Information Technology Reference
In-Depth Information
feedback is effectively used to solve the distributed EV charging problem in the
smart grid. Most recently, Studli et al. [5] propose charging strategies based on
Additive Increase and Multiplicative Decrease (AIMD) algorithms that can be
implemented in a decentralized fashion to maximize power utilization by EVs
while achieving a fair allocation of power across customers. In this paper we in-
vestigate the performance of AIMD based charging strategies for EVs connected
to a low-voltage distribution network with regard to mitigating the impact of
EVs on the grid from the perspective of transformer loading levels and voltage
profiles. In particular, we propose an extension of the AIMD charging strategy
that seeks to reduce peak power by responding to time-of-day pricing. The strate-
gies are evaluated using simulations conducted on a custom OpenDSS-Matlab
platform for a typical low voltage residential feeder network.
2 Methodology
2.1 Assumptions
We make several assumptions in investigating the impact of EV charging on a
LV distribution network, most of which are consistent with previous studies in
[2], [3], [5] and [6]. The assumptions are as follows:(i) All EV batteries have a
capacity of 20 kWh; (ii) Each EV charger is connected to a standard household
outlet at 230V; (iii) The maximum power output from the EV home charger
cannot exceed 3.7kW; (iv) Each EV has the ability to adapt its charge rate in
real-time and continuously; (v) The initial state of charge (SOC) energy request
for each EV ranges from 5kWh to 15kWh uniformly; and (vi) Power flow is
unidirectional from grid to vehicle.
In order to implement smart charging strategies in practice, several specific
assumptions are also required in relation to the communication and sensing
capabilities of the smart grid infrastructure, namely:
1. Each EV charging point is equipped with a communication device and is
able to receive broadcast signals from a local server.
2. Each EV charging point is able to detect its line voltage in real-time.
3. Each EV charging point is able to send the voltage signal back to the local
server and regulate its own charge rate by commands from the server.
4. A centralized server is installed in the substation and is able to sense the
available resource and broadcast signals to the local servers.
2.2 Distributed AIMD Algorithm
The basic idea of AIMD was originally applied in the context of decentralized
congestion control in communication networks [7]. In [5] Studli et al. proposed
applying AIMD to EV charging problems and investigated a number of prac-
tical scenarios. In this paper, the framework we assume is consistent with the
domestic charging scenario demonstrated in [5]. The basic decentralized AIMD
algorithm for EV charging is summarized as follows:
Search WWH ::
Custom Search