ArticlePDF Available

Quantitative Evaluation of Data Centers’ Participation in Demand Side Management

Authors:

Abstract

In recent years, the rapid increase in the number of internet users and widespread usage of internet applications have obliged large servers and networking equipment to manage large data stack and optimize the instantaneous transmission of digital information. The COVID-19 Pandemic has also caused an increase in data exchanges and digital information generation. In order to manage large-scale data, there is a need for gigantic data centers (DCs) which are tremendous energy consumers and have relatively flexible loads that are easier to control by means of shifting in time and space. Therefore, DCs can be regarded as dispatchable loads and are considered good candidates for participating in demand side management (DSM) programs for power curve smoothing and compensation of power fluctuation in electrical power systems. In this paper, the question of why DCs should participate in DSM has been investigated rather than the technical methods used in DSM. The amount of DCs’ participation energy is used by peak shaving/shifting method for power curve smoothing using actual data. The possible environmental and financial effects of it for Turkey and all the world have been carried out. The study results show that DCs’ participation in DSM for Turkey decreases peak load by up to 2.18%, defers up to 34% of the installed power plants launched in 2019, and improves load and loss factors by up to 2.2% and 4.3% respectively. Additionally, global DC’s participation in DSM decreases the peak point by up to 0.77% and reduces CO2 emission by 0.03%.
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2021.3052204, IEEE
Access
VOLUME XX, 2017 1
Date of publication xxxx 00, 0000, date of current version xxxx 00, 0000.
Digital Object Identifier 10.1109/ACCESS.2017.Doi Number
Quantitative Evaluation of Data Centers’
Participation in Demand Side Management
Mehmet Türker Takcı1, Tuba Gözel1, and Mehmet Hakan Hocaoğlu1
1 Department of Electronics Engineering, Gebze Technical University, Kocaeli, TURKEY
Corresponding author: Mehmet Türker TAKCI (turkertakci@gtu.edu.tr).
This work was supported in part by the GREENDC project from the European Union's Horizon 2020 research and innovation program
under the Marie Skłodowska-Curie grant agreement No 734273.
ABSTRACT In recent years, the rapid increase in the number of internet users and widespread usage of
internet applications have obliged large servers and networking equipment to manage large data stack and
optimize the instantaneous transmission of digital information. The COVID-19 Pandemic has also caused an
increase in data exchanges and digital information generation. In order to manage large-scale data, there is a
need for gigantic data centers (DCs) which are tremendous energy consumers and have relatively flexible
loads that are easier to control by means of shifting in time and space. Therefore, DCs can be regarded as
dispatchable loads and are considered good candidates for participating in demand side management (DSM)
programs for power curve smoothing and compensation of power fluctuation in electrical power systems. In
this paper, the question of why DCs should participate in DSM has been investigated rather than the technical
methods used in DSM. The amount of DCs’ participation energy is used by peak shaving/shifting method
for power curve smoothing using actual data. The possible environmental and financial effects of it for Turkey
and all the world have been carried out. The study results show that DCs’ participation in DSM for Turkey
decreases peak load by up to 2.18%, defers up to 34% of the installed power plants launched in 2019, and
improves load and loss factors by up to 2.2% and 4.3% respectively. Additionally, global DC’s participation
in DSM decreases the peak point by up to 0.77% and reduces 𝐶𝑂 emission by 0.03%.
INDEX TERMS Demand Side Management, Electricity Market, Data Center, Energy Efficiency, Peak
Shaving, 𝐶𝑂 Emission Reduction.
I. INTRODUCTION
Over the past twenty years, demand for internet service
providers and cloud services, use of information and
communication devices have reached a significant level. Thus,
the size of generated and processed data has been dramatically
increased. In 2018, the number of global internet users was 3.9
billion and it is projected to reach 5.3 billion users by 2023 [1].
Also, the number of devices connected to the internet has been
estimated to be 29.3 billion by 2023 [1]. According to the
white paper of Seagate [2], the amount of generated digital
information was 33 zettabytes (ZB) in 2018 and it is projected
to reach 175 ZB by 2025. These large data are needed to be
stored and processed in data centers (DCs) which include a
large number of high-performance servers and have the ability
to remote access.
Digital transformation has already been taking place for
many years, yet the COVID-19 pandemic exacerbates it
dramatically. Especially with most companies enabling
employers to work from home, the demand for internet and
online video conferencing software such as Google Meet,
Zoom, Microsoft Teams has skyrocketed. Isolation duration
has also affected the habits of people at home [3]. The ratio of
video call communication with friends and relatives, the time
spent on social media, and the usage of streaming audio and
video services like Netflix, HBO Now, Spotify have increased
so quickly. A survey which has been performed with 2200
adults between March 24 to March 26, 2020, in the U.S.,
shows that 37% of adults preferred to do online shopping more
than before the Pandemic, 41% of adults and 57% of the age
18-29 preferred streaming movies more than before and the
18% of the age 18-29 used FaceTime or Zoom for the first time
with the pandemic [4]. According to [5], the demand for
Google Meet has reached 60% after the COVID-19 and the
mean daily usage is declared as 2 billion minutes per day.
Likewise, the usage rate of Zoom has 10 times increment
during the pandemic [6]. HBO Now observed a 40% increase
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2021.3052204, IEEE
Access
Author Name: Preparation of Papers for IEEE Access (February 2017)
2
in viewing numbers compared to the average for the last 4
weeks before March 14 [7]. All these sudden amendments
have also caused an increase in data transfer traffics an
unprecedented demand for data centers to maintain the
services.
Owing to the growing demand for computational and
storage infrastructures, the size and number of geographically
distributed DCs are drastically increased to provide reliable
low latency services to the customers [8], [9]. DCs consist of
a large number of servers, cooling units, large-scale storage
units, networking equipment, and power infrastructures.
Because of having a huge amount of hosted servers, associated
workload, and cooling units, the energy consumption of DCs
has attained gigantic numbers [10],[11]. According to The
United States (U.S.) Environmental Protection Agency (EPA)
report, the demand in DCs has been increasing by
approximately 12% per year [12]. According to [13], the total
electricity usage of all DCs in the world increased by 56%
between 2005 to 2010. EPA Report mentioned that DCs in the
U.S. consumed about 61 TWh which equal to 1.5% of total
U.S. electricity consumption in 2006 [12]. Moreover, the
worldwide energy consumption of DCs was 198 TWh in 2018
[14]. The reports published in 2013 [15], [16] mentioned that
the IT sector which includes DCs was consuming about 10%
of the world’s electricity generation.
Due to the huge power consumption characteristic of DCs,
different studies [17],[18], [19], [20], [21] about forecasting
and efficiently managing power consumption of DCs have
become more popular in the literature. Also, DCs can be
significant players for the power systems in terms of their
ability to become prosumer and they are good candidates for
demand-side management applications in order to compensate
demand fluctuations and power curve smoothing.
In classical power systems, demand fluctuation is an
important problem with regard to its effects on power loss,
voltage/frequency control, and stability of power systems. It is
generally compensated by dispatching reserve generation
resources using unit commitment techniques. But renewable-
based distributed generators cannot be used for unit
commitment due to their intermittent nature. In the power
system sense, they are non-dispatchable units and further
contribute power fluctuation problems due to the fluctuating
nature of weather conditions. These problems with widespread
usage of renewable energy sources are caused to make energy
management more important.
On the other hand, a tool called demand-side management
has been in place for a time with the development of smart grid
technologies. In demand side management systems, the loads
are dispatched instead of generations. DSM provides benefits
to consumers such as fulfilling required demand, decreasing
electricity bills, and improving lifestyle. DSM also causes
advantages to the supplier such as providing a more efficient
system and requiring less capital for generation, reducing costs
of generation, transmission, and distribution systems [22].
DSM aims to provide required changes on electricity
consumption demand in time through load control systems,
on-site energy storage packages, or encourage non-peak
electricity demand [23],[24]. In that sense, both big and small
DCs can be used as a participant for DSM systems in terms of
their flexible loads which are easier to control and shift in time
than the other big industrial customers. One of the goals of
demand side management for DCs is to efficiently manage
DCs’ energy consumption and generation resources, another
one is to contribute to smoothing demand fluctuations in
power systems. However, the focus of most researchers has
been generally on minimizing energy cost of DCs. The studies
related to DCs implemented with DSM mechanisms can be
examined in two subcategories: Time of Use Pricing and
Market Mechanism. Relevant studies have been summarized
in Table I.
A. TIME OF USE PRICING
Time of Use (ToU) pricing is based on adjustment electricity
price by the time of the day. There is an electricity tariff that
has higher prices at peak hours and lower prices at non-peak
hours. ToU pricing has long been used by most utilities in the
world [25].
B. MARKET MECHANISM
The market mechanism is based on determining electricity
prices according to the participants’ bids for the day-ahead or
intraday electricity market. Bidders can participate in the
market through their generation or consumption amount and
accept to be penalized if they don’t provide their commitment
[11],[26],[27]. According to the literature of DSM market
mechanism implementation to the DC ecosystem, there are
three techniques : (I) load shifting which is shifting energy
consumption in time, (II) load migration which is migrating
the load geographically, and (III) load shedding which is
temporarily enabling the sleep mode or shutting down the
servers to reduce energy consumption while they are in idle
condition [27],[28]. DCs’ workloads are generally divided into
two subcategories, delay-sensitive and delay-tolerant
workloads [9]. The delay-sensitive workload must serve
immediately without any delay. On the other hand, delay-
tolerant workloads can be postponed to another period and can
be used for load shifting technique which is based on shifting
load from the time with higher electricity prices to lower
prices. Delay-tolerant workloads can also be migrated to the
DCs at a different location which has lower electricity prices.
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2021.3052204, IEEE
Access
Author Name: Preparation of Papers for IEEE Access (February 2017)
3
TABLE I
The studies related to DCs implemented with DSM
Existing
Studies
Time of
Use
Pricing
Market Mechanism
Load
Shifting
Load
Migration
Load
Shedding
Objectives
[29]
Mahmoudi et al. proposed a Residential Energy Management system that aims to reduce
customers’ electricity bills and improve the utility load curve through integrating ICT equipment
with electric vehicles to the power grid.
[30]
The authors have performed ToU based optimization methods for DCs in order to reduce
operational expenditures and power consumption. The study shows that there is a reduction in the
electricity bill. However, propagation delays increase.
