Following a rapid increase in the market share of electric vehicles in countries such as Norway (where personal EVs now account for over 90% of new vehicles sold) and multiple manufacturers aiming to enter the market of all types of EVs (including motorbikes, buses, and freight transport), questions on their impact on power systems have been raised.

Overall, the total demand from EVs, including personal use, commercial, and public transport, is not expected to account for a sizeable share of total electricity demand in the short or even medium term as higher efficiency batteries and charging stations are developed. Nonetheless, EVs present significant challenges and opportunities for planning distribution systems, making the exercise of EV load forecasting an essential tool for the growing field of e-mobility.

Total electricity demand from EVs is modest

According to the IEA[1], the total electricity demand from EVs in 2035 is expected to reach 1,400 TWh by 2035. Most of this demand is expected in China and the USA, where in 2023, 38 TWh and 22 TWh, respectively, were attributed to the charging of EVs for personal use. While these numbers look high, it is worth considering their magnitude relative to the total electricity demand of power systems in these countries.

A bottom-up approach for forecasting EV electricity demand can show their magnitude in power systems worldwide. As of 2024, the electricity demand per km required by EVs for personal use in the market ranged from around 0.05 kWh/km for electric motorbikes to 0.2 kWh/km for large automobiles[2][3]. Utilisation varies between countries, with average mileage ranging from 7,800 miles per vehicle in countries such as the UK to over 13,000 miles per year per vehicle in the US[4]. This represents an annual electricity demand of 1.5-2.4 MWh per vehicle; in the UK, where 1.2 million electric cars are circulating (of the total 33 million cars), this represents 2.4 TWh or 0.8% of the total electricity demand. However, in most countries, EV penetration is considerably lower, and their electricity demand is negligible, as shown in the figure below.

Figure 1  Total EV electricity demand and share of overall system demand in 2023 and 2030

Source: ECA analysis based on IEA data

In this regard, forecasting EV electricity demand becomes a question of forecasting the market share of EVs in the current market. At the national level, this can be better forecasted through a bottom-up approach with assumptions regarding utilisation, the capacity of charging stations, and battery efficiency. However, this approach may prove insufficient when assessing the expected impact at an hourly level, and, more importantly, how this demand is spatially distributed.

Forecasting charging patterns and location is key

Forecasting the impact of EVs on the system’s load profile becomes more complex due to the different charging behaviour by customer type and level of any coordination. Households, for example, may charge their vehicles after work or during the night in a highly coordinated manner. In contrast, public transport may stagger their charging periods according to the schedule of vehicles in circulation. The figure below illustrates different charging patterns in EVs for personal use and public transportation.

Figure 2  Example charging patterns on an average day

Source: ECA analysis

Personal use EVs can create challenges for distribution networks in residential areas. However, forecasting their impact at the system peak is more sensitive to behavioural patterns than the number of EVs in circulation. To tackle this, load forecasting ought to consider EV owners’ patterns of use and charging, the type of charging being used, and the elasticity of this demand with respect to prices and availability of public charging stations (mainly in areas with a high concentration of office space).

For example, if 100 car owners[5] opt for nighttime fast charging, this could add over 1,200 kW of demand at the system peak in a relatively small neighbourhood of 400 households. Conversely if 30 car owners opt for slow charging and 20 for work-place charging while the remaining 50 opt for nighttime fast charging, the impact on peak demand would be 34% lower.

Figure 3  Impact of different charging patterns in a residential area

Source: ECA analysis, note that public charging would not be included in the EV demand in the example residential area

In emerging markets with restricted charging station availability, a bottom-up approach based on average utilisation may be enough to capture these effects. In mature markets with multiple charging options, constant tracking of consumption patterns and other real-time factors such as weather, road works, and public transport availability may be needed to forecast EV load for network planning.

Beyond demand

Understanding the impact of EVs on demand forecasting for power systems reveals challenges for generation companies, utilities, system operators, and market operators. These implications reach beyond infrastructure planning; the rollout of EVs also poses questions for regulators on recovering the costs of charging infrastructure from customers and the necessary economic incentives to minimise the challenges of highly coordinated charging.

Finally, it may be worth stepping back and considering the EV category not as merely an additional end-user load but as a potential supplier working in tandem with RES and ESS to ensure the rapid, reliable, and sustainable decarbonisation of electricity supply around the world.

[1] Global EV Data Explorer – Data Tools – IEA

[2] Electric car batteries explained | Octopus EV

[3] Best A1 & A2-friendly electric bikes (2024) | Specs & Prices (bennetts.co.uk)

[4] Fact file: Americans drive the most (frontiergroup.org)

[5] For the purpose of this example, each EV is assumed to have a 50 KWh battery with a peak charging demand of 12KW, corresponding to a commercially available fast charger