A GIS (geographic information system)-based optimization model for estimating the electricity generation of the rooftop PV (photovoltaic) system

Energy xxx (2013) 1e10

Contents lists available at ScienceDirect

Energy journal homepage: www.elsevier.com/locate/energy

A GIS (geographic information system)-based optimization model for estimating the electricity generation of the rooftop PV (photovoltaic) system Taehoon Hong*, Choongwan Koo, Joonho Park, Hyo Seon Park Department of Architectural Engineering, Yonsei University, 262 Seongsanno, Seodaemun-gu, Seoul 120-749, Republic of Korea

a r t i c l e i n f o

a b s t r a c t

Article history: Received 27 June 2013 Received in revised form 28 November 2013 Accepted 30 November 2013 Available online xxx

The global PV (photovoltaic) generation market has been rapidly growing. In the introduction of a PV system, the electricity generation efficiency of the PV system depends on regional factors and on-site installation factors. It has a significant effect on the returns on investment. Therefore, this study conducted a sensitivity analysis on how the impact factors of the rooftop PV system affect its electricity generation. Based on the results of the sensitivity analysis, this study aimed to ultimately develop a GISbased optimization model for estimating the electricity generation of the rooftop PV system. Several impact factors were used in the sensitivity analysis. The result of this study showed that there were 1.12-, 1.62-, and 1.37-fold differences in the annual electricity generation of the rooftop PV system in South Korea due to the regional factor, the azimuth of the installed panel, and the slope of the installed panel, respectively. Using a GIS-based optimization model, final decision-maker could easily and accurately estimate the electricity generation of the rooftop PV system in a preliminary feasibility study. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: Rooftop photovoltaic system Geographical information system Optimization Sensitivity analysis Electricity generation

1. Introduction The global use of fossil fuels has caused grave environmental crises including energy depletion and pollution and is projected to increase by more than one-third by 2035. Despite this rapid increase in energy usage, coal has a reserves-to-production ratio of about 128 years, natural gas 54 years, and oil 41 years, according to the ‘2010 Survey of Energy Resources.’ Given this background, if the annual increase rate of energy consumption from 2008 to 2035 is assumed to be 1.4%, fossil fuel reserves will be fully depleted within 50 years [1e8]. To overcome such a crisis, there has been growing interest in NRE (new renewable energy) [9e12]. As of 2009, NRE accounted for about 18% of global electricity generation and 25% of the world’s electricity generation facilities (1230 GW of 4800 GW). According to the ‘Medium-Term Renewable Energy Market Report 2012,’ global renewable energy generation will increase by 40% over the period from 2011 to 2017 [13]. Since 2008, the U.S. and the EU have been establishing more renewable energy power plants than conventional fossil energy power plants. According to a press release, renewable energy power plants accounted for about 60% of newly

* Corresponding author. Tel.: þ82 2 2123 5788; fax: þ82 2 365 4668. E-mail address: [email protected] (T. Hong).

installed power plants in the EU and more than 50% in the U.S [14e 17]. Also, the International Energy Agency expects to double the share of renewable energy by 2035, compared to in 2008. In other words, it is forecast to increase from 19% in 2008 to almost 33% by 2035, which means renewable energy could catch up with coal [18]. Regarding the global energy and environmental issue, solar energy (including PV (photovoltaic) energy) is recognized to play an important role in the renewable and sustainable development [19e25]. The interest in PV energy has been rapidly increased as the radioactive pollution and energy storage issues were raised in the wake of the nuclear-power-plant accident in Fukushima, Japan [26]. The PV market was only 7.2 GW in 2009, but it was increased more than twofold to 16.6 GW in 2010. As of 2011, the installation capacity of global PV system went up to 40 GW. Especially, with the continuous downward trend in the cost of the PV system, it is expected that the PV market would be expanded to achieve the netzero energy buildings and the carbon emissions reduction target [27e31]. Keeping pace with such global trend, the South Korean government is promoting various incentive policies such as financial support for NRE projects, the 1 Million Green Homes Project, and the Feed-in-Tariff [32e35]. Among NREs, the PV system has the highest potential as a sustainable energy source. Particularly, crystalline silicon, commonly used by the semiconductor industry, is the material used in 94% of all PV modules today. Thus, it is

0360-5442/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.energy.2013.11.082

Please cite this article in press as: Hong T, et al., A GIS (geographic information system)-based optimization model for estimating the electricity generation of the rooftop PV (photovoltaic) system, Energy (2013), http://dx.doi.org/10.1016/j.energy.2013.11.082

2

T. Hong et al. / Energy xxx (2013) 1e10

model, final decision-maker could easily and accurately estimate the electricity generation of the rooftop PV system in a preliminary feasibility study. This study was conducted in three steps: (i) the key factors affecting the electricity generation of the rooftop PV system were selected through an extensive literature review and interviews with experts; (ii) a sensitivity analysis on how the impact factors of the rooftop PV system affect its electricity generation was conducted through an energy simulation; and (iii) using the GIS (geographic information system) and genetic algorithm, the optimal annual electricity generation of the rooftop PV system and the corresponding optimal SoP were visually proposed by region in South Korea.

considered that the PV system has great potential in the export market since South Korea is one of leading countries in semiconductor technology [36]. In spite of the various advantages of the PV system including the government’s financial support, the decrease in the systems’ unit cost, and their high potentials, one major obstacle remains: the high initial investment cost. Therefore, it is crucial to assess the ROI (return on investment) in introducing the PV system by accurately estimating its electricity generation. The electricity generation of the PV system is affected by various impact factors such as (i) regional geographical information; (ii) regional meteorological information; and (iii) on-site installation information. Therefore, a sensitivity analysis should be conducted on how the impact factors of the rooftop PV system affect its electricity generation. In previous studies, various analyses have been conducted on the impact factors of the PV system [37e60]. Zhao et al. [37] analyzed that the optimal installation angle of a PV panel depended on the installation location. Dincer and Meral [38] analyzed the temperature as a factor that impacts the efficiency of a PV solar cell. It was concluded that the temperature of a solar cell should be kept in a lower temperature as the optimal condition. Hwang et al. [39] also conducted an optimization analysis of the building integrated PV system in office buildings. To generate the maximum amount of electricity, they analyzed various factors, including the azimuth and slope of the installed panel, and the installation distance to the module length ratio. Siraki and Pillay [40] focused on the change of the azimuth and slope of the installed panel depending on both the regional latitude and the surrounding buildings. While these previous studies individually analyzed the impact factors of the PV system, they failed to comprehensively analyze them. Therefore, the objective of this study is to conduct a comprehensive sensitivity analysis of how the electricity generation of the PV system would change from the complex interaction of the aforementioned impact factors. Based on the results of the sensitivity analysis, this study aims to ultimately develop a GISbased optimization model for estimating the electricity generation of the rooftop PV system. Using a GIS-based optimization

