Leveraging data to drive your EV charging station strategy

How a data driven approach can make effective public EV charging infrastructure deployment possible

As the world around is quickly transitioning towards electric vehicles, there is an increasing need for an established charging infrastructure network. Many types of organisations can profit from the current emerging demand and targets in the public charging sector. However, it can be a burden identifying the best investment strategy and opportunities. This article focuses on how the data scientists and energy specialists at Sia Partners work together to leverage a solid data-driven approach that facilitates informed decisions in order to help our clients build a charging infrastructure in line with their goals.

EV adoption is taking Europe by storm …

In the last two years, business headlines have reported on record-breaking valuations for Electric Vehicle (EV) companies. Today, 2 out of the 3 most valued car manufacturers are native EV companies(1), in spite of presenting a much lower share of total car sales and less stable cash flows compared to the more established traditional brands.

The number of EV owners has been growing at a rapidly increasing rate over the past decade. More EVs are being sold now on a weekly basis than the full calendar year in 2012. Back then, only 120,000 new EVs entered the market. In 2021, sales of EVs doubled to twice the 2020 estimate, i.e. a staggering 6.6 million. In 2022, sales kept strongly rising with 2 million EVs sold in the Q1. The global EV fleet is now estimated at 18.5 million cars, more than triple that of 2018.

Europe is a leader in the transition towards electric transportation both in terms of volume and market share. In 2021, over 2.2 million xEVs were registered on the continent, of which around 54% battery electric vehicles (BEVs), representing 10.3% of total new vehicle sales. Currently, only China is doing better with a 12.3% market share for BEVs(2). To put things into context, only 0.63 million EVs were sold in the USA, constituting just 4.5% of new car sales(3).

These trends are likely to remain. The two bar charts below show the growth of xEVs as an alternative to ICEVs both in terms of sales and share of the total fleet on the road. The EU targets an electric vehicle stock of 40 million vehicles across the continent by the year 2030(4). And that is only the start. 6 June 2022 the European lawmakers voted to support the proposal of the European commission to ban the sales of fossil-fuel cars from 2035(5).

This switch to EVs can even go quicker than currently foreseen. Next to political ambitions, economic and social factors are influencing the market. While today BEVs are usually still more expensive than ICEVs, on average, their Total Cost of Ownership (TCO) tends to already be lower for most European countries.(6)(7) The major reason for this is the difference in energy costs per km. The graph below clearly shows that these costs are around twice as high for diesel or petrol cars than their electric equivalent in Belgium(8). In general, as fuel costs of an ICEV amount to over 20% of their TCO, the difference is not that insignificant. This also means that ICEVs are more sensitive to energy price fluctuations. Due to the recent supply chain disruptions influenced by covid, the sudden increase in demand post-covid and the ongoing conflicts with Russia, the world’s largest exporter of petroleum, fuel prices are soaring. As a result, the total costs of ownership of an EV keeps getting more and more competitive compared to similar fossil-fuel powered vehicles, making EVs more attractive to the public(9).

… requiring a blooming network of charging stations

Investments in EV infrastructure are required to support both the increasing electrification of the car fleet and further EV adoption in the future. Most importantly, the installation of sufficient car chargers or so-called “electric vehicle supply equipment” (EVSE) is crucial because electric vehicles need access to a sufficiently developed network of charging points to support their further roll-out. The scale of the required transformation is enormous.

The desperate need for (public) charging infrastructure has not gone unnoticed on a legislative level. An amendment to the Alternative Fuel Infrastructure Regulation (AFIR) requires each Member State to expand charging capacity in line with zero-emission car sales, and to install charging and fuelling points at regular intervals on major highways. For electric charging, the maximum interval between charging points is set to 60km. Deployment targets for 2030 aim for a ratio of 1 charger per 10 electric cars. With the current estimates for electric vehicles ranging around 40 million by 2030, that would mean 4 million public charging points need to be installed.

Many consumer, climate and research organisations claim these numbers will not suffice.

“6.8m public charging points are required across the EU by 2030 to reach the proposed 55% CO2 reduction for passenger cars — almost twice the number put forward in the Alternative Fuels Infrastructure Regulation (AFIR) proposal.”

~ Oliver Zipse, ACEA president and CEO of BMW.

Recent cross-industry research by ACEA estimates that no less than 6.8 million (slow) public charging points, or 2.9 million charging points in case of a balanced fast charger utilisation rate, would be required to achieve Europe’s goal of a 55% CO2 reduction for passenger cars. That would roughly translate to installing between 6,000 and 14,000 charging points every week until 2030, in comparison with the current 2,000 per week(10).

… which is both an opportunity and a challenge for various organisations

The development of a charging infrastructure is both a challenge as well as an opportunity to a wide variety of actors, from public authorities and policy makers to companies and entrepreneurs. It is estimated that around €280 billion will need to be invested in charging infrastructure by 2030, of which around €80 billion to buy and install public EV charging infrastructure. Even though this is a huge investment, it seems attainable under the right conditions. For example, this sum would be only 16% of what is invested annually in the roll-out of 5G and glass fiber infrastructure in the EU(11).