[20]
In [20], the authors aim to minimize energy cost of distributed data centers, which have wind
farms and solar panels, using deep learning based optimizer.
[31]
Kantarci et al. extended their previous work [30], by implementing heuristic solutions while
considering inter-DC traffic. Also, the load migration method has been used without migrating
30% of the original workload.
[32]
In this study, a workload shifting mechanism is established to smooth power fluctuations and
reducing power cost of a data center using renewable resources and UPS batteries of DC.
[33]
The authors propose peak shaving strategies include both load shifting and energy storage
alternatives in order to minimize electricity cost of DCs using the convex optimization model.
[34]
In this study, the authors propose a real-time model that controls HVAC, IT workload, and battery
energy storage systems for DC’s participation in demand response through the load shifting
method.
[35]
In this study, the authors developed a system that optimizes the memory of multiple VMs based
on the Xen balloon driver. This system shifts overloaded memory in a VM to another VM
automatically using a global-scheduling algorithm. The authors use load shifting method but do
not intend to participate in DSM
[36]
This study emphasizes that network traffic costs should be considered while the process of load
shifting via VM migration in a DC. The authors solve network-aware VM migration problem
using genetic algorithms and artificial bee colony methods. The authors use load shifting method
without participating in DSM.
[37]
AMMAR et al. propose an algorithm for load shifting via VM migration considering service level
agreement (SLA) violation in a DC. The proposed algorithm balances the consumption of
resources, minimizes resource wastage, SLA violation, and migration cost.
[38]
In this study, an energy management system is proposed considering delay tolerant workload and
maintaining the quality of service to minimize the energy cost. Authors compare different
working scenarios to determine cost effective solutions including load shifting and shedding
methods.
[39]
This study is related to a type of DCs that provides server rental services to the tenants to
participate in emergency demand response by Nash Bargaining Theory without use back up
energy resources. Using an incentive approach, tenants shift or shed their workloads during
emergency situations.
[40]
The authors propose a new method called “usage-based pricing with monetary reward” based on
shifting/shedding tenants’ workloads to reduce energy cost of DCs using decomposition based
algorithm.
[41] In this study, Fridgen et al. present an economic analysis to minimize DC’s electricity cost by
adjusting the optimum load migration schedule.
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2021.3052204, IEEE
Access
Author Name: Preparation of Papers for IEEE Access (February 2017)
4
TABLE I. - CONT
The studies related to DCs implemented with DSM – Cont.
Existing
Studies
Time of
Use
Pricing
Market Mechanism
Load
Shifting
Load
Migration
Load
Shedding
Objectives
[42]
By extending Ref [41]’s approach, Thimmel et al. used load migration method for balancing
demand between different electricity markets via geographically distributed DCs. Their model
achieved to compensate for balancing power demand without a conventional balancing
mechanism.
[43]
In this study, the authors focus on minimizing DC’s electricity cost via load migration method
between multiple cloud providers using coalitional game theory instead of examining the benefits
for the utility side.
[44]
The authors aim to eliminate congestion between cross-region grids using load migration between
geo-distributed data centers via congestion management method. The proposed mechanism
minimizes congestion costs and provides revenue to data centers corresponding to the migrated
load.
[45]
The authors aim to minimize long term operation cost and solve real time energy management
problem of geographically distributed data centers. The lyapunov optimization method is used to
solve optimum energy management problem considering workload shedding and migration.
[46]
The authors developed an online carbon-aware incentive mechanism based on gathering of
tenants' bids correspond to the amount of reduced load in order to minimize the carbon footprint
of DCs and operator’s cost. An online optimizer determines the winning bid that is used to
rewarding tenants.
[47]
The authors devise an incentive-based scheme which enables DCs to participate in emergency
demand response via servers workloads and office buildings’ cooling load instead of backup
generators. During the emergency period, the setpoints of air conditioners are reduced, and the
server workloads are dropped.
C. CONTRIBUTIONS
As seen from the compact literature review above, most
researchers are focused on the technical and mathematical
processes of how the DCs can participate in DSM with the
scope of providing benefits (minimizing electricity cost, etc.)
for solely DCs side instead of the utility. The motivation of
this paper is to draw attention to the prosumer characteristic of
data centers and to raise awareness of its various contributions
to both the power systems and data centers. This paper
investigates the impact of DCs’ participation in DSM on
power systems in terms of economic and environmental
effects, so far lacking in the scientific literature. The peak point
shaving method has been used via DCs’ participation energy
involved in DSM. The possible results such as improving
energy saving capacity of the grid, decreasing power loss,
benefits due to avoidance of building new power plants are
quantitatively analyzed using actual data of Turkey's power
consumption as a case study. Additionally, the economic and
𝐶𝑂 emission impacts of reducing power losses in
transmission systems have been analyzed. The main
contributions of this paper are summarized as follows:
The consequences of peak-point shaving of Turkey
load duration curve have been analyzed. The peak point of
load curve could be reduced by up to 2.18% relative to the
initial case. Correspondingly, the load factor could be
improved by 2.2% while the improvement of loss factor could
be reached up to 4.3%. Thus, up to a 4.13% reduction in
power transmission loss could be achieved.
The effects of global DCs’ participation in DSM for
all around the world have been carried out. The global peak
point of the world could be reduced by up to 0.77%. The
worldwide transmission losses could be reduced by up to
1.5%. Consequently, 0.03% of the world’s total 𝐶𝑂 emission,
equal to the 𝐶𝑂 emission of Senegal, could be reduced.
The global DCs’ participation in DSM could cause
to defer 12.5% of the worldwide installed power plants put
into service in 2018
The remaining part of this paper is structured as follows:
Section II gives information about the status and importance
of DCs in terms of its prosumer characteristics. Additionally,
the main definition and assumption of DCs’ participation
process in DSM have been explained in Section II. The
economic and environmental results of DCs’ participation in
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2021.3052204, IEEE
Access
Author Name: Preparation of Papers for IEEE Access (February 2017)
5
DSM for Turkey and all around the world have been clarified
in Section III. The conclusions have been drawn in Section IV.
II. DATA CENTERS AS A PROSUMER
Besides the global reasons for the growing demand in DCs for
all the countries around the world, the young population is
another reason for the increased requirement of DCs for
Turkey. Broadband internet customers of Turkey, which were
around six million in 2008, exceeded 68 million in 2017 [48].
In the third quarter of 2017, total broadband internet usage
reached 3.5 exabytes. Approximately 91% of this usage was
data download and 9% was data upload [48]. Furthermore, in
2011, Turkey ranked first in the world with a 60% growth rate
in the white area where the servers are located in and ranked
second with a 74% increase in the investment rate of DCs
[49],[50]. The global growth rate of the DCs market is
expected to increase by 15% between 2014 to 2021 while it is
30% for Turkey [50]. All these data show that global DC
needs, and growth rates have increased significantly, while
Turkey has more than the world average. Furthermore, the
amount of consumed energy by DCs will rise enormously with
the development of DC business. Thus, DCs will be good
players in terms of power systems. Moreover, along with the
law enacted in 2018 for Turkey [51], the utility supports large
energy-consuming industries to participate in DSM programs
through ancillary services. Therefore, DCs can be significant
participants for DSM since its ability to consume too much
energy and its prosumer characteristic [18]. That is one of the
reasons why this study was carried out to examine the impact
of the idea of using DCs as prosumers. Furthermore, based on
the analysis conducted for Turkey, the global effects of DCs’
participation in DSM has been carried out. The following
subsection presents analyses of Turkey's electricity
consumption, main definitions, and process of load curve
smoothing.
A. THE ANALYSIS OF TURKEY ELECTRICITY
CONSUMPTION AND DEFINITIONS OF LOAD AND
LOSS FACTORS
The real hourly power consumption data of Turkey for 2019
have been taken from the webpage of Turkey Load Dispatch
Information System (YTBS) [52] via a software developed to
get the data automatically. All analyses and the peak shaving
method are carried out in MATLAB environment.
The actual hourly power consumption has been depicted in
Fig. 1- (a) [52] and the load duration curve, which is obtained
by sorting power consumption values from maximum to
minimum, has been shown in Fig. 1-(b). The integrated areas
in both graphs are the same and represent the annual energy
consumption.
In order to measure the efficiency of electrical transmission
systems, generally, two different indices are used: load factor
and loss factor. The load factor is defined as follows [53]:
Load Factor =
1
8760 *
Pt
8760
t=1
Pmax (1)
The number of total hours of a year is 8760. 𝑃
 represents
the maximum power consumption value over a year and 𝑃
indicates hourly power consumption. According to Gustafson
[54], the loss factor can be approximated as:
Loss Factor =
(
Load Factor
)
1.912 (2)
The loss factor and the load factor values vary from 0 to 1.
The system efficiency rises up if these factors approach 1.
FIGURE 1. Total Power Consumption of Turkey for 2019: (a) Hourly Load Curve; (b) Load Duration Curve
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2021.3052204, IEEE
Access
Author Name: Preparation of Papers for IEEE Access (February 2017)
6
The changes in the load curve affect the loss factor much
more than the load factor because of the nonlinear
characteristic of losses. The relationship between bus power to
total power loss is defined in [55] and shown in (3).
Ploss
=
Rj(ajPi2+bjPi+cj)
nb
j=1 (3)
The impedance of 𝑗 line is indicated by 𝑅, “nb” is the
branch number and 𝑎, 𝑏, 𝑐 represent coefficients. The power
of 𝑖 bus is represented by 𝑃. As can be seen from (3), there
is a nonlinear relationship between loss and power load [55],
[56]. In electric power transmission systems, the load curve is
generally expected as smooth as possible in terms of system
stability, lower loss, and efficient usage of installed power
plants. Therefore, the difference between maximum and
minimum consumption values of load curve is desirable to be
small. Additionally, two different load curves can have the
same total consumed energy, but the load factor and loss factor
can be quite different. For example, Fig. 2 presents two
different curves which have the same energy consumption
value. One of them is the normalized load duration curve of
Turkey which is obtained by dividing actual consumption data
to its maximum value drawn in the black solid line. The other
is the desired load curve of Turkey drawn in a red dotted line.
