Matt Brown examines some of the major uncertainties that exist around electric vehicles: when and where will people charge their cars? What behaviour will we see? And ultimately, what impact may EVs have on the generation and distribution of electricity?
2017 was a major year for announcements on electric vehicles. While there are still hurdles to be overcome, it seems that EVs have already won the battle for the future of transport.
Naturally at Pöyry we don’t rule out the possibility of natural gas, hydrogen and fuel cells playing their part, but in this point of view we consider only an ‘all electric’ future.
Imagine that by 2030 EVs have taken over as the vehicle of choice and rapid EV take-up has transformed our streets making them quieter and with much reduced local air pollution. Still a major uncertainty exists – when and where will people charge their EVs? What behaviour will we see?
Added to this uncertainty is the transformation of the electricity system that is ongoing – decarbonization, decentralization and digitalization. We expect increasing amounts of non-dispatchable renewable generation in the form of wind and solar, and for new technology and innovation to allow for much greater levels of consumer participation in the electricity market.
So the key question we address here is ‘what impact may EVs have on the generation and distribution of electricity?’
In a future with 50 per cent of all cars, buses and motorcycles ‘all electric’ across the EU28, the demand for liquid transport fuels is 68 mtoe p.a. lower (or 24 per cent of the 2016 total) and annual electricity demand is 330TWh higher, which is 11 per cent of EU28 final demand in 2016.
To some this may not seem much, but it is equivalent to adding a country the size of Italy to the electricity demand of the EU28. One reason why the impact is not larger is that EVs are efficient in turning energy into km travelled when compared to reciprocating engines.
It is important though to consider how the electricity is generated and compare primary energy use per passenger km to ensure a like for-like comparison.
What does this 330 TWh mean in terms of additional capacity – how many gigawatts of new plant may be needed as a result of this increased electricity demand? If the 330 TWh was considered on a standalone basis, it would require 45 GW of baseload plant, which is equivalent to 14 Hinkley Point Cs, or 125 GW of onshore wind capacity, which is almost three times the current total onshore wind capacity of Germany.
But what does it mean for investment needs when considered in the existing electricity system? The answer to this question is not at all straightforward, as it depends on how, when and where EV owners choose to charge.
|Figure 1: Filling the overnight trough: A winter day|
When will EVs charge?
Energy versus Capacity. Electricity demand varies across the day and across the year, and storing electricity is currently costly. As a result, traditionally electricity systems have plants that run baseload, mid-merit and peaking duty.
In addition, they hold reserve capacity to deal with unexpected peaks in demand. So, a peaking plant may only run for a small number of hours over the year. In addition, the wires that distribute electricity also have to cope with peaks in demand and be sized appropriately. With spare capacity on the system, additional energy demand could in theory be accommodated without the need for new capacity.
We use the United Kingdom and a simple scenario to demonstrate this. In the UK, peak demand currently occurs in the winter at around 18:00 due to the combination of heating, lighting and cooking demand.
Figure 1 shows demand on a winter’s day together with wholesale prices. Imagine all cars (that are charging) slow charge at the same time overnight in a seven-hour period starting at 23:00.
In this example, it would be possible to accommodate over 21 GW of charging demand before a new peak demand period is created.
However, if charging began earlier in the evening, say at 21:00, then around 4 GW of charging demand creates a new peak.
If charging starts when people return home from work, at say 18:00, then the impact is direct and new capacity is required immediately. Assuming a 50 per cent penetration of EVs in the UK, the demand from charging over these seven hours translates to 20 GW and so can in theory be accommodated within the existing generating capacity.
|Figure 2: Germany generation and prices for week 45, 2017. Sources: 50 Hertz, Amprion, Tennet, TransnetBW, EEX, EPEX|
The energy transition
However, the situation both today (in some countries) and in the future is not well represented by the above characterisation for a number of reasons, not least: the continuing increase in non-dispatchable generation such as solar and wind; and the growing potential of flexible demand from appliances and EVs, to balance supply and demand in a future smart, digitalized, decentralized energy system.
As the amount of wind and solar grows in the electricity system (whether centralized or decentralized) the shape of electricity demand will no longer be the main driver for when to charge an EV, as low electricity prices will not necessarily coincide with periods of low demand overnight.
Rather than charging overnight, it will make sense for EVs to charge (and for other flexible loads to run) during a sunny or windy period. Assuming that the average EV user charges once a week, then as shown in Figure 2 the best day to charge in Germany during Week 45 2017 is the 10th November.
The price of electrical energy on, say, a 15-minute dynamic basis, can provide the right signal about when best to charge an EV.
Consumers will, if so enabled by technology in the future, respond to the price signal, increase aggregate demand and reduce the level of curtailment on zero or negatively priced renewable generation.
Therefore, flexible demand will allow for more wind and solar to be built on a profitable basis as a result.
Consumers will set their preferences and the EV will do the rest. Such preferences may be that they never want less than 40 per cent charge in their EV and are willing to pay a maximum amount per day for their electricity.
In addition to EVs, large controllable loads in the home (washing machines, tumble dryers, immersion heaters) will also be programmed to switch on at such times.
These interactions will likely be automated through a home hub system rather than requiring any human intervention.
The pricing of electricity will also need to be dynamic so that as demand increases, prices respond and additional demand sees its impact on price levels.
