**State-of-the-Art Renewable Energy in Korea**

Editors

**Zong Woo Geem Junhee Hong Woohyun Hwang**

MDPI • Basel • Beijing • Wuhan • Barcelona • Belgrade • Manchester • Tokyo • Cluj • Tianjin

*Editors* Zong Woo Geem Department of Energy IT Gachon University Seongnam Korea

Junhee Hong Department of Energy IT Gachon University Seongnam Korea

Woohyun Hwang Head Office Jeju Energy Corporation Jeju Korea

*Editorial Office* MDPI St. Alban-Anlage 66 4052 Basel, Switzerland

This is a reprint of articles from the Special Issue published online in the open access journal *Applied Sciences* (ISSN 2076-3417) (available at: https://www.mdpi.com/journal/applsci/special issues/Renewable Energy Korea).

For citation purposes, cite each article independently as indicated on the article page online and as indicated below:

LastName, A.A.; LastName, B.B.; LastName, C.C. Article Title. *Journal Name* **Year**, *Volume Number*, Page Range.

**ISBN 978-3-0365-5332-0 (Hbk) ISBN 978-3-0365-5331-3 (PDF)**

© 2022 by the authors. Articles in this book are Open Access and distributed under the Creative Commons Attribution (CC BY) license, which allows users to download, copy and build upon published articles, as long as the author and publisher are properly credited, which ensures maximum dissemination and a wider impact of our publications.

The book as a whole is distributed by MDPI under the terms and conditions of the Creative Commons license CC BY-NC-ND.

## **Contents**


### *Editorial* **Special Issue on State-of-the-Art Renewable Energy in Korea**

**Zong Woo Geem 1,\*, Junhee Hong <sup>1</sup> and Woohyun Hwang <sup>2</sup>**


Nowadays, renewable energy plays an important role in nationwide power systems. We previously dealt with the problem of accepting renewable energy; now we deal with utilizing it. This Special Issue addresses three major aspects of the current trend towards the use of renewable energy in South Korea.

The first aspect is a renewable-based power system, where both main and ancillary supplies are sourced from renewable energies. Ko et al. [1] proposed an incentive model for ESS (energy storage system) utilization in order to reduce the fluctuation of wind power. They applied it to Jeju island which has a very high proportion of renewable energy. Similarly, Lee and Kim [2] proposed an economic model for ESS-based frequency regulation from the electricity market price forecast in Korea. ESS has an advantage in terms of faster response to frequency variation than conventional fossil-fuel generators. Ko et al. [3] developed a demand-side management model using a demand response (DR) aggregator and showed real cases in South Korea. The paper analyzes the economic effect of the DR program.

The second aspect is a distribution network for renewable energy. Kim et al. [4] proposed an optimal operation scheduling model using an energy band in a microgrid. The model operates between a distribution network (DN) and microgrid (MG) while minimizing the cost of the DN and maximizing the profit of MG. A major issue of the DN is a scheme for coordination of the protection relays needed for fault currents. The model proposed by Wadood et al. [5] minimizes the total operating time of all relays to prevent excessive interruptions.

The final aspect is a nano grid network technology. Lee [6] and Shin and Geem [7] show examples of a house while Park and del Pobil [8] show a building. This is a meaningful and timely approach with respect to an ESG (Environment, Social, Governance) trend.

The papers compiled in this special issue do not suggest that the increase in renewable energy is simply the replacement of fossil energy. Renewable energy requires many innovations over existing power infrastructure and regulation. These articles show the changing trend in various sectors in Korea.

**Acknowledgments:** We thank all the authors, reviewers, and staffs (especially Nicole Lian) for their contributions to this special issue.

**Conflicts of Interest:** The authors declare no conflict of interest.

**Citation:** Geem, Z.W.; Hong, J.; Hwang, W. Special Issue on State-of-the-Art Renewable Energy in Korea. *Appl. Sci.* **2021**, *11*, 4401. https://doi.org/10.3390/app11104401

Received: 7 May 2021 Accepted: 8 May 2021 Published: 12 May 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **References**


## *Article* **Assessing the Benefits of Battery Energy Storage Systems for Frequency Regulation, Based on Electricity Market Price Forecasting**

#### **Eunjung Lee and Jinho Kim \***

School of Integrated Technology, Gwangju Institute of Science and Technology, 123 Cheomdangwagi-ro, Buk-gu, Gwangju 61005, Korea; jkl51149@gist.ac.kr

**\*** Correspondence: jeikim@gist.ac.kr; Tel./Fax: +82-62-715-5322

Received: 24 April 2019; Accepted: 20 May 2019; Published: 26 May 2019

#### **Featured Application: Authors are encouraged to provide a concise description of the specific application or a potential application of the work. This section is not mandatory.**

**Abstract:** In electricity markets, energy storage systems (ESSs) have been widely used to regulate frequency in power system operations. Frequency regulation (F/R) relates to the short-term reserve power used to balance the real-time mismatch of supply and demand. Every alternating current power system has its own unique standard frequency level, and frequency variation occurs whenever there is a mismatch of supply and demand. To cope with frequency variation, generating units—particularly base-loader generators—reduce their power outputs to a certain level, and the reduced generation outputs are used as a generation reserve whenever frequency variation occurs in the power systems. ESSs have recently been implemented as an innovative means of providing the F/R reserve previously provided by base-loader generators, because they are much faster in responding to frequency variation than conventional generators. We assess the economic benefits of ESSs for F/R, based on a new forecast of long-term electricity market price and real power system operation characteristics. For this purpose, we present case studies with respect to the South Korean electricity market as well as simulation results featuring key variables, along with their implications vis-à-vis electricity market operations.

**Keywords:** frequency regulation; energy storage system; economic benefits; price forecast; electricity market operation

#### **1. Introduction**

Global electricity markets have started to use energy storage systems (ESSs) to enhance the operational performance efficiency of power systems. Compared to other existing resources such as coal and gas-fueled generators, ESSs respond to changes in demand much more quickly. This feature offers great operational flexibility in the electricity market and in system operations, particularly in the smart operation of frequency regulation (F/R). F/R is an activity through which system operations cope with excessive fluctuations in power system frequency—fluctuations that are caused by real-time mismatches in power supply and demand. Conventional coal and/or gas-type generators have been traditionally used to resolve the F/R problem, by leaving some portion of their generation capacity unused—that is, by procuring generation reserves, and by providing reserved resources in the event of excessive frequency fluctuation.