The energy amount of normalized load duration curve is equal
to the sum of 𝐴 and 𝐴. Desired load curve is obtained by
shifting the energy amount of 𝐴 after the “x” point of the real
load duration curve. The maximum point of the desired load
curve is also the mean value of the normalized load duration
curve and this point is calculated as the peak point that can be
shaved as much as possible. The energy amount of desired
load curve is equal to the sum of 𝐴 and 𝐴. The integrated
areas “𝐴” and 𝐴”, which also means the amount of energy,
are equal. However, both curves have different load and loss
factors. While load and loss factors for the real normalized
load duration curve are 0.7544 and 0.5834, both factors are
1.00 for the desired load curve that means the system has
perfect efficiency. Thus, it is seen that “peak point
shaving/shifting” extremely affects the system efficiency even
if the total consumed energy does not change.
In this study, the relationship between peak point
shaving/shifting to the load and loss factors has been used in
order to improve electrical system efficiency, reduce power
losses in transmission systems, obtain monetary saving via
power curve smoothing, and improve load and loss factors
without changing the amount of total energy consumption.
The overall approach has been demonstrated in Fig. 3.
B. THE PARTICIPATION OF DATA CENTERS IN DSM
In this section, all analyses have been done using Turkey’s
energy consumption data for the year 2019 which was 290.4
TWh [52]. The total energy consumption rate of the DCs for
2019 has been considered as two percent of Turkey’s total
energy consumption and is calculated as 5.8 TWh.
In order to examine the effects of the energy consumed by
DCs participating in DSM at different rates, seven different
ratios ranging between 10 to 70 percent of hourly DCs energy
consumption is assumed to be used in DSM and the analyses
are carried out for each case. While determining the maximum
participating rate (70%), it was taken into consideration that a
data center cannot shift the entire workload instantaneously
and the fact that the backup power capacity of a DC should
have at least twice the average power consumption of a DC
[57]. Since the secondary energy resource amount and the
shiftable workload ratio of each data center are different from
each other, various participation rates allow analyzing the
results of DCs’ participation in DSM at different levels.
FIGURE 2. Desired and Real Normalized Load Duration Curves
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2021.3052204, IEEE
Access
Author Name: Preparation of Papers for IEEE Access (February 2017)
7
FIGURE 3. The schematic view of overall approach.
The peak point of load duration curve has been
shaved/shifted as much as the participated DCs’ energy in
DSM. This process has been depicted in Fig. 4 and the used
mathematical formulation is defined in (4), (5), and (6). DCs’
power consumption has been considered as uniformly
distributed in the analyses due to the lack of power
consumption profile data of individual DCs. Thus, the hourly
power consumption of DCs, 𝐷𝐶𝑠, is calculated by dividing
the total energy consumption of DCs (5.8 TWh) by the total
hours (8760). The hourly total power consumption of all DCs
is calculated as 663MW (𝐷𝐶𝑠). The amount of DCs’
hourly power involved in DSM is represented as
𝐷𝑆𝑀ℎ𝑜𝑢𝑟𝑙𝑦𝑃 and calculated as in (4).
DSMhourlyP = DCsPow * DSMratio (4)
𝐷𝑆𝑀 is the percentage ratio of DCs’ participation in
DSM per hour. 𝐷𝑆𝑀ℎ𝑜𝑢𝑟𝑙𝑦𝑃 is calculated for each value of
𝐷𝑆𝑀 which are 10%, 20%, 30%, 40%, 50%, 60%, and
70%. In Fig. 4, the load duration curve of Turkey is defined as
a time-dependent function, 𝑓(𝑡), where the initial peak point
is “1”. During peak shaving procedure, three unknown
parameters should be calculated for each 𝐷𝑆𝑀. One of
them is "𝑥" indicating how many hours DCs participated in
DSM, the other one is 𝐸 which is indicating the DCs’
energy participated in DSM for each 𝐷𝑆𝑀 also equal to
the energy calculated between the initial peak point to the new
peak point and the last one is the new peak point, 𝑃
,
determined after shaving/shifting the 𝐸 .
In order to determine 𝑃
, firstly, the value of "𝑥" should
be calculated because the 𝑃
 can be obtained using the
equation of
Pnew=f(x)
for each 𝐷𝑆𝑀. The total time of
DCs’ participated in DSM, "𝑥", can be determined using 𝐸
that can be calculated using two different methods.
One of them is multiplying the amount of DCs’ hourly
power involved in DSM (𝐷𝑆𝑀ℎ𝑜𝑢𝑟𝑙𝑦𝑃) by the value of "𝑥"
as shown in (5). The other one is the mathematical calculation
of the integration between the initial peak point to the new
peak point of load duration curve as shown in (6).
EDSM
=
DSMhourlyP
*
x (5)
EDSM
=[ f(
t
) -
f(x) ]dt
x
0 (6)
The value of x is obtained by solving both (5) and (6) which
are used to calculate 𝐸 as 𝑥 dependent. Then, substituting
the value of "𝑥" at the 𝑓(𝑡), 𝑃
 can be calculated. These
steps are repeated for each 𝐷𝑆𝑀. After that, the amount
of 𝐸 is added to the end of the current curve and the new
normalized load duration curve has been created. The new
curves are sketched in Fig. 5. The calculated 𝑃
 for each
percentage of 𝐷𝑆𝑀 can be seen in Fig. 5.
FIGURE 4. The Process of Determining Peak Shaving Point
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2021.3052204, IEEE
Access
Author Name: Preparation of Papers for IEEE Access (February 2017)
8
FIGURE 5. The Normalized Load Duration Curves of Turkey After Peak Shaving for Different Percentages of 𝑫𝑺𝑴𝒓𝒂𝒕𝒊𝒐
III. RESULTS OF ANALYSES
In this section, different effects of peak point shaving of load
duration curve are explained. The main results of DCs’
participation in DSM have been discussed in Subsection A.
The cost saving calculations, which is occurred by deferring
the establishment of new power plants, are described in
Subsection B. The 𝐶𝑂 reduction happened as a result of
avoiding energy generation as much as the amount of
prevented power losses and the economic impact of it have
been clarified in Subsection C. The results of global DCs
participation in DSM for all around the world have been
demonstrated in Subsection D.
A. THE RESULTS OF DCS’ PARTICIPATION IN DSM
The information about how the load and loss factors change
depends on each 𝐷𝑆𝑀, the amount of reduction in peak
point of the load curve and power losses can be interpreted
from Table II. In order to reveal the effects of different
percentages of 𝐷𝑆𝑀, the calculations have been made for
each ratio between 10% to 70% and without DSM. The
corresponding results are shown in Table II. The minimum
and maximum total annual hours of DCs’ participation in
DSM range between 12 to 93 hours and the corresponding
energy amount range between 796 MWh to 43,161 MWh.
The peak point of load curve for 2019 was 43,948MW and
could be reduced to 43,821MW or 42,990MW using the
minimum and maximum 𝐷𝑆𝑀. The peak point of load
duration curve could be decreased between the amount of
127MW to 958MW which correspond to a minimum of 0.28
percent and a maximum of 2.18 percent reduction in the peak
load of Turkey. This reduction has improved Turkey's load
and loss factors even the total consumed energy has not
changed. When the peak point of load curve is reduced from
43,948MW to 42,990MW, the load factor of Turkey could be
improved from 0.7544 to 0.7712 with a 2.2 % increase and the
loss factor could be improved from 0.5834 to 0.6085 with an
increment of 4.3%. This means that the improvement in the
loss factor is almost two times greater than as much as the
improvement in the load factor for the same 𝐷𝑆𝑀. As
explained in the previous section, these results also show that
the load curve smoothing affects the load and loss factors in
different amounts. Loss factor improvement causes a
reduction in power losses of the transmission systems. The
information of power losses rate for 2019 had not been
published yet while this study was being prepared. Thus, the
average rate of power losses between 2006 to 2017 in Turkey,
which is 2.34% of total consumed energy [58], is used to
calculate the power losses for 2019. The total amount of losses
calculated as 6.79 TWh and the corresponding loss factor was
0.5834. Depending on the loss factor improvement, the
minimum and maximum amount of power loss reduction are
calculated as 37 GWh and 280 GWh which equal to 0.54%
and 4.13% reduction, respectively. The corresponding
savings, which are calculated using the conversion rate of
$87,000 per GWh [59], vary between $3.2 million to $24.3
million as seen in Table II. Additionally, improving the loss
factor causes to increase in the carrying capacity of
transmission lines and prevents the construction of new
transmission lines.
B. COST SAVING CALCULATIONS CONSIDERING
DEFERRED NEW POWER PLANTS
In addition to cost saving from the reduction of the losses in
the transmission systems which are caused by DCs’
participation in DSM, a much larger amount of the savings can
occur due to the reduction in peak point of Turkey load
duration curve.
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2021.3052204, IEEE
Access
Author Name: Preparation of Papers for IEEE Access (February 2017)
9
TABLE II
The Results of DCs’ Participation in DSM
DSM

Total Hours
of DCs’
Participation
in a Year (h)
DCs’
Participation
Amount (MWh)
Peak Point of
Load Curve
(MW)
Reduction in
Peak Point
(MW)
Load
Factor
Loss
Factor
Reduction in Power
Transmission Loss
(GWh)
Savings from
Loss Reduction
Without
DSM - - 43,948 - 0.7544 0.5834 - -
%10 12 796 43,821 127 0.7566 0.5866 37 $3.2 million
%20 24 3,182 43,687 261 0.7589 0.5900 76 $6.6 million
%30 36 7,160 43,557 391 0.7612 0.5935 116 $10.1 million
%40 49 12,995 43,421 527 0.7636 0.5970 155 $13.5 million
%50 63 20,885 43,278 670 0.7661 0.6008 197 $17.1 million
%60 77 30,631 43,141 807 0.7685 0.6044 236 $20.5 million
%70 93 43,161 42,990 958 0.7712 0.6085 280 $24.3 million
Reducing the peak point means that deferring the new power
plants which Turkey aims to build in order to meet growing
energy demand and peak point in each year. According to [60],
the installed power capacity of Turkey was 88,442 MW in
2018 and it was 91,060 MW in 2019 [61]. So, new power
plants with a size of 2,799 MW was put into service in 2019.
