- Email: [email protected]
An approach for estimating average daily global solar radiation from cloud cover in Thailand
Available online at www.sciencedirect.com Available online at www.sciencedirect.com
Procedia Engineering
ProcediaProcedia Engineering 00 (2012) 000–000 Engineering 32 (2012) 399 – 406 www.elsevier.com/locate/procedia
I-SEEC2011
An approach for estimating average daily global solar radiation from cloud cover in Thailand P. Nimnuan∗ and S. Janjai Solar Energy Research Laboratory, Department of Physics, Faculty of Science, Silpakorn University,Nakhon Pathom, 73000, Thailand Elsevier use only: Received 30 September 2011; Revised 10 November 2011; Accepted 25 November 2011.
Abstract Global solar radiation is important for solar energy applications. In this work, an approach for estimating monthly average daily global solar radiation in Thailand was proposed. In developing this approach, cloud cover and global solar radiation measured at 24 sites in Thailand were analyzed. According to the analysis, values of normalized monthly average daily global radiation were correlated with those of monthly average cloud cover. Then linear equations relating the normalized global radiation with the cloud cover were formulated. In the next step, the values of the slope and intercept of the equations from all sites were determined. These values were interpolated to cover the entire areas of the country and they were presented as contour maps. In the final step, a general empirical equation for estimating the average daily global radiation from the cloud cover was proposed. The model coefficients for the entire areas of the country were obtained from the maps. For the validation, the general equation was used to calculate average daily global radiation at 4 sites whose data were not used in the model development process. Finally, the results were compared with the measurements and they were in good agreement.
© 2010 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of I-SEEC2011 Keywords: Global solar radiation; cloud cover; model
1. Introduction Information on the amount of solar global radiation at the earth’ surface is essential for designing solar energy systems such as solar photovoltaics and solar water heating systems Ideally, such information should be obtained from a dense network of pyranometer stations. However, in reality the number of stations in
* Corresponding author. Tel.: +66-34-270761; fax: +66-34-271189. E-mail address: [email protected].
1877-7058 © 2012 Published by Elsevier Ltd. doi:10.1016/j.proeng.2012.01.1285
400 2
P. Nimnuan S. Janjai / Procedia Engineering 32 (2012) 399 – 406 P. Nimnuan and and S. Janjai / Procedia Engineering 00 (2012) 000–000
the network of most countries is too sparse to provide sufficient solar radiation data for solar energy applications. An alternative solution to this problem is to estimate solar radiation from other meteorological parameters which routinely measured in most meteorological stations. Among these parameters, cloud cover is considered to be a good estimator of solar radiation since clouds affect strongly solar radiation at the earth’s surface. As a result, a number of models relating global solar radiation with cloud cover have been formulated [1-5]. For the case of Thailand, Exell and Salicali [6] developed an empirical model relating sunshine duration to cloud cover and used sunshine duration to estimate global solar radiation. Kirtikara et al. [7] proposed empirical model to predict global solar radiation from various meteorological parameters including cloud cover data. Janjai and Tosing [8] developed a model for the estimation of global solar radiation from cloud cover using a four-year period (1995-1998) global radiation and cloud cover measured at three meteorological stations in main regions of Thailand. Although a number of models for predicting global radiation from cloud cover have been proposed by using solar radiation from various parts of the world, most of these models were developed for specific areas. For the case of Thailand, the existing cloud-global radiation models were formulated from old and limited solar radiation data. Since 2002, the country has new solar radiation monitoring network with more than 25 pyranometer stations across the country. The objective of this study was to establish a new cloud-global radiation model from the new data set. In addition, an interpolation scheme was proposed to generalize the application of the model for the entire country.
