See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/271617541

Demand response in smart electricity grids

equipped with renewable energy sources: A
review
Article in Renewable and Sustainable Energy Reviews February 2013
DOI: 10.1016/j.rser.2012.09.019

CITATIONS

READS

140

1,099

2 authors, including:
J. Aghaei
Shiraz University of Technology
100 PUBLICATIONS 1,548 CITATIONS
SEE PROFILE

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

Managing Wind Power Uncertainty through Integrated Allocation of Flexibility Resources View project

All content following this page was uploaded by J. Aghaei on 09 June 2015.
The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.

Renewable and Sustainable Energy Reviews 18 (2013) 6472

Contents lists available at SciVerse ScienceDirect

Renewable and Sustainable Energy Reviews

journal homepage: www.elsevier.com/locate/rser

Demand response in smart electricity grids equipped with renewable

energy sources: A review
Jamshid Aghaei n, Mohammad-Iman Alizadeh
Department of Electrical and Electronics Engineering, Shiraz University of Technology, Shiraz, Iran

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 20 July 2012
Received in revised form
8 September 2012
Accepted 15 September 2012
Available online 3 November 2012

Dealing with Renewable Energy Resources (RERs) requires sophisticated planning and operation
scheduling along with state of art technologies. Among many possible ways for handling RERs, Demand
Response (DR) is investigated in the current review. Because of every other year modications in DR
denition and classication announced by Federal Energy Regulatory Commission (FERC), the latest DR
denition and classication are scrutinized in the present work. Moreover, a complete benet and cost
assessment of DR is added in the paper. Measurement and evolution methods along with the effects of
DR in electricity prices are discussed. Next comes DR literature review of the recent papers majorly
published after 2008. Eventually, successful DR implementations, around the world, are analyzed.
& 2012 Elsevier Ltd. All rights reserved.

Keywords:
Renewable Energy Resources
Demand Response Programs
Price responsive demand
Demand Side Management
Energy Management Systems
Advanced metering infrastructures

Contents
1.
2.
3.
4.
5.
6.
7.
8.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Demand response denition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Customer categorization and characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
DRPs classication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Benets and costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
DR Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
Literature review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
8.1.
Demand response and integrating RESs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
9. DR status Quo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
9.1.
USA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
9.2.
Europe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
9.3.
China . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
10. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

1. Introduction
Today, environmental concerns, oil crisis, and economic aspects of
utilizing the available energy resources efciently motivated nations,

n
Correspondence to: Department of Electrical and Electronic Engineering,
Shiraz University of Technology, Modares Blvd. Shiraz, P.O.71555-313, Iran.
Tel.: 98 711 7264121; fax: 98 711 7353502.
E-mail address: aghaei@sutech.ac.ir (J. Aghaei).

1364-0321/$ - see front matter & 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.rser.2012.09.019

governments, technology providers, and academic societies to

substitute conventional resources by Renewable Energy Resources
(RESs) [1,2]. Hence, many attempts are in process over integrating
renewable resources with macro grids. Power systems, however,
have some inherent characteristics, which are needed to be scrutinized. Firstly, electricity supply must always equal demand because
of frequency balance. Secondly, electric power is not economically
storable at the scale of large power systems. Variations in power
consumption come next due to consumer behavior. Finally, power

J. Aghaei, M.-I. Alizadeh / Renewable and Sustainable Energy Reviews 18 (2013) 6472

generation costs may vary remarkably because of generation unit

classications [3]. For many reasons, in a conventional power system
consumers are exposed to xed annual rates without seeing the
whole market conditions. The Independent System Operator (ISO), in
a deregulated power system, has the authority to deal with the
mentioned variations in electricity costs and volatilities in power
consumption in order to operate the power system in safe and secure
conditions [4]. One of the most important features of a deregulated
power system is to avoid large-scale power plants and to decentralize
generation into local generation units called Distributed Energy
Resources (DERs) [5]. Nowadays, DERs are widely used to counteract
the above-mentioned issues regarding power systems. An efcient
approach to implement DERs practically along with High-Tech
communication and control devices is called Micro Grid (MG), which
can cause signicant improvements in deregulated power system
operating conditions. MG, in its general prospective, is an exemplar of
a macro grid. In a MG, local energy potentials are mutually cooperated with each other and upstream network as well [6]. DGs in
both forms of RESs and nonconventional thermal units such as Micro
Turbines (MTs), and Diesel engine Generators (Dgen) are exploited
extensively in a MG. RES Penetration to power systems as dispatchable generation units is accompanied by some difculties such as
uncertainties associated with wind and solar power generations.
These uncertainties pose a challenge while computing optimal bids
necessary for participating in the day-ahead unit commitment
process. Some authors solved this problem by applying fuzzy
optimization techniques, with the aim of both maximizing benets
while minimizing risks considering forecast uncertainties [47].
To overcome the intermittency of wind, solar power generation,
and mentioned power market issues intelligent and automated
control systems along with Demand Response Programs (DRPs) and
storage devices in a micro grid scale are required [7]. Moreover, the
following solutions are presented in [46]:
Power plants providing operational and capacity reserve.







Interconnection with other grid systems.

Curtailment of intermittent technology.
Dispatchable distributed generation.
Use of complementarity between renewable sources.
Demand-side management.

