Page 41 - DP-2-2

Design+ EV charging capacity through queuing model

If the arrival intervals at a location are short, it may be
necessary to increase the capacity of the charging facility
at that location. As shown in Figure 3, charging vehicle
arrival intervals follow a negative exponential distribution.
The negative exponential distribution has a unique
parameter λ. The mathematical expectation of the negative
exponential distribution, 1/λ, implies that 1/λ events occur
per minute, that is, on average, an event occurs every λ
minutes. The time interval for EVs to start charging satisfies
the parameter λ of the negative exponential distribution to
be 9.47, which means that on average every 9.47 min there
is a vehicle charging at this public fast-CS.
The charging habits of different EV users vary greatly,
which is reflected in the charge start state of charge (SOC)
and charge power. Users with EV mileage anxiety tend to
Figure 1. Hourly charging trends at charging stations
start charging when the vehicle has a lot of power left and
choose to charge until the battery is full. Another group
of users tends to spread out the charging time, that is,
each charging time is shorter as long as it can meet their
travel needs. Due to the different usage habits of EV users,
the length of EV charging time also varies. As shown in
Figure 4, the charging time of most EV users is between
40 min and 1 h, and the charging time of some users who
use fragmented time charging is shorter, <30 min. Only
a few users have a charging time longer than 80 min.
The frequency histogram of charging time is fitted with
a normal distribution. It can be seen that the frequency
distribution of EV charging time is basically normal, and
its normal distribution parameter is mean mu = 47.25,
standard deviation std = 17.39. This means that most of
the charging time in this charging time distribution is
concentrated in about 47 min.
Figure 2. Daily charging trends at charging stations
The charging power of an EV reflects the charging
speed: The higher the power, the shorter the charging time.
significantly over different days in a month, with a clear The charging power depends on the SOC of the battery,
pattern of days with more charging orders and days with
fewer charging orders. It can be observed that the days the power of the on-board charger, the output power of the
with fewer charging orders are mostly weekends. We can charging pile, and even the weather can affect the charging
speed. The power of AC charging is mostly 3.3 kW or
conclude that people are more likely to choose public 6.6 kW. DC fast-charging can reach 60 kW or 120 kW,
fast-CS on weekdays. On non-working days, people may while EVs equipped with 800V high-voltage platforms can
choose to charge at private CS more often, or they may reach fast-charging speeds of more than 350 kW. Figure 5
choose not to charge due to reduced commuting needs.
shows a bimodal distribution, with EV charging power
The arrival intervals of charging vehicles can indicate concentrated at two peaks: 18 kW and 40 kW. It can also
the level of user demand for charging facilities and usage be found that although charging piles support charging
patterns. Shorter arrival intervals may indicate higher speeds above 120 kW, EVs with charging power above
demand for charging and more frequent use of charging 100 kW are currently relatively few.
facilities. Observing arrival intervals provides insight into
the use of charging facilities. Shorter intervals may indicate 2.3. Charging service system and operational
that facilities are being used heavily. Understanding the indicators
arrival intervals of charging vehicles can help plan and The queuing theory is an effective approach for the study
optimize the layout and capacity of charging facilities. of the aggregated EV charging behaviors. This approach

Volume 2 Issue 2 (2025) 4 doi: 10.36922/dp.4225