2. Materials and methods 2.1. Definition of the impact factors of the rooftop PV system The following impact factors of the rooftop PV system were derived from an extensive literature review and interviews with experts (i.e., solar photovoltaic staff of Parsons Brinckerhoff, Oerlikon Solar, and Hilti Corporation): (i) regional geographical information; (ii) regional meteorological information; and (iii) on-site installation information. As shown in Table 1, the regional geographical information is classified by latitude and monthly meridian altitude; the regional meteorological information is classified by the MADSR (monthly average daily solar radiation) and the monthly average temperature; and the on-site installation information is classified by the AoP (azimuth of the installed panel), the SoP (slope of the installed panel), and the type of the panel and inverter. The data on the impact factors were collected in 78 regions in South Korea. However, the MADSR data were measured at 24 of the 78 weather stations nationwide (refer to Table S1). The MADSR data at the 54 other weather stations were gleaned from the results of previous studies (refer to Table S2). Koo et al. [61], as in previous research, developed an A-CBR (advanced Case-Based Reasoning) model for MADSR estimation using the monthly geographic and

Table 1 Reviews of previous studies on the impact factors of the rooftop PV system. Variables

Attributes

Independent variable

Regional factors

Geographical information

Meteorological information

On-site installation factor

Target variable

e

On-site installation information

e

Detailed description

Reference

Latitude

( ) N

Monthly meridian altitude Monthly average daily solar radiation Monthly average temperature The azimuth of the installed panel (AoP)

()

()



The slope of the installed panel (SoP)

()



Type of the panel Type of the inverter Electricity generation

() () ( ) kWh

Crook et al. [42], Hummon et al. [56], Levinson et al. [43], Liu et al. [44], Siraki and Pillay [40], Tang and Wu [51], Tiris and Tiris [52], Zhao et al. [37] Kaldellis and Zafirakis [54], Li and Lam [48], Zhao et al. [37] Crook et al. [42], Kaldellis and Zafirakis [54], Li and Lam [48], Siraki and Pillay [40], Braun et al. [57], Dincer and Meral [38], Green [59], Hoffmann and Koehl [58] Asowata et al. [53], Gopinathan et al. [50], Gunerhan and Hepbasli [49], Hummon et al. [56], Hussein et al. [46], Jafarkazemi and Saadabadi [47], Kaldellis and Zafirakis [54], Levinson et al. [43], Li and Lam [48], Siraki and Pillay [40], Tang and Wu [51], Ubertini and Desideri [45] Asowata et al. [53], Boji c et al. [55], Gunerhan and Hepbasli [49], Huld et al. [60], Hummon et al. [56], Hussein et al. [46], Hwang et al. [39], Jafarkazemi and Saadabadi [47], Kaldellis and Zafirakis [54], Li and Lam [48], Liu et al. [44], Siraki and Pillay [40], Tang and Wu [51], Tiris and Tiris [52], Ubertini and Desideri [45], Zhao et al. [37] Li and Lam [48], Ordóñez et al. [41] Li and Lam [48], Ordóñez et al. [41] e



( ) kWh/m2/day ( ) C

Please cite this article in press as: Hong T, et al., A GIS (geographic information system)-based optimization model for estimating the electricity generation of the rooftop PV (photovoltaic) system, Energy (2013), http://dx.doi.org/10.1016/j.energy.2013.11.082

T. Hong et al. / Energy xxx (2013) 1e10

meteorological data measured in 15 regions for 10 years. Also, Hong et al. [62] estimated the MADSR at the 54 unmeasured locations in South Korea by using the A-CBR model developed by Koo et al. [61]. A total of 78 regional data (consisting of the 24 locations with the measured MADSR data and the 54 locations with the estimated MADSR data) were used as the data for the MADSR in this study. For detailed information on the algorithms for the A-CBR model, refer to the Koo et al. [61] and Hong et al. [62]. Using the meteorological data in 78 weather stations nationwide (which area measured by the Korea Meteorological Administration), the monthly average temperature was collected (refer to Table S3). The latitudes in 78 regions were collected from the geographical information offered by Google EarthÔ. Also, the monthly meridian altitude was calculated using Eq. (1) (refer to Table S4) [63].

s ¼ 90+  4  ε

(1)

where, s is the minimum incidence angle of the sun (monthly meridian altitude by region), 4 is the regional latitude, and ε is the tilt angle of the Earth’s axis by month (e.g., the tilt angles at the summer and winter solstice are 23.5 and 23.5 , respectively). 2.2. Sensitivity analysis on the impact factors of the rooftop PV system A sensitivity analysis on how the impact factors of the rooftop PV system affect its electricity generation was conducted using the software program called “RETScreen,” which was co-developed by specialists from the Department of Natural Resources in Canada and the United Nations Environment Programme. First, a sensitivity analysis was conducted depending on the regional factor, the AoP, and the SoP. Next, a sensitivity analysis was conducted by simultaneously considering the AoP and the SoP. 2.3. Geographical analysis of the optimal annual electricity generation of the rooftop PV system 2.3.1. Optimization process using a genetic algorithm To identify the optimal annual electricity generation of the rooftop PV system in 78 regions in South Korea, a genetic algorithm was applied to establish the optimization process. The following basic information was used to conduct the energy simulation using the software program called “RETScreen”: (i) regional geographic and meteorological information in 78 regions in South Korea (refer to Tables S1eS4); and (ii) regarding the on-site installation information, the AoP set at 0 (southward) and the types of the panel and the inverter set at the specifications shown in Table S5. On the basis of the above information, the annual electricity generation of the rooftop PV system was set as the optimization goal, and the SoP affecting the optimization goal was set as the optimization parameter. Accordingly, the optimization process was established (refer to Table 2). 2.3.2. Mapping of the optimal annual electricity generation of the rooftop PV system The software program called ‘ArcMap 10.1’ from ‘ArcGIS 10.1’ can be used to develop the map of the optimal annual electricity generation of the rooftop PV system [61,64]. Particularly, this study used kriging, among the various spatial interpolation techniques offered by ‘ArcGIS 10.1.’ Reflecting not only the distance from the measured value but also the statistical correlation between the measured values, kriging is useful in determining the overall tendency of a given region. The map was presented with an interval of 15 colors so that the final decision-maker could easily use it.