(
Fig. 1. Positions of the stations where solar radiation and cloud cover data were used in this work stations used for the development of the approach, stations employed for the validation of the approach)
P.P.Nimnuan Engineering 3200 (2012) 399 – 406 Nimnuanand andS.S.Janjai Janjai/ Procedia / Procedia Engineering (2012) 000–000
3
2. Measurement and data Global solar radiation data used in this study were obtained from 25 pyranometer stations of the Department of Alternative Energy Development (DEDE). The pyranometer stations were established by our research group during 2000-2001. The systematic measurements of solar radiation were started at the beginning of 2002. We have taken care of the measurements and processed the data since the beginning of the measurements. Most stations are located at the existing meteorological stations of the Thai Meteorological Department where cloud cover is routinely observed. Additionally, global radiation data from 3 stations of our laboratory in Chiang Mai, Ubon Ratchathani and Songkhla were also used in this work. For all stations of DEDE, global solar radiation was measured by using a Kipp&Zonen pyranometer (model CM11). Global radiation at 3 stations of our laboratory was measured by using pyranometer of Kipp&Zonen (model CM21). The output voltage of all pyranometers was captured every 1 second by a Yokogawa datalogger (model DC100). The values of the voltage were averaged over a 10minute period and the average value was recorded by the datalogger. The average voltage data were sent to our laboratory once a month. At the laboratory, the voltage data were converted into solar irradiance by using a conversion factor or pyranometer sensitivity of each station. The 10-minute average irradiance from each station was integrated over a day to obtain daily global radiation in MJ/m2. All pyranometers were calibrated once a year by using a traveling standard pyranometer recently calibrated by Kipp&Zonen. For the quality control of solar radiation data, the abnormal data or data which violate physical laws were discarded from the data set. For the cloud cover data from each station, they were obtained from visual observation of a welltrained meteorologist. The observation at each station was carried out at 3 hour time interval. In this work the data from cloud observation at 7:00h, 10:00h, 13:00h and 16:00h local time were used. The cloud cover was quantified as tenths of the sky covered by clouds. This means that values of cloud cover vary between 0 and 10. Cloud data 7:00h, 10:00h, 13:00h and 16:00h local time were averaged over the month to obtain monthly average cloud cover. Both solar radiation and cloud cover data for the period of 4-8 years (Table 1) from 28 stations (Fig. 1) were used in this work. They were separated into 2 data sets, one set from 24 stations for the model formulation, the other 4 stations for testing the model (see Table 1) 3. Development of the approach for estimating global radiation from cloud cover As shown in our previous work [8], global solar radiation correlated well with cloud cover for the long-term monthly average data. For this reason, in this work we put emphasis on the development of a general model for the estimation of the monthly average daily global