See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/337510444

Demand Side Management- Literature Review and Performance Comparison

Conference Paper · January 2019

DOI: 10.1109/CICN.2019.8902364

CITATIONS READS
10 807

2 authors:

Hayder Alwan Sherif Abdelwahed

Virginia Commonwealth University Virginia Commonwealth University
16 PUBLICATIONS 96 CITATIONS 260 PUBLICATIONS 2,828 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Big Data Benchmarking Suite for Cyber-Security Analytics View project

Three recently Submitted Research Proposals (Currently Under Review) View project

All content following this page was uploaded by Hayder Alwan on 26 September 2021.

The user has requested enhancement of the downloaded file.

 2019 11th International Conference on Computational Intelligence and Communication Networks

Demand Side Management- Literature Review and

Performance Comparison
Hayder O. Alwan, Sherif Abdelwahed
Department of electric and computer Engineering
Virginia Commonwealth University
Richmond, VA, USA
Email: {alwanho, sabdelwahed} @ vcu.edu

Abstract – Several demand side management (DSM) techniques scheme as proposed in [3]. These references present a DSM
and algorithms used to show that by adopting DSM and Time-of- strategy based on load shifting techniques in three services
use (TOU) price tariffs, electricity cost significantly decreases, which are residential, commercial, and industrial. The authors
and optimum load scheduling is achieved. In the first part, this take into account the priority of the operation time for each
paper gives a comprehensive literature review on DSM’s that are
appliance by a proposed time delay function to minimize the
related to load scheduling, Direct Load Control (DLC), and
Demand Response (DR). In the second part, two algorithms are customer discomfort. It is worthy to mention here that the
chosen to compare performance in terms of load consumption authors used the same TOU price for commercial, residential,
profile, Photovoltaic (PV) utilization efficiency, and power loss. and industrial; while in our work the electrical price signal used
These algorithms are implemented to find the optimal electric in the residential load differs from the electrical price signal
load consumption profile with presence of local PV generation. used in the commercial load. In [5], a multi stage optimization
Furthermore, this paper aims to present two approaches for DSM for a typical home energy with assumed rooftop solar PV is
for a residential home. These approaches can be used in response proposed. The proposed algorithm assumed that the surplus
to changes in the price of electricity overtime and in the presence power is injected to the grid with a reward. The second thing
of PV generation to minimize the consumption cost and change
the case study of this paper considered is the appliances
the consumption pattern by shifting part of the load to off-peak
hours. In addition, a case study of a single household with a single preference and no penalty assumed for the shifted appliances.
line is considered under the assumptions of its participation in a In [6], a residential load scheduling approach is proposed to
DSM program. Results show that the proposed scheduling manage the operation time and to achieve optimal daily usage
algorithms can effectively reflect and affect user’s energy of DGs. That locally available, genetic algorithm is designed
consumption behavior and achieve optimal time distribution of and implemented with a two level of optimization base on the
electricity usage. Numerical results show the impact of applying DG-based scheduling and RTP-based scheduling. In our work
DSM algorithms on total power losses of the feeder. The proposed the objective function minimizes the cost by finding optimal
algorithms are implemented based on the Clonal Selection load scheduling and makes the best use of PV generation. Our
Algorithm (CSA).
objective function includes penalty cost and power loss.
Keywords: Demand Side Management, Distributed Generation, References [7] – [15] illustrate that DSM can be performed
Rooftop Photovoltaic, Energy Utilization, Optimization algorithm. by using Direct Load Control (DLC). Utility companies
I . INTRODUCTION manage the customer consumption using this method, with the
customer’s appliance controlled by networked technology
The major goal of Demand Side Management (DSM) [1] –- embedded system. Different type of algorithms are used for
[6] is to efficiently manage loads in such a way that will DLC. For instance, in [7] and [8], the author proposes an
improve efficiency of the grid, reduce costly generation, and optimization algorithm that can apply with either ON or off
decrease excessive load pressure, and increase power system time to manage and control the load in residential and
stability and sustainability. This is done by maximizing system industrial locales. DLC can also be used to reduce the peak
capacity without changing the entire physical infrastructure of load for large-scale residential demand response implantation
the power system. For instance, researchers in [1] and [2] [9]. In [10] an optimization technique based on dynamic
focused on household DSM, neglecting photovoltaic (PV) programming is used to achieve the optimal DLC strategy. A
utilization efficiency, maximum demand limit, and customer’s genetic algorithm is used to optimize the scheduling of direct
welfare. While load scheduling schemes for scheduling load control (DLC) strategies in [11], and integer linear
residential loads consumption proposed in [3] – [5]. Ref.[4] A programming is used for the same problem in [12]. It is notable
combination of DSM and Time-Of-Use (TOU) tariffs to mention that some DLC optimization algorithm is used for
significantly decreases the cost of energy with high utilization certain appliances in a household, for example refrigerator and
of PV generation by using a heuristic–based load scheduling