3

3. Results and discussion 3.1. Sensitivity analysis depending on the regional factor The sensitivity analysis depending on the regional factor was conducted in terms of two aspects: (i) regional geographical information and (ii) regional meteorological information. Since these factors were selected as independent variables, the on-site installation factor was set as the control variable. In other words, the AoP and the SoP were set at 0 (southward) and 30 , respectively, and the types of the panel and the inverter were set according to the specifications shown in Table S5. To determine the trend in the monthly electricity generation of the rooftop PV system by region, Seoul, Daejeon, and Busan were selected, which are in the northern, central, and southern parts of South Korea, respectively. As shown in Fig. S1, the monthly electricity generation tends to increase from the northern part to the southern part. In Seoul, the northern part of South Korea, the monthly electricity generation was lowest among the three regions in all 12 months. The monthly electricity generation of Busan, the southern part of South Korea, was higher than that of Daejeon, except for the electricity generation in May, June, and October. The annual electricity generation of Seoul was predicted at 245.51 (kWh/EA) and that of Busan was estimated at 275.19 (kWh/EA). A 1.12-fold difference in the annual electricity generation was shown depending on the changes in the regional factors. Such result was determined to have occurred due to the complex interaction among the regional geographic and meteorological information. In greater detail, the difference in the latitude by region causes the difference in the monthly meridian altitude (refer to Table S6). Such a difference causes the variation in the incidence angle of the sun. In other words, the difference in the latitude causes the difference in the MADSR (refer to Tables S1eS2 and Fig. S2) and the monthly average temperature (refer to Table S3 and Fig. S3). Ultimately, the difference in the monthly electricity generation by region is caused by the complex interaction among such regional factors. To determine the correlation between the aforementioned regional factors and the monthly electricity generation by region as well as between the regional factors, a statistical analysis was conducted. Table 3 shows the results of the correlation analysis of these factors using a total of 192 cases from the database, which covered 12 months in 2011 in 16 administrative divisions in South

Table 2 Establishment of the optimization process using a genetic algorithm. Classification

Variables

Detailed description

Optimization goal

Annual electricity generation SoP

Target: maximization

Optimization parameter Constraint parameters

Latitude Monthly meridian altitude MADSR Monthly average temperature AoP Type of the panel Type of the inverter Number of the panels

Adjustable parameter Regional value (refer to Table S4) Regional value (refer to Table S6) Regional value (refer to Tables S1 and S2) Regional value (refer to Table S3) S: 0 (southward) SM-200 PD0 (refer to Table S5) PV-C340 S/H (refer to Table S5) 1

Note: SoP stands for the slope of the installed panel; MADSR stands for the monthly average daily solar radiation; and AoP stands for the azimuth of the installed panel.

Please cite this article in press as: Hong T, et al., A GIS (geographic information system)-based optimization model for estimating the electricity generation of the rooftop PV (photovoltaic) system, Energy (2013), http://dx.doi.org/10.1016/j.energy.2013.11.082

4

T. Hong et al. / Energy xxx (2013) 1e10

Table 3 Correlation analysis on the impact factors of the rooftop PV system. Variables MEG Lat. MMA MADSR MAT

Pearson correlation Sig. (2-tailed) Pearson correlation Sig. (2-tailed) Pearson correlation Sig. (2-tailed) Pearson correlation Sig. (2-tailed) Pearson correlation Sig. (2-tailed)

MEG

Lat.

MMA

MADSR

MAT

1 e .049 .501 .254** .000 .258** .000 .134 .064

.049 .501 1 e .080 .270 .083 .252 .093 .202

.254** .000 .080 .270 1 e .915** .000 .757** .000

.258** .000 .083 .252 .915** .000 1 e .737** .000

.134 .064 .093 .202 .757** .000 .737** .000 1 e

Note: MEG stands for the monthly electricity generation (kWh/EA); Lat. stands for the latitude ( N); MMA stands for the monthly meridian altitude ( ); MADSR stands for the monthly average daily solar radiation (kWh/m2/day); and MAT stands for the monthly average temperature ( C). The shaded areas stand for the cells that have been explained in detail in the body of this study. The asterisks (**) mean that the correlation coefficient is significant at 0.01 level (both sides).

Korea. Some of the more notable results are as follows: (i) the greatest impact factor of the monthly electricity generation was the MADSR, which was positively correlated with it (correlation coefficient: 0.258); (ii) the monthly average temperature was negatively correlated with the monthly electricity generation (correlation coefficient: 0.134); and (iii) the MADSR and the monthly average temperature were positively correlated with the monthly meridian altitude (correlation coefficients: 0.915 and 0.757, respectively). According to the above results, the MADSR and the monthly average temperature were inversely correlated with the monthly electricity generation (refer to the results (i) and (ii)), however, were proportional correlated with the monthly meridian altitude (refer to the result (iii)). In summary, each regional factor has a very complex correlation with the monthly electricity generation. Thus, the regional factors such as latitude, monthly meridian altitude, MADSR, and monthly average temperature should be considered comprehensively. By considering these regional factors, the optimal location that could maximize the ROI could be determined. 3.2. Sensitivity analysis depending on the on-site installation factor The sensitivity analysis depending on the on-site installation factors was conducted in terms of two aspects: (i) the AoP and (ii)