Given the technical advantage of ESSs in terms of their prompt responsiveness to frequency fluctuation, electricity markets in the United States—such as Pennsylvania-New Jersey-Maryland(PJM) Interconnection, Midcontinent Independent System Operator(MISO), and New York Independent System Operator(NYISO)—have already designed ESS practices in F/R markets, and they have attracted

the entry of ESSs through multiple incentive mechanisms [1]. For example, some F/R markets have introduced an incentive mechanism divided into a capacity market and an energy market, to offer more benefits to those resources that respond accurately and rapidly [2–4]. Moreover, because power system frequency signals can be more frequently transmitted to ESS than conventional generators, PJM and NYISO operate single transmission systems that are separated into fast and slow-response resources [5,6]. In addition, ESSs are being more broadly applied to electricity systems: ESSs, for example, are typically associated with connecting variable renewable energy sources to enhance the power of battery charging, as well as to effectively operate and utilize electric vehicles; both are typical recent examples of ESS applications in power system operations [7–11]. In some electricity markets, including that of South Korea, electricity utility companies have undertaken large-scale ESS deployment plans for F/R.

To date, various studies have been conducted on F/R ESSs, covering topics such as optimal ESS capacity estimation and economic benefit assessments. Some studies [12–14] discuss the optimal capacity estimation of ESS for frequency control and evaluate the benefits thereof. Other studies [15,16] suggest the economic dispatch methodology and the optimal sizing of ESSs from a utility operation perspective. In determining benefits, an economic assessment should precede ESS installation. Hur et al. [17] propose economic analysis when an ESS is introduced as an F/R resource in an electricity market. Some studies discuss, from a utility perspective, economic benefit analyses in accordance with price arbitrage as a result of ESS application [18–20]. In economic analyses of electricity markets, the long-term system marginal price (SMP) estimation is the most important factor; however, most studies conduct short- and medium-term SMP estimation. Conejo et al. [21] conducted short-term SMP estimations using 24-h electricity price predictions for the day-ahead energy market, by applying various methodologies (i.e., neural network, time series, and wavelet models). Paraschiv et al. [22] propose a regime-switching model for short- and medium-term electricity price forecasting and show the superiority of the proposed model compared to an autoregressive integrated moving average and generalized autoregressive conditional heteroscedasticity models. Nowotarski et al. [23] discuss a long-term seasonal component that considers annual seasonality and estimates a future (one-year) electricity spot price by applying a wavelet-based model.

The current study proposes an analyticalmethod by which to assess the benefit of ESS implementation for F/R in electricity markets. First, to capture the basic benefit of ESS for F/R, we developed a method of predicting the SMP, which is the weighted mean of the fuel cost of a marginal plant. Second, we proposed a new scenario-based method to forecast utility economic benefits; this method considers both the electricity market structure and power system operations. The case study results show the diverse profile of the economic aspects of ESS implementation; one can readily infer from the results economic insights pertaining to large-scale ESS implementation. This study contributes to the literature on economics analysis and long-term SMP estimation. this study contributes to the literature on two perspectives. First of all, a benefits analysis of ESS for F/R is conducted in accordance with electricity market in South Korea. And the proposed methodology, Long-term SMP estimation doesn't require large time series data, therefore there shouldn't be too much difficulty with respect to data collection.

The remainder of this paper is organized as follows. In Section 2, we develop a novel methodology for assessing the economic benefits of ESS implementation in the F/R electricity market, based on South Korea's national plan for long-term electricity supply and demand [24]. Section 3 addresses the simulation results by using the methods proposed in this study, while our conclusions are presented in Section 4.

#### **2. Benefits Assessment of ESS Introduction in the F**/**R Market**

#### *2.1. ESS Introduction: Benefits Overview*

The economic benefits of ESSs for F/R derive primarily from the difference in generation cost (i.e., fuel cost in \$/kWh) between base and peak-loader generators. To balance the mismatch in supply and demand in the real-time operation of a power system, a certain amount of a base-loader generator's capacity (typically 5%) is reserved for power system F/R. Instead, to meet load demands, expensive peak-loader generators produce the required electric power not otherwise supplied by base load generators. In this way, use of this F/R reserve causes an increase in the power system operational fuel cost.

However, the reserved generation amount offered by base-loader generators can be replaced by introducing an ESS for F/R. In other words, the reserved amount from base-loader generators—which are cheaper than peak-loaders—can be supplied to power systems to meet load demand. From the standpoint of power system operation, the use of ESSs for F/R facilitates the replacement of expensive peak-loader generators with cheap base-loaders.

Figure 1 depicts the basic economic benefit of an ESS for F/R in the electricity market. One of the key factors in assessing the economic benefit of ESSs for F/R in power system operation is the estimation of future prospects for the SMP in electricity markets. The SMP is the spot market price used in electric power transactions, and it is determined by considering the most expensive generation cost of the marginal generating unit that meets the marginal demand of electricity markets. When it comes to a base load generator's reserved generation associated with F/R, the revenue lost by not selling the reserved generation can be compensated for by offering the opportunity cost (COFF), which is defined as the difference between the SMP and the base-loader generator's fuel cost. Because the SMP, or the generation cost of a marginal generator, is typically decided by the peak load generator's fuel cost, the COFF offered to base load generators for F/R can be redefined as the difference between the generation costs of peak and base-loaders. In this regard, the economic benefit of ESSs for F/R can be captured by the replacement benefit—that is, the benefit that derives from fuel cost savings on account of replacing expensive peak-loader generators with cheaper base loaders in F/R. The benefit can, therefore, be assessed primarily by forecasting future SMP (i.e., the generation cost of peak load) and base load fuel costs.

**Figure 1.** Economic benefits of ESSs for F/R in the electricity market [25,26].

#### *2.2. Probabilistic Long-Term SMP Forecast*

To assess effectively the economic benefit of ESS for F/R over a given horizon (typically 10–15 years), we propose the novel probabilistic weighted average to predict future annual SMP profiles for the horizon. Because the SMP is the most expensive fuel cost of the generator that is last committed to meet the forecasted demand at a given hour, the estimation for an hourly marginal plant profile across the operating horizon is the key element in predicting annual SMPs. To this end, the current study proposes a probabilistic method by which to forecast a long-term marginal plant profile associated with hourly SMP, to assess the economic benefit of ESS for F/R.

To obtain the annual SMP profiles—that is, annual marginal plant probability profiles in electricity markets—we used a 15-year national long-term supply and demand projection plan published by a South Korean energy agency [24]. Detailed descriptions of the development of a probabilistic annual SMP forecast, based on the estimation of a long-term marginal plant profile, are given below.