978-1-5386-8439-9/19/$31.00 ©2019 IEEE 93

DOI: 10.1109/CICN.2019.18

Authorized licensed use limited to: Virginia Commonwealth University. Downloaded on May 27,2021 at 17:44:18 UTC from IEEE Xplore. Restrictions apply.
air conditioning in [13] – [15]. However, in [5] the developed management of local energy resources. A high penetration of
algorithm covers a variety of appliances in different types of distributed generation can lead to problems in low voltage
loads; commercially and residentially. distribution networks. A a good review of the literature is
presented in [26] that is focused on PV penetration limit due to
In [16] – [24], a Demand Response (DR) program, where voltage violation in low-level networks. In [27], the author
users are motivated to be an active participant in managing proposed an optimization technique base real-time demand
their loads by reducing their consumption at peak load hours response using renewable energy. Optimal load scheduling of
can be the alternative to DLC. In this regard, the most common household’s appliances and management of local resource
DR programs include critical-peak pricing (CPP), time-of-use approach assuming a rooftop solar PV. As mentioned in [28]
pricing (TOUP), and real-time pricing (RTP). For example, in the proposed techniques do not consider the importance of
RTP tariffs, the price of electricity changes at different hours balance between the objectives, energy cost minimization,
of the day; while the TOUP electricity prices are previously peak load minimization and user comfort maximization. Also
determined and the customer shifts operation time of the load the voltage monitoring part was absent from the constraints.
accordingly. In real time pricing, it is difficult and confusing Additionally, the voltage limitation has to be taken into
for the customer to respond to the variation in the price every account. This shows that a good DSM algorithm should
hour. Also, peak load may occur in the low-price time period minimize the electricity cost and maximize the convenience
[18]. Peak load may cause instability in the system; therefore, for the customer.
it would be important to deploy TOUP price with max demand
limit as block rates. In fact, this is included in the second The authors in [29] – [38] focused on load scheduling
optimization algorithm in this paper. Several methods and methods based on a day-ahead optimization process to reduce
studies have been implemented over the past two decades in the customers’ electricity bill. This is achieved by producing
order to minimize the electricity cost-based day head price. In or storing energy to lessen their energy purchased from the
[19], a DLC algorithm using linear programming is developed; grid. For instance, in [35], the author proposed a load
the results showing the electricity cost charged to the management technique for air conditioners to enable
customers reduced after participating in this DLC program. customers with small air conditioner appliances to participate
Smart pricing adopted in [20] to achieve lower electricity cost in various load management programs. This motivates the
proposes a power scheduling for demand response in smart customer by receiving incentives from the utilities to lower
grid system; with results showing the proposed scheme leads their monthly electricity bills. The work in [36] utilizes smart
to reduction in the peak demand. In [21], a proposed direct load load management for coordinating the schedule of electric
control based on a linear programming algorithm to manage vehicles, taking into account the grid performance such as
large number of customers controlling appliances in order to voltage limit and the total system power loss. In [37] and [38],
achieve maximum load reduction. Another linear the authors proposed model, which aims to find the optimal
programming model was developed in [22] to optimize system starting time for charging the battery to minimize the cost
peak period load reduction in commercial and residential load. delivered by the transformer to minimize distribution charging
Heuristic-based evolutionary algorithm (EA) used for energy problems in electric vehicles. The DSM techniques used in
management in [23], is applied for load scheduling for DLC [39] – [42] reduce the peak load of the grid proposed. The idea
based on load shifting, this algorithm easily adapts in this technique is that the load demand information of the
heuristically in the problem. In this paper, a Clonal Selection user side using the smart power system is provided to the
Algorithm (CSA) method based on the biological immune energy provider and the energy price is updated accordingly.
system and the natural defense mechanism of human body is The outcomes of these references illustrate that the daily power
used. In CSA, the limitation is that the quality of the results consumption pattern can be smoother and more controlled
depends highly on initial population and probabilities of through different DSM schemes. For example, the Demand
mutations. Population size selection in the simulation process response scheduling under load uncertainty based on real-time
makes the algorithm converge towards high quality solutions pricing in a residential grid is illustrated in [40, 41, and 42]. In
within a few generations. In addition, it is easy to implement. these references authors used game theory to obtain the
The main difference between Clonal Selection algorithm optimal load consumption pattern based on provided price. The
(CSA) and Particle swarm optimization (PSO) is that PSO authors model the interaction between the energy provider and
doesn’t have genetic operators such as mutation, but both CSA the customer using this theory. In [43-49] researchers covered
and PSO share one aspect which is memory. This means they optimization of the performance of the micro grid to make the
save the last iterations and update during the optimization best use of renewable generation resources and to reduce
process. According to [24], PSO is similar to the genetic dependency on grid energy providers. For example, in [43] and
algorithm (GA) as they both are population-based searches; [45], an energy management system model proposed including
with PSO having memory as important to its algorithm. PV generators with storage units. While in [44], the author
Moreover, the optimal solutions in references [25] – [28] to the introduces a residential PV generation energy storage system
problem of scheduling household DSM in the presence of PV considering the pattern of daily operation load of the
generation took more precedence. For instance, in [25], an homeowner. The benefit of a storage system in local residential
optimal scheduling and controlling approach has been studied PV installation presents in [46] and [47], wherein the authors
which performs scheduling of household appliances and study the impact of PV on the quality of the low voltage network.

94
Authorized licensed use limited to: Virginia Commonwealth University. Downloaded on May 27,2021 at 17:44:18 UTC from IEEE Xplore. Restrictions apply.
 The works [48] – [56] proposed scheduling schemes in a smart 2-The DSM techniques and algorithms used in [12], [14], and
grid environment to improve the quality of the power grid and to [27] propose a consumption scheduling mechanism for home area
enable the customer to reshape his production and consumption load management in smart grid using an integer linear
pattern. Moreover, in t hese references an optimal residential load programming (ILP) techniques. Most of them are system specific
scheduling models uses mixed integer linear programming. The [12], [13], [10], and some of which are not applicable to practical
proposed model also presents the integration of renewable systems that have a wide variety of independent appliances.
generation with a battery storage system. The model aims to help Moreover, the techniques were developed in [10], [19] using a
the customer to reschedule the appliances to get optimal benefit linear program. These algorithms and techniques cannot handle a
and to minimize their electricity bill. The author in [57] – [71] large number of controllable devices from several types’ loads
proposed an optimal scheduling technique for residential which have several consumption patterns.
appliances in smart homes with local PV generation. Results 3-The TOUP methods proposed in [38], [40] and [46] were
indicate that the proposed strategy has the capability of applied to achieve low electricity payment. However, the purpose
maximizing the savings in electricity cost. The work in [72] – [76] of the DSM is not only to achieve optimal electricity cost, but also
studies scheduling of different types of appliances by adopting a to prevent higher power demand peaks even if the electricity price
dynamic programming-based game theory approach. It is assumed is low. From this point of view, TOUP applied in these references
that customers with extra power generation can inject their unused still has limits causing the demand to be shifted to hours with low
power and receive revenue to reduce their electricity bill. electricity price and would lead to a higher peak electricity
demand and peak-to-average ratio. During the low-price time, a
The previous work in [5], develops an optimization model for a combination of TOUP with a fixed threshold which represents
single household DSM model. This model contains a single max demand is necessary.
distribution feeder line supplying a small community of thirteen 4-Most DLC optimization algorithms are used for certain
houses based on a day-ahead household DSM system with local appliances in a household, for example refrigerator, air
PV generation. The proposed algorithm searches for the optimal conditioning [34] – [37] which are categorized as residential load.
solution to the problem of scheduling household DSM in the However, in our work the developed algorithm is expanded to
presence of PV generation under a set of technical constraints such cover a variety of appliances in different types of loads,
as dynamic electricity pricing and voltage deviation. The commercial or residential. Even if they have different
proposed solution is implemented based on the Clonal Selection characteristic in terms of load profile, electricity price, appliances,
Algorithm (CSA). and their customer willingness for DSM participation.
This paper gives a brief survey of DSM techniques and 5-In [17], appropriate TOUP profiles were used for both
algorithms and then compares our previous work [the second residential and commercial loads, where the profiles were
algorithm] with another optimization problem. The two different for each load [49]. Our algorithm takes into account the
algorithms mainly aim to reduce the cost by optimizing load
voltage fluctuation evaluation and the power loss of the entire
scheduling. One of the main advantages of our proposed algorithm
system while optimizing the electricity cost. The goal is to
is the ability to monitor the voltage fluctuation and power loss of
the entire system while optimizing the electricity cost and reschedule the energy consumption, taking into account the day-
avoiding the peak load occurrence by the shift of the operation ahead dynamic electricity price and the real production of the
time. This is done by inserting a block rate on the load represented photovoltaic system.
as threshold Maximum Demand (MD). In addition, the proposed 6-According to [24], the limitation of PSO is that quality of the
algorithm can handle large number of controllable appliances. results depends highly on the initial population and the
The rest of this paper is structured as follows. Section II gives probabilities of mutation. The population size in my work makes
overview of the system. Section III gives mathematical the algorithm converge towards high quality solutions within a
description of the DSM problem. Section IV provides and few generations, while also easy to implement. The main
analyzes some numerical simulation results. Finally, section V difference between CSA and PSO is that PSO doesn’t have
concludes the paper and discusses future works. genetic operators such as mutation, but both CSA and PSO share
 Limitations of the Current Micro grid Literature. one aspect that is they both have memory. This means they save
Based on the conducted literature review, we review in the rest of the last iterations and updates during the optimization process.
this section related current work and discusses their limitations. PSO is similar to the genetic algorithm (GA) as they are both
population-based searches, where in PSO the memory is
1-The DSM programs that have been used (e.g. [33], [34], important to the algorithm.
[41]), proposed an efficient energy management system to
schedule the electricity use of appliances to achieve maximum II. SYSTEM MODEL
benefits for customers considering all three types of appliances;
This section describes the system model with input
base, interruptible, and curtailable appliances. The proposed
algorithms do not consider the importance of balance between the parameters. This model consists of a single line feeder
objective’s energy cost minimization and peak load minimization. supplying a small community of thirteen households based on
While in our work, a DSM model is proposed and mathematical day-ahead pricing. The proposed algorithms used for
models for the grid, renewable energy resource represented by comparison search for the optimal solution for load scheduling
rooftop PV generation are presented as well as for different type with presence of local PV generation. The two models used for
of electrical appliances in different types of loads commercial and comparison will have the same input parameters. In other
residential. This model can effectively minimize the energy words, the same PV generation profile, the same power
consumption cost for day-ahead time. consumption profile, and the same electricity price signal.
These parameters are used to show the impact of the proposed