The rst assessment of DR and Advanced Metering (AM) was

released in 2006 and the modied version was published in 2008
consecutively. Signicant changes in classications, however,
took place from twelve in 2008 to fteen in 2010. These noticeable modications in DR area between 2008 and 2010 motivated
authors to review recent contributions and modications presented in literatures and to extend [8] as the main framework of
the present review.

2. Demand response denition

In the strategic plan of International Energy Agency (IEA) for
20082012, demand side activities have been in warmth of spot
light in all energy policy decisions because of their signicant
benets both at economic and operational levels [9]. A comprehensive denition for DSM can be stated as what was expressed
in [14]. The DMS optimizes the power ows in the network,
regulates the voltage proles, acting on reactive ows and tap
changers in substation, minimizes the energy losses, recongures
the network, exploits storage devices and responsive loads in an
integrated way. Considering overall purpose of the Load management (LM) programs, three types of Demand Side Management
(DSM) are denable [9,10]:

65

(1) Economic/Market-driven: The purpose of this program is to

reduce general costs of energy supply, increase the reserve
margin, and mitigate price volatility by means of smart-term
responses to electricity market conditions.
(2) Environmental-driven: Provides environmental and/or social
purposes by decreasing energy usage, dening commitment
of not environmentally friendly generation units, leading to
energy efciency augmentation, and/or reduction in greenhouse
gas emissions.
(3) Network-driven: The aim of this program is maintaining the
system reliability by decreasing demand in a short period
of time and reducing extra generation/transmission capacity
enhancement.
In addition to the above classications, a more precise denition
of DRP seems to be crucial. One of the basic comprehensive DR
denitions, as a subsequent part of DSM programs, announced by
FERC says that DR is the ability of customers to respond to either a
reliability trigger or a price trigger from their utility system operator,
load-serving entity, regional transmission organization (RTO)/ISO, or
the demand response provider by lowering their power consumption [11]. For many years, according to the term, DR means just peak
clipping approach for specic hours of a year. Order no. 719,
however, dened DR to mean a reduction in consumption of electric
energy by customers from their normal consumption pattern in
response to a price augmentation or incentive payment designed to
encourage lower consumption of electric energy [11]. As what can
be inferred from the recent denition, DR includes incentive payment actions in addition to peak clipping actions. Meanwhile, Nordic
Power Market (NORDEL) dened DR in a similar frame but with
modied details. The NORDEL uses DR to refer to a voluntary
temporary adjustment of electricity demand in response to a price
signal or a reliability-based action.

 DR can be implemented in a short period of time (Capacity) or

medium-term (Energy).

 The price signals can be sent from the power market, intraday



market, and regulatory power market after a Transmission

System (TSO) Call, balancing markets or from tariffs.
TSO along with distribution companies have the authority to
provide reliability-based actions, which can be activated manually or automatically.
DGs can be considered as a DR from consumption prospective [3].

Responsive demand, however, refers to those changes applied by

customers to their expected load pattern in response to energy price
signals for improving the economic efciency of their energy.
The mechanism discourages the energy load when real-time price
is high and vice versa. Consequently, peak reduction and eco-friendly
standard of life is obtained by performing this mechanism [12,13]
Eventually, the recent denition of DR used in [24] and report is:
Demand Response: Changes in electric use by demand-side
resources from their normal consumption patterns in response to
changes in the price of electricity, or to incentive payments
designed to induce lower electricity use at times of high wholesale market prices or when system reliability is jeopardized.
This denition substitutes demand-side resources for the phrase
end-use customers used in previous surveys, to conform to the
denition in use by NERCs Demand Response Data Task Force in its
development of a Demand Response Availability Data System (DADS)
to collect demand response program information [24].
3. Requirements
FERC staff recently modied denition for advanced meters to be
consistent with that used by the Energy Information Administration

66

J. Aghaei, M.-I. Alizadeh / Renewable and Sustainable Energy Reviews 18 (2013) 6472

(EIM). The report says Advanced Meters are Meters that provide
usage data, measured and recorded at hourly intervals or more
frequently, to both consumers and energy companies at least once
daily. Data are used for billing and other purposes. Advanced meters
include basic hourly interval meters, one-way communication
meters, and real-time meters with built-in two-way communication
cable of recording and transmitting instantaneous data.
The modied denition resulted in changes in advanced meter
counts by some entities responding to the 2010 Survey compared
to the 2008 and 2006 FERC Surveys. In follow-up calls, staff
learned that three respondents reclassied meters previously
reported as advanced meters to non-advanced meters because
of the new requirement of providing customers with usage data
at least once daily. [24].
An efcient infrastructure to implement a DRP practically is
Energy Management System (EMS) in a Smart Grid (SG) infrastructure. The term SG is a fully automated electric power system which
has the authority to control and optimize the operation of all its
interconnected elements, in order to operate generation, transmission, and distribution safely and efciently [16,17]. Today, many
countries exercise SGs, including, for example, U.S., Canada,
Germany, Japan, India, and Australia [18,19]. Further initiatives
towards the future SGs are concerned with the so called Virtual
Power Plants (VPPs) which means aggregating interconnect DGs
placed in different locations by managing them to work as a unique
power plant. This method allows even the smallest DG to participate
in electric market and contribute to energy cost reduction process
[20]. In [21] a VPP with combined heat and power micro-units is
presented. VPP cooperated with DR and wind power generation is
also investigated in [22]. A possible scenario to achieve a dynamic
control of all interconnected elements is as follows:







supervisory control and data acquisition (SCADA);

remote terminal units (RTUs);
advanced metering infrastructure (AMI);
state estimation algorithms (SEAs);
generation and load forecast system (GLFS).