First, the annual generation capacity of each fuel type for the next 10 years can be obtained from the national plan for long-term electricity supply and demand. However, this capacity cannot be identified as real generation capacity, because it does not take into account the operational unavailability of generators owing to events such as forced and maintenance outages. The forced outage rate (FOR) and maintenance outage rate (MOR) speak to the unavailability of generating units associated with unplanned and planned outages, respectively. When assessing the annual generation capacity for each fuel type, the unavailable generation capacity should therefore be extracted from the nominal generation capacity. In addition, given the fluctuating output of renewable energy, we use estimations of actual generation capacity from the national plan, rather than installed capacity data.

Second, we obtain from the hourly demand average the past demand profiles that are assumed to be identical to estimated peak demand. However, these profiles can also be divided into two different demand profiles—weekdays and weekends. Annual peak demand for the next 10 years is used to determine the annual hourly demand. The methodology is as follows. Equations (1) and (2) represent the idea that the sum of the hourly average demand is the product of hourly peak demand and α*<sup>t</sup>* (rate of hourly demand on the basis of peak demand), which transform to Equation (3). Applying this notion, annual peak demand satisfying average demand is calculated by Equation (4). Peak demand is calculated as:

$$d\_l = d\_{peak} \times \alpha\_l \tag{1}$$

$$\sum\_{t=1}^{24} d\_t = \sum\_{t=1}^{24} d\_{\text{peak}} \times \alpha\_t = d\_{\text{peak}} \sum\_{t=1}^{24} \alpha\_t \tag{2}$$

$$d\_{\text{peak}} \sum\_{t=1}^{24} \frac{\alpha\_t}{24} = d\_{\text{average}}\tag{3}$$

$$d\_{\text{peak}} = \frac{d\_{\text{average}}}{\sum\_{t=1}^{24} \alpha\_t} \tag{4}$$

where *dt*(0 ≤ *t* ≤ 24), *daverage*, and α*t*(0 ≤ α*<sup>t</sup>* ≤ 1) are the demand at each time, peak demand, and rate of past demand for each time on the basis of peak demand, respectively.

Third, the annual demand clustering pattern can be obtained from the peak demand for each year, as drawn from past data. This means that estimated demand is equal to the movement of the past demand pattern, in line with peak demand. Therefore, the annual demand pattern is estimated by multiplying annual peak demand by each value of the percentage of demand for every hour, based on peak demand from the past demand clustering pattern. We compare the annual generation mix from generation capacity and the demand clustering for every hour to identify the marginal plant resources used on weekdays, weeknights, and weekends. Marginal plant profiles for daytime, nighttime, and weekends are realized by designating daytime as 16 h, nighttime as 8 h (i.e., 12 AM–8 AM), and weekends as 24 h. This can be used to count numbers determined as SMP for the specific resource.

Fourth, this study assigns weighting for generation costs, such that they are allocated a heavier weight when they are closer to the present; it is assumed that future generation costs will be similar to past generation costs. The SMP for weekdays, weeknights, and weekends is estimated by using the weighted average of the marginal plant profile. Future SMP is calculated as follows.

$$\text{SMP} = \mathbb{C}\_{\text{nucher}} \times P(\text{X}\_{\text{nucler}}) + \mathbb{C}\_{\text{coal}} \times P(\text{X}\_{\text{coal}}) + \mathbb{C}\_{\text{LNG}} \times P(\text{X}\_{\text{LNG}}) + \mathbb{C}\_{\text{oil}} \times P(\text{X}\_{\text{oil}}) \tag{5}$$

where the subscripts Cnuclear, Ccoal CLNG, and Coil denote nuclear power, coal, liquefied natural gas (LNG), and oil generation costs, respectively; subscripts *X*nuclear, *X*coal, *X*LNG, and *X*oil denote nuclear power, coal, LNG, and oil variables, respectively; and *P*(*X*i) denotes the marginal plant profile of *X*i.

Using this function, the SMP for daytime, nighttime, and weekends can be determined. The annual SMP contains the rates for daytime (approximately 16 h per day for five days per week), nighttime (approximately 8 h per day for five days per week), and weekend (24 h per day for two days per week). In accordance with supposition, the outcomes of rate calculation are 0.476, 0.238, and 0.286, respectively, on the basis of one year (8760 h); we assign these rates as a calculus in Equation (6). The annual SMP associated with these rates is defined as

$$\rm{SMP\_A} = \rm{SMP\_d} \times 0.476 + \rm{SMP\_n} \times 0.238 + \rm{SMP\_W} \times 0.286.\tag{6}$$

SMPA, SMPd, SMPn, and SMPw are annual, daytime, nighttime, and weekend SMPs, respectively. The SMPA formula consists of SMPd, SMPn, and SMPw, with their weights calculated by using the duration rate in the year. The long-term SMP estimation framework is illustrated as Figure 2.

**Figure 2.** Schematic diagram of the SMP forecast methodology.

#### *2.3. Assessment of Economic Benefits from ESS for F*/*R in the Electricity Market*

ESS is introduced in a power system for F/R. If implemented in the South Korean electricity market, it will change the overall demand placed on coal and LNG supply capacity generators and modify electricity costs.

Currently, coal generators need to secure a reserve for F/R through a 5% reduced operation in the electricity market. Although this generation constraint is not included in the price-setting schedule, it is used in the operation schedule that is produced following the creation of the price-setting schedule. Therefore, reduced coal capacities receive an opportunity cost payment known as a constrained-off energy payment (i.e., the aforementioned COFF), which is calculated based on the minimum SMP and

coal fuel cost. To meet the shortfall, LNG generators increase generation and then receive compensation, known as a constrained-on energy payment (CON); CON is calculated based on the maximum SMP and LNG fuel cost. However, the settlement would differ when operating ESS for F/R: because F/R ESSs can alleviate the constraint, coal capacities can generate more, and be compensated for this increased generation in the form of a scheduled energy payment (SEP). This SEP is calculated based on the SMP in the price-setting schedule, rather than the COFF. Furthermore, the LNG generators need not generate more to compensate for the shortfall, and so they do not receive the CON payment. As a result, each participant (i.e., Coal Gen., LNG Gen., and Utility) would then be compensated as in Table 1.

**Table 1.** Payment changes deriving from frequency regulation energy storage systems introduction to the South Korean energy market.


We propose a method by which to estimate the utility benefits (UBs) of introducing into electricity markets ESSs for F/R. We consider the benefits in the energy and ancillary service (A/S) markets, based on generation constraints and when considering F/R in the South Korean electricity market.