95
Authorized licensed use limited to: Virginia Commonwealth University. Downloaded on May 27,2021 at 17:44:18 UTC from IEEE Xplore. Restrictions apply.
DSM on a single household with more diversified combination of
applicances. Also, appliances can be classified into three types:
interruptible, base line, and uninterruptible appliances. More
details and definition of each type of household are found in our
previous paper [4]. In order to test the DSM models to manage set
of appliances a TOUP environment is chosen to test our model as
depicted in Table 1. 𝑢𝑎,𝑚 Represents a binary status of appliance
a; 0 = off, 1 = on at Household m, with following format:
[𝑢𝑎,𝑚 (𝑡)]𝐴×𝑇 = [𝑢11 , 𝑢12 …, 𝑢1𝑇 ; 𝑢12 , 𝑢22 , … 𝑢2𝑇 ;…...; 𝑢𝐴1 , 𝑢𝐴2 ,...,𝑢𝐴𝑇 ] (1)
Where T is total number of time slots, T=48, and t is the index of
the time slots. Each appliance is modeled using four
parameters 𝑠𝑎 , 𝑓𝑎 , 𝑟𝑎 and 𝐷𝑎 , where [𝑠𝑎𝑚, 𝑓𝑎,𝑚 ] defines the
allowable operating time during which the appliance a in the
household m may be switched on, 𝑟𝑎,𝑚 and 𝐷𝑎,𝑚 denote the power
rating and the total number of operating time slots as requested in Fig.4. Original Load Profile (kW)
the household m respectively. Table.1 No. of appliances of the household with the agreed MD
Household Index Interruptible Uninterruptible MD(kW)
Appliance No Appliance No.
1 21 7 12.4

III. Mathematical Formulation

The work in [4] shows the household DSM model is aimed

to minimize the electricity cost by scheduling the on/off status
of domestic appliances over the operational periods,
considering the dynamic electricity prices, locally available PV
generation, and the penalty prices of appliance operation time-
shifting which are included to manage the customer
inconvenience caused by the proposed DSM algorithms.
The optimization model developed is based on this system.
The performance of the DSM algorithms will be tested and
Fig.1. the model of the domestic energy system with solar penal. compared based on the load scheduled and power loss. First,
the model we talk about now aims to make the best of the PV
power generation. This reduces the demand on the power grid
and reduces the electricity purchased from the grid. The
implemented algorithm in this work tries to arrange the
operation of the appliance through the day and to find the
optimal electric consumptions pattern and hence, reduce the
demand during the peak load.

A. The first optimal energy management approach

used for the performance comparison [29]
formulated as:
Fig.2. Time-of-Use price(¢/kWh) 48 𝑚+𝑛
ℎ ℎ
𝑚𝑖𝑛 ∑ |𝑃𝐷𝐺 − 𝑃𝑚𝑢𝑠𝑡 − ∑ 𝑥𝑎ℎ 𝑃𝑎 | (1)
ℎ 𝑎=1

In the case, the power used by the customer is smaller than

the PV generated power, and therefore, no electricity purchase.
However, when the demand exceeds the PV generated power
the customer needs to reschedule the electricity load to a lower
electricity price to reduce the electricity bill.

𝑥𝑎ℎ = {0,1}, h∈ ∀ ℎ𝑎
Fig.3. PV generation (kW) 𝛽𝑎 -𝛼𝑎 ≥ 𝑑𝑎 (2)

96
Authorized licensed use limited to: Virginia Commonwealth University. Downloaded on May 27,2021 at 17:44:18 UTC from IEEE Xplore. Restrictions apply.
ℎ𝑎 , is the time that appliance a may operate. In this work, the domestic appliances are categorized into
the following typing based on their operation characteristics
𝑥𝑎ℎ , is appliance a on/off condition.
{see Appendix -A}.
𝑑𝑎 , length of operation time
(1) Baseline load: it is the must-run service that needs to
𝑃𝑎 , Rated power of appliance a be served immediately when it is requested by the
[𝛼𝑎 , 𝛽𝑎 ] Allowable operation time range. residents, e.g. lighting, fridge, computer, television.
The energy supply to such must-run services is
𝑡𝑎𝑠𝑡𝑎𝑟𝑡 , opeation statim slot
considered not schedulable and needs to be included
𝑡𝑎𝑒𝑛𝑑 , operation end time into the load demand as the baseline load.
(2) Uninterruptible flexible load: it refers to the domestic
m, No. uninterruptible appliance
appliances (e.g. rice cooker, dish washer, washing
n, No. of interruptible appliance machine) that require to be operated continuously
ℎ
𝑃𝐷𝐺 , DG power generation at time h until completion of the task, and their starting and
ending times can be flexibly set.
𝑖
𝑃𝑓𝑙𝑒𝑥𝑖𝑏𝑙𝑒 Total flexible loads consumed at i (3) Interruptible flexible load: it refers to the appliances
𝑅𝑇𝑃𝑗 /𝑅𝑇𝑃𝑘 The time of use pricing at time j or k which are allowed to run and can be shut down at any
time in the given time interval, e.g. air conditioner,
𝑖
𝑃𝑚𝑢𝑠𝑡 Total baseline loads consumed at time h clothes dryer, pool pump, floor cleaning robot,
𝑗 electric radiator.
𝐶𝑎 Energy to buy after the curtailment of appliance a at time j
In the case where the required power demand is smaller than
Table 2. Parameters of appliances used in the simulation
the DG outputs during the operational periods, no electricity
purchase is needed. Categorization Appliance Original 𝐷𝑎 Power (kw)
of appliances Index Operation time (30)min
slot
Baseline 1 1 ~ 48 48 0.15
In the next stage of optimization, the real time pricing appliances 2 1 ~ 48 48 1.60
scheme (RTP) will be considered 3 1 ~ 48 48 0.15
And the optimization problem which is formulated as flows: Uninterruptible 4 14 ~ 17 4 0.73
Appliances 5 23 ~ 25 3 0.73
6 24 ~ 26 3 0.80
𝑗
min {0.5× 𝑐𝑎 × 𝑅𝑇𝑃𝑗 + 0.5 𝑐𝑎𝑘 × 𝑅𝑇𝑃𝑘 } (3) 7 20 ~ 22 2 0.80
8 21~ 22 2 0.8
𝑖 𝑖 𝑖 9 12~ 15 4 0.38
𝑗 ∈ {𝑖|𝑃𝐷𝐺 <𝑃𝑚𝑢𝑠𝑡 +𝑃𝑓𝑙𝑒𝑥𝑖𝑏𝑙𝑒 , i =1……..., 48} 10 6~ 9 4 1.26
Interruptible 11 21 ~ 24 4 1.26
s.t Appliances 12 39 ~ 42 4 1.64
13 43 ~ 46 4 1.26
k∈ {k|𝑥𝑎𝑘 =0, k=[𝛼𝑎 , 𝛽𝑎 ]} 14 24 ~ 27 4 0.70
15 19 ~ 22 4 0.74
𝑗 𝑗 𝑗 𝑗 16 29 ~ 32 4 0.64
𝑗
0, 𝑃𝐷𝐺 ≥ 𝑃𝑚𝑢𝑠𝑡 + 𝑃𝑓𝑙𝑒𝑥𝑖𝑏𝑙𝑒 − 𝑥𝑎 𝑃𝑎 17 39 ~ 42 4 1.60
𝑐𝑎 = { 𝑗 𝑗 𝑗 𝑗 𝑗 𝑗 𝑗
(4) 18 15 ~ 18 4 1.90
𝑃𝑚𝑢𝑠𝑡 + 𝑃𝑓𝑙𝑒𝑥𝑖𝑏𝑙𝑒 −𝑃𝐷𝐺 , 𝑃𝐷𝐺 < 𝑃𝑚𝑢𝑠𝑡 + 𝑃𝑓𝑙𝑒𝑥𝑖𝑏𝑙𝑒 − 𝑥𝑎 𝑃𝑎
19 22 ~ 25 4 1.64
20 36 ~ 39 4 1.50
21 13 ~ 16 4 1.50
22 35 ~ 39 4 1.50
𝑘 𝑘 𝑘
0, 𝑃𝐷𝐺 ≥ 𝑃𝑚𝑢𝑠𝑡 + 𝑃𝑓𝑙𝑒𝑥𝑖𝑏𝑙𝑒 + 𝑃𝑎 23 36 ~ 39 4 1.10
𝑐𝑎𝑘 { 𝑘 𝑘 𝑘 𝑘 𝑘 𝑘
(5) 24 28 ~31 4 2.00
𝑃𝑚𝑢𝑠𝑡 + 𝑃𝑓𝑙𝑒𝑥𝑖𝑏𝑙𝑒 + 𝑃𝑎 − 𝑃𝐷𝐺 , 𝑃𝐷𝐺 < 𝑃𝑚𝑢𝑠𝑡 + 𝑃𝑓𝑙𝑒𝑥𝑖𝑏𝑙𝑒 + 𝑃𝑎
25 26 ~ 29 4 1.80
26 23 ~ 26 4 0.25
Where j is the time when the energy requirement exceeds 27 35 ~ 38 4 1.00
the output of DGs after full utilization of DG generation. And 28 20 ~ 23 4 1.20
29 27 ~ 30 4 1.20
k the time slot to which appliance a is shifted. It is important 30 10 ~ 13 4 1.20
to mention here that each appliance is only periodically 31 30 ~ 33 4 1.20
scheduled every half an hour.