the SoP. Since these two factors were selected as independent variables, the regional factors were set as the control variables. In other words, Seoul, the capital of South Korea located in the northern part of the country, was set as the installation region, and the types of the panel and the inverter were set according to the specifications shown in Table S5. 3.2.1. Sensitivity analysis depending on the AoP The sensitivity analysis was conducted depending on the AoP. Since the AoP was selected as an independent variable, the SoP was set as the control variable. Considering the results of the study of Ju et al. [65], the SoP was set at 30 in Seoul. The AoP was divided by 15 from 0 (southward) to 180 (northward) (because the monthly electricity generation in the range of 0 e180 is identical to that in the range of 180 e360 , the AoP between 180 and 360 was excluded from the sensitivity analysis). Fig. 1 shows the trend in the monthly electricity generation depending on the AoP. The detailed analysis results based on the interval of 15 azimuth are shown in Table S7 and are summarized as follows. First, the annual electricity generation was highest at 245.51 (kWh/EA) when the AoP was set at 0 (southward). On the other hand, it was lowest at 151.49 (kWh/EA) when the AoP was set at 180 (northward). There was a 1.62-fold difference in the annual electricity generation depending on the AoP. The latitude of South Korea is between 33 N and 43 N; and Jeju, its southernmost region that is at the latitude of 33.29 N, has the highest monthly meridian altitude, which is 82.21 at noon of the summer solstice (refer to Table S6). In other words, the path of the sun lies entirely in the southern sky throughout the year in South Korea. Accordingly, under the same condition, the PV system could acquire the most MADSR when the AoP is 0 (southward), and therefore, it has the highest annual electricity generation. Second, the deviation in the monthly electricity generation in the summer season (especially from June to August) depending on the AoP was shown to be smaller than that in the winter season. This result is related to the monthly meridian altitude, which in Seoul at noon of the summer solstice is 76.20 and at noon of the winter solstice, 29.20 (refer to Table S6). In other words, while the MADSR in summer is sufficient, the MADSR in winter is not. Accordingly, the monthly electricity generation in winter is very sensitive to the AoP. Therefore, it is determined that the ROI of the rooftop PV system could be maximized by considering the AoP based on a given region’s monthly meridian altitude.

Fig. 1. Monthly electricity generation of the rooftop PV system depending on the AoP (in Seoul as the northern part).

Please cite this article in press as: Hong T, et al., A GIS (geographic information system)-based optimization model for estimating the electricity generation of the rooftop PV (photovoltaic) system, Energy (2013), http://dx.doi.org/10.1016/j.energy.2013.11.082

T. Hong et al. / Energy xxx (2013) 1e10

According to the above results, it is considered that the annual electricity generation would be highest when the AoP is 0 (southward) regardless of the region. To verify this assumption, a sensitivity analysis was conducted based on the identical process, targeting Busan, the second largest city in South Korea and located in the southern part. The result was very similar to that in Seoul (refer to Fig. S4 and Table S8), which shows that the trend of the monthly electricity generation of the rooftop PV system based on the AoP will appear similarly in most regions in South Korea. Fig. 2 shows the trends in the annual electricity generation depending on the AoP in Seoul (northern part) and Busan (southern part), simultaneously. As in the preceding analysis results, it was determined that the trend of the annual electricity generation depending on the AoP in the two regions were very similar. Moreover, the annual electricity generation tends to increase from the northern to the southern part. 3.2.2. Sensitivity analysis depending on the SoP The sensitivity analysis was conducted depending on the SoP. Since the SoP was selected as an independent variable, the AoP was set as the control variable. As shown in Section 3.2.1, the AoP was set at 0 (southward), at which the electricity generation efficiency of the rooftop PV system was highest. Table 4 shows the trend of monthly electricity generation depending on the SoP. The SoP was divided by 5 from 0 to 90 . Between 30 and 40 , however, where it is deemed that a detailed analysis would be required, the SoP was divided further by 1. The key findings are as outlined below. First, the annual electricity generation was highest at 246.13 (kWh/EA) when the SoP was set at 35 . On the other hand, it was lowest at 179.87 (kWh/EA) when the SoP was set at 90 . There was a 1.37-fold difference in the annual electricity generation depending on the change in the SoP. Second, the monthly electricity generation differed depending on the SoP. That is, in summer, when the incidence angle of the sun rises, the SoP was smaller. Such a result is

5

related to the monthly meridian altitude; and the monthly electricity generation is generally highest when the incidence angle of the sun is perpendicular to the panel. Therefore, it was determined that the ROI of the rooftop PV system would be maximized by considering the SoP based on the monthly meridian altitude of a given region. According to the above results, it is considered that such a trend in the monthly electricity generation depending on the SoP would be identical in most of the regions. To verify this, a sensitivity analysis was conducted based on the identical process, targeting Daejeon (the central part) and Busan (the southern part). As shown in Table 5, the optimal SoP decreased from the northern part to the southern part. That is, the annual electricity generation is at its maximum (246.27 kWh/EA) in Seoul (the northern part) when the SoP was set at 35.327. The annual electricity generation is at its maximum (269.61 kWh/EA) in Daejeon (the central part) when the SoP was set as 34.779 while the annual electricity generation is at its maximum (275.33 kWh/EA) in Busan (the southern part) when the Sop was set as 33.029 . 3.2.3. Sensitivity analysis depending on both the AoP and the SoP The optimal SoP might change depending on the AoP. A sensitivity analysis on the optimal SoP in Seoul was conducted on the condition that the AoP was divided by 15 from 0 to 90 . From the perspective that annual electricity generation was maximized, Table 6 shows the optimal SoP depending on the AoP. As the AoP increases from 0 (southward) to 90 (eastward), the SoP decreases from 35.327 to 0 . When the AoP was 0 (southward) and the SoP was 35.327, the annual electricity generation was highest at 246.27 (kWh/EA). On the other hand, when the AoP was 90 (eastward) and the SoP was 0 , the annual electricity generation was lowest at 217.39 (kWh/EA). There was a 1.13-fold difference in the annual electricity generation depending on both the AoP and SoP. Such a result was due to the absolute time in which the rooftop PV system can receive the MADSR. In other words, as the AoP increases from 0 (southward) to 90 (eastward), the absolute time in which the rooftop PV system can receive the MADSR decreases, subsequently reducing the annual electricity generation. Therefore, in introducing the rooftop PV system, the AoP and SoP should be simultaneously considered. For example, if the shape of the rooftop area is not quadrangular or if the orientation of the given building is not 0 (southward), the AoP should be adjusted. In this case, it is determined that the ROI of the rooftop PV system would be maximized by considering both the AoP and the SoP. 3.3. Geographical analysis of the optimal annual electricity generation of the rooftop PV system

Fig. 2. Annual electricity generation of the rooftop PV system depending on the AoP in Seoul and Busan (by region).