Given that ESSs can be used as reserves, the implementation of an ESS alleviates the base load generation constraint and can produce benefits similar to ESS capacity; this is because the utility need not pay additional costs with respect to the coal and LNG generators. In addition, an A/S payment would be provided to the utility because the F/R ESS, which the utility needs to plan to own, replaces the conventional generation role. Therefore, the UBs increase in terms of the energy and A/S aspects. Equations for calculating the energy market price (EP) and the A/S price (ASP) benefits are as follows.

$$\text{EPP} = \text{Availability}\_{ESS} \times \left(\text{CON} + \text{COFF} - \text{SEP}\right) \times 8760 \times \text{R}\_{\text{Power}} \tag{7}$$

where *AvailabilityESS* and *RPower* are the coefficient of utilization for ESS and the generation operation rate, respectively, and

$$\text{ASP} = \text{Capacity}\_{\text{ESS}} \times \text{Arailability}\_{\text{ESS}} \times \text{W}\_{\text{ESS}} \times \text{W}\_{\text{Droop}} \times \text{W}\_{\text{Deailbound}} \times \text{LI}\_{\text{FR}} \times \text{R}\_{\text{ESS}} \times 8760 \tag{8}$$

where *CapacityESS*, *WESS*, *WDroop*, and *WDeadband* are the practical ESS capacity, weighted values of ESS, droop, and dead band, respectively. Furthermore, *UFR* is the unit cost for F/R, and *RESS* is the ESS operation rate. ESS compensation should be differentiated from conventional resource compensation, because it provides outstanding F/R performance; therefore, resource weighting was added through *WESS*—which has a value exceeding 1 in the ESS settlement of the A/S market—to provide a larger payment than conventional resources. Both the droop and dead band demonstrate the performance of resources in F/R, and thus, these factors should also be considered in the A/S settlement by using *WDroop* and *WDeadband*. These have values in the range of 0.8–1.05 and 0.85–1.05, respectively, and resources with lower values in them are set so as to have heavy weighting. The sum of energy (EP) and A/S benefits (ASP) equals the UB, given by Equation (9).

$$\text{UB} = \text{EP} + \text{ASP} \tag{9}$$

Although we actually assume that UBs reflecting the current electricity market need to come about during the daytime of a weekday, we also consider two other cases to make a total of three. In the first case, UBs occur during the daytime of a weekday, because coal generators generate more electricity during the day than at other times. In the second, UBs are derived during the nighttime of a

weekday, because the upward generation of coal generators deepens during that time. In the third, UBs come about all day, on account of a stable trend of reserves and little upward generation among coal generators. For each of these cases, we present below equations by which to calculate EP and ASP benefits.

1. Case A: benefits occur during the daytime

$$\text{EP} = \text{Availability}\_{\text{ESS}} \times \left( \text{CON} + \text{COFF} - \text{SEP} \right) \times 8760 \times 47.6\% \tag{10}$$

$$\text{ASP} = \text{Capacity}\_{\text{ESS}} \times \text{Availability}\_{\text{ESS}} \times \text{W}\_{\text{ESS}} \times \text{W}\_{\text{Droop}} \times \text{W}\_{\text{Dailband}} \times \text{LI}\_{\text{FR}} \times R\_{\text{ESS}} \times 8760 \times 66.7\% \quad \text{(11)}$$

2. Case B: benefits occur during the nighttime

$$\text{EP} = \text{Availability}\_{\text{ESS}} \times \left( \text{CON} + \text{COFF} - \text{SEP} \right) \times 8760 \times 28.8\% \tag{12}$$

ASP = *CapacityESS* × *AvailabilityESS* × *WESS* × *WDroop* × *WDeadband* × *UFR* × *RESS* × 8760 × 33.3% (13)

3. Case C: benefits occur all day

$$\text{EP} = \text{Availability}\_{\text{ESS}} \times (\text{CON} + \text{COFF} - \text{SEP}) \times 8760 \times 100\% \tag{14}$$

$$\text{ASP} = \text{Capacity}\_{\text{ESS}} \times \text{Anailability}\_{\text{ESS}} \times \text{W}\_{\text{ESS}} \times \text{W}\_{\text{Drap}} \times \text{W}\_{\text{Denhaal}} \times \text{LI}\_{\text{FR}} \times \text{R}\_{\text{ESS}} \times 8760 \times 100\% \tag{15}$$

#### **3. Case Study: ESS Participation Benefits Assessment in the F**/**R Market, Based on the Long-Term SMP Forecasting Methodology**

#### *3.1. Comparison of Real and Estimated SMPs*

We measure real SMP against estimated SMP and use past data from the Korean Power Exchange (KPX) to assess the results of our proposed methodology. To compare the SMP based on real data to the estimated SMP, we use real SMP for each month, forecasted SMP using marginal plant probability, and generation cost, using data from the 2001–2015 period. Figure 3 presents the results.

**Figure 3.** Comparison of real and estimated SMP, 2001–2015.

The results show that the estimations are similar to the real SMP values, with an average error of approximately 3.5%. Therefore, we determined that it is possible to predict the SMP by using the methodology presented herein.

#### *3.2. Long-Term SMP Forecasting of the South Korean Electricity Market*

We conducted a case study to investigate UBs and examine how ESS for F/R affects efficiency in South Korea. To undertake this investigation, we must first estimate the SMP according to the method proposed in Section 2.1. First, we considered the auxiliary power consumption ratio, the FOR, and the MOR to determine the actual supply capacity; this was obtained by deducting these rates for each of the generation resources. We assumed that the auxiliary power consumption ratio was 5%; additionally, for FOR and MOR, we used the average from the 2014–2015 period (Table 2).

**Table 2.** Average forced outage rate (FOR) and maintenance outage rate (MOR) values of South Korean nuclear and coal power plants.


By substituting these figures and applying the renewable energy availability in resource capacity [24], we can determine the actual supply capacity and generation mix. The generation mix of the base load is based on the constructed actual supply capacity (Table 3). As part of the renewable energy penetration policy, the capacities of nuclear and coal resources will be reduced, while those of renewable energy resources will be increased. According to the national plan [24], renewable resources consist of photovoltaic power, wind power, tidal power, and by-product gas.



Once we estimate the actual supply capacity, we can then estimate the annual demand on weekdays and weekends and compare these values to the actual supply capacity. Demand is estimated based on [24], and we use weekday and weekend patterns from 2016 to estimate the annual demand pattern ratio for 24 h; future demands are projected using the annual electricity consumption projection of [24], based on the load-pattern ratio. The hourly load profiles of weekdays and weekends, which are estimated as the average of the 2016 electricity load, are illustrated in Figure 4.