𝑗
𝑐𝑎 , is energy to buy at the interruptible appliance a at time j,
𝑐𝑎𝑘 , is energy to buy after the increased shifted appliance a at
tome k. the flow chart of the algorithm described in Fig.5.

97
Authorized licensed use limited to: Virginia Commonwealth University. Downloaded on May 27,2021 at 17:44:18 UTC from IEEE Xplore. Restrictions apply.
 Start

Input Data
1-Load consumption profile 2-TOUP profile (30 mins/Slot)
3- PV generation profile 4 –appliances setting a [𝛼𝑎 , 𝛽𝑎 , 𝑑𝑎 , 𝑟𝑎 ]

Initialization Population (N=10)

N=schedule

Update the schedule 𝑥𝑎ℎ

Y
48 No
𝑚+𝑛
ℎ ℎ
∑ |𝑃𝑝𝑣 − 𝑃𝑚𝑢𝑠𝑡 −∑ 𝑥𝑎ℎ 𝑃𝑎 |
𝑎
ℎ=1

ℎ ≥ 𝑃ℎ ℎ ℎ
𝑗 0, 𝑖𝑓𝑃𝑝𝑣 𝑚𝑢𝑠𝑡 + 𝑃𝑓𝑙𝑒𝑥𝑖𝑏𝑙𝑒 − 𝑥𝑎 𝑃𝑎
Cost per slot 𝐶𝑎 ={ ℎ ℎ ℎ |𝑒𝑙𝑒𝑠𝑒
|𝑃𝑚𝑢𝑠𝑡 + 𝑃𝑓𝑙𝑒𝑥𝑖𝑏𝑙𝑒 − 𝑃𝑝𝑣
Choose a schedule 𝑥𝑎ℎ

Best Cot calculation =min {Cost, 𝐶𝑜𝑠𝑡𝑏𝑒𝑠𝑡 }

Shift the time operation of the appliances a

[ 𝑡𝑎𝑠𝑡𝑎𝑟𝑡 , 𝑡𝑎𝑒𝑛𝑑 ] in [ 𝛼𝑎 , 𝛽𝑎 ] Y No
𝑗
𝐶𝑎 >0

Y
48 Shift the appliance a to the time slot
∑ 𝑥ℎ𝑎 = 𝑑𝑎 k
𝑘=1

No ℎ ≥ 𝑃ℎ ℎ
𝑗 0, 𝑖𝑓𝑃𝑝𝑣 𝑚𝑢𝑠𝑡 + 𝑃𝑓𝑙𝑒𝑥𝑖𝑏𝑙𝑒 + 𝑃𝑎
Cost per slot 𝐶𝑎 ={ ℎ ℎ ℎ
Choose another appliance with 𝑃𝑚𝑢𝑠𝑡 + 𝑃𝑓𝑙𝑒𝑥𝑖𝑏𝑙𝑒 + 𝑃𝑎 − 𝑃𝑝𝑣 𝑒𝑙𝑒𝑠𝑒
(on/off)

𝑗
0.5× 𝐶𝑎 × 𝜋𝑒 (𝑗) + 0.5 × 𝐶𝑘𝑎 × 𝜋𝑒 (𝑘)