The distribution of new power plants by energy resource is
shown in Table III [58], [60], [61]. Natural gas includes LNG,
the energy resource of coal consists of hard coal, imported
coal, lignite, and asphaltite. Renewables comprise geothermal,
wind, solar, biomass. The liquid fuel, naphtha, and oil are
involved under the Fuel Oil category.
TABLE III
Distribution of Turkey Installed Power by Energy Resources and Years
Energy
Resources
Installed Power
Capacity (MW)
2018
Installed Power
Capacity (MW)
2019
Difference
2018-2019
(MW)
Natural Gas 25,909 26,186 277
Coal 19,742 20,284 542
Hydro 28,293 28,505 212
Renewables 14,005 15,773 1,768
Fuel Oil 493 312 -
Total 88,442 91,060 2,799
According to Table II, if the DCs had participated in DSM
for 2019, new power plants would not have required the
amount of installed power between 127MW to 958 MW which
are equal to the amount of decreased peak points of Turkey.
Thus, Turkey could have deferred the installed power plants
put into service in 2019 by 4.5% at the minimum and 34.3%
at the maximum. Therefore, Turkey could gain huge monetary
savings. The capital and operations and maintenance (O&M)
cost of power plants are calculated according to [62] and
shown in Table IV. Total cost calculations of avoided new
power plants are given in Table V.
New coal based power plants with the size of 542 MW
have been put into service in 2019 as shown in Table III.
Because the coal based power plants produce more pollution
than others, in this paper, building new coal power plants is
preferred to be deferred first, then the natural gas and
hydropower plants are preferred. Thus, the avoidable power
plants with the size of 127MW are chosen to be coal for 10%
of 𝐷𝑆𝑀. Turkey could have obtained $472 million due
to avoid building coal based installed power plants for the
minimum 𝐷𝑆𝑀. With the maximum 𝐷𝑆𝑀, Turkey
could have avoided building 542 MW of coal, 277 MW of
natural gas, and 139 MW of Hydro based installed power
plants as shown in Table V. Hereby, $3.276 billion could have
been saved.
TABLE IV
Unit Values of Capital and O&M Costs by Power Plant Resources
Power Plant
Resources Capital Cost O&M Cost
Coal 3,676 $/kW 40.58 $/kW-year
Natural Gas 1,810 $/kW 35.16 $/kW-year
Hydro 5,316 $/kW 29.86$/kW-year
TABLE V
Cost Calculation of Deferred New Power Plants
𝐷𝑆𝑀

Power
Plant
Resources
Deferred
Installed
Power (MW)
Capital
Cost
O&M
Cost
Total
Cost
10%
Coal
127
$467
million
$5
million
$472
million
70%
Coal 542 $2
billion
$22
million
$3.276
billion
Natural
Gas 277 $501
million
$10
million
Hydro 139 $739
million
$4
million
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2021.3052204, IEEE
Access
Author Name: Preparation of Papers for IEEE Access (February 2017)
10
C. THE EFFECTS OF DCS’ PARTICIPATION IN DSM ON
𝑪𝑶𝟐 EMISSIONS
The rapid development of technology in the 21st Century
provides many benefits that will make people’s life easier. At
the same time, it causes irreversible damage to nature in many
ways. Currently, one of the most crucial issues for humanity
is global warming which is caused by raised greenhouse gas
emissions due to the increasing consumption of fossil fuels
and the destruction of forests. Among the greenhouse gases,
carbon dioxide (𝐶𝑂) has the biggest responsibility with a
72% ratio of the greenhouse effect. Methane (𝐶𝐻) follows it
with a 19% ratio after that nitrous oxide (𝑁𝑂) and fluorinated
gases (F-gases) share the responsibility by 6% and 3%,
respectively [63].
DCs’ participation in DSM can contribute to preventing
𝐶𝑂 emissions thanks to reducing the power losses in
transmission systems. According to conversion rates in [64],
[65], coal based power plants produce 908 t𝐶𝑂 per GWh,
natural gas power plants emit 400 t𝐶𝑂 per GWh and hydro
power plants produce 21 t𝐶𝑂 per GWh. If the energy between
37 GWh to 280 GWh, which correspond to power loss
reduction at the minimum and maximum rate as seen in Table
II, had not generated, the 𝐶𝑂 emission could have decreased
as between 33,596 t𝐶𝑂 to 254,240 t𝐶𝑂. Coal based power
plants have been used for the calculation due to their more
polluting characteristic than other power plants Since the
negative effects of carbon emissions on health, water, food
production, and environment, economic loss has been also
occurred. According to [66], per ton of 𝐶𝑂 can cause an
economic loss of up to $100. Hereby, DCs’ participation in
DSM could have additionally made a profit for Turkey
between $3.3 million to $25.4 million in 2019 by reducing the
𝐶𝑂 emission.
D. GLOBAL EFFECTS OF DCS’ PARTICIPATION IN DSM
It can be inferred from the above results that DCs in all around
the world can provide more gains if they can participate in
DSM globally in terms of their flexible and huge power
consumption which was 198 TWh in 2018 [14].As a result of
DCs’ participation in DSM at certain times for peak
shaving/shifting, there may not need to build new power
plants, especially for countries that have large populations and
mainly generates electricity from coal. Therefore, both
monetary and environmental gains could have reached a
significant amount.
Since global power consumption and generation data for
2019 has not been published yet, in this section, all
calculations and analysis for environmental and economic
effects of global DCs’ participation in DSM have been carried
out for the year 2018. Additionally, the world load duration
curve has been assumed to be the same as Turkey’s profile,
which is shown in Fig. 3, because there is no hourly power
consumption data for all the world. The only known value is
the total energy consumption of the world that was 26,615
TWh in 2018 [67]. To reveal the effects of global DCs’
participation in DSM, similar steps and methods in Section II
are used. The peak point of the world load duration curve has
been shaved/shifted as much as the energy in different
participation ratios that vary %10 to %70 of DCs’ hourly
power consumption. The amount of energy involved in
DSM, 𝐸, is shifted to the end of the curve for each
𝐷𝑆𝑀, thereby new normalized load duration curves of
the world have been created as shown in Fig. 6. The calculated
new peak points of load curves for each percentage of
𝐷𝑆𝑀 can be seen in Fig. 6.
FIGURE 6. The Normalized Load Duration Curves of the World After Peak Shaving for Different Percentages of 𝑫𝑺𝑴𝒓𝒂𝒕𝒊𝒐
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2021.3052204, IEEE
Access
Author Name: Preparation of Papers for IEEE Access (February 2017)
11
In order to calculate required real values such as DC’s
participation amount, the amount of decreased peak point, etc.,
the total energy amount (𝐸) has been calculated in per unit
by integral of the load duration curve without DSM in Fig. 6.
Then, (7) has been used to calculate the base value.
Ereal
=
Epu*
Ebase (7)
𝐸 represents the calculated energy amount in per unit
which is 6,609PU and 𝐸 is the real energy consumption of
the world which is 26,615 TWh [67]. According to the (7), the
base value (𝐸 ) is calculated as 4,027GWh. All other
requirement values are calculated using the base value. Similar
to Section II, the main purpose is to decrease the peak point of
the world duration curve by peak shaving. Thus, (4), (5), and
(6) are used to calculate a new peak point after peak shaving
likewise in Section II. Accordingly, 𝐷𝐶𝑠 is calculated as
22.6GW. The results of peak shaving for each 𝐷𝑆𝑀 are
summarized in Table VI. The peak point of the world power
consumption is calculated as 4,027 GW for 2018. This amount
could be reduced to 4,024 GW, which equals a decrease of
0.08%, if the DCs in all around the world were participating in
DSM with 10 percent of its hourly power consumption. The
maximum reduction in peak point, which is 30.8 GW and
equals a decrease of 0.77%, could be obtained with 70 percent
of 𝐷𝑆𝑀.
Decreasing peak point leads to an improvement in load and
loss factors. Based on the World Bank data [68], the world
average electric power transmission and distribution (T&D)
losses between 2010 to 2014 is %8.23 of the world's total
electricity generation. According to The Regulatory
Assistance Project (RAP) report [69], the rate of T&D losses
varies from 6% to 11%. Additionally, the National
Association of Clean Air Agencies (NACAA) report [70] says
that transmission line losses range between 2% to 5%. Since
T&D losses include non-technical and transformers losses, we
focus on only transmission losses related to loss factor for
calculating worldwide power loss reduction that happened due
to peak shaving. Consequently, the amount of worldwide
transmission losses for the year 2018 have been considered as
3% of worldwide energy generation. So, the worldwide
transmission losses are calculated as 798.45 TWh that
corresponds to the loss factor value of 0.5834. After peak
shaving for each 𝐷𝑆𝑀, minimum and maximum amounts
of loss reduction have been calculated as 1,230GWh and
11,865 GWh which are corresponded to the loss factors of
0.5843 and 0.5922 respectively. This means that worldwide
energy generation could be reduced between 1,230GWh to
11,865 GWh via global DCs’ Participation in DSM for 2018.
These amounts are equal to 0.15% and 1.5% of worldwide
transmission losses of 2018. Thus, the global 𝐶𝑂 emission
could be reduced between 1.2 Mt of 𝐶𝑂 to 10.78 Mt of
𝐶𝑂 and it is caused to monetary savings between $120
million to $1.078 billion due to negative effects of 𝐶𝑂 on
nature. These calculations are determined based on the
conversion rates explained in Section III-C. The maximum
avoidable 𝐶𝑂 amount is equal to the 𝐶𝑂 emission of
Senegal [71], which is also equal to 0.03 percent of the world
total 𝐶𝑂 emission (37,887Mt) for 2018.
The peak point reduction could have also led to deferring
build new power plants. In order to decide which type of
power plants could be deferred, the amount of world total
installed power capacity [72] by resource between 2017 to
2018 is shown in Table VII. The Renewables category
includes geothermal, wind, solar, biomass, nuclear, and
battery. As can be seen in Table VI, if the DCs all around the
world had participated in DSM in 2018, new power plants
would not have required in the amount of installed power
between 3.2 GW to 30.8 GW corresponding to the minimum
and maximum 𝐷𝑆𝑀 . These amounts are equal to 1.3
percent and 12.5 percent of the total global capacity of
installed power plants launched in 2018. In other words, 1.3%
to 12.5% of global installed power plants could be deferred if
the DCs participated in DSM globally. When the power plants
to be built are allocated to energy resources, coal, and natural
gas based power plants primally preferred to be deferred in
terms of their more polluting nature than others.