According to the optimization process mentioned in Section 2.3.1, the optimal annual electricity generation of the rooftop PV system in the 78 regions in South Korea was identified (refer to Table S9). Fig. 3 shows the map of the optimal annual electricity generation of the rooftop PV system by region in South Korea. Fig. 4 shows the map of the corresponding optimal SoP by region in South Korea. The regions in red (in web version) are the 24 regions with the measured MADSR data, while the regions in pink are the 54 regions where the MADSR data was not measured. In the 54 regions, the MADSR estimated by previous studies (Koo et al. [61] and Hong et al. [62]) were used. The overall trend of optimal annual electricity generation, as shown in Fig. 3, has a broad spectrum of 250e270 (kWh/EA). With Seoul, Daejeon, and Busan as the key axes, the optimal annual electricity generation increased from the northwestern part to the southeastern part. Under the identical condition, the ROI could be maximized if the rooftop PV system is

Please cite this article in press as: Hong T, et al., A GIS (geographic information system)-based optimization model for estimating the electricity generation of the rooftop PV (photovoltaic) system, Energy (2013), http://dx.doi.org/10.1016/j.energy.2013.11.082

6

T. Hong et al. / Energy xxx (2013) 1e10

Table 4 Monthly electricity generation of the rooftop PV system depending on the SoP (in Seoul as the northern part). SoP ( )

0 5 10 15 20 25 30 31 32 33 34 35 36 37 38 39 40 45 50 55 60 65 70 75 80 85 90

AEG (kWh/EA)

217.41 224.86 231.23 236.49 240.62 243.65 245.49 245.71 245.89 246.02 246.10 246.13 246.12 246.06 245.95 245.79 245.59 243.84 240.92 236.84 231.71 225.51 218.25 209.96 200.73 190.68 179.87

Monthly electricity generation (kWh/EA) Jan.

Feb.

Mar.

Apr.

May

Jun.

Jul.

Aug.

Sep.

Oct.

Nov.

Dec.

15.95 17.74 19.43 21.04 22.54 23.93 25.20 25.43 25.67 25.90 26.12 26.34 26.55 26.76 26.96 27.16 27.35 28.22 28.95 29.53 29.95 30.23 30.35 30.31 30.12 29.77 29.26

15.17 16.14 17.02 17.81 18.50 19.08 19.57 19.65 19.73 19.80 19.87 19.94 20.00 20.06 20.11 20.16 20.20 20.35 20.39 20.32 20.14 19.84 19.44 18.93 18.31 17.60 16.78

24.73 25.74 26.62 27.36 27.95 28.39 28.68 28.72 28.76 28.78 28.81 28.82 28.83 28.83 28.83 28.82 28.81 28.64 28.33 27.86 27.24 26.49 25.59 24.55 23.39 22.10 20.70

24.00 24.39 24.67 24.83 24.87 24.79 24.60 24.54 24.49 24.42 24.36 24.29 24.21 24.13 24.04 23.95 23.86 23.32 22.67 21.91 21.05 20.10 19.06 17.92 16.75 15.51 14.20

24.46 24.54 24.50 24.36 24.14 23.82 23.40 23.30 23.20 23.09 22.98 22.87 22.75 22.63 22.50 22.37 22.24 21.51 20.69 19.78 18.81 17.80 16.72 15.57 14.37 13.12 11.91

22.39 22.37 22.25 22.04 21.73 21.37 20.91 20.81 20.71 20.60 20.48 20.37 20.25 20.13 20.00 19.87 19.74 19.03 18.24 17.39 16.51 15.57 14.58 13.54 12.46 11.43 10.36

15.61 15.63 15.59 15.47 15.30 15.08 14.80 14.73 14.67 14.60 14.53 14.46 14.38 14.30 14.22 14.14 14.06 13.60 13.10 12.54 11.96 11.35 10.69 10.01 9.29 8.58 7.87

17.14 17.26 17.31 17.30 17.22 17.06 16.84 16.78 16.73 16.67 16.61 16.54 16.47 16.40 16.33 16.26 16.18 15.75 15.27 14.72 14.12 13.46 12.78 12.05 11.28 10.48 9.65

20.45 21.00 21.45 21.80 22.03 22.16 22.18 22.18 22.16 22.15 22.12 22.10 22.07 22.03 21.99 21.95 21.90 21.60 21.20 20.69 20.08 19.37 18.57 17.69 16.72 15.67 14.55

17.23 18.13 18.93 19.63 20.23 20.72 21.11 21.17 21.23 21.28 21.33 21.38 21.42 21.46 21.49 21.52 21.54 21.59 21.52 21.35 21.06 20.66 20.15 19.54 18.82 18.01 17.10

9.30 9.85 10.36 10.81 11.21 11.55 11.84 11.89 11.94 11.98 12.02 12.06 12.10 12.14 12.17 12.20 12.23 12.33 12.36 12.34 12.25 12.10 11.89 11.62 11.29 10.90 10.46

10.99 12.07 13.09 14.03 14.90 15.68 16.38 16.50 16.63 16.75 16.87 16.98 17.09 17.20 17.30 17.40 17.49 17.90 18.21 18.42 18.53 18.53 18.43 18.23 17.92 17.52 17.01

Note: SoP stands for the slope of the installed panel; AEG stands for the annual electricity generation per unit. The shaded areas stand for the cells that have been explained in detail in the body of this study.