Y
𝐶𝑜𝑠𝑡 < 𝑐𝑜𝑠𝑡𝑏𝑒𝑠𝑡

No Cost =0

Update 𝑎𝑛𝑑 𝑠𝑎𝑣𝑒 𝑡ℎ𝑒 𝑛𝑒𝑤 𝑐𝑜𝑠𝑡𝑏𝑒𝑠𝑡

Fig.5.The flow chart of the energy dispatch

98
Authorized licensed use limited to: Virginia Commonwealth University. Downloaded on May 27,2021 at 17:44:18 UTC from IEEE Xplore. Restrictions apply.
 B. The Second Approach at the bus, both of which contribute to the total power loss of
the system. The math representation of the 𝑃𝑙𝑜𝑠𝑠 (𝑡):
The proposed household DSM model is aimed to minimize the 𝑏
electricity cost by scheduling the on/off status of domestic 𝑃𝑙𝑜𝑠𝑠 (𝑡) = ∑|𝐼𝑙 |2 𝑅𝑙 , (17)
appliances over the operational periods, considering the 𝑙=1
dynamic electricity prices, locally available PV generation, and
the penalty prices of appliance operation time-shifting which I𝑙 represent the current betwen two nodes
are included to manage the customer inconvenience caused by
the proposed DSM program. Assume that the proposed DSM 𝑉𝑖 −𝑉𝑗
|𝐼𝑙(𝑖−𝑗) |=| |. (18)
program is scheduled day-ahead over a 24-h (30 min per slot) 𝑍𝑖𝑗
period. The decision variables are the operational status of
appliances 𝑢𝑎 (𝑡) over the next 24 hours and a typical schedule Constraint (12) gives the voltage constraints in the
[𝑢𝑎 (𝑡)]𝐴×𝑇 is a binary (0/1) matrix with following format: distribution network. Constraint (13) indicates the Maximum
Demand (MD) that the aggregate appliance power of a
[𝑢𝑎 (𝑡)]𝐴×𝑇 = [𝑢11, 𝑢12,…, 𝑢1𝑇 ; 𝑢21, 𝑢22 , … 𝑢2𝑇 ;…..; 𝑢𝐴1 , 𝑢𝐴2,...,𝑢𝐴𝑇 ]
household cannot exceed at any time. This upper limit is to
The objective function and constraints of the second proposed prevent super-high-power demand peak even during the hours
DSM model can be presented as below: when day-ahead electricity price is low because the utilities do
not want to have “new” peak load created by the DSM load-
min 𝒞𝑒 + 𝒞𝑝 (6) shifting or because the distribution feeders have capacity
𝑢𝑎 (𝑡)
constraints. Constraint (14) and (15) indicate the total
Subject to:
operation duration and the allowable turn-on time of an
𝒞𝑒 = 0.5 × ∑𝑇𝑡=1 ((∑𝐴𝑎=1 𝑟𝑎 × 𝑢𝑎 (𝑡) − 𝑃𝑝𝑣 (𝑡)) + 𝑃𝑙𝑜𝑠𝑠 (𝑡)) × 𝜋e (𝑡)
appliance. Constraints (14) – (16) specify the original and the
𝑠𝑡 𝑠𝑡
𝒞𝑝 = 0.5 × 𝜋𝑝 ∑𝐴𝑎=1 ∆𝑇𝑎 𝑟𝑎 (8) new starting point, 𝑡𝑎 𝑜𝑙𝑑 and 𝑡𝑎 𝑛𝑒𝑤 respectively, to capture
the duration of time-shifting for flexible appliances.
0 = 𝑓𝐴𝐶 {𝑃(𝑡), 𝑄(𝑡), 𝑃𝑝𝑣 (𝑡), 𝑉(𝑡)} (9)
Where:
𝐴𝑚
𝑃 𝑚 𝑙𝑜𝑎𝑑 (𝑡) = max((∑𝑎,𝑚=1 𝑟𝑎,𝑚 × 𝑢𝑎,𝑚 (𝑡) − 𝑃 𝑚 𝑝𝑣 (𝑡)) , 0) (10) 𝑢𝑎 Binary status of appliance a; 0 = off, 1 = on
𝑟𝑎 Rated power in kW for appliance
𝑃𝑙𝑜𝑠𝑠 (𝑡) = 𝑔𝐴𝐶 {𝑃(𝑡), 𝑄(𝑡), 𝑃𝑝𝑣 (𝑡), 𝑉(𝑡)} (11)
𝐴 Total number of Interruptible and uninterruptible.
0.95 ≤ |𝑉𝑏 (𝑡)| ≤ 1.05 ∀ 𝑏 ∈ {1 𝑡𝑜 𝐵} (12) T Total number of time slots, T=48
∑𝐴𝑎=1 𝑟𝑎 × 𝑢𝑎 (𝑡) ≤ 𝑀𝐷 ∀𝑎 ∈ {1 𝑡𝑜 𝐴} (13) 𝑡 Index for time slots
∑𝑇𝑡=1 𝑢𝑎 (𝑡) = 𝐷𝑎 ∀𝑎 ∈ {1 𝑡𝑜 𝐴} (14) 𝐷𝑎 During of operation time for appliance a
𝑢𝑎 (𝑡) = 0 ∀𝑡 < 𝑠𝑎 or ∀𝑡 > 𝑓𝑎 (15) [𝑠𝑎 , 𝑓𝑎 ] Allowable time range for appliance a
𝑇
∆𝑇𝑎 = 𝟏 . |𝑡𝑎
𝑠𝑡𝑛𝑒𝑤 𝑠𝑡𝑜𝑙𝑑
− 𝑡𝑎 | {1
∀𝑎 ∈ 𝑡𝑜 𝐴} (16) 𝑡𝑎𝑠𝑡 Operation start time of appliance a
𝑠𝑡 𝑡𝑎𝑒𝑛 Operation end time of appliance a
𝑡𝑎 𝑛𝑒𝑤 = [𝑡|𝑢𝑎𝑛𝑒𝑤 (𝑡) = 1]1×𝐷𝑎 ∀𝑎 ∈ {1 𝑡𝑜 𝐴𝑖𝑛𝑡 } (17) 𝑠𝑡𝑛𝑒𝑤
𝑠𝑡 𝑡𝑎,𝑚 Start time of appliance a after DSM shifting
𝑡𝑎 𝑜𝑙𝑑 = [𝑡|𝑢𝑎𝑜𝑙𝑑 (𝑡) = 1]1×𝐷𝑎 ∀𝑎 ∈ {1 𝑡𝑜 𝐴𝑖𝑛𝑡 } (18) 𝑠𝑡𝑜𝑙𝑑
As we can see the objective function in Eq. 6 consists of 𝑡𝑎 Start time of appliance a before shifting
two parts, electricity cost, 𝒞𝑒 , and penalty cost, 𝒞𝑝 . 𝑃𝑝𝑣 (𝑡) PV output in kW at time slot t
Where (7) – (8) define the electricity cost and the penalty 𝜋e (𝑡) Electricity price (¢/kWh) at time slot t
cost. Normally when there is no shift the penalty cost is zero 𝜋𝑝 Penalty price (¢/kWh)
(𝐶𝑝 = 0 $/𝑑𝑎𝑦). As there is no time shift for the operated ∆𝑇𝑎 Time shifting for appliance a
appliances and (∆𝑇𝑎 = 0), in this case, the customer has to pay 𝑀𝐷 Max demand of household consumption
for the consumed energy kW/h as electricity 𝒞𝑒 is based on the 𝒞𝑒 Total electricity cost ($/day)
TOUP. With optimal shifting, the total cost that the customer 𝒞𝑝 Total penalty cost ($/day).
has to pay will be reduced due to the cost saving ( 𝐶𝑠𝑎𝑣𝑖𝑛𝑔 ) in
IV. NUMERICAL SIMULATIONS AND RESULTS
¢/kWh that comes after shifting the operation time from high
price to a low-price period. In this case, the customer will pay As we can see from the first algorithm there is no MD limit on
an extra penalty cost. It is important to mention that the the load consumption. Therefore, the first algorithm tested at
algorithm makes a shift whenever there is cost saving in MD = 0 kW where the customer is free to shift the load without
another world (𝐶𝑠𝑎𝑣𝑖𝑛𝑔 > 𝐶𝑝 ) and the total cost paid by any concern about the maximum demand limit. Also, the
objective function of the first algorithm Eq. 1 neglects the
customer is (𝒞𝑒 − 𝐶𝑠𝑎𝑣𝑖𝑛𝑔 + 𝐶𝑝 ). (9)-(11) represent the
power loss of the feeder. While in the second approach, the
distribution network constraints of AC power flow balance and
algorithm in Eq. 7 considers that additional cost will be added
real power loss and to avoid the violation which may happen
due to power loss of the system. Moreover, the first approach
in the distribution network. We impose a 5% limit on each bus,
has no penalty portion in
where node voltages should be in their normal limit (+/- 5%). the objective function (see Eq. 1). While in Eq.6 the objective
Power loss is due to the high resistance of the line and the load function takes into consideration the customer inconveniences
caused by DSM scheduling which will be translated into