TABLE VI
The Results of Global DCs’ Participation in DSM
𝐷𝑆𝑀

Total Hours of
DCs’ Participation
in a Year (h)
DCs’ Participation
Amount (GWh)
Peak Point of
Load Curve
(GW)
Reduction in
Peak Point
(GW)
Load Factor Loss Factor
Reduction in Power
Transmission Loss
(GWh)
Without
DSM - 0 4,027 0 0.7544 0.5834 -
%10 4 9.1 4,024 3.2 0.7550 0.5843 1,230
%20 9 40.7 4,019 8.4 0.7560 0.5858 3,271
%30 13 88.2 4,015 12.6 0.7568 0.5870 4,897
%40 17 153.7 4,010 16.7 0.7576 0.5882 6,516
%50 22 248.6 4,005 21.8 0.7585 0.5895 8,262
%60 26 352.6 4,001 25.8 0.7593 0.5907 9,867
%70 31 490.5 3,996 30.8 0.7603 0.5922 11,865
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2021.3052204, IEEE
Access
Author Name: Preparation of Papers for IEEE Access (February 2017)
12
Considering the values in Table VII, the amount of
deferring coal power plants is determined as 3.2GW for 10%
𝐷𝑆𝑀 while it is determined as 8GW for 70% 𝐷𝑆𝑀.
The rest amount (22.8GW) is decided to be as natural gas
based power plants for 70% 𝐷𝑆𝑀. Therefore, huge
monetary savings and environmental benefits could be
obtained.
TABLE VII
Distribution of World Installed Power by Energy Resources and Years
Energy
Resources
Installed Power
Capacity (GW)
2017
Installed Power
Capacity (GW)
2018
The Capacity of
Installed Power
Plants
Established in
2018 (GW)
Natural Gas 1,693 1,745 52
Coal 2,071 2,079 8
Hydro 1,269 1,290 21
Renewables
1,488 1,654 166
Oil 450 450 0
Total 6,971 7,218 247
The avoided cost has been calculated using capital and
O&M costs of power plants defined in Table IV [62], and the
results are shown in Table VIII.
TABLE VIII
Cost Calculation of Deferred New Power Plants for all the World
DSM

Power
Plant
Resources
Deferred
Installed
Power
(GW)
Capital
Cost
O&M
Cost
Total
Cost
10% Coal 3.2 $11.7
billion
$129.8
million
$11.83
billion
70%
Coal 8 $29.4
billion
$324.6
million $71.72
billion
Natural
Gas 22.8 $41.2
billion
$801.6
million
As a result, if the DCs all around the world had participated
in DSM for the year 2018, the minimum monetary saving
could have been $11.95 billion for minimum 𝐷𝑆𝑀 and
$72.8 billion with 70% 𝐷𝑆𝑀 in maximum by deferring
the establishment of new power plants and preventing the
harmful effects of 𝐶𝑂 emission.
IV. CONCLUSIONS
In this paper, the reasons and the significance of using DCs as
participants of DSM in classical electrical systems have been
clarified. At first, the needs for DCs and effects of COVID-19
Pandemic on it have been described. Additionally, it is
explained why DCs can be used as a prosumer. Then, the
consequences of DCs’ participation in DSM for Turkey have
been examined with the actual data and corresponding benefits
are analyzed. The effects of DCs’ participation in DSM on
load and loss factor, power losses in the transmission systems,
avoided cost of new power plants planned to be established,
and reduction in 𝐶𝑂 emissions of Turkey have been
examined. Furthermore, all these analyses have been carried
out for global DCs’ participation in DSM for all around the
world as well.
This study exemplifies that the values of load and loss
factors for two different load curves might not be the same
even if they have the same energy consumption. This enables
us to reduce losses and increase monetary savings just by using
peak point shaving method without changing overall energy
consumption. It is clearly shown that the ratio of DCs’
participation in DSM affects the loss factor more than the load
factor. Seventy percent of 𝐷𝑆𝑀 improves the loss factor
two times greater than the improvement ratio of the load
factor. These advancements lead to a reduction in power losses
in transmission systems, an increase in the carrying capacity
of transmission lines, the prevention of constructing new
transmission lines, and new power plants.
The most obvious findings to emerge from this study are as
follows:
DCs’ participation in DSM could cause the peak
point of Turkey to decrease by between 0.28% to 2.18%.
DCs’ participation in DSM could cause a 2.2%
improvement in the load factor and 4.3% improvement in the
loss factor.
DCs’ participation in DSM could cause power
transmission loss to decrease by between 0.54% to 4.13%
which could provide to save between $3.2 million and $24.3
million.
DCs’ participation in DSM could cause 4.5% to
34.3% of the installed power plants put into service in 2019 to
defer. Thus, the cost of new power plants between $472
million to $3.276 billion could be avoided.
DCs’ participation in DSM might have provided to
reduce 254,240 t 𝐶𝑂 emissions of Turkey for maximum
𝐷𝑆𝑀. Corresponding cost saving could be $25.4 million
in terms of the harmful effects of 𝐶𝑂 on nature.
The results of global DCs’ participation in DSM for
all around the world have been examined using similar steps
and analyses carried out for Turkey.
Global DCs’ participation in DSM for all around the
world could cause the peak point of the world to decrease by
between 0.08% to 0.77%.
Global DCs’ participation in DSM for all around the
world could cause worldwide power transmission loss to
decrease by between 0.15% to 1.5%.
Global DCs’ participation in DSM for all around the
world 1.3% to 12.5% of the installed power plants put into
service in 2018 to defer. The monetary saving of $11.83
billion could be obtained due to the deferring establishment of
new power plants for 10% of 𝐷𝑆𝑀, while it could be
$71.72 billion for 70% of 𝐷𝑆𝑀 .
The global 𝐶𝑂 emission could be reduced by 1.2 Mt
of 𝐶𝑂 emission for minimum global DCs’ participation ratio
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2021.3052204, IEEE
Access
Author Name: Preparation of Papers for IEEE Access (February 2017)
13
(%10). The maximum amount could be 10.78 Mt of
𝐶𝑂 which is equal to the total carbon emission amount of
Senegal and 0.03% of the worldwide emission in 2018. Due
to the harmful effects of 𝐶𝑂 on nature, it would provide
financial gains between $120 million to $1.078 billion.
As a result, it is clearly shown that DCs are significant
players in terms of power systems operations by compensating
demand fluctuations and power curve smoothing. Besides the
mentioned benefits, DCs’ participation in DSM has great
potential to provide financial gains. This study encourages
similar studies by drawing attention to the importance of DCs
in terms of power systems by examining the participation of
DCs in DSM with a holistic view and highlighting its
contribution to power systems. The future work is planned to
examine the effects and contribution of small data centers’
participation in local DSM for both themselves and
distribution systems from a win-win perspective.
REFERENCES
[1] “Cisco Annual Internet Report (2018-2023),” Cisco, San Jose,
CA, USA, 2020, [Online]. Available:
https://www.cisco.com/c/en/us/solutions/collateral/executive-
perspectives/annual-internet-report/white-paper-c11-
741490.html, Accessed on: Nov. 25, 2020.
[2] D. Reinsel, J. Gantz, and J. Rydning, “The Digitization of the
World - From Edge to Core,” International Data Corporation
(IDC) & SEAGATE, Framingham, MA, USA, Nov. 2018.
[Online]. Available: https://www.seagate.com/files/www-
content/our-story/trends/files/idc-seagate-dataage-
whitepaper.pdf, Accessed on: Nov. 25, 2020.
[3] R. Y. Kim, “The Impact of COVID-19 on Consumers: Preparing
for Digital Sales,” IEEE Eng. Manag. Rev., vol. 48, no. 3, pp. 212-
218, Sept. 2020, doi: 10.1109/EMR.2020.2990115.
[4] “National Tracking Poll #200394 March 24-26, 2020
Crosstabulation Results,” Morning Consult, Washington, D.C,
USA, 2020. [Online]. Available:https://morningconsult.com/wp-
content/uploads/2020/03/200394_crosstabs_CORONAVIRUS_
CONTENT_Adults_v4_JB-1.pdf, Accessed on: Nov. 25, 2020.
[5] Q. Hardy, “COVID-19 And Our Surprising Digital
Transformation,” Forbes, Apr. 2020. [Online]. Available:
https://www.forbes.com/sites/googlecloud/2020/04/08/covid-19-
and-our-surprising-digital-transformation/#4495673420e7,
Accessed on: Nov. 25, 2020.
[6] R. De, N. Pandey, and A. Pal, “Impact of Digital Surge during
Covid-19 Pandemic: A Viewpoint on Research and Practice,” Int.
J. Inf. Manage., vol.55, Dec. 2020, doi:
10.1016/j.ijinfomgt.2020.102171.
[7] T. Spangler, “HBO Now Streaming Has Ballooned 40% in Past
Week,” Variety, Mar. 2020. [Online]. Available:
https://variety.com/2020/digital/news/hbo-now-streaming-
growth-stats-coronavirus-1203543848/, Accessed on: Nov. 25,
2020.
[8] G. Jung, M. a. Hiltunen, K. R. Joshi, R. D. Schlichting, and C. Pu,
“Mistral: Dynamically managing power, performance, and
adaptation cost in cloud infrastructures,” in Proc. 30th Int. Conf.
on Distr. Comp. Syst.(ICDS), Genova, Italy, 2010, pp. 62–73.
[9] Y. Yao et al., “Data centers power reduction: A two time scale
approach for delay tolerant workloads,” in Proc. IEEE Int. Conf.
on Comp. Comm. (INFOCOM), Orlando, FL, 2012, pp. 1431–
1439.
[10] L. Rao et al., “Minimizing electricity cost: Optimization of
distributed internet data centers in a multi-electricity-market
environment,” in Proc. IEEE Int. Conf. on Comp. Comm.
(INFOCOM), San Diego, CA, 2010, pp. 1–9.
[11] J. Li, Z. Bao, and Z. Li, “Modeling demand response capability
by internet data centers processing batch computing jobs,” IEEE
Trans. Smart Grid, vol. 6, no. 2, pp. 737–747, 2015, doi:
10.1109/TSG.2014.2363583.