**Figure 4.** Hourly load profiles of weekday and weekend power usage.

We assume that the 2019–2028 load patterns are similar to the 2016 load patterns. To estimate the 2019–2028 load profile, the load-pattern ratio was obtained by dividing the load profile by the load profile peak demand. Annual peak load was calculated from the electricity consumption and the sum of the load-pattern ratio for 24 h (Table 4). We were then able to estimate the annual load profile of weekdays and weekends for the 2019–2028 period.


**Table 4.** Annual total consumption and peak load for 2019–2028, based on 2016 data.

Following our 2019–2028 load profile estimation, we verified which resource would be selected in each hour by comparing the load profile and generation mix. Figure 5 shows the weekday and weekend load profiles, as well as generation, in 2024. Renewable resources do not comprise a single resource; rather, they are derived from multiple sources. Nonetheless, the national plan does not provide planned capacity for each type of renewable resource. This means that we encountered difficulties in generating actual hourly generation projections for renewable resources. Therefore, despite the inherent flexibility, we assumed that the single renewable resource is constant over time.

From these demand- and supply-side processes, we can obtain the marginal plant probability for each generation resource. Comparing the generation mix and demand allows us to see how frequently specific resources are selected as the marginal plant. Table 5 shows the marginal plant probability profile of coal generation.

**Figure 5.** Hourly load profile of weekdays and weekends, and generation mix, in 2014.


**Table 5.** Marginal plant probability profile of coal generators, 2019–2028.

We applied the average value of the annual generation cost from the 2016–2017 period. The SMP is estimated by using the marginal plant probability profile and generation cost and is shown in Table 6.



The SMP estimation results show a gradual decrease until 2023, and then a steady increase. The major driver of the SMP projection trend is recent energy policy that works to reduce South Korea's reliance on nuclear and coal power plants and expand its reliance on renewable resources.

#### *3.3. Economic Assessment of ESS in F*/*R the Market, with Respect to UBs*

Following the SMP estimation, we determined UBs as a function of introducing F/R ESS. The UBs consist of benefits in the energy and A/S markets. Simulations were conducted for each of case A, B, and C, using two ESS capacity scenarios that consider existing installed capacity (i.e., 52 MW) and planned future capacity (i.e., 500 MW). To calculate UBs in each scenario, the value of UBs is derived by using the following settings: *AvailabilityESS* = 48%, *WESS* = 1.1, *WDroop* = 1.05, *WDeadband* = 1.05 and *UFR* = 2.53 \$/kWh. Furthermore, the daytime, nighttime, and annual average SMPs are applied to each equation in a regular sequence. Results pertaining to UBs during the 2019–2028 period are presented in Table 7.

**Table 7.** Projected Energy Market Price (EP) Benefits, Ancillary Service Price (ASP) Benefits, and Utility Benefits (UBs), 2019–2028 (Unit: Thousands of \$).


In case A, over the 10-year period, for a 52-MW ESS (500-MW ESS), we could anticipate ASP benefits of about \$3 million (\$31 million), EP benefits of about \$33 million (\$318 million), and UBs of about \$36 million (\$348 million). When considering other cases based on case A, we can see that case B (case C) is 26% (155%) the scale of case A. In cases A, B, and C, the difference between the two capacity scenarios was approximately \$31, \$8, and \$48 million, respectively—demonstrating an 862% increase for an ESS capacity increase from 52 MW to 500 MW. This finding demonstrates that UBs are directly proportional to the size of the ESS. The benefit scale for these cases differs very much from the simulation results; this divergence derives from the fact that SMP change depending on the time involved and the times in which ESS benefits are generated. Consequently, changes made to the electricity industry—such as changes to energy policy, fuel costs, and demand among others—will decide the benefit level.

#### **4. Conclusions**

Recently, ESSs for electricity systems have been utilized in numerous ways. (For example, they are connected to renewable resources and used to discharge large quantities of electricity at peak usage times.) A plan to implement ESSs for F/R has recently been introduced in South Korea; other countries have already implemented them, because they allow for the stabilization of electricity systems in a way that compensates for its higher costs and encourages more efficient fuel use. We carried out a benefit estimation in anticipation of the introduction in South Korea of an F/R ESS.

We present a novel methodology for assessing the anticipated UBs. First, we extrapolated the future SMP by using a weighted average of marginal plant probability and fuel cost for each resource. We then calculated the UB as the sum of the energy and A/S market benefits, as determined by the electricity market and industrial structure. In the case study, we found the scale of benefits to range from \$91 million to \$540 million for 500 MW, and noted that among the three cases, case C—in which ESS for F/R is operated all day—offers the largest benefit. Although the results of the simulation models present different benefit levels, all cases show large and positive benefits; none show a negative result. Thus, we conclude that the introduction of ESSs for F/R in South Korea would enhance power system stability and bring about substantial UBs.

**Author Contributions:** Both authors contributed to this work. E.L. undertook related research, performed the analysis, and wrote the paper. J.K. designed the study and led and supervised the research.

**Funding:** This work was funded by the Korea Institute of Energy Technology Evaluation and Planning (KETEP) and the Ministry of Trade, Industry & Energy (MOTIE) of the Republic of Korea (grant number 20171210200810).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Implementation of a Demand-Side Management Solution for South Korea's Demand Response Program**

#### **Wonsuk Ko 1, Hamsakutty Vettikalladi 1, Seung-Ho Song 2,\* and Hyeong-Jin Choi <sup>3</sup>**


Received: 31 December 2019; Accepted: 26 February 2020; Published: 4 March 2020

**Abstract:** In this paper, we show the development of a demand-side management solution (DSMS) for demand response (DR) aggregator and actual demand response operation cases in South Korea. To show an experience, Korea's demand response market outline, functions of DSMS, real contracted capacity, and payment between consumer and load aggregator and DR operation cases are revealed. The DSMS computes the customer baseline load (CBL), relative root mean squared error (RRMSE), and payments of the customers in real time. The case of 10 MW contracted customers shows 108.03% delivery rate and a benefit of 854,900,394 KRW for two years. The results illustrate that an integrated demand-side management solution contributes by participating in a DR market and gives a benefit and satisfaction to the consumer.

**Keywords:** demand response; demand-side management solution; electricity market; energy management

#### **1. Introduction**

In a power system, electricity demand changes constantly. Power suppliers need to generate more power generation when demand is high, and less when demand is low. Traditionally, to coordinate supply and demand has been the supplier's responsibility and the demand side has been considered secondary. The power system supplier predicts the demand and, then, generates the supply capacity to meet the demand. After that, market price changed to supply capacity serves as criteria for deciding the facility capacity. The capacity that cannot be adjusted on the supply side is supplemented by DSM (demand-side management) such as temporarily reducing or moving the load on the demand side [1,2].