99
Authorized licensed use limited to: Virginia Commonwealth University. Downloaded on May 27,2021 at 17:44:18 UTC from IEEE Xplore. Restrictions apply.
additional compensation cost in the optimization objective
function. This additional cost represents penalty due to shift of the Results in Figs 9, 10, and 11 are from the second algorithm.
time operation slots of appliances so that it includes penalty part, Fig.9 illustrates the load consumption profile of the households
𝒞𝑝 . It is important to mention that to make a fair comparison, it is where the red solid line represents the original energy
assumed that the amount of 𝒞𝑝 equal to zero which means 𝜋𝑝 =0 consumption while blue bars represent the load scheduled after
¢/kWh when applied to the simulation. In this case, the customer applying the second DSM algorithm. Fig.9 also shows that the
can shift the time slots of the appliances. Results in figs. 6, 7, and cost increase to 4.46 ($/day) with just 5¢/day difference as
8 are from the first algorithm. As we can see, Fig.6 illustrates the compared with the first DSM algorithm. However, there are some
typical load profile of the residential household seen in Fig. 2 with advantages in the second algorithm. The first advantage is that
most of the load shifted to the low-price period. The red solid line there is no overload after applying the DSM algorithm, and the
represents the original energy consumption with two peak loads reason is that this algorithm includes a fixed power consumption
from 6:00AM - 10:00 AM and from 4:00PM - 10:00 PM while represented by MD (see Eq.13) this constraint indicates the MD
blue bars represent the load scheduled after applying the first that the aggregate appliance power of a household cannot exceed
DSM approach. Fig.6 also shows that the cost reduced from 18.4 at any time; with 12.4 kW assumed as MD in the simulation. The
to 4.41 ($/day). However, there is one defect here which is the second advantage is that this algorithm takes into consideration
occurrence of high load during the low-price time between 5- 8 the additional cost due to the feeder loss, which makes the
AM. The PV Utilization efficiency was 98.1%. Fig. 7 represents algorithm close to the realistic costs (See Eq.7). It is also important
the power loss of the feeder as its found equal to 103 kW. Fig.8. to mention here that there is one more advantage in the second
illustrates the voltage profile of the bus node with the household algorithm and that is the constraints included the voltage limit
connected. As we can see there is voltage drop during the peak (See Eq.12) to maintain the voltage profile swing between the
load time and voltage increased during the maximum PV accepted values as illustrated in Fig 10. The voltage increased
generated time. from 8:00AM to 4:30 PM . Fig. 11 shows the system power loss
after applying DSM algorithm, this figure clarifies the behavior of
the line loss with load consumption, the total power loss decreased
to 99.3kW as the PV utilization efficiency increased to 98.8%.