As the transition towards EVs is increasingly becoming an essential mobility goal for governmental instances as part of their broader sustainable strategies, governments are putting measures into place to facilitate these required investments. As a lot of different actors need to be involved, these measures can take a lot of different shapes. They range from tax breaks and subsidies for opening up private charging equipment to the general public, to personally subcontracting charging point operators. Those players can then each decide where they fit best in the supply chain.

The opportunity is not only big, but also very diverse. Not all chargers are equal. A wide range is needed for a well functioning charging network. Depending on the location and the related use cases, different types of chargers will be a better fit. In general three big public charging use case segments can be identified:

• Routine (or On-Street): long-duration charging close to home or work.
• Activities (or Destination): medium-duration (1–3h) charging while the user is occupied with activities like shopping, exercising, visiting, eating out, …
• Travel (or Transit): (extremely) short-duration charging, when the user is interrupting a long distance trip, not unlike in the case of a classic gas station.

These segments are illustrated on the horizontal axis of graph below

A second distinction that can be made, is one based on the charging speed of the charging point (illustrated below on the vertical axis):

• Level 1 (L1): lowest speed, typically 3–7kW
• Level 2 (L2): higher speed, typically 11–50kW
• Level 3 (L3): highest speed, > 50kW, yet typically > 150 kW. A new generation of “ultra-fast chargers” can even boast a charging power ranging between 350 and 480kW

However, both dimensions are not completely unconnected from one another. Each use case is best connected with the appropriate charging speed. A home- or work-based charger doesn’t need to charge at a high speed as time spent at these locations is typically quite long. Activity-based charging typically requires a higher charging speed to accommodate the medium-duration of these activities. On the opposite end of the spectrum, travel-based charging requires the absolute fastest charging speed possible. One major benefit is that this minimises the impact on the travel duration by reducing the time waiting until the car is charged. Another clear advantage is that faster charging ensures a higher availability at the charging station which maximises the number of users and reduces the time they need to wait in queue or even search for another station.

Players like Ionity, TotalEnergies, DATS 24 and Eneco have understood that an efficient charging network requires all types of use cases, allowing for co-existence. They have defined a strategy that allows them to enter this challenging market by focusing on a segment that naturally fits their existing capabilities. To be able to define such a strategy, a dynamic data-driven approach is recommended to minimise risk and optimise returns.

… That can be best tackled by using a dynamic data-driven approach

The EV market as an optimization problem to match demand and supply

In essence, the optimization of EV charging infrastructure is not that different from other market optimization problems. As hinted in the previous section, everything starts with effectively matching supply and demand under a set of constraints. The key variable that greatly influences both sides of the equation is the chosen geographical location.

• The demand will have specific characteristics at different locations. The use cases (routine, activity and travel) will influence factors such as the best fitted type of charging point and the optimal business model of the location.
• The extent of the existing charging infrastructure will need to be taken into account when estimating the unfulfilled supply. The marginal value of a certain charging station will vary between locations depending on what is already in place.
• Each location will come with its own set of constraints or bottlenecks that influence the feasibility of the operation. One can think of the amount of available space, building regulations and the capacity of the local electricity grid.

An enormous variety of quickly changing and correlated parameters as input for a dynamic and smart data-driven approach can help players optimise their strategy. It allows them to pick the optimal investment locations that fit both internal as well as external constraints. The figure below provides a non-exhaustive overview on data that can be gathered on demand, supply and feasibility of a certain location. It is essential to understand that the quality of the optimization depends both on a smart methodology and the access to the latest available data.

The demand for EV charging infrastructure

Several determinants are influencing the demand for EV charging infrastructure. These can mainly be grouped into three major categories of data: territory data, mobility data and demographic data. It is the combination of all of these vectors that will result in a trustworthy estimation of the demand.

• Territory data links geographical locations to different types of activities. For example, territory data could label a location as a commercial hub, a business park, a parking centre, an office, a transit hub or a residential area. Furthermore, the data takes into account the specific environment of these activities. As such, it indicates differences between residential areas with individual housing and those with collective housing (e.g. apartment blocks). These distinctions are vital since demand for communal slow chargers will be higher close to an apartment block or areas with limited private garage ownership (thus limited possibilities for installing private charging points). Another example of high potential areas for public charging points are transit hubs used for commuting by public transport. At these hubs, cars typically stay parked for more than 8 hours per day.
• Mobility data reflects how people move around. This type of data entails inventories of different traffic zones, evolutions and projections of car travel records and adoption rate of EVs in a specific area. Often mobility data is linked to territory data. For example, utilisation of locations identified as parking sites can be included in mobility data.
• Demographic data provides insight into the population itself, its structure, evolution and income. This is vital input to model, for example, the expected rate at which the population will be switching to EVs in the short and long term and willingness to pay for charging, which in turn has a big impact on the need for chargers.

The supply of EV charging infrastructure

The existing supply of a certain location can be estimated by dynamically capturing existing infrastructure and its main characteristics, such as:

• Number of charging locations, charging poles and actual charging points
• Type and associated power of the charging stations (Speed, AC vs. DC)
• Compatibility of the plug
• Payment method
• Charging Point Operator (CPO)
• Accessibility

Combined, these features could enable a model to estimate what sort of supply already exists within the “influence zone” of a potential charging location. As charging points are popping out of the ground at an ever-increasing speed, an optimal model should constantly scrape for the latest state of affairs in the market and recalculate supply.