installed in the southeastern part. The overall trend of the optimal SoP, as shown in Fig. 4, has a broad spectrum of 23.3 e38.2 . It was shown that the optimal SoP decreased from the northern part to the southern part. As explained in Section 3.2.2, it was determined that the meridian altitude by region was considered in such a result. Meanwhile, without energy simulation, the optimal annual electricity generation of the rooftop PV system in regions other than the 78 regions shown in Table S9 can be easily determined using Fig. 3. If the power capacity of the rooftop PV system will improve, the result can be converted by considering that the power capacity of the rooftop PV system used in this study was 200 W (refer to Table S5). For example, if the power capacity of the rooftop PV system improves to 300 W, the ratio of 1.5 (300 W: 200 W) can be applied to the annual electricity generation by region, as shown in Fig. 3. It can easily determine the annual electricity generation of a given region where the rooftop PV system will be installed. In addition, the corresponding optimal SoP can be easily determined using Fig. 4. Such an analysis can be used in a preliminary feasibility study for the implementation of the rooftop PV system. 4. Conclusions This study aimed to conduct a sensitivity analysis on how the impact factors of the rooftop PV system affect its electricity

generation. Based on the results of the sensitivity analysis, this study aimed to ultimately develop a GIS-based optimization model for estimating the electricity generation of the rooftop PV system. The results of this study are summarized as follows.  First, it was found that the higher the MADSR is, the higher the annual electricity generation of the rooftop PV system becomes. In other words, the southern part with a low latitude showed a higher MADSR, and thus a higher annual electricity generation by the rooftop PV system under the identical condition. There was a 1.12-fold difference in the annual electricity generation depending on the regional factor.  Second, the annual electricity generation of the rooftop PV system changes depending on the AoP. For example, in Seoul (the northern part), there was a 1.62-fold difference in the annual electricity generation depending on the AoP. When the AoP was 0 (southward), it can receive the MADSR for the longest time under the identical condition, and therefore, the annual electricity generation of the rooftop PV system was highest. Moreover, the region with a higher monthly meridian altitude (i.e., the southern part) showed a wider-ranging change in the annual electricity generation of the rooftop PV system depending on the AoP. In other words, it was shown to be more sensitive.

Table 5 The optimal SoP and annual electricity generation by region. Region

SoP ( )

AEG (kWh/EA)

Monthly electricity generation (kWh/EA) Jan.

Feb.

Mar.

Apr.

May

Jun.

Jul.

Aug.

Sep.

Oct.

Nov.

Dec.

Seoul Daejeon Busan

35.327 34.779 33.029

246.27 269.61 275.33

26.44 27.00 27.41

19.96 20.87 20.00

28.84 30.79 30.98

24.25 26.02 27.57

22.82 23.89 22.98

20.34 22.95 20.87

14.42 17.54 23.34

16.50 18.13 19.39

22.10 22.47 22.37

21.44 23.92 22.86

12.07 16.48 15.80

17.09 19.56 21.77

Note: SoP stands for the slope of the installed panel; AEG stands for the annual electricity generation per unit panel. The optimization process can be found in Section 2.3.1. The shaded areas stand for the cells that have been explained in detail in the body of this study.

Please cite this article in press as: Hong T, et al., A GIS (geographic information system)-based optimization model for estimating the electricity generation of the rooftop PV (photovoltaic) system, Energy (2013), http://dx.doi.org/10.1016/j.energy.2013.11.082

T. Hong et al. / Energy xxx (2013) 1e10

7

Table 6 The optimal SoP and annual electricity generation depending on the AoP (in Seoul as the northern part). AoP ( )

0 15 30 45 60 75 90

SoP ( )

35.327 34.364 32.092 28.598 21.981 10.650 0.000

AEG (kWh/EA)

246.27 244.48 240.06 233.27 225.64 219.18 217.39

Monthly electricity generation (kWh/EA) Jan.

Feb.

Mar.

Apr.

May

Jun.

Jul.

Aug.

Sep.

Oct.

Nov.

Dec.

26.44 25.82 24.20 22.00 19.35 16.81 15.95

19.96 19.66 18.87 17.77 16.54 15.44 15.16

28.84 28.59 28.03 27.09 25.98 25.03 24.73

24.25 24.21 24.17 24.01 23.89 23.87 23.98

22.82 22.91 23.20 23.48 23.84 24.27 24.45

20.34 20.46 20.78 21.14 21.62 22.18 22.40

14.42 14.48 14.65 14.84 15.13 15.47 15.60

16.50 16.52 16.60 16.67 16.81 17.02 17.12

22.10 22.00 21.75 21.32 20.86 20.50 20.45

21.44 21.20 20.60 19.67 18.60 17.60 17.26

12.07 11.90 11.44 10.81 10.09 9.47 9.28

17.09 16.73 15.77 14.48 12.94 11.51 11.01

Note: AoP stands for the azimuth of the installed panel; SoP stands for the slope of the installed panel; and AEG stands for the annual electricity generation per unit panel. The optimization process can be found in Section 2.3.1. The shaded areas stand for the cells that have been explained in detail in the body of this study.

 Third, the annual electricity generation of the rooftop PV system changes depending on the SoP. For example, in Seoul (the northern part), there was a 1.37-fold difference in the annual electricity generation depending on the SoP. The region with a higher monthly meridian altitude (i.e., the southern part)

showed a lower SoP, which can maximize the annual electricity generation of the rooftop PV system.  Fourth, even in the same region, the optimal SoP, which can maximize the annual electricity generation of the rooftop PV system, was found to change depending on the AoP. For

Fig. 3. Map of the optimal annual electricity generation of the rooftop PV system by region in South Korea.

Please cite this article in press as: Hong T, et al., A GIS (geographic information system)-based optimization model for estimating the electricity generation of the rooftop PV (photovoltaic) system, Energy (2013), http://dx.doi.org/10.1016/j.energy.2013.11.082

8

T. Hong et al. / Energy xxx (2013) 1e10

Fig. 4. Map of the optimal SoP by region in South Korea.

example, in Seoul (the northern part), there was a 1.13-fold difference in the annual electricity generation depending on both the AoP and the SoP. Such a result was found to be related to the absolute time in which the rooftop PV system can receive the MADSR.  Fifth, it was shown that using the map of the optimal annual electricity generation of the rooftop PV system by region in South Korea, the final decision-maker could easily and accurately estimate the optimal annual electricity generation in a preliminary feasibility study without conducting an energy simulation. If the power capacity of the rooftop PV system improves, the result can be converted by considering that the power capacity of the rooftop PV system that was used in this study was 200 W. In conclusion, the results of this study showed that the southern part of South Korea (which has a lower latitude) has a higher monthly meridian altitude, which results in a higher MADSR and

ultimately, higher annual electricity generation of the rooftop PV systemdwhich translates to a superior ROI in the introduction of the rooftop PV system. Meanwhile, this study has the following limitations: (i) the electricity price is an extremely important determinant in the preliminary feasibility study on the rooftop PV system. It should be considered to develop an optimization model that is capable of maximizing the ROI in the introduction of the rooftop PV system in the future research. The electricity price is the same everywhere in South Korea. Regardless of the region, the electricity price is depending on the customer type (residential, industrial, and commercial); and (ii) the mandatory amount of solar energy supply should be considered in terms of the political feasibility. According to the ‘Mandatory Renewable Energy Installation Program [66]’ in South Korea, the public building should supply over 10% of energy consumptions in the operation and maintenance phase by introducing an NRE system. This mandatory amount can be realistically feasible. It should be considered to conduct the analysis of the