Fig.6. Scheduled Load Profile

Fig.9 Scheduled Load Profile

Fig.7. Power Loss

Fig. 10. Voltage Profile

Fig.8. Voltage Profile

100
Authorized licensed use limited to: Virginia Commonwealth University. Downloaded on May 27,2021 at 17:44:18 UTC from IEEE Xplore. Restrictions apply.
 [8] Abolfazl Salami, Farsi Mehdi “Demand side management using direct
load control for residential and industrial Areas” 2015 International
Congress on electric industry automation (ICEIA).
[9] Chen Chen, Jianhui Wang “A Distributed Direct Load Control Approach
for Large-Scale Residential Demand Response” IEEE Trans. Power
system. 29, No. 5, September 2014.
[10] Yuan-Yih Hsu, “Dispatch of Direct load control using dynamic
programming” IEEE Translon power system, Vol.6, pp.1056-1061, 1991.
[11] Y. Lehtr C. Wen-Chi “An Iterative Deepening Genetic Algorithm for
Scheduling of Direct Load Control” IEEE Transaction on Power System,
vol. pp. 1414-1421, 2005.
[12] Zhu, Ziming, et al. "An integer linear programming-based optimization
for home demand-side management in smart grid." Innovative Smart
Grid Technologies (ISGT), 2012 IEEE PES. IEEE, 2012.
[13] Zehir, M. Alparslan, and Mustafa Bagriyanik. "Demand side
management by controlling refrigerators and its effects on
Fig. 11. Power Loss customers." Energy Conversion and Management 64 (2012): 238-244.
[14] M. A. A. Pedrasa, T. D. Spooner, and I. F. Mac Gill, “Scheduling of
demand side resources using binary particle swarm optimization,” IEEE
V. Conclusion Trans. Power Syst., vol. 24, no. 3, pp. 1173–1181, Aug. 2009.
Several demand side management (DSM) techniques and [15] C. M. Chu, T. L. Jong, and Y. W. Huang, “A direct load control of air-
conditioning loads with thermal comfort control,” in Proc. IEEE PES
algorithms has been used in the literature. These algorithms show Gen. Meet., San Francisco, CA, Jun. 2005.
that by adopting DSM and Time-of-Use (TOU) price tariffs; [16] Herter, Karen. "Residential implementation of critical-peak pricing of
electricity cost significantly decreases, and optimal load electricity." Energy Policy 35.4 (2007): 2121-2130.
scheduling is achieved. However, the purpose of the DSM is to [17] Gomes, A., C. H. Antunes, and A. G. Martins. "A multiple objective
approach to direct load control using an interactive evolutionary
not only lower the electricity cost, but also to avoid the peak load algorithm." IEEE Transactions on Power Systems22.3 (2007)
even if the electricity prices low. To address this concern, this [18] Mohsenian-Rad, Amir-Hamed, and Alberto Leon-Garcia. "Optimal
paper starts with a brief literature review on the existing DSM residential load control with price prediction in real-time electricity
algorithms and schemes. These algorithms can be suitable for pricing environments." IEEE transactions on Smart Grid 1.2 (2010)
[19] Ng, K-H., and Gerald B. Sheble. "Direct load control-A profit-based load
Direct Load Control (DLC) schemes, Demand Response (DR), management using linear programming." IEEE Transactions on Power
and load scheduling strategies. Secondly, comparison of two DSM Systems 13.2 (1998): 688-694.
algorithms to show the performance based on cost minimization, [20] Xiong, Gang, et al. "Smart (in-home) power scheduling for demand
voltage fluctuation, and system power loss. The results show the response on the smart grid." Innovative smart grid technologies (ISGT),
2011 IEEE PES. IEEE, 2011.
importance of balance between objectives such as electricity cost [21] Ruiz, Nerea, Iñigo Cobelo, and José Oyarzabal. "A direct load control
minimization, peak load occurrence, and voltage fluctuation model for virtual power plant management." IEEE Transactions on
evolution while simultaneously optimizing the cost. The second Power Systems 24.2 (2009): 959-966
[22] Kurucz, C. N., D. Brandt, and S. Sim. "A linear programming model for
optimization algorithm has potential to provide more value to the reducing system peak through customer load control programs." IEEE
customer in reducing the cost. This is done by evaluating the Transactions on Power Systems 11.4 (1996): 1817-1824.
voltage of the entire system and subsequently avoiding the use the [23] Yao, Leehter, Wen-Chi Chang, and Rong-Liang Yen. "An iterative
appliances at high peak load during the time when the electricity deepening genetic algorithm for scheduling of direct load control." IEEE
Transactions on Power Systems 20.3 (2005): 1414-1421.
price periods is high. In other words, a good DSM algorithm [24] Hassan, Rania, et al. "A copmarison of particle swarm optimization and
should have the objective to minimize the electricity cost for the the genetic algorithm." American Institute of Aeronautics and
customer and maximize the customer convenience to handle large Astronautics (2004).
number of appliances of several types. [25] Beaudin, Marc, and Hamidreza Zareipour. "Home energy management
systems: A review of modelling and complexity." Renewable and
REFERENCES Sustainable Energy Reviews 45 (2015): 318-335.
[26] Aziz, Tariq, and Nipon Ketjoy. "PV Penetration Limits in Low Voltage
[1] M. Pipattanasomporn, M. Kuzlu, and S. Rahman, “An algorithm for Networks and Voltage Variations." IEEE Access 5 (2017): 16784-16792.
intelligent home energy management and demand response analysis,” [27] Luna, Adriana C., et al. "Mixed-integer-linear-programming-based
IEEE Transactions on Smart Grid, vol. 3, no. 4, pp. June. 2012. energy management system for hybrid PV-wind-battery microgrids:
[2] Y. H. Lin and M. S. Tsai, “An advanced home energy management Modeling, design, and experimental verification." IEEE Transactions on
system facilitated by nonintrusive load monitoring with automated multi Power Electronics 32.4 (2017): 2769-2783.
objective power scheduling,” IEEE Transactions on Smart Grid, vol. 6, [28] Beaudin, Marc, and Hamidreza Zareipour. "Home energy management
no. 4, pp. 1839-1851, Jan. 2015. systems: A review of modelling and complexity." Renewable and
[3] Logenthiran, Thillainathan, Dipti Srinivasan, and Tan Zong Shun, Sustainable Energy Reviews 45 (2015): 318-335.
“Demand side management in smart grid using heuristic optimization.” [29] Dominion Virginia Power:
IEEE transactions on smart grid 3.3(2012): 124-1252. https://www.dom.com/residential/dominionvirginia-power.
[4] Ruzbahani, Hossein Mohammedi, and Hadis Karimipor. “Optimal [30] Ruan, Bingjie, et al. "Demand response under real-time pricing for
incentive-based demand response management of smart households.” domestic energy system with DGs." Power System Technology
Industrial and commercial power system technical conference (I&CPS), (POWERCON), 2014 International Conference on. IEEE, 2014.
2018 IEEE/IAS 54th. IEEE, 2018. [31] C. Ibars, M. Navarro, and L. Giupponi, “Distributed demand
[5] Alwan, Hayder O., and Noor M. Farhan. "Load Restoration management in smart grid with a congestion game,” in Proc. IEEE Int.
Methodology Considering Renewable Energies and Combined Heat Conf. Smart Grid Comm. (Smart Grid Comm), Oct. 2010, pp. 495–500.
and Power Systems." arXiv preprint arXiv: 1806.01789(2018). [32] Atzeni, Italo, et al. "Noncooperative and Cooperative Optimization of
[6] .Ruan, Bingjie, et al. "Demand response under real-time pricing for Distributed Energy Generation and Storage in the Demand-Side of the
domestic energy system with DGs." Power System Technology Smart Grid." IEEE Trans. Signal Processing 61.10 (2013): 2454-2472.
(POWERCON), 2014 International Conference on. IEEE, 2014. [33] M. Iqbal, M. Azam, M. Naeem, A. Khwaja, and A. Anpalagan,
[7] Mohsenian-Rad, Amir-Hamed, et al. "Autonomous demand-side “Optimization classification, algorithms and tools for renewable energy:
management based on game-theoretic energy consumption scheduling A review,” Renewable Sustain. Energy Rev., vol. 39, pp. 640–654, 2014.
for the future smart grid." IEEE transactions on Smart Grid 1.3 (2010) [34] Alwan, Hayder E., and Qais S. Al-Sabbagh. "Detection of Static Air-
Gap Eccentricity in Three Phase Induction Motor by Using Artificial

