Applied Engineering Sciences

Applied Engineering Sciences – Deng (Ed.)

© 2015 Taylor & Francis Group, London, 978-1-138-02649-0

Optimal scheduling of LTC and switched shunt capacitors in smart grid

concer ning overnight charging of Plug-in Electric Vehicles

Sara Deilami, Amir S. Masoum, Mohammad A.S. Masoum, Ahmed Abu-Siada & Syed Islam

Department of Electrical and Computer Engineering, Curtin University, WA, Australia

ABSTRACT: It is well-known that load variation and nonlinearity have detrimental impacts on the operation

and performance of conventional power systems and future smart grids (SGs), including their voltage profiles,

power quality, losses, and efficiency, particularly during the peak load hours. This paper will illustrate the

performance of optimal scheduling of transformer Load Tap Changers (LTC) and Switched Shunt Capacitors

(SSCs) in Smar t Grids with nonlinear loads and Plug-in Electric Vehicle (PEV) charging activities to improve

voltage profile, reduce grid losses, and control the Total Harmonic Distortion (THD). An established genetic

algorithm (GA) for the dispatch of LTC/SSCs and a recently implemented algorithm, based on Maximum

Sensitivity Selections (MSS) optimization for the coordination of PEVs, are used to perform detailed simulations

and analyses.

Keywords: Smart grid, PEV coordination, charging, optimal dispatch, LTC and switched capacitors.

1 INTRODUCTION

The changing nature of linear loads, harmonic cur-

rent injections of nonlinear loads and the intermit-

tent behavior of renewable distributed generation

(DG) systems have detrimental impacts on the oper-

ation and performance of the conventional aging

power network and the innovative smart g rids (SGs)

[1–3]. Examples of nonlinear loads are variable speed

drives, energy-efficient lights, switching converters,

smart appliances, and plug-in electric vehicles (PEVs).

Load variations change the balance and flow of active

and reactive power that can cause voltage regulation

problems while harmonic injections increase the total

harmonic distortion (THD), reduce efficiency, and

force premature aging of power systems components.

Possible solutions to solve the voltage and power

quality problems are the installation of passive, active,

and hybrid filters, as well as the utilization of custom

power devices which tend to be expensive alternatives

[2]. Recently, the possibility of rescheduling LTCs and

the existing Switched Shunt Capacitors (SSCs) to mit-

igate harmonic distortion in conventional power grids

have been proposed and implemented [1].

One of the main sources of load variation in SG

is random charging of PEVs, particularly during the

evening hours [4–6]. Two main approaches have been

proposed in the literature to prevent these problems:

i) consumers can be motivated to charge their vehicles

during the off-peak hours by offering price incentives

and implementing dynamic energy prices, ii) PEV

charging can be coordinated [5–7]. There are a num-

ber of offline and online PEV charging/discharge

coordination algorithms that are classified to decen-

tralize (distributed) and centralized strategies [5–7].

This paper will use the proposed genetic algorithm

(GA) of [1] to perform optimal scheduling of LTCs and

SSCs in Smart Grids with nonlinear loads and PEV

charging activities to improve voltage profiles, reduce

grid losses, and control THD. There charging PEV

scenarios are considered, simulated and compared:

i) uncoordinated (random), ii) the online maximum

sensitivities selection based coordination algorithm

(OL-MSSCA) of [7] and iii) an inexpensive overnight

MSS based coordination algorithm (ON-MSSCA).

Detailed simulations are presented and compared for

optimal GA scheduling of LTC/SSCs of a 449 node

SG network, populated with PEVs.

2 PROBLEM FORMULATION FOR OPTIMAL

DISPATCH

The LTC/SSCs scheduling problem is minimization of

energy loss over a 24-hour period [1]:

where P

loss

is total power loss at hour t as a function of

Q

t

(status of SSCs) and T

t

(LTC tap position), t =1

hour is the time inter val, while H , m, i and R

i,i+1

are the

71