Please cite this article in press as: Hong T, et al., A GIS (geographic information system)-based optimization model for estimating the electricity generation of the rooftop PV (photovoltaic) system, Energy (2013), http://dx.doi.org/10.1016/j.energy.2013.11.082

T. Hong et al. / Energy xxx (2013) 1e10

potential of the rooftop PV system to achieve the net-zero energy solar building in the future research. To address these limitations, the research teams have currently conducted the follow-up studies: (i) to develop an economic and environmental optimization model for determining the optimal solution for the rooftop PV system [67]; and (ii) to develop a framework for the analysis of the potential of the rooftop PV system to achieve the net-zero energy solar building, which aims to help policymakers or facility managers to conduct an energy supply and demand analysis as well as to propose an energy supply and demand strategy [68]. Acknowledgments This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIP; Ministry of Science, ICT & Future Planning) (No. NRF2012R1A2A1A01004376 & No. 2011-0018360). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.energy.2013.11.082. References [1] Kyoto protocol to the United Nations framework convention on climate change. UN: United Nations (UN); 1998. [2] Climate change 2007. Intergovernmental Panel on Climate Change (IPCC); 2007. Synthesis report. [3] 2010 survey of energy resources. London: World Energy Council (WEC); 2010. [4] Hong T, Kim J, Koo C. LCC and LCCO2 analysis of green roofs in elementary schools with energy saving measures. Energy Build 2012;45(2):229e39. [5] Replacement fossil fuel by solar energy. Korea: Korea Electric Association (KEA); 2000. [6] Greenhouse gas emission trends and projections in Europe 2011: tracking progress towards Kyoto and 2020 targets. Copenhagen: European Environment Agency (EEA); 2011. [7] U.S. climate action report 2010: fifth national communication of the United States of America under the United Nations Framework Convention on Climate Change. Washington, DC: U.S. Department of State (DOS); 2010. [8] Jones R, Yoo B. Korea’s green growth strategy: mitigating climate change and developing new growth engines. OECD Economics Department Working Papers; 2010. p. 54. [9] Global market outlook for photovoltaics until 2016. Brussels: European Photovoltaic Industry Association (EPIA); 2012. [10] Renewable energy market and policy trends in IEA countries. Paris: International Energy Agency (IEA); 2009. [11] Urban BIPV in the new residential construction industry. Ottawa: International Energy Agency (IEA); 2008. [12] PV status report 2011. Italy: Joint Research Centre (JRC); 2011. [13] Medium-term renewable energy market report 2012. Paris: International Energy Agency (IEA); 2012. [14] The fourth carbon budget: reducing emissions through the 2020s. Committee on Climate Change (CCC); London: 2020. [15] Annual energy review 2011. Washington, DC: U.S. Energy Information Administration (EIA); 2012. [16] UK emissions statistics. London: Department of Energy & Climate Change (DECC); 2012. [17] Energy technology perspectives 2012. Paris: International Energy Agency (IEA); 2012. [18] World energy outlook 2010. Paris: International Energy Agency (IEA); 2010. [19] Li DHW, Yang L, Lam JC. Impact of climate change on energy use in the built environment in different climate zones e a review. Energy 2012;42(1):103e12. [20] Zhai P, Larsen P, Millstein D, Menon S, Masanet E. The potential for avoided emissions from photovoltaic electricity in the United States. Energy 2012;47(1):443e50. [21] Li DHW, Yang L, Lam JC. Zero energy buildings and sustainable development implications e a review. Energy 2013;54(1):1e10. [22] Milan C, Bojesen C, Nielsen MP. A cost optimization model for 100% renewable residential energy supply systems. Energy 2012;48(1):118e27. [23] Paudel AM, Sarper H. Economic analysis of a grid-connected commercial photovoltaic system at Colorado State University-Pueblo. Energy 2013;52(1): 289e96. [24] Lam KH, Lai TM, Lo WC, To WM. The application of dynamic modelling techniques to the grid-connected PV (photovoltaic) systems. Energy 2012;46(1):264e74.