Recently, there has been a growing interest in considering the demand as the same as the supply side. Technological changes are occurring both on the supply side and on the demand side. Demand response is an alternative to additional infrastructure to maintain the safe margin between generation or distribution capacity and demand. The definition of demand response from the United State Department of Energy says, "Changes in electric usage by end-use customers from their normal consumption patterns in response to changes in the price of electricity over time, or to incentive payments designed to induce lower electricity use at times of high wholesale market prices or when system reliability is jeopardized" [3]. The most significant technological change on the demand side is the spread of smart meters and advanced metering infrastructure. According to this environment, traditional DSM programs should redesign to an automated market-based mechanism. The responses from the demand-side resources should also be reliable, fast, flexible, and large enough to compete with the supply-side resources. DSM programs can be classified into load management (LM) types and energy efficiency (EE) types [4–8].

In order to deal with peak load conditions, electric utilities have to invest in system capacity, which is underutilized during most times. Not surprisingly, utilities have been seeking methods to improve capacity utilization. Demand response (DR) is one mechanism utilities use to curtail or shift peak customer load [9–11].

Regarding the remuneration, a price-based and an incentive-based DR program are used. In a price-based DR program, consumers reduce their power consumption by responding to the electricity tariff set by the electricity market. In an incentive-based DR program, consumers are contracted by individuals or groups to reduce their power consumption for a certain period that the economic transactions requested in the electricity market.

For a price-based DR program, researchers investigated a multi-agent modeling and optimization algorithms under DR programs for real-time prices [12]; a coordination strategy between a micro network and a price-based demand response program for adjusting loads [13]; a bidirectional communication smart meter design between the household smart meters and the distribution management system [14]; a dynamic price scheme for electricity in a smart network, by analyzing the behavior and the possible demand response of the consumer [15]; and a modified real-time price model that encourages customer choice in electricity rates and is based on the amount of risk customers are willing to take and a real-time grid condition index developed by the California Independent Service Operator [16].

For an incentive-based DR program, research has shown that a fuzzy-based dynamic incentive scheme for residential customers can effectively incorporate the influences of socioeconomic conditions, expected curtailment, probability of curtailment, and notice period [17]. The dynamic adjustment of the offered prices is analyzed to reduce the demand and maximize its performance within T days [18]. The uncertainty of the demand in the network planning is modeled and includes the integral control of the load disconnection in search of the minimum cost [19]. The participation of smart homes with the help of a controller is capable of managing the electrical installation and restructuring the demand profile by changing the operation of the flexible loads [20]. The incentives motivate clients to participate in automatic DR programs with the purpose of compensating imbalances between offer and demand [21], and a DR program is shown from the economic point of view based on optimal incentives [22].

Demand resources have played an important role in Korea for more than 20 years. To reduce peak demand during summer and winter, DR programs and the operating system have been researched and implemented as a demonstration since 2010. In 2014, any customers who joined the DR market were able to sell their reduced demand in the electricity market as supply resources [23,24].

In this paper, a development and case studies of demand-side management solution (DSMS) in South Korea is presented. After that, the DSMS is verified with a one-year experience of the Korea DR program. To implement this solution, a design structure of the DSMS is proposed and tested. Customer baseline load (CBL), relative root mean squared error (RRMSE), load curtailment value, and money-saving of contracted customer's data are also displayed from the DSMS. To calculate CBL, a short historical period close to the event day was chosen, then, the CBL was calculated by merely averaging the data of the previous not-event days. After deciding the CBL, the assessment of the estimated CBL is needed. In order to verify the accuracy of the calculated CBL, RRMSE is used to assess the estimation error by comparing the actual electricity load and the estimated CBL. If the estimation error is close to zero, it means there is the high accuracy of the estimated CBL; if it is greater than zero, it means overestimated CBL; and if it is less than zero, it means underestimated CBL [25]. The calculation process of CB L and RRMSE are explained in Section 2.

This paper is composed as follows: Section 2 shows Materials (status of demand resource market in South Korea) and Methods (development of DSMS); Section 3 presents the money savings and energy conservation results using DSMS; finally, the discussion and conclusions are given in Sections 4 and 5.

#### **2. Materials and Methods**

#### *2.1. Status of Demand Resources Market, South Korea*

The role of the Korea Power Exchange (KPX) is to control the operation of South Korea's electricity market and the power system, as well as execute real-time dispatch and establish the basic plan for supply and demand. Every year, the KPX has issued an annual report for the electricity market trend and analysis. In this section, the status of the demand resource market from the 2016 annual report is summarized. Demand response refers to a suite of policies and institutions to provide efficient and stable electric power service at the lowest cost by helping consumers change their consumption pattern. Under the current contract-based utility rate schemes in South Korea, consumers have a very weak incentive to voluntarily participate in the demand response [26].

The introduction of the demand response program can effectively stabilize the electricity market and the operation of its system. The demand response can decrease investment in the power system including generation, transmission, and transformation networks; enhancing reliability in electric-power supply. Consumers can take part in the demand response programs by reducing their electricity usage at critical times through monitoring demand or securing a load that can be shut down by KPX and then making a bid on a load.

In early 2008, when the demand response market opened, it was bidding-based sponsored by the Electricity Industry Fund and, now, is making a transition into an advanced demand response market where market price and real-time system operation are linked with each other for resource transaction. In 2012, a smart demand response market opened where demand resources in small and medium quantity were traded in an effort to tap into the highly reliable and easily accessible demand resources using smart grid technology. The smart demand response market makes payment at a fixed rate on the condition of keeping the capacity to be curtailed unchanged. A payment is also made to those who cut down demand at the system operator's request for load curtailment. As the Electricity Business Act was revised to allow demand resource trading in the electricity market, in 2014, the demand response market and smart demand response market sponsored by the Electricity Industry Fund were abolished in late 2014, and the elimination of the two-month-ahead and week-ahead programs followed in late 2015 [26].

At the end of 2015, a new demand response market replaced these previous programs and was integrated into the electricity market. Figure 1 shows the process of the trading mechanism of a demand response market. The demand response market trades demand resources arranged by retailers, each demand resource is required to come from more than ten end users, and must be valued at above 10 MW. The DR aggregator collects consumers to organize demand resources. After registering with KPX, these resources are certified for trading under the same rules governing the centrally dispatched generators. Demand resources are put on a bid against power generation resources on a daily basis, and when sold, demand curtailment begins. In the system operation process, consumers are required to cut down on demand within an hour of a dispatch order.