[12] R. Brown et al, “Report to Congress on Server and Data Center
Energy Efficiency: Public Law 109-431,” U.S. Environmental
Protection Agency and Lawrence Berkeley National Laboratory,
Berkeley, CA, USA, Aug. 2007. [Online]. Available:
https://eta.lbl.gov/publications/report-congress-server-data-
center, Accessed on: Nov. 25, 2020.
[13] Jonathan Koomey, “Growth in Data Center Electricity use 2005
to 2010,” Analytics Press., Oakland, CA, 2011. [Online].
Available:
https://www.missioncriticalmagazine.com/articles/82420-
growth-in-data-center-electricity-use-2005-to-2010, Accessed
on: Nov. 25, 2020.
[14] “Data centres and data transmission networks,” International
Energy Agency (IEA), Paris, Jun. 2020. [Online]. Available:
https://www.iea.org/reports/data-centres-and-data-transmission-
networks, Accessed on: Nov. 25, 2020.
[15] A. Qureshi, R. Weber, H. Balakrishnan, J. Guttag, and B. Maggs,
“Cutting the electric bill for internet-scale systems,” in ACM
SIGCOMM Comput. Commun. Rev, vol. 39, no. 4, pp. 123–134,
Aug. 2009, doi: 10.1145/1594977.1592584.
[16] M. P. Mills, “The cloud begins with coal,” Digital Power Group,
Aug. 2013. [Online]. Available: https://www.tech-
pundit.com/reports/page/3/, Accessed on: Nov. 25, 2020.
[17] P. Huang et al., “A review of data centers as prosumers in district
energy systems: Renewable energy integration and waste heat
reuse for district heating,” Appl. Energy, vol. 258, Jan. 2020, doi:
10.1016/j.apenergy.2019.114109.
[18] M. T. Takcı et al., “Designing of A Simulator Architecture For
Greener Data Center Through Knowledge Transfer By Partners In
Different Sectors,” in Proc. 12th Int. Tech. Edu. and Devlp. Conf.
(INTED2018), Valencia, Spain, 2018, pp. 4069–4076.
[19] M. T. Takcı, T. Gozel, and M. H. Hocaoglu, “Forecasting Power
Consumption of IT Devices in a Data Center,” in Proc. 20th Int.
Conf.on Int.Syst. Appl. to Pow. Syst. (ISAP), New Delhi, India,
2019, pp. 1–8.
[20] D. K. Kang, E. J. Yang, and C. H. Youn, “Deep Learning-Based
Sustainable Data Center Energy Cost Minimization with
Temporal MACRO/MICRO Scale Management,” IEEE Access,
vol. 7, pp. 5477–5491, Jan. 2019, doi:
10.1109/ACCESS.2018.2888839.
[21] P. Wang, Y. Cao, and Z. Ding, “Flexible Multi-Energy
Scheduling Scheme for Data Center to Facilitate Wind Power
Integration,” IEEE Access, vol. 8, pp. 88876-88891, May. 2020,
doi: 10.1109/access.2020.2990454.
[22] S. Ahmad, A. Ahmad, M. Naeem, W. Ejaz, and H. S. Kim, “A
compendium of performance metrics, pricing schemes,
optimization objectives, and solution methodologies of demand
side management for the smart grid,” Energies, vol. 11, no. 10,
pp. 1–33, Oct. 2018, doi: 10.3390/en11102801.
[23] M. a. Al-Iriani, “Climate-related electricity demand-side
management in oil-exporting countries - The case of the United
Arab Emirates,” Energy Policy, vol. 33, no. 18, pp. 2350–2360,
Dec.2005, doi: 10.1016/j.enpol.2004.04.026.
[24] P. Yilmaz, M. Hakan Hocaoglu, and A. E. S. Konukman, “A pre-
feasibility case study on integrated resource planning including
renewables,” Energy Policy, vol. 36, no. 3, pp. 1223–1232, Mar.
2008, doi: 10.1016/j.enpol.2007.12.007.
[25] R. Verzijlbergh, L. J. De Vries, and Z. Lukszo, “Renewable
Energy Sources and Responsive Demand. Do We Need
Congestion Management in the Distribution Grid?,” IEEE Trans.
Power Syst., vol. 29, no. 5, pp. 2119–2128, Sept. 2014, doi:
10.1109/TPWRS.2014.2300941.
[26] J. M. Arroyo and A. J. Conejo, “Multiperiod auction for a pool-
based electricity market,” IEEE Trans. Power Syst., vol. 17, no.
4, pp. 1225–1231, Nov. 2002, doi:
10.1109/TPWRS.2002.804952.
[27] A. Borghetti, G. Gross, and C. A. Nucci, “Auctions with explicit
demand-side bidding in competitive electricity markets,” in The
Next Generation of Electric Power Unit Commitment Models, vol.
36, Springer, Boston, MA., 2002, pp. 53-57.
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2021.3052204, IEEE
Access
Author Name: Preparation of Papers for IEEE Access (February 2017)
14
[28] G. Ghatikar, V. Ganti, and N. Matson, “Demand Response
Opportunities and Enabling Technologies for Data Centers :
Findings from Field Studies,” Lawrence Berkeley National
Laboratory, Berkeley, CA, USA, Aug. 2012. [Online]. Available:
https://gridintegration.lbl.gov/publications/demand-response-
opportunities-and, Accessed on: Nov. 25, 2020.
[29] N. Mahmoudi, T. K. Saha, and M. Eghbal, “A new trading
framework for demand response aggregators,” IEEE Power
Energy Soc. Gen. Meet., National Harbor, MD,USA, 2014, pp. 1-
5.
[30] B. Kantarci and H. T. Mouftah, “The impact of time of use (ToU)-
awareness in energy and opex performance of a cloud backbone,”
in Proc. IEEE Global Telecom. Conf. (GLOBECOM), Anaheim,
CA, 2012, pp. 3250–3255.
[31] B. Kantarci and H. T. Mouftah, “Time of use (ToU)-awareness
with inter-data center workload sharing in the cloud backbone,”
in Proc. 2013 IEEE International Conference on Communications
(ICC), Budapest, 2013, pp. 4207-4211.
[32] T. Yang, Y. Zhao, H. Pen, and Z. Wang, “Data center holistic
demand response algorithm to smooth microgrid tie-line power
fluctuation,” Appl. Energy, vol. 231, pp. 277–287, Apr. 2018, doi:
10.1016/j.apenergy.2018.09.093.
[33] M. Dabbagh, B. Hamdaoui, and S. Member, “Shaving Data
Center Power Demand Peaks Through Energy Storage and
Workload Shifting Control,” IEEE Trans. Cloud Comput., vol. 7,
no. 4, pp. 1095–1108, Oct.-Dec. 2019, doi:
10.1109/TCC.2017.2744623.
[34] L. Cupelli, T. Schutz, P. Jahangiri, M. Fuchs, A. Monti, and D.
Muller, “Data center control strategy for participation in demand
response programs,” IEEE Trans. Ind. Informatics, vol. 14, no.
11, pp. 5087–5099, Nov. 2018, doi: 10.1109/TII.2018.2806889.
[35] W. Z. Zhang, H. C. Xie, and C. H. Hsu, “Automatic Memory
Control of Multiple Virtual Machines on a Consolidated Server,”
IEEE Trans. Cloud Comput., vol. 5, no. 1, pp. 2–14, Jan.-Mar.
2017, doi: 10.1109/TCC.2014.2378794.
[36] W. Zhang, S. Han, H. He, and H. Chen, “Network-aware virtual
machine migration in an overcommitted cloud,” Futur. Gener.
Comput. Syst., vol. 76, pp. 428–442, 2017, doi:
10.1016/j.future.2016.03.009.
[37] A. M. Ammar, J. Luo, Z. Tang, and O. Wajdy, “Intra-Balance
Virtual Machine Placement for Effective Reduction in Energy
Consumption and SLA Violation,” IEEE Access, vol. 7, pp.
72387–72402, Jun. 2019, doi: 10.1109/ACCESS.2019.2920010.
[38] J. Li and W. Qi, “Toward optimal operation of internet data center
microgrid,” IEEE Trans. Smart Grid, vol. 9, no. 2, pp. 971–979,
Mar. 2019, doi: 10.1109/TSG.2016.2572402.
[39] Y. Guo, H. Li, and M. Pan, “Colocation Data Center Demand
Response Using Nash Bargaining Theory,” IEEE Trans. Smart
Grid, vol. 9, no. 5, pp. 4017–4026, Sept. 2018, doi:
10.1109/TSG.2016.2647246.
[40] Y. Zhan, M. Ghamkhari, D. Xu, S. Ren, and H. Mohsenian-Rad,
“Extending Demand Response to Tenants in Cloud Data Centers
via Non-Intrusive Workload Flexibility Pricing,” IEEE Trans.
Smart Grid, vol. 9, no. 4, pp. 3235–3246, Jul. 2018, doi:
10.1109/TSG.2016.2628886.
[41] G. Fridgen, R. Keller, M. Thimmel, and L. Wederhake, “Shifting
load through space–The economics of spatial demand side
management using distributed data centers,” Energy Policy, vol.
109, pp. 400–413, Mar. 2017, doi: 10.1016/j.enpol.2017.07.018.
[42] M. Thimmel, G. Fridgen, R. Keller, and P. Roevekamp,
“Compensating balancing demand by spatial load migration – The
case of geographically distributed data centers,” Energy Policy,
vol. 132, pp. 1130–1142, May 2019, doi:
10.1016/j.enpol.2019.06.063.
[43] M. M. Moghaddam, M. H. Manshaei, W. Saad, and M. Goudarzi,
“On data center demand response: A cloud federation approach,”
IEEE Access, vol. 7, pp. 101829–101843, Aug. 2019, doi:
10.1109/ACCESS.2019.2928552.
[44] S. Chen, “Congestion Management in Cross-Region Grid
Considering Spatially Transferable Characteristic of Data Center
Load,” in Proc. IEEE Conference on Energy Internet and Energy
System Integration (EI2), Beijing,China, 2018, pp. 1-6.
[45] L. Yu, T. Jiang, and Y. Zou, “Distributed Real-Time Energy
Management in Data Center Microgrids,” IEEE Trans. Smart
Grid, vol. 9, no. 4, pp. 3748–3762, Jul. 2018, doi:
10.1109/TSG.2016.2640453.
