Date of Award

9-7-2024

Document Type

Thesis

School

School of Electrical & Electroncis Engineering

Programme

Ph.D.-Doctoral of Philosophy

First Advisor

Dr.D.Karthikaikannan

Keywords

Charging Station, Electric Vehicle, Facility Location Problem, Particle Swarm Optimization, Direct Search, Renewable Energy Sources

Abstract

Electric vehicles (EVs) are recognized as a potentially effective solution to the environmental crisis and fuel deficit that modern metropolises are currently experiencing. Meanwhile, the expansion of EVs and their haphazard charging would negatively affect the road transportation network (RTN) as well as the electric distribution network (EDN). The proper infrastructure for the electric vehicle charging station (EVCS) will enhance the effective utilization of energy of both the networks.

This research work proposes a two-phase planning strategy for deploying Electric Vehicle Charging Stations (EVCS), considering both static and dynamic factors. The strategy uses a hybrid PSO-DS algorithm which is a combination of particle swarm optimisation algorithm and direct search method to optimize multiple objectives. The effectiveness of the proposed algorithm is validated using benchmark functions facilitates in solving of proposed planning strategy. This approach can help decision-makers in efficiently deploying EVCS infrastructure.

In the first phase, a static framework model is developed considering the nodes and distance of the traffic network and distribution network. The EV charging station placement problem concerns the total coverage in the traffic network, the system losses, and node voltage deviations in the electric distribution system. The distribution systems are typically equipped with shunt capacitors for reactive power compensation to address the loss reduction and voltage profile improvement.

In this work, a mathematical model comprising three objective functions, maximisation of coverage and minimisation of loss and node voltage deviations subjected to constraints, is proposed for the simultaneous placement of EV charging stations and shunt capacitors. The control variables for optimisation are the rating and location of charging stations and shunt capacitors. The placement of EVCS on the distribution network would increase the power loss and total voltage deviation. These issues are addressed by the placement of the shunt capacitor in the optimal location. To verify the model, simulations are carried out on an IEEE33-bus distribution system and a 25-node traffic network system to determine the different planning strategies for the placement of charging stations.

In the second phase, a dynamic framework model is proposed which considers the dynamic behavior of traffic network and distribution network. A combined model of road transport and electric distribution network (CoRTED) is proposed in this work to place the charging station on the urban road traffic network with the minimal travel expense of the EV user’s and minimal power loss, voltage deviation in the distribution network. The suggested model calculates the dynamic user equilibrium (DUE) on the traffic network based on the placement of charging stations and simultaneously takes the grid network’s economic operation into account.

The appropriate injection of active and reactive power by PV panels and shunt capacitors resolves the grid difficulties caused by the EVCS. The proposed model is resolved by a bi-level optimization algorithm, where the user equilibrium of the traffic network is resolved by the Combined convex optimisation (CCO) method and the economic operation of the grid is formulated as alternating current optimal power flow, which is solved by a primal-dual interior point method. Finally, an integrated system comprising modified Nguyen-Dupius RTN and IEEE 33 bus EDN is used to validate the effectiveness of the proposed CoRTED model.

Finally, a hybrid short-term solar energy prediction model is suggested to enhance grid safety and reliability, particularly with the rising demand for energy due to the increasing number of electric vehicles (EVs). This necessitates the utilization of renewable energy sources (RES), which emit zero carbon. The intermittency nature of RES makes the grid unstable. Due to the stochastic nature of the sources, it is essential to predict solar power for effective integration. The evolution of deep learning makes the prediction simple with complex and non-linear temporal data.

In this work, a hybrid Vector Auto Regressive (VAR) model, Convolutional neural network (CNN) and Longshort term memory (LSTM) are proposed to predict solar power using real-time data from the Adirampattinam station, Tamilnadu, India for the period of 2014 to 2020. The VAR-CNN-LSTM model employs the VAR model and deep learning network to capture linear and non-linear data features. In this proposed hybrid method, the VAR model capture linear feature of the data, the CNN layer will extract the hierarchical structure from the various weather parameters, and LSTM will extract the long-term temporal characteristics of the data. The results obtained in this work for the proposed model is then compared with distinct VAR, CNN, LSTM, and hybrid CNN-LSTM models to verify the performance of the model.

Download

Share

COinS