**Figure 1.** Trading mechanism of a demand response market.

Capacity available for

Curtailed energy

The KPX calculates curtailed energy and makes payment to retailers who allocate the profits to consumers. As indicated in Tables 1 and 2, the demand response market has seen significant progress from both quantity and quality perspectives. The number of consumers has grown from 90 to 3592 and curtailed energy amounts to 175,771 MWh, up by 342.6 times from 513 MWh, becoming a world-renowned global demand response market. The transition to an advanced demand response market is also politically compelling. As the program matures along with demand resources integrated into the electricity market, market participants will feel easy to understand the system and the market will flourish. Furthermore, lower resistance towards generators and greenhouse gases are expected along with eased market exploitation and a more efficient market [23].


**Table 1.** Consumers and curtailed power before market opening [23].


Curtailment (MW) 1520 2444 2889 3272 3885 4352 4271 4222

(MWh) 117,075 91,034 98,898 293,955 113,661 62,110 121,206 -

**Table 2.** Consumers and curtailed power after market opening (provided by KPX, July 2018).

#### *2.2. Development of Demand-Side Management Solution (DSMS)*

To realize the trading mechanism of a demand response market, as shown in Figure 1, the structure of a demand-side management solution is designed, as shown in Figure 2. The DSMS directly captures usage data by sensor every 5 min from demand resources such as a house, building, apartment, and factory. After that, these usage data are sent to the KPX server every 5 min. If a power shortage occurs, the KPX sends a reduction order to the DSMS. When the request is generated from KPX, the DSMS calculates the CBL of the DR duration time and, then, contacts the DR resources to request the contracted power reduction amount, and after that, demand curtailment begins. In the system operation process, consumers are required to cut down on demand within an hour of the dispatch order. The DSMS also calculates the RRMSE (relative root mean square error), the CBL (customer baseline load) of the DR resources, regularly to keep the DR resources in the Korea electricity market.

**Figure 2.** The structure of a demand-side management solution.

The CBL is the prediction of the amount of electricity that would generally be used if electricity consumption had not been reduced by the KPX directive [27]. Figure 3 shows the power consumption and CBL of the desired date and time. The bar graph illustrates power usage at 5 min, 10 min, and 30 min, respectively and the line means the CBL of every hour. At the bottom, the dialog box shows contracted curtail power, CBL, load, result of curtailed power, and remained contracted curtail power in order.

**Figure 3.** Dashboard of power usage.

Figure 4 illustrates a customer baseline load (CBL) calculation. The CBL is an average hourly energy consumption calculated as follows: According to the KPX guideline, the CBL calculation method is either MAX 4/5 or Mid 6/10. The Max 4/5 method is calculated using the electricity usage in normal working for 5 consecutive days. To calculate the CBL, first, the smallest electricity usage day of the 5 days is excluded, then, the average usage for 4 days is used as Max 4/5 CBL. The Mid 6/10 method is calculated based on the power electricity usage in normal working for 10 consecutive days. Two days are excluded from the top and bottom usage of the 10 days, respectively. The average usage of the remaining six days is used as Mid 6/10 CBL. Table 3 explains how to calculate the CBL as MAX 4/5. First, D-2 is the smallest electricity usage day, therefore, this day is excluded. Then, the average usages of the remained 4 days are added, and then divide by 4 [27].

**Figure 4.** Customer baseline load calculation.


**Table 3.** Calculation example of the customer baseline load (CBL) (Max4/5 Method).

To estimate the DR reduction value, the pattern of regular power use should be fairly uniform, and an objective evaluation technique is required. For this objective evaluation, the RRMSE was used as an index to specify the uniformity of the pattern of power usage. Figure 5 shows the RRMSE for the customer who wants to participate in demand response market and Equation (1) shows that RRMSE calculation [26]. In this equation, *D* is an investigation day, *D*(*n*) is the number of investigation days, *T* is a time duration of an investigation day, *T*(*n*) is number of time durations of an investigation day, *CBLd*,*<sup>t</sup>* is a customer baseline load at *t* hour on *d* day, and *Loadd*,*<sup>t</sup>* is an electricity usage at *t* hour on *d* day.

$$\sqrt{\frac{\sum\_{d \in D, t \in T} \text{(CBL}\_{d,t} - \text{Load}\_{d,t}\text{)}^2}{D(n) \times T(n)}} \div \frac{\sum\_{d \in D, t \in T} \text{Load}\_{d,t}}{D(n) \times T(n)}\tag{1}$$

*D*(*n*): Number of investigation days

*T*: Time duration of investigation day

*T*(*n*): Number of time durations of investigation day

*CBLd*,*t*: Customer baseline load at *t* hour on *d* day

*Loadd*,*t*: Electricity usage at *t* hour on *d* day

**Figure 5.** Relative root mean squared error (RRMSE) for a customer.

The RRMSE is calculated by dividing the RMSE (root mean square error) with the average value of electricity usage data. The fluctuation between the CBL and actual electricity usage is a critical judgment criterion as a reliable DR resource. To register as a DR resource in the Korea electricity market, the RRMSE value must be less than 30%; if the value exceeds 30%, it is not allowed to join the DR market. If the RRMSE value becomes more increased, conformity of power usage pattern decreases, which makes it difficult to judge the reduction value accurately. Although the U.S. PJM power DR market is set at less than 20%, the Korea DR market is set to within 30% at the beginning

of the system. To determine the suitability of the DR customer, the KPX enforces an annual RRMSE assessment, and then that result determines whether the DR customer can participate in the DR market for one year. This calculation is based on data from 45 weekdays from the date of verification [26]. For example, the flat line of the electricity usage graph in Figure 5 indicates 10%, that means Figure 5 is a reliable DR resource, and therefore this customer can participate in the DR market.

The incentive for participating in the DR market can be classified into basic settlement and performance settlement, and the monthly basic settlement payment is as shown in Table 4. According to the reduction duration time, the actual performance-settlement payment is different under the KPX condition [27].