[46] M. A. Islam, H. Mahmud, S. Ren, and X. Wang, “A Carbon-
Aware Incentive Mechanism for Greening Colocation Data
Centers,” IEEE Trans. Cloud Comput., vol. 8, no. 1, pp. 4–16,
Jan.-Mar. 2020, doi: 10.1109/TCC.2017.2767043.
[47] Y. Yi et al., “Joint adjustment of emergency demand response
considering data center and air-conditioning load,” in Proc. IEEE
Int. Conf. Energy Internet( ICEI 2018), Beijing, China, 2018, pp.
72–76.
[48] Ö.F. Sayan, “Türkiye Elektronik Haberleşme Sektörü Pazar
Verileri Raporu 2017 Yılı 3. Çeyrek”, Bilgi Teknolojileri ve
İletişim Kurumu, Ankara, Türkiye, 2017, [Online]. Available:
https://www.btk.gov.tr/duyurular/turkiye-elektronik-haberlesme-
sektoru-3-aylik-pazar-verileri-raporu-yayimlandi, Accessed on:
Nov. 25, 2020.
[49] İ. E. Uyanık, “Türk Telekom Veri Merkezleri,” presented at Bgd
Veri Merkezleri Konferansi, 2013, [Online]. Available:
https://www.bilgiguvenligi.org.tr/faaliyetler/etkinlikler/2013-
2/veri-merkezi-calistayi-2013/ , Accessed on: Nov. 25, 2020.
[50] E. B. Eymirli and C. ÇİFTÇİ, “Veri Merkezi Yatırımları
Açısından TRA1 Bölgesinin Değerlendirilmesi,” Kuzeydoğu
Anadolu Kalkınma Ajansı (KUDAKA), Erzurum, Türkiye, 2016,
[Online]. Available: https://www.kudaka.gov.tr/veri-merkezi-
yatirimlari-acisindan-tra1-bolgesinin-degerlendirilmesi/,
Accessed on: Nov. 25, 2020.
[51] “Elektrik Piyasasi Yan Hizmetler Yönetmeliği,”T.C. Enerji
Piyasası Düzenleme Kurumu (EPDK), Ankara, Türkiye, Nov.
2017. [Online]. Available:
http://www.resmigazete.gov.tr/eskiler/2017/11/20171126-8.htm,
Accessed on: Nov. 25, 2020.
[52] “Yük Tevzi Bilgi Sistemi (YTBS)”, [Online]. Available:
https://ytbsbilgi.teias.gov.tr/ytbsbilgi/frm_istatistikler.jsf,
Accessed on: Nov. 25, 2020.
[53] K. M. and P.K.Sen, “A Better Understanding of Load and Loss
Factors,” in Proc. IEEE Industry Appl. Society Ann.Meeting
(IAS), Edmonton, AB, Canada, 2008, pp. 1–6.
[54] J. S. B. and S. S. M. M. W. Gustafson, “The equivalent hours loss
factor revisited (power systems),” IEEE Trans. Power Syst., vol.
3, no. 4, pp. 1502–1508, Nov. 1988, doi: 10.1109/59.192959.
[55] A. A. Şeker, T. Gözel, and M. H. Hocaoğlu, “An Analytic
Approach to Determine Maximum Penetration Level of
Distributed Generation Considering Power Loss,” in Proc.
Twentieth Int. Middle East Power Syst. Conf. (MEPCON), Cairo,
Egypt, 2018, pp. 956–961.
[56] T. Gözel and M. H. Hocaoglu, “An analytical method for the
sizing and siting of distributed generators in radial systems,”
Electr. Power Syst. Res., vol. 79, no. 6, pp. 912–918, Jun. 2009,
doi: 10.1016/j.epsr.2008.12.007.
[57] N. Rasmussen, “Calculating Total Cooling Requirements for Data
Centers,” Schneider Electric, May 2011, [Online]. Available:
https://www.se.com/us/en/download/document/SPD_VAVR-
5TDTEF_EN/, Accessed on: Nov. 25, 2020.
[58] “Türkiye Elektrik Enerjisi 5 Yillik Üretim Kapasite Projeksiyonu
(2018-2022),” T.C Enerji Piyasası Düzenleme Kurumu (EPDK),
Ankara, Türkiye, May 2018. [Online]. Available:
https://www.epdk.org.tr/Detay/Icerik/3-0-66/elektrikuretim-
kapasite-projeksiyonlari, Accessed on: Nov. 25, 2020.
[59] “Elektrik Fiyatlarına Esas Tarife Tabloları - Nihai Tarife
Tablosu,” T.C. Enerji Piyasası Düzenleme Kurumu (EPDK),
Ankara, Türkiye, Sept. 2019. [Online]. Available:
https://www.epdk.org.tr/Detay/Icerik/3-1327/elektrik-
faturalarina-esas-tarife-tablolari, Accessed on: Nov. 25, 2020.
[60] M.Yilmaz, “Elektrik Piyasası Sektör Raporu 2018,” T.C Enerji
Piyasası Düzenleme Kurumu (EPDK), Ankara, Türkiye, 2019,
[Online]. Available: https://www.epdk.org.tr/Detay/Icerik/3-0-
24-3/elektrikyillik-sektor-raporu, Accessed on: Nov. 25, 2020.
[61] “Elektrik Piyasası Sektör Raporu- Aralık 2019,” T.C Enerji
Piyasası Düzenleme Kurumu (EPDK), Ankara, Türkiye, Dec.
2019. [Online]. Available:
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2021.3052204, IEEE
Access
Author Name: Preparation of Papers for IEEE Access (February 2017)
15
https://www.epdk.org.tr/Detay/Icerik/3-0-23-3/elektrikaylik-
sektor-raporlar, Accessed on: Nov. 25, 2020.
[62] “Capital Cost and Performance Characteristic Estimates for
Utility Scale Electric Power Generating Technologies,” U.S.
Energy Information Administration (EIA), 2020. [Online].
Available:https://www.eia.gov/analysis/studies/powerplants/capi
talcost/, Accessed on: Nov. 25, 2020.
[63] J. G.J. Olivier et al., “Trends in Global CO2 Emissions: 2016
Report,” PBL Netherlands Environmental Assessment Agency,
Hague, Netherlands, Rep. No: 103428, 2016. [Online].
Available: https://www.pbl.nl/en/publications/trends-in-global-
co2-emissions-2016-report, Accessed on: Nov. 25, 2020.
[64] F. Birol, "CO2 emmisions from fuel combustion - Highlights
2019,” International Energy Agency, Paris, France, 2019.
[Online]. Available: https://webstore.iea.org/co2-emissions-
from-fuel-combustion-2019-highlights , Accessed on: Nov. 25,
2020.
[65] T. Jiang, Z. Shen, Y. Liu, and Y. Hou, “Carbon footprint
assessment of four normal size hydropower stations in China,”
Sustainability, vol. 10, no. 6, pp. 1–14, Jun. 2018, doi:
10.3390/su10062018.
[66] “Carbon Dioxide Emissions From Power Plants Rated
Worldwide,” Center for Global Development ScienceDaily,
Nov.2007.[Online].Available:www.sciencedaily.com/releases/20
07/11/071114163448.htm, Accessed on: Nov. 25, 2020.
[67] B. Dudley, “BP Statistical Review of World Energy-2019 68th
edition,” BP p.l.c., 2019, [Online]. Available:
https://www.bp.com/en/global/corporate/energy-
economics/statistical-review-of-world-energy/downloads.html,
Accessed on: Nov. 25, 2020.
[68] “Electric power transmission and distribution losses (% of
output)”, World Bank- IEA World Energy Outlook, Washington
DC, USA, [Online]. Available:
https://data.worldbank.org/indicator/EG.ELC.LOSS.ZS,
Accessed on: Nov. 25, 2020.
[69] J. Lazar and X. Baldwin, “Valuing the Contribution of Energy
Efficiency to Avoid Marginal Line Losses and Reserve
Requirements,” The Regulatory Assistance Project (RAP), Aug.
2011.[Online]. Available: https://www.raponline.org/knowledge-
center/valuing-the-contribution-of-energy-efficiency-to-avoided-
marginal-line-losses-and-reserve-requirements/, Accessed on:
Nov. 25, 2020.
[70] G. Aburn , M. Hough,“Implementing EPA’s Clean Power Plan:A
Menu of Options,” National Assocation of Clean Air Agencies
(NACAA), Arlington, VA, USA, May 2015. [Online]. Available:
http://www.4cleanair.org/NACAA_Menu_of_Options, Accessed
on: Nov. 25, 2020.
[71] E. Crippa et al., “Fossil CO2 and GHG emissions of all world
countries-2019 Report,” EUR 29849 EN, Publications Office of
the European Union, Luxembourg, 2019, [Online]. Available:
https://edgar.jrc.ec.europa.eu/archived_datasets.php#reports,
Accessed on: Nov. 25, 2020.
[72] “Installed power generation capacity in the Stated Policies
Scenario, 2000-2040,” International Energy Agency (IEA), Paris,
Nov . 2019. [Online]. Available: https://www.iea.org/data-and-
statistics/charts/installed-power-generation-capacity-in-the-
stated-policies-scenario-2000-2040, Accessed on: Nov. 25, 2020.
MEHMET TÜRKER TAKCI received the
B.Sc. and M.Sc. degrees in electronics
engineering from Gebze Technical University,
Kocaeli, Turkey, in 2011 and 2015, respectively.
He is currently pursuing the Ph.D degree with
Electronics Engineering Department of Gebze
Technical University. Since 2013, he has been a
research assistant with the same University. His
research areas include demand side
management, energy forecasting and
optimization, power quality analysis.
TUBA GÖZEL. received the B.Sc. degree from
Selcuk University, Konya, Turkey, in 1994 and
the M.Sc. and Ph.D. degrees from Gebze Institute
of Technology, Kocaeli, Turkey, in 2002 and
2009, respectively. She was a Research Associate
at The University of Manchester, U.K. Currently,
she is a faculty member with Gebze Technical
University, Kocaeli, Turkey. Her research
interests include power system analysis and
distribution networks.
M.HAKAN HOCAOĞLU received the B.Sc. and
M.Sc. degrees from Marmara University, Turkey.