radiation from cloud cover. The formulation of the model started with the calculation of long-term monthly average daily global radiation ( H ) and cloud cover ( C ). To eliminate the effect of the variation of the sun-earth distance, the global radiation was divided by monthly average daily extraterrestrial radiation ( H 0 ). H 0 was calculated using the formula reported in Iqbal [9] with the new solar constant of 1366.1 W/m2. Then values of H / H 0 were plotted against C and the results are shown in Fig. 2.
401
402 4
P. Nimnuan and S. Janjai / Procedia Engineering 32 (2012) 399 – 406 P. Nimnuan and S. Janjai / Procedia Engineering 00 (2012) 000–000
0.5
cloud cover
8
10
0
2
4
6
cloud cover
8
0
10
Tak
0.3 R2 = 0.95 0
2
4
0.3 0.2 0.1
H/H0 = -0.0418C + 0.7497
0
0.4
R2 = 0.92
0 6
cloud cover
8
0
10
0.8
2
4
6
cloud cover
8
0.2 H/H0 = -0.0174C + 0.6351 0.1 R2 = 0.78 0 0 2 4 6
cloud cover
0.4 0.3 R2 = 0.95
0 0
0.8
2
4
6
cloud cover
8
0.2 0.1
0
2
4
0.3 R2 = 0.98
0
6
cloud cover
8
10
0
0.8
2
4
cloud cover
8
0.3 0.2 0.1
R2 = 0.87 0
2
4
0.3 0.2 0.1
H/H0 = -0.0231C + 0.6204
0
0.4
R2 = 0.93
0 6
cloud cover
8
10
0
Songkhla
0.7
2
4
6
cloud cover
8
H/H0
0.5 0.4 0.3 0.2
H/H0 = -0.0635C + 0.962
0.1
2
R = 0.93
0 0
2
4
0.3 H/H0 = -0.0496C + 0.8139 2
R = 0.89
0 6
cloud cover
8
10
0
2
4
cloud cover
8
10
10
8
10
8
10
8
10
0.4 0.3 H/H0 = -0.0578C + 0.8935 R2 = 0.93 2
4
6
cloud cover
2
Kanchanaburi
4
0.6 0.5 0.4 0.3 0.2 0.1
R2 = 0.92
H/H0 = -0.0191C + 0.6246 R2 = 0.80
0 6
cloud cover
8
10
0
2
4
6
cloud cover
0.8
Chumphon
R2 = 0.91 2
4
0.6 0.5 0.4 0.3 0.2 0.1
H/H0 = -0.0425C + 0.751
0
Loei
0.7
0.3
H/H0 = -0.0238C + 0.6122 R2 = 0.87
0 6
cloud cover
8
0
10
2
4
6
cloud cover
0.8
Chonburi
0.4 0.3 2
R = 0.77 2
4
0.6 0.5 0.4 0.3 0.2 0.1
H/H0 = -0.0166C + 0.5939
0
Trad
0.7
0.6 0.5
0 6
8
Ranong
0.6 0.5
0.7
0.4
0.2 0.1
10
0.3
0
H/H0 = -0.0658C + 0.9620
0.7
0.4
8
0.4
10
0.8
0.6 0.5
0.2 0.1
cloud cover
8
0.6 0.5
10
Suratthani (punpin)
0.7
0.6
6
0.3
0
0.8
0.8
4
0.4
0.2 0.1
H/H0 = -0.0449C + 0.8477
0.2 H/H0 = -0.0400C + 0.725 0.1 R2 = 0.96 0 0 2 4 6
Nakhonpanom
0.6 0.5
0.8
0.7
0.6 0.5
H/H0
H/H0
0.4
2
0.8
Phrae
0.7
0.6 0.5
10
0
Narathiwat
0
0.8
Chiang Mai
0.7
8
0.2 0.1
0.6 0.5
10
6
0.7
R2 = 0.86
0 6
4
cloud cover
cloud cover
H/H0 = -0.0210C + 0.6484
0.2 0.1
H/H0 = -0.0660C + 0.9598
0.1
R2 = 0.90
0
0.4 0.2
H/H0 = -0.0580C + 0.8859
10
0.3
0.7
H/H0
H/H0
0.3
8
0.8
0.5
2
0.8
0
0.6
0.4
6
0.4
10
Phuket
0.7
0.6 0.5
4
cloud cover
Nakhonratchasima
0
0.8
Koh Samui
0.7
2
0.6 0.5
0.2 0.1
H/H0 = -0.0518C + 0.8867
0.1
10
R2 = 0.93
0.7
0.5
0.2
8
H/H0 = -0.0269C + 0.6664
0.8
H/H0
H/H0
0.3
R2 = 0.96 0
0.7
0.3
0
0.6
0.4
10
0.4
10
Ubon Ratchathani
0.7
0.6 0.5
cloud cover
8
Khonkhan
0
0.8
Surin
0.7
6
0.6 0.5
0.2 0.1
H/H0 = -0.0269C + 0.6608
H/H0 = -0.0392C + 0.7839
0
0.8
0.7
0.6 0.5
H/H0
H/H0
H/H0
0.4 0.2 0.1
Phetchabun
0.7
0.6 0.5
4
0.3 0.1
0.8
0.8
0.7
2
0.4 0.2
R2 = 0.89
0
H/H0
6
0.8
H/H0
R2 = 0.95
0
H/H0= -0.0271C + 0.6448
H/H0
4
0.1
H/H0
2
0.2
H/H0 = -0.0523C + 0.8849
H/H0
0
H/H0
0.3
0.1
R2 = 0.76
0
H/H0
0.4
0.2
H/H0 = -0.0152C + 0.6042
H/H0
0.6
0.5
H/H0
0.6
0.5 0.3
Prachuabkhirikhan
0.7
0.6 0.4
0.3
0.8
Thongphaphum
0.7
H/H0
H/H0
0.4
H/H0
Bangkok
0.7
0.6 0.5
0.2 0.1
0.8
0.8
Nakhonsawan
H/H0
0.7
H/H0
0.8
H/H0 = -0.0509C + 0.795 R2 = 0.97
0 6
cloud cover
8
10
0
2
4
6
cloud cover
