SMART CHARGING OF PLUG-IN ELECTRIC VEHICLES IN RENEWABLE INTEGRATED DISTRIBUTION SYSTEMS
Loading...
Date
2026
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Department of Electrical Engineering, NITS
Abstract
The rapid adoption of plug-in electric vehicles (PEVs) coupled with the increasing
penetration of renewable energy sources (RESs) in distribution networks offer significant
opportunities for reducing greenhouse gas (GHG) emissions and fossil fuel dependence.
While these trends support global decarbonization goals, they also introduce new
complexities in grid management due to the intermittent nature of RESs and the stochastic,
high-power demand from uncoordinated PEV charging. This PhD research presents an
integrated approach combining multi-objective optimization and machine learning to
address these emerging challenges and enable intelligent, environmentally conscious, and
technically robust smart charging strategies in modern distribution networks.
The thesis begins by exploring smart charging of PEVs in distribution systems,
highlighting the implications of uncoordinated charging on PEV charging cost, load
variance, and power losses. A methodology is developed to model the stochastic charging
demand of PEVs using probability distribution functions (PDFs) for charging start time,
initial state of charge (SOC), and daily driving distance. A centralized multi-objective smart
charging strategy is formulated under time-of-use (TOU) pricing. The optimization problem
addressed using Ant Lion Optimization (ALO). To ensure robustness, the results are
validated against Grey Wolf Optimization (GWO) and Non-dominated Sorting Genetic
Algorithm II ((NSGA-II). Simulation studies are conducted on IEEE 33 and 69 bus
distribution networks to demonstrate the proposed strategy.
A central contribution lies in the development of a smart grid framework that
coordinates RESs and PEVs to simultaneously minimise both fuel cost and emission. RESs
particularly wind and solar can significantly reduce emissions from electricity generation
while PEVs with vehicle-grid (V2G) capability can mitigate emissions from the
transportation sector. The study highlights the proposed smart grid model enabling
maximum utilization of renewable resources for cost and emission reduction. Particle Swarm
Optimization (PSO) and Ant Lion Optimization (ALO) are employed to solve the
optimization problem and generate pareto fronts. To validate the proposed methodology,
three case studies are analysed on a hybrid system comprising a ten-unit generation system
and a 33-bus distribution network
To further improve system intelligence and responsiveness, the thesis integrates
machine learning (ML) models for wind energy forecasting and SOC estimation. An
adaptive coulomb counting method combined with SOC-OCV look up table incorporating
temperature effects is presented for more accurate SOC estimation. Additionally, given the
intermittent nature of wind, an empirical mode decomposition (EMD) machine learning
based is proposed for wind energy forecasting. The developed model reduces the prediction
errors and improves reliability in renewable based smart grids.
The work also addresses congestion and emission management in power systems. A
congestion management framework is developed for distribution system with high PEV
penetration, leveraging machine learning methods (Adaboost and XGBoost) to forecast SOC
for optimal selection of participating vehicles. The coordination problem is solved using
particle swarm optimization (PSO) and tested on a 38-bus radial distribution system
integrated with a solar powered charging and parking lot (SPCPL). The proposed work
demonstrates effective congestion management and CO2 emission reduction.
The proposed approaches collectively demonstrate that smart charging strategies
coupled with RESs integration, optimization techniques, and ML methods can significantly
improve the technical, economic, and environmental performance of distribution systems.
The findings of this research contribute toward realizing the vision of sustainable, resilient,
and intelligent power systems enabled by PEVs and RESs.