In the context of DR, there is a need to develop a consistent

approach for integrating the communication backbone for providing price signals or notication of system emergencies with the
Advanced Metering Infrastructure (AMI) system. Advanced
metering systems (AMS) are comprised of state-of-the-art electronic/digital hardware and software, which combine interval
data measurement with continuously available remote communications. These systems enable measurement of detailed, timebased information and frequent collection and transmittal of such
information to various parties. AMI typically refers to the full
measurement and collection system that includes meters at the
customer site, communication networks between the customer
and a service provider and data reception and management
systems that make the information available to the service
provider [15]. The SCADA system transmits the measurement
data, provided by an AMI and a set of remote collecting data
devices (RTUs) placed in strategic positions along the SG, to the
EMS. The latter determines the actions required for the optimum
state of the SG by using SEAs and GLFS [23].
It is using new communication technology innovations and
policy directions that many DR actions can change any part of the
load prole of a utility or region, not just the period of peak usage.

All of the above mentioned customers have their own specic

characteristics to be participated in DRPs. Customers can participate in DRPs by decreasing their usage during critical peak
periods when prices are high without changing their consumption pattern as illustrated in Fig. 1. Another alternative can be a
reduction in customers normal consumption pattern by shifting
some of their peak demand usage to off-peak hours as shown in
Fig. 2.
The last choice for customers is employing DGs to supply their
demand in high price hours. As reported in [24] commercial and
industrial customers, though fewer in number than residential
customers, provide a higher total level of load reduction potential
than residential customers. Because of the available systems and
technology in place commercial and industrial customers are
more likely to facilitate demand response program participation.
Moreover, many demand response programs are available only to
customers above a certain size cut-off. As it can be inferred from
Fig. 3, potential peak load reduction rose rapidly from just below
15,000 MW in 2006 to far more than 20,000 MW in 2010 for
commercial and industrial customers.
In contrast, potential peak load reduction for residential
customers showed a slight upward trend from about 6000 MW
in 2006 to near 7500 MW in 2010 respectively. This means that
commercial and industrial customers, though fewer in number

Fig. 1. Curtailing load without changing load pattern.

Fig. 2. Load shifting.

4. Customer categorization and characterization

Electricity customers can be classied into the three main
parts namely, industrial, commercial, and residential customers.

Fig. 3. Potential peak load reduction.

J. Aghaei, M.-I. Alizadeh / Renewable and Sustainable Energy Reviews 18 (2013) 6472

than residential customers, provide a higher total level of load

reduction potential than residential customers.

5. DRPs classication
DRPs classication has been changed a lot since 2006, when
the rst Assessment of Demand Response and Advanced Metering
was announced by FERC. Previously, DRPs had been classied into
two major categories namely, Incentive-based programs (IBP) and
Price-based programs (PBP). Ref. [24], however, presented all
programs without previous clustering. It is worth noting that
the number of the programs, specied by FERC Surveys, was
twelve in 2008 and fteen in 2010 respectively. This part is
dedicated to the recent program categorizations and denitions.
Direct Load Control as one of the most efcient and popular
DRPs is dened as an activity by which the sponsor (i.e. RTO, ISO,
or local provider) turns off or cycles a customers electrical
equipment (e.g. air conditioner, water heater) remotely on short
notice. This program is basically offered to low consumption
customers (i.e. residential and small commercial customers).
Interruptible Load Program has the responsibility to curtail or
interrupt an electric consumption under tariffs or contracts that
provide a rate discount or bill credit for agreeing to regulate load
in case of system contingencies. In some cases, System Operator
might reduce load after noticing customer in accordance with
contractual provisions. One of the recently added programs is
Critical Peak Pricing with Control, which is a combination of
direct load control programs and a pre-specied high price during
allocated critical peak hours, triggered by system contingencies or
high wholesale market price. Load as a Capacity Resource, the
latest terminology of Capacity program, is a Demand Side
Resource (DSR) that commits to reduce the pre-specied amount
of load in case of contingencies. Spinning reserve (responsive
reserve), previously included in ancillary services, is a DRP that is
synchronized and always prepared to deal with any demand and
supply imbalance as a cure within the few minutes when an
emergency event occurs. Non-spinning reserve, previously
counted as an ancillary service, can be activated with a ten
minutes delay to solve energy supply and demand imbalance.
Emergency DRP is a DSR that encourages customers to reduce
their loads during emergency DR events by offering incentive
payments. Regulation Service, prior to the latest FERC Survey
counted as one of the ancillary service, is a DR service by which an
SO can increase or decrease a customers load in response to real
time signals. This DRP is preformed continuously during a
commitment period. Providing normal regulating margin is the
main responsibility of the service as an Automatic Generation
Control (AGC) provider. Some literature papers named this service
as regulation or regulating reserves, up-regulation, and downregulation. Demand Bidding and Buy-back as a DRP gives the
authority to a demand resource to offer load reductions associated with a price or to specify how much load can be curtailed
at a given price. This service can be employed in both retail and
wholesale market. Time of Use (TOU) is a service, which provides
different electricity prices associated with different periods. It is
worth noting that the rates specied by this service reect the
average cost of power generation and delivery during each time
intervals, which are typically longer than one hour within a
24-hour day. Critical Peak Pricing program allocates high electricity prices/rates during restricted number of days or hours in
case of high wholesale market prices or contingencies to discourage electricity consumption during mentioned time intervals.
The Real Time Pricing (RTP) reects the changes in wholesale
price of electricity by uctuating hourly or more often retail
prices/rates. Economists believe that RTP programs are the most