For this to work, of course, consumers will need smart meters that can record demand on this 15-minute basis and retail prices that reflect the changing value of electricity in each 15-minute period. In a world in which flexible demand responds to changes in the price of electrical energy, what implication will this have for the distribution of electricity?
Can EVs solve the gird problems they cause?
Distribution and Diversity. In practice, our electricity systems rely on diversity of demand to hold down costs. The fact that people use electricity at different times means that the capacity of the system is lower than it would need to be if they used it at the same time.
If everyone in a street put their electric ovens on at the same time, then the low voltage fuse at the street substation would blow and supply would be lost to everyone on the street. If people were willing to pay more so they could all have their ovens on at the same time and not lose supply, the distribution company could come and put in a bigger cable and potentially a bigger transformer at quite a significant cost (tens of thousands of euros per street).
By dint of natural diversity, the cost of distributing electricity to consumers is kept lower when we share assets. With non-smart systems it doesn’t matter to the residential consumer when their electricity demand occurs as settlements are based on half-hourly or hourly profiles rather than on actual demand.
In a 50 per cent EV penetration scenario, if all the EVs in a city street slow-charged at the same time, major investments in the electricity distribution system would be required.
In a world in which flexible demand is chasing low electricity prices, there is an incentive for consumers to charge their vehicles at the same time. Natural diversity will reduce and distribution systems will need even greater levels of investment.
The cost of distributing electricity will be low most of the time and then increase significantly when grid capacity grows scarce. There will exist at times a tension between the cost of electrical energy and the cost of distribution. The cost of delivered electricity will vary significantly with time and location.
The impact that this has on the electricity system will depend on the underlying characteristics of the system. In systems with high levels of hydro storage, the variation in electricity prices driven by wind and solar will be low.
The incentive to all charge at the same time will be reduced. In systems built to cope with mainly electric heating, home-based slow-charging demand is proportionally less important as the distribution system is already built for larger loads (as long as one avoids having the heating on at the same time as the EV is charging).
One solution is a system of dynamic pricing that reflects the cost of electricity at a specific location. The pricing option could be a variation on nodal pricing, common in many electricity markets, but which is extended down to the local distribution level, even to a price at the top of a city street.
Whatever the form, the key will be reflecting the cost of electrical energy and the cost of distributing electricity to an appropriate degree of temporal and geographic resolution.
Unless customers see the cost of their actions through locational dynamic pricing of electricity, it is likely that very significant investments in electricity distribution infrastructure will be made unnecessarily.
In the interim, a system of pricing distribution use on a kW capacity rather than kWh energy basis to reduce individual consumer peaks may alleviate the issue. Some trials of command and control by distribution companies, in which the distribution company controls the charging time, have taken place but it is difficult to see how this is consistent with a smart energy future.
One of the key questions that remain with EVs is their ability to inject energy back into the grid economically. With current technology, the received wisdom is that cycling of the EV battery has too great a cost (in reduced battery performance and early replacement) for injecting back into the grid to be economic for much of the time.
But if the scarcity of the wires were priced accurately, the economics will change. In addition, as the number of EVs increases, this will lead to more periods of grid scarcity with greater value.
Battery technology for EVs will no doubt improve over time and, if re-injecting from an EV creates a value that can be captured, then developments will likely lead to a lower cost of re-injecting.
Alternatively, static batteries in the home or in the local grid may be the answer to reducing congestion on the local distribution wires. Evidence from Norway suggests that avoiding grid capacity fees is a major driver of residential battery deployment.
The economics of EVs reinjecting electricity into the system could end up being based on the cost of storing energy in, and re-injecting energy from, an EV (or static) battery versus the cost of grid reinforcement. So, when you want to charge your EV at a specific time and there is local grid congestion, you will charge from other EVs that are discharging energy in your local street or area.
The available re-injection capacity from EV batteries will be limited by the connection to the grid and by the ability of the grid to distribute electrical energy. An EV battery can deliver a large amount of power to the motor by comparison to its grid connection
Even with this limit, the GW of capacity that could be delivered by a 50 per cent penetration of docked EVs is large and this could lead the way to an electricity system premised on renewables and EV battery storage (as long as the issue of rate of change of frequency can be addressed).
So we may find that the problems that EVs cause in the future are actually solved by EVs themselves (either directly or indirectly through advancement in battery technology in static battery applications).
This will be the case as long as the correct price signals are seen. And this may mean having a new electricity market design fit for the future that prices not only the electrical energy dynamically within day, but also the grid congestion on the same basis down to a local level.
It is uncertain exactly how EVs may develop and given this uncertainty a flexible pricing system may be the best solution to make the most of the flexibility they will bring.
The truth is that EVs will fundamentally challenge the whole of the electricity industry – from the approach and remit of regulators, to the licences that define the activities of companies, as well as the settlement processes. It will also impact business models across the industry as we move behind the meter and allow for multiple suppliers to supply each home.
Developments are already being seen in some markets but a huge amount of work remains to be done.
If we are too slow to bring about these changes, we risk making generating capacity and grid investments in the shorter term that become unnecessary in the long-term and that burden consumers with higher costs for years to come.
Matt Brown is Vice-President of Energy for Western Europe, Middle-East and Americas, at Pöyry