Fig. 2. Correlation between the ratio of monthly average daily global to monthly average daily extraterrestrial radiation ( H / H0 ) and monthly average daily cloud cover ( C ) for 24 stations
P.P.Nimnuan Engineering 3200 (2012) 399 – 406 Nimnuanand andS.S.Janjai Janjai/ Procedia / Procedia Engineering (2012) 000–000
5
From Fig. 2, for each station it was observed that H / H 0 correlated well to C with a straight line fit. Additionally, the slope and y-intercept of the graphs varied with locations of the stations (Table 1). As a result, the correlation between H / H 0 and C can be written in a general form as:
H = a 0 + a1C H0
(1)
where a0 and a1 are the y-intercept and slope, respectively. The values of a0, a1 and the square of the correlation coefficient (R2) for all stations are shown in Table 1. Table 1. Positions of solar monitoring stations, data periods, model coefficients (a0,a1) and values of the square of correlation coefficient (R2) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Station
Nakhon Sawan Bangkok Thong Phaphum Prachuabkhirikhun Tak Phetchaboon Khonkhan Nakhon Panom Surin Ubon Ratchathani Nakhonratchasima Ranong Koh Samui Phuket Narathiwat Kanchanaburi Chiang Mai Phrae Chumphon Loei Songkhla Surat Thani Chonburi Trad Mae Sariang* Lopburi* Ubon Ratcahthani(sub)* Trung*
Period of data Jan, 2006-Dec, 2009 Jan, 2003-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2006-Jan, 2008 Jan, 2002-Dec, 2009 Mar, 2006-Dec, 2009 Jun, 2006-Dec, 2009 Jan, 2006-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2005-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2005-Dec, 2009 Feb, 2005-Dec, 2009 Feb, 2006-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2006-Dec, 2009 Mar, 2005-Dec, 2009 Mar, 2006-Dec, 2009 Jan, 2005-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2009-Dec, 2009 Mar, 2005-Dec, 2009
Latitude (degree) 15.80 13.67 14.75 11.83 16.77 16.43 16.19 17.42 14.88 15.25 14.97 9.98 9.47 8.13 6.42 14.02 18.78 18.16 10.48 17.40 7.20 9.13 13.37 11.77 18.17 14.83 15.23 7.52
Longitude (degree) 100.17 100.62 98.63 99.83 98.93 101.15 102.83 104.78 103.50 104.87 102.08 98.62 100.03 98.30 101.82 99.53 98.98 100.17 99.18 101.73 100.60 99.35 100.98 102.88 97.93 100.67 105.03 99.53
a1
a0
R2
-0.0152 -0.0523 -0.0271 -0.0392 -0.0418 -0.0269 -0.0269 -0.0400 -0.0174 -0.0518 -0.0210 -0.0578 -0.0580 -0.0660 -0.0658 -0.0191 -0.0231 -0.0449 -0.0425 -0.0238 -0.0636 -0.0496 -0.0166 -0.0509 -0.0288 -0.0313 -0.0378 -0.0630
0.6042 0.8849 0.6448 0.7839 0.7497 0.6608 0.6664 0.7250 0.6351 0.8867 0.6484 0.8935 0.8859 0.9598 0.9620 0.6242 0.6204 0.8477 0.7510 0.6122 0.9642 0.8139 0.5939 0.7950 0.6586 0.7189 0.7843 0.9378
0.7643 0.9501 0.8910 0.9626 0.9466 0.9171 0.9313 0.9637 0.7835 0.9475 0.8630 0.9316 0.8990 0.9756 0.9177 0.7983 0.8701 0.9347 0.9121 0.8701 0.9303 0.8881 0.7670 0.9667 0.7849 0.8080 0.8160 0.9500
* Stations used for model validation
In order to generalize Eq (1), the values of the y-intercept (a0) and slope (a1) were plotted as contours in the maps of Thailand. The values of a0 and a1 in the maps were interpolated to cover the entire areas of the country by using the minimum curvature spline surface method (MCS) [10] (Fig. 3). These maps allow users to obtain the values of a0 and a1 for all locations in Thailand.
