DC Quick Charging and the Grid
When production EVs were first introduced in 2008-2012, the primary concern was limited range. Low-power AC charging was fine for local travel, but typically inadequate for trips exceeding one or two hundred miles in length.
Rapidly-expanding DC quick-charging networks are beginning to mitigate range concerns along selected travel corridors. CHAdeMO, J1772 DC, and Tesla’s Supercharger offer charge rates between 50-135 kW, typically providing an 80% charge within 30 minutes.
However, what happens to the grid when hundreds or thousands of DC quick chargers come online, and EV fleets increase by one or two orders of magnitude?
First: some baseline numbers.
The US generated approximately 3800 billion kilowatt-hours of electricity in 2011, varying from 300 to 400 billion watt-hours per month. Usage is not constant, but relatively predictable when considered as an aggregate. Grid operators are fairly sophisticated with teasing out usage patterns and adapting plant output to match anticipated demand based upon temperature and historical usage patterns.
Why is EV adoption, and high-power DC charging in particular, unlikely to be a concern for the national and regional grids?
Even in a best-case scenario, EV adoption rates will be relatively slow, allowing utilities time to upgrade infrastructure.
The US purchased approximately 15 million passenger vehicles in the last year, approximately 100k of which were EVs (0.7%). EV adoption is growing, but in the short term is constrained both by demand and supply of basic components, most notably lithium-ion batteries. Tesla recently announced a battery Gigafactory that can equal the current global lithium-ion cell production rate, and that will only be enough for 500k EVs per year. Suppose by 2020 Tesla’s Gigafactory is online and fully operational, and the US adds 500k EVs per year. At an average of 12k miles/year and 400 Wh/mile, this means the US is adding approximately 2.4 billion kWh of energy demand per year due to EVs, or increasing the total grid demand by 0.06% per year.
For comparison, the US grid has grown by about 1.4% per year for the last 20 years, or an average of 45 billion kWh/year.
As EVs saturate the transportation segment, the energy consumed will not dominate national grid usage.
Assume after a decade of this production – 2025 – we might have 5 million EVs on the road (50x higher than 2013), traveling a total of 60 billion miles. Vehicles in the US travel approximately 3 trillion miles per year, so EVs would represent approximately 2% of annual miles driven. Those 60 billion electric miles would consume 24 billion kWh, or 0.6% of annual demand.
Suppose EV adoption takes off, huge material and other technical breakthroughs are made, and the nation shifts approximately 50% of its travel to electric miles over a decade or two. 1.5 trillion annual EV miles will require around 600 billion kWh, or a 15% increase in total US grid demand. This is a non-trivial increase, but bear in mind the US grid grew 36% over the last 20 years.
The vast majority of EV miles are covered by low-power charging, often at night or at the workplace.
Tesla Model S is the best approximation of the future EV, which will have plenty of range for daily needs and can charge at special high-power DC chargers for long-distance trips. Tesla Model S vehicles have traveled 200 million miles in total, of which 10 million miles (5%) have been charged on the rapidly expanding Supercharger network.
This sounds surprisingly low, but bear in mind that the Model S has plenty of range for the majority of daily trips, and even on long trips the start and end will be charged on low-power AC.
An example 400 mile trip might look like so:
- mile 0: overnight low-power AC charge sets range to 220 miles
- mile 180: 135 kW DC charge, 20 minute charge increases remaining range from 40 to 160 miles
- mile 300: 135 kW DC charge, 20 minute charge increases remaining range from 40 to 160 miles
- mile 400: destination reached, 60 miles remaining. 160 miles added overnight on low-power AC
In this example, 40 minutes of high-power DC charging supply 240 miles of the 6.5 hour trip.
The Supercharger network will almost certainly benefit from the network effect, but even doubling the high-power portion of utilization as the network rolls out will only increase the portion of total vehicle miles driven to 10%. In our 2025 scenario with 5 million EVs, annual high-power DC charging represents 0.06% of total grid consumption (2.4 billion kWh).
Increasing levels of EV adoption will increase burst demand, but can also increase the grid’s tolerance for bursty demand and generation.
I wrote previously about the potential for using EVs as grid energy storage; utilities have been talking about vehicle-to-grid (V2G) technologies for a long time, but the benefit of V2G will increase even as EV burst demand grows proportionally and as intermittent renewable resources increase in scale.
Look at the 2025 scenario again. 2.4 billion kWh annually is 46 million kWh per week used for high-power DC charging. 5 million EVs at an average of 50 kWh apiece, there’s 250 million kWh of storage capacity in vehicles. If only 20% of vehicles are connected to V2G EVSE at any given time, and V2G can only access 20% of the vehicle’s capacity, then at any given point V2G grid storage will typically hold 150% of typical weekly energy demand for high-power DC charging. Further, if these 1 million connected EVs can provide up to 10 kW per vehicle, then total V2G power is 10 GW. The US has approximately 21 GW of storage capacity available today, so even over a decade this would represent a huge increase in grid-connected storage power.
Atypical demand patterns, such as vehicles departing from large sporting events or holiday travel, will place unusual localized demands upon the grid. Solving these challenges is a topic for another day : )