101
Authorized licensed use limited to: Virginia Commonwealth University. Downloaded on May 27,2021 at 17:44:18 UTC from IEEE Xplore. Restrictions apply.
 Neural Network (ANN)." Journal of Engineering15, no. 4 (2009): Residential and Commercial Load." In 2018 IEEE International
4176-4192. Conference on Electro/Information Technology (EIT), pp. 0712-
[35] SC Lee, SJ Kim, and SH Kim, “Demand side management with air 0717. IEEE, 2018.
conditioner loads based on the queuing system model,” IEEE [58] K. Clement-Nyns, E. Haesen, and J. Driesen, “Coordinated charging of
Transactions on Power Systems, vol. 26, no. 2, pp. 661–668, 2011. multiple plug-in hybrid electric vehicles in residential distribution grids,”
[36] Zehir, M. Alparslan, and Mustafa Bagriyanik. "Demand side 2009 IEEE/PES Power Systems Conference and Exposition, vol. 25, no.
management by controlling refrigerators and its effects on customers." 1, pp. 1–7, 2009.
Energy Conversion and Management 64 (2012): [59] Nguyen, Hung Khanh, Ju Bin Song, and Zhu Han. "Demand side
[37] C. M. Chu, T. L. Jong, and Y. W. Huang, “A direct load control of air- management to reduce peak-to-average ratio using game theory in smart
conditioning loads with thermal comfort control,” in Proc. IEEE PES grid." Computer Communications Workshops (INFOCOM WKSHPS),
Gen. Meet., San Francisco, CA, Jun. 2005. 2012 IEEE Conference on. IEEE, 2012.
[38] P. Yang, G. Tang, and A. Nehorai, “A game-theoretic approach for [60] P. Yang, G. Tang, and A. Nehorai, “A game-theoretic approach for
optimal time-of-use electricity pricing,” Power Systems, IEEE optimal time-of-use electricity pricing,” Power Systems, IEEE
Transactions on, vol. 28, no. 2, pp. 884–892, 2013. Transactions on, vol. 28, no. 2, pp. 884–892, 2013.
[39] S. Bu, F. R. Yu, and P. X. Liu, “A game-theoretical decision-making [61] S. Bu, F. R. Yu, and P. X. Liu, “A game-theoretical decision-making
scheme for electricity retailers in the smart grid with demand-side scheme for electricity retailers in the smart grid with demand-side
management,” in 2011 IEEE International Conference on Smart Grid management,” in 2011 IEEE International Conference on Smart Grid
Communications (Smart Grid Comm), 2011, pp. 387–391. Communications (Smart Grid Comm), 2011, pp. 387–391.
[40] Fan-Lin Meng and Xiao-Jun Zeng, “An optimal real-time pricing for [62] Fan-Lin Meng and Xiao-Jun Zeng, “An optimal real-time pricing for
demand side management: A stackelberg game and genetic algorithm demandside management: A stackelberg game and genetic algorithm
approach,” in Neural Networks (IJCNN), 2014 International Joint approach,” in Neural Networks (IJCNN), 2014 International Joint
Conference on, 2014, pp. 1703–1710. Conference on, 2014, pp. 1703–1710.
[41] Kanchev, Hristiyan, et al. "Energy management and operational planning [63] Kanchev, Hristiyan, et al. "Energy management and operational planning
of a microgrid with a PV-based active generator for smart grid of a microgrid with a PV-based active generator for smart grid
applications." IEEE transactions on industrial electronics 58.10 (2011): applications." IEEE transactions on industrial electronics 58.10 (2011):
[42] S. J. Chiang, K. T. Chang, and C. Yen, “Residential photovoltaic energy [64] S. J. Chiang, K. T. Chang, and C. Y. Yen, “Residential photovoltaic
storage system,” IEEE Trans. Ind. Electron., vol. 45, no. 3, Jun. 1998. energy storage system,” IEEE Trans. Ind. Electron., vol. 45, no. 3, pp.
[43] B. Yang, W. Li, Y. Zhao, and X. He, “Design and analysis of a grid 385–394, Jun. 1998.
connected photovoltaic power system,” IEEE Trans. Power Electron., [65] B. Yang, W. Li, Y. Zhao, and X. He, “Design and analysis of a grid
vol. 25, no. 4, pp. 992–1000, Apr. 2010. connected photovoltaic power system,” IEEE Trans. Power Electron.,
[44] X. Vallvé, A. Graillot, S. Gual, and H. Colin, “Micro storage and demand vol. 25, no. 4, pp. 992–1000, Apr. 2010.
side management in distributed PV grid-connected installations,” in Proc. [66] X. Vallvé, A. Graillot, S. Gual, and H. Colin, “Micro storage and demand
9th Int. Conf. Elect. Power Quality Utilization, Barcelona, Spain, Oct. 9– side management in distributed PV grid-connected installations,” in Proc.
11, 2007, [CD-ROM]. 9th Int. Conf. Elect. Power Quality Utilization, Spain, Oct. 9–11, 2007.
[45] G. Delille and B. Francois, “A review of some technical and economic [67] G. Delille and B. Francois, “A review of some technical and economic
features of energy storage technologies for distribution systems features of energy storage technologies for distribution systems
integration,” Ecol. Eng. Environ. Protect. No. 1, pp. 40–49, 2009. integration,” Ecol. Eng. Environ. Protect. No. 1, pp. 40–49, 2009.
[46] Melhem, Fady Y., et al. "Optimal residential load scheduling model in [68] Melhem, Fady Y., et al. "Optimal residential load scheduling model in
smart grid environment." Environment and Electrical Engineering and smart grid environment." Environment and Electrical Engineering and
2017 IEEE Industrial and Commercial Power Systems Europe 2017 IEEE Industrial and Commercial Power Systems Europe
(EEEIC/I&CPS Europe), 2017 IEEE International Conference on. IEEE. (EEEIC/I&CPS Europe), 2017 IEEE International Conference on. IEEE.
[47] Sadeghian, H. R., and M. M. Ardehali. "A novel approach for optimal [69] Zhu., Won and Yoan Shin “An Optimal power Scheduling Method for
economic dispatch scheduling of integrated combined heat and power Demand Response in Home Energy Management System” IEEE Trans.
systems for maximum economic profit and minimum environmental Smart Grid, Vol.4, no.3, Sep.2013.
emissions based on Benders decomposition." Energy 102 (2016): 10-23. [70] Arora, Monika, and Saurabh Chanana. "Residential demand response
[48] Arora, Monika, and Saurabh Chanana. "Residential demand response from PV panel and energy storage device." Power Electronics (IICPE),
from PV panel and energy storage device." Power Electronics (IICPE), 2014 IEEE 6th India International Conference on. IEEE, 2014.
2014 IEEE 6th India International Conference on. IEEE, 2014. [71] Samadi, Pedram, Vincent WS Wong, and Robert Schober. "Load
[49] Samadi, Pedram, Vincent WS Wong, and Robert Schober. "Load scheduling and power trading in systems with high penetration of
scheduling and power trading in systems with high penetration of renewable energy resources." IEEE Transactions on Smart Grid 7.4
renewable energy resources." IEEE Transactions on Smart Grid 7.4 (2016): 1802-1812.
(2016): 1802-1812. [72] Alwan, Hayder O., and Noor M. Farhan. "Design and Application of
[50] A. Alessandro, P. de Gianluca, D. Paolo, V. Antonio, “Load Sceduling Data Aquistion Interface Circuit." arXiv preprint
for Household Energy Consumption Optimization”, IEEE Transaction on arXiv:1806.08721 (2018).
Smart Grid, pp. 2364-2373, vol. 4, no. 4, 2013. [73] S. Shengan, P. Manisa, R. Saifur. “Demand Response as a Load Shaping
[51] M. Corradni, G Ippoliti, and G. Orlando, “An aerodynamic torque Tool in an Intelligent Grid with Electric Vehicles”, IEEE Transactions
observer for the robust control of variable-speed wind turbines,” in IEEE on Smart Grid, pp. 624-631, vol.2, no.4, 2011.
51st Annual Conference on Decision and Control, 2012, pp. 2483-2488. [74] Cohen, Arthur I.; Wang, C.C., "An optimization method for load
[52] Atzeni, Italo, et al. "Noncooperative and Cooperative Optimization of management scheduling," Power Systems, IEEE Transactions on, vol.3,
Distributed Energy Generation and Storage in the Demand-Side of the no.2, pp.612,618, May 1988doi: 10.1109/59.192913.
Smart Grid." IEEE Trans. Signal Processing 61.10 (2013): 2454-2472. [75] Y. Yan, Y. Qian, H. Sharif, and D. Tipper, “A survey on smart grid
[53] M. Iqbal, M. Azam, M. Naeem, A. Khwaja, and A. Anpalagan, communication infrastructures: Motivations, requirements and
“Optimization classification, algorithms and tools for renewable energy: challenges,” Communications Surveys Tutorials, IEEE, vol. 15, no. 1,
A review,” Renewable Sustain. Energy Rev., vol. 39, pp. 640–654, 2014. pp. 5–20, First 2013.
[54] Liam Paull, Howard Li, and Liuchen Chang, “A novel domestic electric [76] Alwan, Hayder O., Hamidreza sadeghian, and sheriff Abdalwahed."
water heater model for a multi-objective demand side management Optimal Energy scheduling of an off-grid reneable system used for
program,” Electric Power Systems Research, vol. 80, no. 12, pp2010.
typical residential household." In 2019 IEEE International
[55] SC Lee, SJ Kim, and SH Kim, “Demand side management with air
Conference on Computational Intelligence and Communication
conditioner loads based on the queuing system model,” IEEE
Transactions on Power Systems, vol. 26, no. 2, pp. 661–668, 2011. Networks. IEEE, 2019.
[56] Elyas, Seyyed Hamid, Hamidreza Sadeghian, Hayder O. Alwan,
and Zhifang Wang. "Optimized household demand management
with local solar PV generation." In Power Symposium (NAPS), 2017
North American, pp. 1-6. IEEE, 2017.
[57] Alwan, Hayder O., Hamidreza Sadeghian, and Zhifang Wang.
"Decentralized Demand Side Management Optimization for

102

Authorized licensed use limited to: Virginia Commonwealth University. Downloaded on May 27,2021 at 17:44:18 UTC from IEEE Xplore. Restrictions apply.
View publication stats