67

efcient and direct DR programs for being spotted in competitive

markets [37]. Peak Time Rebate (extreme day pricing) is a service
by which customers can earn a rebate by consuming electric energy
less than a baseline during a determined number of hours on
critical peak days which is typically capped for a calendar year
because of reliability concerns over very high supply prices.
Eventually, a System Peak Response Transmission Tariff, newly
added program, functions during peak periods to specify terms,
conditions, and rates and/or prices for customers with interval
meters that decrease load as a way of abating transmission changes.

6. Benets and costs

DR benets as shown in Fig. 4 are assessed in seven categories
in this paper:
Economic, pricing, risk management and reliability, market
efciency impacts, lower cost electric system and service, customer services, and environmental benets.
Economic benets can be classied among the most important DR
advantages. From customers point of view, acceptable rebates can be
achieved in electricity bills if customers abate their consumption
pattern as stated in contractual provisions [2529]. Moreover,
customers can obtain rebates even without reducing their total usage
by shifting their consumptions from peak hours with high prices to
off-peak hours. Savings are available for those customers who
consume less than their class average during peak hours. Successful
DR implementations revealed the acceptable effects of DR in electricity price reduction [24,30,31]. Implementing DRPs, in long term
horizon, can cause wholesale market price abatements because of a
more sophisticated utilization of available infrastructures, as in, for
example a reduction of demand means a reduction in committing
expensive generation units [8,29,32]. Besides, performing DRPs can
increase short-term capacity costs. Consequently, avoided or deferred
capacity costs such as need for distribution and transmission infrastructure upgrades are obtainable [8]. One of the other advantages of
DR in pricing area is price volatility and hedging cost reduction.
Price volatility is explicitly shown in Fig. 5, at the spot market
and consumption in Eastern Denmark, for two Mondays in

Fig. 4. DR benets.

68

J. Aghaei, M.-I. Alizadeh / Renewable and Sustainable Energy Reviews 18 (2013) 6472

Fig. 5. DR costs.

loads. Although it sounds that individual demand resources may

have a high failure rate to curtail load on short notice, the
aggregation of several smaller resources into one larger resource
makes it probabilistic that the assigned response will be achieved.
This characteristic of demand resources potentially makes demand
response resources more reliable than conventional generation
resources. For generating resources that are called upon and therefore produce energy for a period of up to 30 min, the resource is
guaranteed recovery of its costs to provide the energy, including
incremental and startup costs. Demand resources would similarly be
reimbursed for any costs incurred during a spinning event. As a
regulation provider, demand response can respond quicker to a
regulation signal than a conventional, fossil fueled, steam generator
could. Besides, DRPs have additional advantages on market power
mitigation based on the concept that greater demand response and
increased transparency will make it easier for individual players to
monitor and respond to each others behavior. Considering DRPs
effects on electric system cost, many benets are obtainable such as
reducing peaking capacity requirements, decreasing transmission
capital and operating expense, deducting distribution capital and
operating expense, reducing in Load Serving Entities (LSE) commodity costs, and reducing in long-term resource adequacy requirements. Customers choices in receiving electric power in proportion
to their preferences have been increased by introducing DRPs.
Eventually, FERC emphasizes on demand side resources instead of
customers motivates penetrating Renewable Energy Resources
(RERs) more and more in SGs. Consequently, lower released emission is obtainable.

7. DR Costs

Fig. 6. Price volatility reduction with DRPs.

November 2005. On November 28 the price reached up to an

extreme peak of DKK13, 469/MW, about 60 times more than the
normal price level, but compared to a week before, consumption
was almost unchanged. Demand responsiveness causes a large
reduction in price volatility by displacing demand curve from
peak demand to off-peak demand. As illustratively shown in
Fig. 6, displacing demand curve caused a smaller deviation from
the cost points B and B to the points of intersection C and C.
DRPs have also remarkable inuences on risk reduction and
reliability enhancement. Reliability advantages can be extracted
from Ancillary Service DRPs (AS-DRPs). The basic concept of
demand providing AS-DRPs is similar to that of a generator,
except that rather than changing the level of generation output
to the grid, the demand resource is changing its level of consumption from the grid. The Notice of proposed rulemaking
(NOPR) identies three ancillary services (regulation and frequency response, operating reservesspinning, and operating
reservessupplemental) that might conceivably be provided by

Implementing and operating DRPs impose expenses to both

demand side participants and service operators. Demand side participants have to buy and install smart thermostats, communication
infrastructures, peak load control, and Energy Management Systems
(EMSs). In addition to the capital costs, operating and maintenance
costs have to be paid by participants as continuous expenses.
Investments in Renewable Energy Resources (RERs) in demand side
can be considered as participant costs. System operators or any other
DRP provider such as Transmission System Operator (TSO) (e.g.
NORDEL power system) or ISO are imposed some costs. These costs
are communication infrastructures in control side and AMIs, which
are responsible for measuring, gathering, storing, and transmitting
energy consumption/supply information and can be considered as
initial costs imposed to DR providers. Eventually, an important cost
component before deploying DRP is educating eligible customers
about DRP advantages.

