance, the installation of smart EV chargers connected to cloud-based algorithms (AI-led EV charging). These generate useful data for consumers – such as the identification of network fluctuations. Left to our own devices, most people would not know what – if any – action to take presented with such extensive and complex data.  AI and smart chargers can detect, interpret and optimise data to initiate appropriate responses, therefore making car charging active and dynamic.

AI-led EV charging is thus a holy grail for the triumvirate of energy security, sustainability and affordability. As such it is central to governmental and market plans to achieve these requirements, as well as climate change emission obligations. But to what extent will the British public accept handing over control of their car to AI? And how often will they override what the AI wants? How to best incentivise smart charging and demand response? And what will be the implications of such AI and technologies in domestic life?

The answers to these questions are critical to the implications of EVs in domestic life and sustainable, affordable energy systems, as well as matters of privacy, autonomy, wellbeing and trust. With EV policy and AI already showing signs of favouring the already privileged, there are also questions of equity and vulnerability. Yet they remain unanswered and under-researched.

Domesticating Electric Vehicle Charging (DEVeC) aims to start addressing these questions.

Blog by: Lewis Cameron, Horizon Research Associate


Tags: , , , , , , , , , , , ,