**Table 4.** Monthly basic settlement payment 2016–2018.

*Appl. Sci.* **2020**, *10*, 1751

To calculate the basic settlement money, Equations in (2) are applied to the integrated demand-side management solution [26].

$$\begin{aligned} \text{DRBP}\_{i.m} &= \text{ORC}\_{i.m} \times \text{BP}\_{m} \times 1,000\\ \text{IBPC}\_{i.m} &= \frac{\text{TDRBP}\_{i.m}^{\*}}{\text{ORC}\_{i.m} \times \text{MRT}} \times \sum\_{t}^{m} \text{DRD}\_{i.l} \times 2 \times \text{DF}\_{i.l} \\ \text{DRD}\_{i.t} &= \text{Max}(\text{RSO}\_{i.t} \times 0.97 - \text{DR}\_{i.t} \text{ } 0) \\ \text{BPC}\_{i.m} &= \text{Min}(\text{DRBP}\_{i.m\prime}^{\*} \text{IBPC}\_{i.m}) \\ \text{FDRBP}\_{i.m} &= \text{DRBP}\_{i.m} - \text{BPC}\_{i.m} \end{aligned} \tag{2}$$

*DRBPi*.*<sup>m</sup>* Demand response basic payment by monthly (KRW) *ORCi*.*<sup>m</sup>* Obligation reduction capacity (MW) *BPm* Basic price by monthly (KRW/kW) *IBPCi*.*<sup>m</sup>* Initial basic penalty charge (KRW) *TDRBPi* Total basic settlement money during the contract period (KRW) *MRT* Maximum reduction time (Max 60 h) *DRDi*.*<sup>t</sup>* Dispatch reduction deficiency (kWh) *RSOi*,*<sup>t</sup>* Reduction ordered by system operator (MWh) *DRi*,*<sup>t</sup>* Dispatched reduction (kWh) *DFi*,*<sup>t</sup>* Dispatch flag (1 for active, 0 for non-active) *BPCi*.*<sup>m</sup>* Basic penalty charge by monthly (KRW) *FDRBPi*.*<sup>m</sup>* Final demand response basic payment by monthly (KRW)

To verify the DSMS functional test, the sampled data is divided into the following three categories: (1) large amount, (2) medium amount, and (3) small amount. Table 5 shows sampled data.


**Table 5.** The data for functional verification of demand-side management solution (DSMS).

Baseline load, peak, average power consumption, and CBL are calculated based on the customer's power usage which is monitored and recorded from an electricity smart meter. Figures 6–8 show the data on 26 June 2017. The large amount, N company, is a chemical company located in the southwestern area of South Korea. The power consumption pattern of N company is a typical factory type. Figure 6 illustrates the power consumption pattern baseline load and peak. The medium amount, a provincial government building, is located in the central area of South Korea. The CBL of this customer is 1938.56 kW at 13:00~14:00 on 26 June 2017, as shown in Figure 7. Lastly, the small amount, G store, is a retail store located in the southeastern area of South Korea. Figure 8 shows the RRMSE value as 9.815%, less than 30%.

**Figure 6.** N company power consumption and baseline load.


**Figure 7.** Provincial government building CBL calculation.


**Figure 8.** G store power RRMSE.

#### **3. Results**

Table 6 shows the DR benefit and reduction rate results for three contracted customers after participating in the DR program. The results show the money-saving and energy conservation using the DSMS. The data used in this paper was collected in two years (from December 2016 to November 2018) from selected customers in South Korea. The N company case is selected to demonstrate the result. In the case of N company, the contracted capacity is 10 MW, participating in the 20 h DR dispatch during the two years. As a result, the DR delivery rate which is an average 108.03% and the benefit from the DR participation which is 854,900,394 KRW (basic settlement benefit is 844,014,870 KRW and actual settlement money is 10,885,524 KRW) are occurring, and the savings in electricity cost over the two years is about 2,160,560 KRW


**Table 6.** Participation result of annual demand response market.


#### **Table 6.** *Cont.*

To verify the capacity of the demand response from the customer, the request from KPX on 20 July 2017, 14:00~17:00, is displayed in Table 7. The large customer, N company, a big factory which has a contracted capacity of 10,000 kW delivered 97%, 93%, and 96% for each period and the average delivery rate was 95%. Figure 9 Illustrates the CBL load, the DR reduction result, and the delivery rate of each period. For example, Figure 9a shows the CBL is 18,059 kW, the real load is 8378 kW, the DR reduction is 9,681 kW, and the contracted capacity is 10,000 kW at 14:00~15:00.

**Figure 9.** N company demand response result (data of 20 July 2017, 14:00~17:00).


**Table 7.** Result of demand response from customer.

#### **4. Discussion**

Unlike the traditional energy management models that focus on the supply side, the DSMS considers the energy demand and control on the interactions between customers and supplier to manage electricity usage reduction and money saving. Through the implementation of the DSMS technology, the end user can automatically check necessary information, such as the CBL, the RRMSE value, and the amount of DR reduction, without the need for complex formulas and contents. Furthermore, the CBL and current usage can be checked in real time by monitoring power usage every 5 min. The DSMS operated in the actual South Korea DR market for a year and based on these results the developed solution was verified. In addition, the result illustrates that the integrated demand-side management solution contributes by participating in the DR market and provides a benefit and satisfaction to the consumer.

The researcher and stakeholders of the DR market should consider the criteria value of the RRMSE. As previously mentioned in Section 2, the U.S. PJM power suggests a RRMSE value less than 20%, but the South Korea DR market sets the value within 30%. This consideration helps an effective DR operation and derives a successful outcome. This unique demand-side manage experience in South Korea could provide the rest of the world with a model to efficiently maintain a national power grid and potentially suggest the development of novel energy managing plans for local situations and policy.

#### **5. Conclusions**

In this paper, a case study of the demand-side management solution in South Korea is introduced and explained. The experience from the consumer and the DR aggregator shows that the integrated demand response solution technologies is a fast-responding approach in a cost-effective way. The curtailed energy from contracted customers contributed by reducing peak power in the national power grid and therefore can effectively provide a reliable power system. The demand resource can be an alternative to the redundant generation in short term such as 5 min. The case of a 10 MW contracted customer shows average 108.03% delivery rate and the total benefit of 854,900,394 KRW for two years. It also shows that all customers regardless of the amount of participation have responded well to 20-h DR dispatch during the two years. The results illustrate that the integrated demand-side management solution contributes by participating in the DR market and gives a benefit and satisfaction to the consumer.

**Author Contributions:** Methodology, W.K. and H.V.; software, H.-J.C. and S.-H.S.; validation, W.K. and H.-J.C.; formal analysis, S.-H.S.; investigation, W.K. and H.V.; resources, W.K., and H.-J.C.; data curation, W.K., and H.-J.C.; writing—original draft preparation, W.K., and H.-J.C.; writing—review and editing, W.K., and H.-J.C.; visualization, H.-J.C.; supervision, W.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors extend their appreciation to the Deanship of Scientific Research at King Saud University for funding this work through research group no. (RG-1439-028).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