The feasibility for investments in EV charging infrastructure

Additionally, certain constraints need to be taken into account. Grid and connection capacity are typical bottlenecks for the installation of charging points. Hence data on the local electricity network is required to estimate the potential for specific locations to host an additional charging station. Not to be underestimated is that (medium-)fast EV chargers can easily require a power of 50kW per plug or more. Load limits of typical end-user connections, such as supermarkets, event centres and parking lots, do not allow for the usage of multiple 50kW chargers at the same time. This could lead to grid congestion and results in locations for these types of chargers being hard to come by, especially in bigger cities.

Combining supply, demand and feasibility to determine the optimal investment locations

Based on the input parameters for supply, demand and feasibility, the potential for charging point locations can be modelled. For different scenario parameters, the potential can be illustrated on a heatmap, considering the scores for different types of chargers.

As additional charging points are placed virtually, the “optimal” locations for other chargers are shifted due to the introduction of the “influence zone” of these new chargers. Taking this into account, an iterative process is required to select each subsequent charging station to be as effective and provide as much value as possible.

Challenges ahead

A wide range of challenges lay ahead for large-scale EV charging infrastructure deployment. First and foremost on the technical side, as the existing low-voltage grid and grid connections used by supermarkets, parking lots, event locations and so on are not dimensioned for typical EV charging rates. With a single rapid charger requiring at least a power input of 50kW, a couple of these chargers would easily saturate the current distribution grid or the customer connections.

Thankfully, this requirement coincides with the necessity of implementing more small-scale distributed renewable energy sources (RES) on the distribution grid. Both additional EV chargers and small-scale distributed RES require significant investment into re-dimensioning the distribution grid by the Distribution System Operators (DSOs) which can thus be seen as a “double win”.

Secondly, the deployment of EV chargers also poses governance issues. The deployment policy should follow the long term vision, taking into account current trends towards both public transport and alternative mobility such as bikes, steps and car sharing at the expense of personal cars. Next, the cost-benefit for a private organisation (and thus the business case) is significantly different from that of public authorities. The latter should analyse the trade-off between citizen service and profitability.

Finally, complex economic decision making is an inherent part of the deployment process. When costs for charging stations range from a few thousand euros for an activity charger to tens of thousands of euros for fast chargers, the main challenge is balancing these costs with the benefits and needs. This is the main challenge Sia Partners wants to address with this model, providing a data-driven approach to identify the most efficient and cost-effective way to roll-out the infrastructure required for the future of electric mobility.

Annex

1 Top 3 is Tesla, Toyota, BYD; data on 2022–06–14
2 JATO (2022), JATO European EV Press Release, JATO
3 IEA (2022), Global EV Data Explorer, IEA, Paris, https://www.iea.org/articles/global-ev-data-explorer
4 https://cdn.eurelectric.org/media/3805/charging-infrastructure-factsheet-2019-h-1BC6E8C5.pdf
5 https://www.reuters.com/business/autos-transportation/eu-lawmakers-support-effective-ban-new-fossil-fuel-cars-2035-2022-06-08/
6 https://www.leaseplan.com/en-ix/blog/tco/tco-ev/
7 https://www.volkswagenag.com/en/news/stories/2020/03/the-big-cost-comparison--e-car-vs--combustion-engine.html#
8 https://economie.fgov.be/sites/default/files/Files/Energy/comparatif-des-prix-2021-4-f-n.pdf
9 https://www.theguardian.com/commentisfree/2022/mar/20/high-petrol-prices-tempt-us-to-buy-electric-cars-but-were-less-put-off-by-surging-electricity-cost#:~:text=Using%20data%20from%20California%2C%20it,the%20rate%20of%20electricity%20prices
10 ACEA. (2022, March). European EV Charging Infrastructure Masterplan. https://www.acea.auto/files/Research-Whitepaper-A-European-EV-Charging-Infrastructure-Masterplan.pdf
11 Ibid.

Lexicon

Charging infrastructure: Charging point vs EVCS vs EVSE vs charge pool

EVSE (electric vehicle supply equipment) and charging point refer to the same thing. It is a piece of equipment at a certain location connected to an energy source where one electric vehicle can connect to charge its battery. EVCS (electric vehicle charging station) is the physical equipment that connects multiple charge points into one interface.

A charge pool in turn links multiple charging stations to a location hosted by the same CPO (charge point operator)

EV related

• BEV = battery electric vehicle
• EV = electric vehicle
• FCEV = fuel cell electric vehicle
• ICEV = internal combustion engine vehicle
• PHEV = plug-in hybrid electric vehicle
• xEV = any electric vehicle

Other

• ACEA = European Automobile Manufacturers’ Association
• AFIR = Alternative Fuels Infrastructure Regulation
• DSO = Distribution System Operators
• RES = Renewable Energy Sources
• TCO = Total Cost of Ownership
• V2G = Vehicle To Grid