Hogeschool van Amsterdam

Intelligent Data-driven Optimisation of Charging Infrastructure

The big pictures of public charging infrastructure 1

The basics of charging speed
Posted at: 2 Jun 2016 | IDO-Laad

It has been a few weeks since the AVERE conference was held in Amsterdam [1] and in a few weeks the 29 th Electric Vehicle Symposium will be held in Montreal[2]. The topic of smart charging and sustainable charging infrastructure was one of the hot topics at the AVERE conference and will most likely be one of the hot topics at EVS as well. The sustainability of charging infrastructure in particular is a hot topic in both academics and journalism.

However, the problem with the journalistic discourse on this topic seems to be that it is politically motivated and driven by industry lobbyists. [3]. It therefore emphasizes the negative aspects of charging infrastructure without proper fact checking and the use of real data calculations. An example is the exaggerated and incorrect article in the NRC of April 14 th that stated charging infrastructure will be outdated in five years[4]–[6]. Worse still, is that other news sources accepted this information without question. [7]. As an academic researcher in the field of public charging infrastructure I believe it is time to advance the discussion on charging infrastructure with information based on real data. Therefore, I will publish several blogposts in a series on big pictures on charging infrastructure. These big pictures are to provide insights in the optimization of rollout and use of charging infrastructure for policy makers, journalists and other stakeholders within the EV value chain. I sincerely hope that the discussion on charging infrastructure will be elevated to an objective and professional and expert level in the near future.

In this first blog I will elaborate on how to determine charging speed and why the EV, and not the charging infrastructure, functions as the bottleneck in performance. .

The basics of charging speed

A charging station generally exists of a fused main connection to the electricity grid. There is an operation unit with usually 2 devices (EVSE) that control the power supply to a single EV during a single session , see Figure 1[8]. The charging speed of a charging session is in general determined by two factors, (1) the capacity (max flow of electricity) of the EVSE measured in Ampère and (2) the amount of phases available for charging in the EV (1-phase is comparable with household electricity (230Volt) and 3-phase is 380Volt. For readability concerns I will skip the explanation of the different charging modes.

Figure 1 Charging point elements according to OCPI [8]

The process of car charging can best be explained by the parable of water flowing through pipes. The diameter of each pipe is comparable to the ampere of an EVSEand the amount of pipes with water flowing from the charging station to the EV is comparable with the amount of phases that are present. The vast majority of EV’s currently charging at the public charging infrastructure contain single phase charging technology, whereas all EVSE in public charging points have three phase capabilities. Therefore, by definition single phase EV’s do not use the maximum capacity of the charging station.

Table 1 shows an overview of charging speeds for a given capacity of charging stations and number of phases in an EV. An electric vehicle is able to drive about 5.6 kilometer on 1 kWh. From Table 1 we can conclude that the slowest combination of EVSE and EV results in about 20 km charged per hour while the fastest combination results in roughly 220 km/hour charged. The rapid development of new EV models and their charging abilities reveal a trend towards 3-phase charging.

Table 1 overview of charging speeds

Charging speed in kW with 1-phase charging

Charging speed in kW with 3-phase charging

13 3 9
16 3.7 11
20 4.6 13.8
32 7.4 22
63 NA 43.5

The percentage of EV’s that are able to charge in three phases can be calculated from the data of the current EV population in the Netherlands, provided by RVO[9]. The current models that (to my knowledge) allow three phase charging are: Tesla model S/ Roadster, Renault Zoe, Smart fortwo, BYD, Mercedes B class electric. Adding up the numbers of these models we see that at most ~7% of the current total population has 3-phase charging properties, while 100% of the public charging infrastructure is able to provide so. This means that in 93% of cases the EV is the limiting factor for the charging speed.

Figure 2 Overview 1phase 3phase charging infrastructure and EV

Smart charging – a solution for a problem or a problem by definition?

The public charging infrastructure rolled out in the large cities of the Netherlands consists predominantly of a non-smart charging type. This implies that charging starts at the moment of connection of a session and ends when the battery is full or when the user stops the connection. Idle time of charging sessions is found at the end of the session after 100% state of charging.

Let us define the charging time ratio as the relative amount of slack of a specific charging session in time, calculated by dividing the charging time by the connection time per charging session[10][11]. If the charging time ratio is high this means that there is limited idle time and if the charging time ratio is low this means more idle time. From the definition above it appears that low time ratios have high potential for smart charging, since slack means room for transposing charging moments over the total connection of the session. From a previous research it appeared that the mean charging time ratio in 2014 in Amsterdam was 20-40% [12].

It is also possible to calculate the capacity overhead in terms of kW by dividing uptake of kW over time by the EV by the offered kW of the charging station. This number is important to calculate the growth capacity of the current technology. I will revisit this in a later blogpost in which I calculate the overcapacity of the current charging infrastructure in the G4 cities.

From the perspective of a Charging Point Operator (CPO) a low time ratio means that the effective use of a charging station is low. In other words a connector is occupied and not charging while the service is being delivered no energy is transferred and thus no money is earned by the CPO[13][14]. Having a high smart charging potential on a population of charging points is having suboptimal customer buying behavior from business perspective. An occupied charging station by a non-charging user could lead to missed sales for users willing to charge. Of course it is difficult to measure loss of revenue from the charging point transaction data, but simulation models might reveal this. Our research group is currently working on several agent-based models for charging behavior in metropolitan areas. The results are expected in the first quarter of 2017.

Recently several apps have been launched that enable conversing between users at the same charging point in order to induce social behavior like making space for other EV users as soon as an EV is fully charged. Note that this might result in an increased effective use of the Charging point, but also a decrease of smart charging potential!