9

[25] Trends in photovoltaic applications: survey report of selected IEA countries between 1992 and 2011. Photovoltaic Power Systems Programme (PVPS). International Energy Agency (IEA); 2012. [26] Holt M, Campbell RJ, Nikitin MB. Fukushima nuclear disaster. Congressional Research Service (CRS); 2012. [27] Clean energy progress report. Paris: International Energy Agency (IEA); 2011. [28] 2010 solar technologies market report. National Renewable Energy Laboratory (NREL): U.S. Department of Energy (DOE); 2011. [29] Rooftop photovoltaic market penetration scenarios. Colorado: National Renewable Energy Laboratory (NREL); 2008. [30] Community-scale solar photovoltaics: housing and public development example. Japan: International Energy Agency (IEA); 2008. [31] Renewable energy technology cost review. Melbourne: Melbourne Energy Institute (MEI); 2011. [32] New Renewable Energy Center in the Korea Energy Management Corporation (KEMCO). Available from: http://www.energy.or.kr. (Jun. 27, 2013). [33] Renewables 2012: global status report. Paris: Renewable Energy Policy Network for the 21st Century (REN21); 2012. [34] Wand R, Leuthold F. Feed-in tariffs for photovoltaics: learning by doing in Germany? Appl Energy 2011;88(12):4387e99. [35] Cherrington R, Goodship V, Longfield A, Kirwan K. The feed-in tariff in the UK: a case study focus on domestic photovoltaic systems. Renew Energy 2013;50:421e6. [36] Exploration of new regeneration energy. Korea: Korea Mech. Const. Contractors Association; 2011. [37] Zhao Q, Wang P, Goel L. Optimal PV panel tilt angle based on solar radiation prediction. In: Probabilistic Methods Applied to Power Systems (PMAPS), 2010 IEEE 11th International Conference 2010. pp. 425e30. [38] Dincer F, Meral ME. Critical factors that affecting efficiency of solar cells. Smart Grid Renew Energy 2010;1:47e50. [39] Hwang T, Kang S, Kim J. Optimization of the building integrated photovoltaic system in office buildings: focus on the orientation, inclined angle and installed area. Energy Build 2012;46:92e104. [40] Siraki AG, Pillay P. Study of optimum tilt angles for solar panels in different latitudes for urban applications. Solar Energy 2012;86(6):1920e8. [41] Ordóñez J, Jadraque E, Alegre J, Martínez G. Analysis of the photovoltaic solar energy capacity of residential rooftops in Andalusia (Spain). Renew Sustain Energy Rev 2010;14(7):2122e30. [42] Crook JA, Jones LA, Forster PM, Crook R. Climate change impacts on future photovoltaic and concentrated solar power energy output. Energy Environ Sci 2011;4:3101e9. [43] Levinson R, Akbari H, Pomerantz M, Gupta S. Solar access of residential rooftops in four California cities. Solar Energy 2009;83(12):2120e35. [44] Liu G, Rasul MG, Amanullah MTO, Khan MMK. Techno-economic simulation and optimization of residential grid-connected PV system for the Queensland climate. Renew Energy 2012;45:146e55. [45] Ubertini S, Desideri U. Performance estimation and experimental measurements of a photovoltaic roof. Renew Energy 2003;28(12):1833e50. [46] Hussein HMS, Anmad GE, El-Ghetany HH. Performance evaluation of photovoltaic modules at different tilt angles and orientations. Energy Convers Manag 2004;45(15e16):2441e52. [47] Jafarkazemi F, Saadabadi SA. Optimum tilt angle and orientation of solar surfaces in Abu Dhabi, UAE. Renew Energy 2013;56:44e9. [48] Li DHW, Lam TNT. Determining the optimum tilt angle and orientation for solar energy collection based on measured solar radiance data. Int J Photoenergy 2007;2007:1e9. [49] Gunerhan H, Hepbasli A. Determination of the optimum tilt angle of solar collectors for building applications. Build Environ 2007;42(2):779e83. [50] Gopinathan KK, Maliehe NB, Mpholo MI. A study on the intercepted insolation as a function of slope and azimuth of the surface. Energy 2007;32(3):213e20. [51] Tang R, Wu T. Optimal tilt-angles for solar collectors used in China. Appl Energy 2004;79(3):239e48. [52] Tiris M, Tiris C. Optimum collector slope and model evaluation: case study for Gebze, Turkey. Energy Convers Manag 1998;39(3e4):167e72. [53] Asowata O, Swart J, Pienaar C. Optimum tilt and orientation angles for photovoltaic panels in the Vaal triangle. In: Power and Engineering conference (APPEEC), 2012 Asia-Pacific 2012. pp. 1e5. [54] Kaldellis J, Zafirakis D. Experimental investigation of the optimum photovoltaic panels’ tilt angle during the summer period. Energy 2012;38(1):305e14. [55] Boji c M, Bigot D, Miranville F, Parvedy-Patou A, Radulovi c J. Optimizing performance of photovoltaics in Reunion Island-tilt angle. Progr Photovolt Res Appl 2012;20(8):923e35. [56] Hummon M, Denholm P, Margolis R. Impact of photovoltaic on orientation on its relative economic value in wholesale energy markets. Progr Photovolt Res Appl 2012. http://dx.doi.org/10.1002/pip.2198. [57] Braun A, Katz EA, Gordon JM. Basic aspects of the temperature coefficients of concentrator solar cell performance parameters. Progr Photovolt Res Appl 2012;21(5):1087e94. [58] Hoffmann S, Koehl M. Effect of humidity and temperature on the potentialinduced degradation. Progr Photovolt Res Appl 2012. http://dx.doi.org/ 10.1002/pip.2238. [59] Green MA. General temperature dependence of solar cell performance and implications for device modelling. Progr Photovolt Res Appl 2003;11(5):333e 40. [60] Huld T, Cebecauer T, Suri M, Dunlop ED. Analysis of one-axis tracking strategies for PV systems in Europe. Progr Photovolt Res Appl 2010;18:183e94.

Please cite this article in press as: Hong T, et al., A GIS (geographic information system)-based optimization model for estimating the electricity generation of the rooftop PV (photovoltaic) system, Energy (2013), http://dx.doi.org/10.1016/j.energy.2013.11.082

10

T. Hong et al. / Energy xxx (2013) 1e10

[61] Koo C, Hong T, Lee M, Park HS. Estimation of daily solar radiation using geographic information system and advanced case-based reasoning. Environ Sci Technol 2013;47(9):4829e39. [62] Hong T, Lee M, Koo C, Park HS. Framework for the mapping of monthly average daily solar radiation using an advanced case-based reasoning. Renew Sustain Energy Rev 2013. submitted for publication. [63] Google Earth. Available from: http://www.google.com/earth/. (Jun. 27, 2013). [64] Johnston K, Ver Hoef J, Krivoruchko K, Lucas N. ArcGIS 9: using ArcGISÔ geostatistical analyst. ESRI; 2001. [65] Ju J, Kim H, Oh S, Lee M, Choi J. A study on the difference of regional electricity and economic comparative valuation of the photovoltaic system. Korean

Institute of Architectural Sustainable Environment and Building Systems; 2008. pp. 137e40. [66] New Renewable Energy Center in the Korea Energy Management Corporation (KEMCO). Available from: http://www.energy.or.kr. (Oct. 9, 2013). [67] Hong T, Koo C, Park J, Park HS. An economic and environmental optimization model for a rooftop photovoltaic system: focused on energy efficiency in the built environment. Progr Photovolt Res Appl 2013. submitted for publication. [68] Koo C, Hong T, Park HS, Yun G. Framework for the analysis of the potential of the rooftop photovoltaic system to achieve the net-zero energy solar buildings. Progr Photovolt Res Appl 2013. in press, http://dx.doi.org/10.1002/pip. 2448.

Please cite this article in press as: Hong T, et al., A GIS (geographic information system)-based optimization model for estimating the electricity generation of the rooftop PV (photovoltaic) system, Energy (2013), http://dx.doi.org/10.1016/j.energy.2013.11.082