8. Literature review
One of the most recent challenging issues regarding RERs
(i.e. wind power generation, photovoltaic resources, solar systems)
is that highly protable resources are inherently intermittent and
cannot be employed as dispatchable resources. Many attempts have
been recently done to mitigate the intermittency behavior of RERs.
Energy Storage Systems (ESSs) and DRPs are two major applicable
solutions presented in recent papers. In [7] many recent problem
formulations in case of integrating wind power generation and
price responsive demand are compared with respect to information
exchange requirements, computational complexity, and physically
implementable dynamic model-predictive dispatch. Authors proposed a novel multi-directional interactive dynamic monitoring
and decision-making systems (DYMONDS) algorithm to implement near-optimal predictive dispatch, instead of commonly used

J. Aghaei, M.-I. Alizadeh / Renewable and Sustainable Energy Reviews 18 (2013) 6472

Security-Constrained Economic Dispatch (SCED). To provide the

input data, the proposed method requires information exchange in
the function spaces, instead of only communicating (quantity, price)
data points. They claim that mitigating the intermittency of RERs can
be achieved by implementing (1) demand response, (2) predictive
wind power model, and (3) the use of physically implementable
dynamic model-predictive dispatch of responsive demand, wind
power, and conventional power plants. Same authors in [33] showed,
in a case study that the look-ahead dispatch of the proposed method
is physically implementable and when used with the elastic demand,
wind power generation up to 50% can be accommodated. They
showed that the most efcient way to implement the existing
infrastructure is to enable current available SCADA with a multidirectional information exchange between the control center and the
distributed decision maker, use this information interactively to
manage inter-temporal constraints by distribution decision makers
to create their supply and demand functions, and nally have system
operator clear the supply and the demand functions by running
todays SCED with its special constraints. Utilizing DR in wind
penetrated networks in order to mitigate the randomized behavior
of the RERs has been investigated in many other papers [34,35].
Strengths and weaknesses of the recently announced DRPs in
[24] have been scrutinized in many papers. Real time pricing, as
one of the most protable DRPs, is investigated in [36]. In this
paper, it is supposed that ISO has the authority to pile up bids and
offers from Generation Companies (GENCOs), Transmission Companies (TRANSCOs), Load Serving Entities (LSEs), and Curtailment
Service Providers (CSPs) and then by solving a SCUC considering
hourly DR with inter-temporal load characteristics, both electricity price and scheduling supply and demand are optimally
determined. The proposed method implements the SCUC in order
to maximize the social welfare by adding realization DR constraints (i.e. minimum up/down time limits, load pickup/drop
rates, minimum hourly curtailment, and maximum daily curtailment) in the optimization problem. The results show that (1) peak
load reduction would mitigate price spikes and enhance economical dispatch (2) changes in hourly demand prole can cause
reduction in the average system Local Marginal Price (LMP), (3)
DR application in some buses in the grid could provide benets to
the entire power system and all market participants, (4) SCUC is
preferred instead of SCED because SCUC would enhance the
exibility and the efciency of market operations, (5) higher DR
means not only more at demand curve but also lower fuel
consumption and less carbon footprint in power system. A novel
approach for real time DR model along with electricity price
uncertainty is proposed in [37]. Price uncertainty is modeled
through robust optimization using duality properties and exact
linear equivalences [38]. Meanwhile, the presented model is
expressed in linear programming algorithm in order to be easily
implementable in EMSs.
The concept of price elasticity of demand is widely investigated in [4,3941]. Price elasticity of demand can be dened as
the ratio of the relative change in demand to the relative change
in price [4]. Interruptible/curtailable load as one of the most
benecial DRPs based on price elasticity of demand has been
investigated in [40]. Price elasticity in [40] is bisected in selfelasticity and cross-elasticity. Self-elasticity represents the characteristics of such loads which are not able to move from one
period to another (e.g. illuminating loads), and they always have
negative elasticity. Cross-elasticity, however, is related to those
loads, which can be transferred from the peak period to the offpeak, or low periods (process loads). In order to model single
period elastic loads, the procedure proposed in [40,41] has to be
implemented. Authors showed that the following equation
declares that how much should be the customers consumption
to achieve maximum benet in a 24 h interval while participating

in I/C and CAP programs.




