Electric Vehicle Charging Control in Smart Grid Applications
Problem Statement: To optimize the time to reach from one destination to another destination by
an Electric Vehicle taking various constraints into considerations.
Electric vehicle (EV) charging in smart grids has become a
popular topic due to the increasing number of EVs on the
road. Smart grid technologies can help mitigate the load
impacts and protect the power grid from instability due to
rapid charging of EVs. This paper studies EVs and their
applications in the smart grid and provides practical
solutions for EV charging strategies in a smart grid. It
outlines three basic electric vehicle charging technologies
and examines their impacts on the grid stability and power
quality. The paper also presents an event-driven model
predictive control (MPC) framework for managing charging
operations of EVs in a smart grid. The framework provides a
way to maintain power balance between generation and
demand while ensuring maximum efficiency of the charging
operations. Additionally, the paper investigates the charging
control problems of EVs in smart grids, where electricity
transactions exist, and suggests methods to optimize the
charging process. Finally, the paper studies the challenges
that need to be addressed in order to successfully manage
the EV charging process in a smart grid.
Various Constraints:
1. First check the state of charge (SoC) of battery so that it can be understood how long the battery
back-up is available.
2. According to back-up, check if the vehicle can reach the destination without charging (Time taking
to reach the destination and KMs to be travel by the vehicle).
3. If battery needs to be charged, then location of various EV charging stations along the route of the
destination to be evaluated.
4. We need to evaluate the location of various charging stations depending upon various routes
available to reach the destination.
Few constraints:
A. We may need to choose the EV charging station which is near to our starting point so that we can
recharge our EV, if it is not properly charge in the beginning of the journey. This condition may need
to choose irrespective of optimum time. The charging station may be located at some other route
other than the optimize (tme) route.
B. If our EV is charged in such a way that it can go few other miles with optimum time, we need to
choose the charging station ON the route of optimum route.
5. We also need to consider few other constraints:
A. The time taken to recharge at a particular charging station should be minimum.
B. The Harmonics generated by EV charging station (As a whole on Grid) should be optimal in such a
way that grid must not be overburdened. In this case, the charging pattern of different 2W and 4W
need to be evaluated. It can be done considering in a decentralized micro grid. Demand side
management and scheduling of (charging of) EV is important in this aspects.
C. The running cost/ charging cost of EV from one destination to another destination can also be an
important aspect. We can propose/ implement to kind of strategies-
(i) Centralized strategy where an EV aggregator get the data from the grid and directly communicate
from the EV user. It creates a charging pattern of EV load from EV user and try to avoid maximum load
demand from grid in peak time.
(ii) Decentralized strategy where the governing body proposes the charging cost according o peak
demand in different weather and load conditions.
This can be done by creating an optimal profit of an aggregator.
D. The charging strategies also depends upon the type of charger I.e. if DC fast charger (250kWh), Fast
AC charger (high rating- 24kWh) or slow AC charger (low rating- 16kWh). The load profile of particular
EV and the availability of that type of charger at a particular charging station is an important aspect.
In the first case, it can be done for Unit commitment method and then it can be done for the system
in large. In the same way, first it can be done for small Micro grid and the system can be applied for
the system in large.
Strategies Implemented in literature:
1. Various soft computing methods used:
(i) Mathematical optimization ex. quadratic programming, dynamic programming, and mixed integer
linear programming, Game theory, .
(ii) Genetic algorithm, particle swarm optimization, and evolutionary algorithms.
(iii) ANN, Markov process
2. Machine Learning models/ Deep learning models
Fig. Calculation process to search for the optimal solutions by Model (k)
Time Lines:
By April, submission of first Review paper (Scopus/ Conference) communicated
By April, preparation of different data sets based on different loads from different resources
(Google/ IIT Delhi Lab- Prof. Panigrahi/ DTU lab- Prof. Vinod Yadav/ OLA charging stations- Lalit
mohan Goel)
By August, first Scopus paper communicated
By November 2023, 2nd paper in conference communicated
By March 2024, second Scopus paper communicated