From a different perspective smart charging potential parameters such as charging time ratio and capacity overhead give insight in the life time of the current rolled out charging infrastructure technology. Both the charging time ratios as well as the capacity overhead are based on (1) the amount of kWh charged related to (2) the amount of connection time and (3) charging capacity of the EV. The first aspect is related to battery technology of the EV, the second is related to the charging behaviour of EV users [11] and the third is related to the current rolled out charging infrastructure technology.

For example, imagine that all sessions on the charging infrastructure would be like a typical charging session of a Plugin Hybrid Electric Vehicle (PHEV). Let us say that this has a size of 10 kWh, a single phase charging capacity and a connection time of 8 hours. This would result in a charging time ratio of ~40%. Simple calculations show that if the battery of the EV for this specific charging session had a capacity of 24kWh, no slack would be available at all. In other words, the current technology is able to supply PHEV’s up to 24kWh without any change of user charging behaviour properties[11]. So even if batteries become 2,5x or even 15x larger, nothing would change unless the car increases its charging speed.


In this blogpost I briefly explain factors determining charging speed. From there I conclude that the bottleneck in charging speed is the EV rather than the charging point with high capacity.

From the charging time ratio and slack on charging sessions I conclude that the current rolled out infrastructure is capable of coping with developing technology. By illustration I show why charging infrastructure may be able to keep up with battery size increase of 250% without any change in charging infrastructure or user charging behaviour properties. The battery size can even increase by 1500% if the charging speed of the EV increases. In short, that charge station will persevere.

Therefore, I recommend that the discussion on sustainability of charging infrastructure should start at the EV itself and not necessarily at the sustainability of charging points.

Our research team will monitor the developments of battery size in relation to charging infrastructure performance. In the next blogpost I will display why current charging is able to supply for at least 100 million more emission free kilometers without replacing any charging points.


As a scientific researcher in this subject, I must note that this is a blogpost and not a scientific paper. It is my personal vision on charging infrastructure combined with real data from charging infrastructure. This implies that the use for inspiration and education is allowed but it does not contain a scientifically backed and nuanced truth. The graphs shown support the statements that I want to make on this subject. A thousand more graphs could have been shown, but I specifically chose these to support my personal and scientific thoughts.

[1] “AVERE E-mobility Conference.” [Online]. Available: http://www.aec.amsterdam/. [Accessed: 18-May-2016].

[2] “EVS29 - 2016 Electric Vehicle Symposium & Exhibition 29.” [Online]. Available: http://www.evs29.org/. [Accessed: 18-May-2016].

[3] B. van Zoelen, “‘De stad staart zich blind op laadpaal voor elektrische auto’ - Binnenland - PAROOL,” 04-Aug-2015.

[4] J. Verlaan, “‘Laadpalen elektrische auto’s binnen vijf jaar achterhaald’ - NRC,” 14-Apr-2016.

[5] “Iedereen elektrisch rijden? Eerst snellere laadpalen - NRC.” [Online]. Available: http://www.nrc.nl/next/2016/04/14/iedereen-elektrisch-rijden-eerst-snellere-laadpa-1611573. [Accessed: 26-May-2016].

[6] “We hebben alle soorten laadpalen nodig! En veel! | Steinbuch on WordPress.com.” [Online]. Available: https://steinbuch.wordpress.com/2016/04/14/we-hebben-alle-soorten-laadpalen-nodig-en-veel/. [Accessed: 26-May-2016].

[7] “‘Publieke laadpalen over vijf jaar al verouderd.’” [Online]. Available: http://www.nu.nl/amsterdam/4246644/publieke-laadpalen-vijf-jaar-al-verouderd.html. [Accessed: 26-May-2016].

[8] NKL, “Open Charge Point Interface OCPI, Projecten - Nationaal Kennisplatform Laadinfrastructuur,” 2016. [Online]. Available: http://www.nklnederland.nl/projecten/onze-lopende-projecten/open-charge-point-interface-ocpi/. [Accessed: 30-Apr-2016].

[9] “Cijfers elektrisch vervoer | RVO.nl.” [Online]. Available: http://www.rvo.nl/onderwerpen/duurzaam-ondernemen/energie-en-milieu-innovaties/elektrisch-rijden/stand-van-zaken/cijfers. [Accessed: 28-Apr-2016].

[10] M. V. Rivera, R. Van Den Hoed, and J. Helmus, “Charging in the city of Amsterdam : Data Monitoring of charge point performance,” no. December, pp. 1–9, 2014.

[11] J. Helmus and R. van den Hoed, “Unraveling User Type Characteristics : Towards a Taxonomy for Charging Infrastructure,” in EVS28 International Electric Vehicle Symposium and Exhibition, 2015, pp. 1–16.

[12] R. Van Den Hoed, J. R. Helmus, R. De Vries, and D. Bardok, “Data analysis on the public charging infrastructure in the city of Amsterdam,” pp. 1–10, 2014.

[13] E. Paffumi, M. De Gennaro, G. Martini, and H. Scholz, “Assessment of the potential of electric vehicles and charging strategies to meet urban mobility requirements,” Transp. A Transp. Sci., vol. 11, no. 1, pp. 22–60, 2015.

[14] C. Madina, H. Barlag, G. Coppola, I. Gomez, and R. Rodriguez, “Economic assessment of strategies to deploy publicly accessible charging infrastructure,” in EVS28 International Electric Vehicle Symposium and Exhibition, 2015, pp. 1–11.