rir0 i Ai peni
di d0 i: 1 Ei,i:
r0i 9


24
X
rjr0 j Aj penj =

Ei,j:
;
r0 j

69

j1

where d(i) and d0(i) are the optimal curtailed load and the initial
load respectively. r0 (i) is the initial price and r(i) is the secondary
price, which has to be determined during optimization solving. E(i, i)
and E(i, j) are the self-elasticity and the cross-elasticity of demand in
different hours. Depending on incentive based or penalty based
DRPs, A(i) and pen(i) can be accommodated in the above formula.
The achieved formula has been employed in many recent papers to
model the responsiveness of the demand. Aggregating all four
possible demand vs. price formulations (i.e. linear, logarithmic,
potential, exponential) by weighting coefcients, based on optimal
curtailed load, caused a comprehensive model of DR proposed in
[39]. Moreover, dynamic price elasticity is investigated instead of
xed price elasticity. In [32] a Demand Response Unit Commitment
(DRUC) is solved considering non-penalized DRPs (i.e. DLC, EDRP)
using the optimal curtailed load as a proportion of the initial load.
Comparing many versatile case studies showed that DR as a virtual
power plant can signicantly inuence on both fuel cost and
emission cost reduction. In [45] curtailed demand in DRPs is treated
as virtual generation unit. In this paper, it is assumed that the
difference between initial and curtailed demand would have a
marginal cost and a cost function respectively. By adding realization
constraints such as ramping rates, minimum up/down time, and the
constraints regarding the customers information, the virtual generation would be more applicable.
8.1. Demand response and integrating RESs
As stated in the previous sections, integrating renewable
resources into the grid in large scales is along with some restrictions,
which are scrutinized in this section. Moreover, to reconcile
between conventional generating units and the renewable power
resources multi-operation, special DRPs are also investigated.
RESs are intermittent resources and integrating them into the
grid means dealing with many uncertainties. The uncertainties
with each time intervals and the proper services to deal with each
time interval uncertainty are illustrated in Table 1.
Prior to announce the required DRPs in order to counteract the
effects of uncertainties, introducing general additional required
infrastructures for a RES integrated system seems crucial.
To prepare the power system for integrating RESs, the reliability of the power system has to be increased through the
following requirements:
increase the need for regulation, spinning reserve, and load
following resources;
resulting in steeper system ramping requirements;
increase the frequency and magnitude of over-generation
events;
resulting in less efcient dispatch of conventional resources.
As the rst alternative says, power system operator requires
many different ancillary services in order to maintain the balance
between demand and supply. The suitable services to deal with
the different uncertainties are as follows:

 Spinning reserve: Spinning reserve is the portion of unloaded

capacity from units already connected or synchronized to the
grid and that can deliver their energy in 10 min and run for at
least two hours.

70

J. Aghaei, M.-I. Alizadeh / Renewable and Sustainable Energy Reviews 18 (2013) 6472

Table 1
Uncertainties in different time intervals.

 Non-spinning (or supplemental) reserve: Non-spinning (or

supplemental) reserve is the extra generating capacity that is

not currently connected or synchronized to the grid but that
can be brought online and ramp up to a specied load within
ten minutes.
Regulation up and regulation down: Regulation energy is used
to control system frequency that can vary as generators access
the system and must be maintained very narrowly around
60 Hz. Units and system resources providing regulation are
certied by the ISO and must respond to automatic generation
control (AGC) signals to increase or decrease their operating
levels depending upon the service being provided, regulation
up or regulation down.

The second requirement addresses the rapid changes in the

generation of the renewables. The integration of such resources
requires dening proper ramp rates and load following services:

 Maximum continuous ramping: Maximum continuous ramp-

ing is the megawatt amount by which the net load (load minus
wind and solar) is expected to change in either an upward or a
downward direction continuously in a given month.
Load following: load following is the ramping capability of a
resource to match the maximum megawatts by which the net
load is expected to change in either an upward or a downward
direction in a given hour in a given month.

Generally, the required DRPs as ancillary services, as stated in

Table 1, are:

 Frequency response: immediate response (up to 20 seconds) in





response to contingencies;
Regulation: manage uncertainty in 510 min ahead forecasts;
Following/ramping: manage remaining intra-hour uncertainty;
Additional resource commitment: manage deviations between
day-ahead and hour-ahead schedules.
These programs are briey tabulated in Fig. 7 [48].

Fig. 7. Resources needed to satisfy grid reliability and exibility requirements.

9. DR status Quo
In this section, three main implementations of DRPs, in USA,
Europe, and China, are discussed in many aspects.

9.1. USA
The 2010 FERC Survey [24] results indicate that the majority of
customers enrolled in a direct load control program are in
programs offered by their Investor-Owned Utilities (IOU), reecting the larger number of customers in IOUs. After IOUs, cooperatives have the most customers enrolled in direct load control
programs.