He obtained the Ph.D. degree in 1999 from
Cardiff School of Engineering, UK. From 1988 to
1993, he worked at Gaziantep University, Turkey
as a Lecturer. Since 1999, he has been with the
Electronics Engineering Department of Gebze
Technical University, Turkey. He is, currently, a
full Professor at the same University. His research
interests include Power Systems, Power Quality,
Earthing, Renewable Energy.
... In terms of technologies, at the top of the list and with special emphasis, are renewables (mainly solar PV), energy storage systems (mainly batteries and ultra-capacitors), and all aspects of electric mobility that require the power grid for charging purposes. Specifically, contemplating these technologies in future power grids, it is important to establish advanced metering infrastructure and demand-side management strategies (e.g., to deal with the uncertainties of power production from renewables, the consumption profiles of loads in different sectors, and cooperation of variable production of renewables with energy storage [44]), including the participation of buildings, the residential sector, and even data centers, as well as the impact of prediction errors in scenarios of demand-side management [45][46][47][48][49][50]. Specifically, the contextualization of a High Variable Renewable Energy Penetration in hybrid AC/DC grids is presented in [51]. ...
Article
Full-text available
Power grids are evolving with the aim to guarantee sustainability and higher levels of power quality for universal access to electricity. More specifically, over the last two decades, power grids have been targeted for significant changes, including migration from centralized to decentralized paradigms as a corollary of intensive integration of novel electrical technologies and the availability of derived equipment. This paper addresses a review of a disruptive vision of future power grids, mainly focusing on the use of hybrid AC/DC grids and solid‑state transformers technologies. Regarding hybrid AC/DC grids in particular, they are analyzed in detail in the context of unipolar and bipolar DC grids (i.e., two‑wire or three‑wire DC grids), as well as the different structures concerning coupled and decoupled AC configurations with low‑frequency or high‑frequency isolation. The contextualization of the possible configurations of solid‑state transformers and the different configurations of hybrid transformers (in the perspective of offering benefits for increasing power quality in terms of currents or voltages) is also analyzed within the perspective of the smart transformers. Additionally, the paper also presents unified multi‑port systems used to interface various technologies with hybrid AC/DC grids, which are also foreseen to play an important role in future power grids (e.g., the unified interface of renewable energy sources and energy storage systems), including an analysis concerning unified multi‑port systems for AC or DC grids. Throughout the paper, these topics are presented and discussed in the context of future power grids. An exhaustive description of these technologies is made, covering the most relevant and recent structures and features that can be developed, as well as the challenges for the future power grids. Several scenarios are presented, encompassing the mentioned technologies, and unveiling a progressive evolution that culminates in the cooperative scope of such technologies for a disruptive vision of future power grids.
Technical Report
Full-text available
The Emissions Database for Global Atmospheric Research provides emission time series from 1970 until 2019 forfossil CO2 for all countries. This report is contributing to the Paris Agreement process with an independent andquantitative view of global fossil CO2 emissions
Article
Full-text available
Due to the intermittent and uncontrollable nature of wind resources and inflexible operation of conventional generation units, they present challenges for the power system to integrate more wind power. With its unique flexibility on the demand side, the data center can be considered as an effective solution to relieve wind curtailment. Moreover, with the help of waste heat recovery module, the data center can reduce the utilization of conventional thermal units especially in the residential heating sector which increases the flexibility of system operation and facilitates more renewable integration. In this paper, a flexible workload management and resource scheduling model are proposed to achieve a multi-energy co-optimization for data center and enhance the integration of wind power. A two-stage stochastic programming model is formulated to address the uncertainties involved in this process. The proposed model is examined by a simulative data center microgrid and the numerical results demonstrate its effectiveness and robustness.
Conference Paper
Full-text available
In recent years, estimation algorithms become more popular in terms of forecasting customer behavior or any required data for IT companies. Forecasting results can be used in different purposes such as improving the quality and capacity of production and ser-vices, reducing to greenhouse gas emissions, and minimizing the power consumption. The accurate forecasting results are also beneficial for data centers which are the significant participants in the electricity market in terms of consuming huge power demand and have a chance to reduce consumed power, electricity costs by rescheduling their flexible loads for the future period. In this pa-per, power-consuming devices and variables affecting power con-sumption are explained. Also, the brief information about artificial neural network and regression analysis methods has been provid-ed. The power consumption of Information Technology devices is forecasted by nonlinear regression analysis and artificial neural network methods. The forecasting results show that artificial neu-ral network method is more successful.
Article
Full-text available
The significantly high energy consumption of data centers constitutes a major load on the smart power grid. Data center demand response is a promising solution to incentivize the cloud providers to adapt their consumption to the power grid conditions. These policies not only mitigate the operational stability issues of the smart grid but also potentially decrease the electricity bills of cloud providers. Cloud providers can improve their contribution and reduce their energy cost by collaboratively managing their workload. Through cooperation in the form of cloud federations, providers can spatially migrate their workload to better utilize the benefits provided by demand response schemes over multiple locations. To this end, this work considers an interaction system between the independent cloud providers and the corresponding smart grid utilities in the context of a demand response program. Leveraging the cooperative game theory, this paper presents a federation formation among the cloud providers in the presence of a location-dependent demand response program. A distributed algorithm which is coupled with an optimal workload allocation problem is applied. The effect of the federations formation on the clouds’ profits and on the smart grid performance is analyzed through simulation. Simulation results show that cooperation increases the clouds’ profits as well as the smart grid performance compared to the noncooperative case.
Article
Full-text available
Cloud computing has emerged as one of the most important technological revolutions globally. However, the rapid growth of cloud computing has imposed a massive financial burden and resulted in environmental side effects due to excessive energy consumption. The high power consumption is not only attributed to the size of data centers, but also the ineptitude of resource usage. Most of the extant research has focused on reducing power consumption by an aggressive VM consolidation, which leads to the violation of the Service Level Agreement (SLA). Furthermore, the unbalanced resource consumption exacerbates the unavailable wasted resources that are referred to as unavailable resource fragmentation. In this paper, we propose the use of a balanced resources consumption algorithm called (BRC-IBMMT) in order to enhance the efficiency of resource consumption whilst achieving an acceptable balance between conflicting correlation objectives of power consumption as well as SLA violation. Extensive simulation results of different types of workload validate and lend credence to the significance of the proposed method in reducing power consumption and SLA violation of the cloud data center.
Article
The Covid-19 pandemic has led to an inevitable surge in the use of digital technologies due to the social distancing norms and nationwide lockdowns. People and organizations all over the world have had to adjust to new ways of work and life. We explore possible scenarios of the digital surge and the research issues that arise. An increase in digitalization is leading firms and educational institutions to shift to work-from-home (WFH). Blockchain technology will become important and will entail research on design and regulations. Gig workers and the gig economy is likely to increase in scale, raising questions of work allocation, collaboration, motivation, and aspects of work overload and presenteeism. Workplace monitoring and technostress issues will become prominent with an increase in digital presence. Online fraud is likely to grow, along with research on managing security. The regulation of the internet, a key resource, will be crucial post-pandemic. Research may address the consequences and causes of the digital divide. Further, the issues of net neutrality and zero-rating plans will merit scrutiny. A key research issue will also be the impact and consequences of internet shutdowns, frequently resorted to by countries. Digital money, too, assumes importance in crisis situations and research will address their adoption, consequences, and mode. Aspects of surveillance and privacy gain importance with increase digital usage.
Article
COVID-19 has affected everyone's daily lives. At least 316 million people in 42 states have been asked to stay at home to slow down the pandemic. In this aspect, businesses have been susceptible to make substantial transformations. Workplace operations of many businesses went virtual. The effect of the digital transformation on productivity and corporate culture has been studied extensively. Meanwhile, how COVID-19 has influenced consumers, and the consumption culture has received relatively limited attention. Managers often take a wait-and-see approach on the impact of COVID-19 on sales. It is often uncertain whether and how many customers will return after the pandemic passes. Consumers live through the pandemic, and some changes might be long-lasting even after the situation eases. We examine the pandemic as an accelerator of the structural change in consumption and the digital transformation in the marketplace. Managers might adapt to the digital transformation in the market to recover or even grow further the sales after COVID-19.
Article
As large energy prosumers in district energy systems, on the one hand, data centers consume a large amount of electricity to ensure the Information Technologies (IT) facilities, ancillary power supply and cooling systems work properly; on the other hand, data centers produce a large quantity of waste heat due to the high heat dissipation rates of the IT facilities. To date, a systematic review of data centers from the perspective of energy prosumers, which considers both integration of the upstream green energy supply and downstream waste heat reuse, is still lacking. As a result, the potentials for improving data centers’ performances are limited due to a lack of global optimization of the upstream renewable energy integration and downstream waste heat utilization. This study is intended to fill in this gap and provides such a review. In this regard, the advancements in different cooling techniques, integration of renewable energy and advanced controls, waste heat utilization and connections for district heating, real projects, performance metrics and economic, energy and environmental analyses are reviewed. Based on the enormous amount of research on data centers in district energy systems, it has been found that: (1) global controls, which can manage the upstream renewable production, data centers’ operation and waste heat generation and downstream waste heat utilization are still lacking; (2) regional climate studies represent an effective way to find the optimal integration of renewable energy and waste heat recovery technologies for improving the data centers’ energy efficiency; (3) the development of global energy metrics will help to appropriately quantify the data center performances.
Article
The increasing share of renewables confronts existing power grids with a massive challenge, stemming from additional volatility to power grids introduced by renewable energy sources. This increases the demand for balancing mechanisms, which provide balancing power to ensure that power supply always meets with demand. However, the ability to provide cost-efficient and eco-friendly balancing power can vary significantly between locations. Fridgen et al. (2017) introduce an approach based on geographically distributed data centers, aiming at the spatial migration of balancing power demand between distant locations. Although their approach enables the migration of balancing demand to cost-efficient and/or eco-friendly balancing mechanisms, it will come up against limits if deployed on a global scale. In this paper, we extend Fridgen et al. (2017)'s approach by developing a model based on geographically distributed data centers, which not only enables the migration of balancing demand but also compensates for this migration when it is contradictory between different balancing power markets without burdening conventional balancing mechanisms. Using a simulation based on real-world data, we demonstrate the possibility to exploit the potential of compensation balancing demand offered by spatial load migration resulting in economic gains that will incentivize data center operators to apply our model.