J. Aghaei, M.-I. Alizadeh / Renewable and Sustainable Energy Reviews 18 (2013) 6472

It is noteworthy that over 10 percent of customers in MRO and

FRCC regions participate in direct load control programs. In FRCC,
Florida Power and Lights On Call direct load control program is
the largest such program nationally, with over 800,000 customers
enrolled. Besides direct load control programs, participation in
interruptible/curtailable programs showed an upward trend from
248 in 2008 to 265 in 2010 where more than 85 percent of the
reporting entities are IOUs and Cooperatives [11]. Next comes
TOU programs, contributing to an overall decline from 241
entities in the 2008 to 169 entities in the 2010 while a reported
1.1 million US residential customers are enrolled in time-of-use
rates in the US [24].
Despite time-of-use offerings by dozens of entities across the US,
and particularly in MRO, most residential customers with time-ofuse rates are in WECC. Wholesale market participants also had the
largest increase in potential peak reductions: from 12,656 MW in
the 2008 to 22,884 MW in the 2010 [24]. Much of the increase in the
wholesale class is due to the growth of demand response in RTOs,
from a reported 9060 MW in 2008 to 20,533 MW in 2010. Since
2007, ISO New England and PJM Interconnection have commenced
long-term forward capacity markets that have attracted signicant
amounts of demand response.
Commercial and industrial customers potential load reduction
increased by 23 percent from 2008 to 2010. The top four demand
response programs, emergency response, interruptible load,
direct load control, and load as capacity resource account for 79
percent of the total US peak load reduction potential. Except for
direct load control, these programs predominantly enroll wholesale, commercial, and industrial customers that bring high per
participant peak load reductions. Direct load control programs are
typically targeted to residential and small nonresidential customers; the controllable technologies are relatively homogeneous
across these customers. Radio or other communication signals
sent by the program sponsor are necessary for effective control of
the large numbers of small loads [11].
9.2. Europe
In this subsection DR experiences in three European countries
namely, Norway, Italy, and Spain are reported. In Norway specic
programs have been undertaken with the objectives of postponing
expansion of grid capacity: 10% reduction in peak demand in the
Oslo area; increasing knowledge of customer behavior; and developing a motivation model for DSM [31]. Pilot studies show that
using this DR program the peak load for commercial end-users was
reduced by 4.5 MW and the energy saving was around 15%. In Italy,
Interruptible Programs, applied to very large industries, only represent 6.5% of peak power and Load Shedding Programs initiate
automatic load shedding in emergencies [42]. Load Shedding
Programs are divided into real time programs (without notice) and
15 min notice programs. The size ranges from 1200 MW for real
time programs to 1750 MW for notice programs. Participants in
these programs have to install and maintain Load Shedding Peripheral Units and will be compensated according to a non-market price
dened in regulation. The size of curtailable power is of 10 MW for
programs without notice and 3 MW for programs with notice.
Currently, DR programs in Spain focus mainly on industry, as most
European utilities include Direct Load Control programs as part of
their DSM strategies, with xed compensations attributed to small
numbers of large industrial end-users. Interruptible Programs for
large industries are very frequent, although the mechanisms for
compensating industries vary signicantly. The extent to which such
system-led programs can be labeled as authentic DR initiatives is
debatable. In the future, however, large numbers of end-users,
including commercial customers and households could be involved
in DR programs with compensations consisting of prices and

71

deliberate shifts in electricity demand in correspondence with peak

loads [43].
9.3. China
The DRPs in China have been ourished mainly through TOU
and interruptible load pricing and deployment of energy storage
devices [30]. Based on the mentioned DSM, just in Beijing,
signicant load factor enhancement is obtained, at about 81
percent in 1997. It is worth noting that the load factor would
have decreased to 76.59 percent if no DSM had been used [44].
The following actions have been taken up to 2003.
The difference between the on peak and off-peak electricity
price was risen. By the end of 2003, more than 77,431 consumers,
participated in TOU programs, represent 61.69% of the total consumption.
Interruptible load contracts have been signed with major industrial loads such as Yanshan Petrochemical Corporation and Shougang
Group. The result was 100 MW shifted peak load each year.
Loads were controlled at Electric Load Management Center
(ELMC). Wireless communication infrastructure connects more than
5000 consumers. In 2003, 1600 locations with 2800 MW of load were
connected, and approximately 500 MW could be directly controlled.

10. Conclusion
In this DR comprehensive review the latest modications in
DR denition and classications are scrutinized. Moreover, benets and related expenditures are categorized extensively.
A literature review of the very recently published papers about
DRPs is presented. Eventually, successful DR experiences throughout the world are assessed.
References
[1] Sioshansi R, Short W. Evaluating the impacts of real-time pricing on the usage
of wind generation. IEEE Transactions on Power Systems 2009;24(2):51624.
[2] Ipakchi A, Albuyeh F. Grid of the future. IEEE Power Energy Magazine
2009;7(2):5262.
[3] Andersen FM, Jensen SG, Larsen HV, Meibom P, Ravn H, Skytte K, et al.
Analyses of demand response in Denmark. Report of Ris National Laboratory
2006; Ris-R-1565(EN).
[4] Kirschen D, Strbac G. Fundamentals of power System economics. John Wiley
& Sons Ltd; 2004.
der L. Distributed generation: a denition.
[5] Ackermann T, Andersson G, SO
Electric Power Systems Research 2001;57(3):195204.
[6] Hernandez, Aramburo CA, Green TC, Mugniot N. Fuel consumption minimizationof a microgrid. IEEE Transactions on Industry Application 2005;41(3):
67381.
[7] Ilic MD, Xie L, Joo J. Efcient coordination of wind power and priceresponsive demandPart I: Theoretical foundations. IEEE Transactions on
Power Systems 2011;26(4):187584.
[8] Albadi MH, El-sadany EF. A summary of demand response in electricity
markets. Electric Power System Research 2008;78(11):198996.
[9] Demand-side management program. Strategic plan for the IEA 20082012.
Available from: /http://www.ieadsm.orgS.
[10] Worldwide survey of network-driven demand-side management projects.
Task XV, IEA-DSM Res, Report no. 1, 2006.
[11] National action plan on demand response, The federal energy regulatory
commission, 2010.
[12] Chao H. Price responsive demand management for a smart grid world.
Electricity Journal 2010;23(1):720.
[13] Nilsson H. The many faces of demand-side management. IET Power
Engineering 1994;8(5):20710.
[14] Pilo F, Pisano G, Soma G. Advanced DMS to manage active distribution
networks. IEEE bucharest powertech conference. Bucharest: Romania; 2009.
[15] Advanced metering infrastructure (AMI), Electric Power Research Institute
2007.
[16] Smart grids european technology platform smart grids. Available from:
/http://www.smartgrids.euS.
[17] Ergeg S. ERGEG position paper on smart metering with a Focus on electricity
regulation. Ref_R07-RMF-04-03, 2007.
[18] Morozumi S. Micro-grid demonstration projects in Japan. Proceedings of the
power conversion conference. (Nagoya, PCC07), 2007. p. 63542.

72

J. Aghaei, M.-I. Alizadeh / Renewable and Sustainable Energy Reviews 18 (2013) 6472

[19] Balijepalli S, Khaparde VSKM, Dobariya SA. Deployment of microgrids in

India. IEEE Power and Energy Society General Meeting 2010:17.
[20] First virtual power plant operated by siemens and RWE energy siemens and RWE
energy 2008. Available from: /http://w1.siemens.com/press/en/pressreleaseS.
[21] Schulz C, Roder G, Kurrat M. Virtual power plants with combined heat and
power micro-units. International Conference of Future Power Systems 2005:5.
[22] Mohammadi J, Rahimi-Kian A, Ghazizadeh MS Joint operation of wind power
and exible load as virtual power plant. Proceedings of the international
conference on environment and electrical engineering (EEEIC). 2011. p. 14.
[23] Muller H, Rudolf A, Aumayr G Studies of distributed energy supply systems
using an innovative energy management system. 22nd IEEE power engineering society. Proceedings of the international conference on power industry
computer applications, (PICA 2001). 2001. p. 8790.
[24] Assessment of demand response and advanced mitering. Federal energy
regulatory commission reports 2010.
[25] Bloustein E. Assessment of customer response to real time pricing, The State
University of New Jersey 2005. Available from: /http://www.policy.rutgers.eduS.
[26] Albadi MH, El-Saadany EF. Demand response in electricity markets: an
overview. IEEE Power Engineering Society General Meeting 2007:15.
[27] Gellings C, Chamberlin J. Demand side management: concepts and methods.
US: The Fairmont Press Inc; 1988.
[28] Kirschen DS. Demand-side view of electricity markets. IEEE Transactions on
Power Systems 2003;18(2):5207.
[29] Violette DM. A comprehensive/ Integrated DR value framework. Summit Blue
Consulting 2006.
[30] Wang J, Bloyd Cary N, Hu Z, Tan Z. Demand response in China. Energy
2010;35(4):152997.
[31] Torriti J, Hassan MG, Leach M. Demand response experience in Europe:
Policies, programs. Energy 2009;35(4):157583.
[32] Abdollahi A, Parsa Moghaddam M, Rashidinejad M, Sheikh-El-Eslami M.
Investigation of economic and environmental-driven demand response
measures incorporating UC. IEEE Transactions on Smart Grid 2012;3(1):
1225.
[33] Ilic MD, Xie L, Joo J. Efcient coordination of wind power and price-responsive
demandPart II: Case studies. IEEE Transactions on Power Systems 2011;26(4):
188593.
[34] Papavasiliou A, Oren SS. Coupling wind generators with deferrable loads.
Energy 2030 Conference 2008:178.

View publication stats

[35] Roscoe AJ, Ault G. Supporting high penetrations of renewable generation via
implementation of real-time electricity pricing and demand response. IET
Renewable Power Generation 2010;4(4):36982.
[36] Khodaei A, Shahidehpour M, Bahramirad Sh. SCUCwith hourly demand
response considering inter-temporal load characteristics. IEEE Transactions
on Smart Grid 2011;2(3):56472.
[37] Conejo AJ, Morales JM, Baringo L. Real-time demand response model. IEEE
Transactions on Smart Grid 2010;1(3):23643.
[38] Bertsimas D, Sym M. Robust discrete optimization and network ows.
Mathematical Programming 2003;98(1-3):4971.
[39] Youse Sh, Parsa Moghaddam M, Johari Majd V. Optimal real time pricing in
an agent-based retail market using a comprehensive demand response
model. Energy 2011;36(9):571627.
[40] Aalami HA, Parsa Moghaddam M, Youse GR. Demand response modeling
considering interruptible/curtailable loads and capacity market programs.
Applied Energy 2010;87(1):24350.
[41] Aalami HA, Parsa Moghaddam M, Youse GR. Modeling and prioritizing
demand response programs in power markets. Electric Power Systems
Research 2010;80(4):42635.
[42] Grattieri W. Demand side management priorities in Italy. Load Management
and Demand Response Programs 2006 London.
[43] Tembleque SL, Arriaga PI, Pardo M. Electric load management, report for
Fundacion Alternativas. Madrid 2005.
[44] Hu Z, Moskovitz D, Zhao J. Demand-side management in Chinas restructured
power industry. Energy Sector Management and Assistance Program 2005
World Bank.
[45] Kwag H, Kim J. Optimal combined scheduling of generation and demand
response with demand resource constraints. Applied Energy 2012;96:16170.
[46] Moura PS, bal AT, de Almeida. Multi-objective optimization of a mixed
renewable system with demand-side management. Renewable and Sustainable Energy Reviews 2010;14(5):14618.
[47] Banosa R, Manzano-Agugliarob F, Montoyab FG, Gila C, Alcaydeb A, Gomezc J.
Optimization methods applied to renewable and sustainable energy: A
review. Renewable and Sustainable Energy Reviews 2011;15(4) 1753176.
[48] Potential role of demand response resources in maintaining grid stability and
integrating variable renewable energy under californias 33 percent renewable portfolio standard. Californias demand response measurement and
evaluation committee (DRMEC). July 20, 2012.