403
404 6
P. Nimnuan S. Janjai / Procedia Engineering 32 (2012) 399 – 406 P. Nimnuan and and S. Janjai / Procedia Engineering 00 (2012) 000–000
(a)
(b)
Fig. 3. Maps showing the values of the slope (a) and the y-intercept (b)
4. Performance evaluation of the approach To evaluate its performance, this approach was used to estimate monthly average daily global radiation at 4 stations, namely Mae Sariang, Ubon Ratchathani(sub), Lopburi and Trung (Fig. 1). As the data from these stations were not involved in the development of the approach, they are independent data set. Cloud cover observed at these stations were used for the estimation. The values of a0 and a1 were retrieved from the maps in Fig. 3. With these values of input parameter, the values of monthly average daily global radiation at these stations were calculated from Eq (1). Then, they were compared with those obtained from solar radiation measured at these stations and the results are shown in Fig. 4. The discrepancy in term of root mean square difference (RMSD) and mean bias difference (MBD) are shown in Table 2. 30
4 stations 25
Mae Sariang Ubon Ratchathani (sub)
Hmodel
20
Lopburi Trung
15 10 RMSD = 7.5% MBD = 0.1% N = 48
5 0 0
5
10
15
20
25
30
H m eas
Fig. 4. Comparisons between measured ( H meas ) and calculated ( H mod el ) monthly average daily global radiation at four stations (N is total number of data)
P.P.Nimnuan Engineering 3200 (2012) 399 – 406 Nimnuanand andS.S.Janjai Janjai/ Procedia / Procedia Engineering (2012) 000–000
7
Table 2. Root mean square difference (RMSD) and mean bias difference (MBD) between solar radiation calculated from the model and the measurements Station Mae Sariang Ubon Ratchathani Lopburi Trung Combined data from four stations
MBD (%) 1.5 -4.5 -1.5 5.0 0.1
RMSD (%) 8.5 6.7 6.0 8.6 7.5
From Fig. 4 and Table 2, it was observed that the values of RMSD and MBD from 4 stations are in the range of 6.0-8.6% and -1.5-5.0%, respectively. For the combined data from four stations, RMSD was 7.5% and MBD was 0.1%. These results indicated that values of monthly average global radiation estimated from this approach agreed well with those obtained from the measurement. 5. Conclusion An approach for estimating monthly average global solar radiation from cloud cover in Thailand has been proposed. Global radiation and cloud cover at 24 station across the country were used to formulated a model relating global radiation to cloud cover. The model expressed normalized global radiation ( H / H 0 ) as a linear empirical function of cloud cover ( C ). Maps of the values of the model coefficients covering the entire areas of the country were constructed. From the values of cloud cover together with the model coefficients from the maps, monthly average global radiation could be estimated from the model. The performance of this approach was tested against an independent data set. It was found that the estimated and measured global radiation were in good agreement. Acknowledgements We would like to thank Department of Alternative Energy Development and Efficiency (DEDE) for inviting Silpakorn University to carry out the project on the establishment of a new solar monitoring network of Thailand. The authors are grateful to the Thai Meteorological Department for providing cloud cover data. References [1] Kimball H.H. Variations in the total and luminous solar radiation with geographical position in the United States. Mon. Weather Rev. 1919;47(11):769-793. [2] Black J.N. The distribution of solar radiation over the earth’s surface. Arch. Meteorol. Geophys. Bioklimatol 1956;7:165189. [3] Kasten F. and Czeplack G. Solar and terrestrial radiation dependent on the amount and type of cloud. Solar Energy 1979; 24: 177-189. [4] Gul, M., Muneer, T., Kambezidis, H. Models for obtaining solar radiation form other meteorological data. Solar Energy 1998;64:99–108. [5] Zhang Q Y., Joe H., Yang H. and Lou C. Development of Models to Estimate Solar Radiation for Chinese Locations. Journal of Asian Architecture and Building Engineering 2003; 2(2): 35-41. [6] Exell R H. B. and Salicali K. The Availability of solar energy in Thailand. Research Report 1976; 63. Asean Institude of Technology. Bangkok. Thailand. [7] Kirtikara K, Siriprayuk T, Namprakai P. The use of regression equation for estimating solar radiation from meteorological data. Proceedings of The Third Conference on Renewable Energy and Applications KMITT. 4-6 November 1981. Bangkok, Thailand.
405
406 8
P. Nimnuan and S. Janjai / Procedia Engineering 32 (2012) 399 – 406 P. Nimnuan and S. Janjai / Procedia Engineering 00 (2012) 000–000
[8] Janjai S. and Tosing K. A new model for calculating global radiation from cloud cover data for Thailand. Proceedings of The First Regional Conferrence on Energy Technology Towards a Clean Environment. JGSEE. 1-2 December 2000. Chiang Mai, Thailand. [9] Iqbal, M. An Introduction to Solar Radiation. New York. Academic Press; 1983. [10] Briggs, I. C. Machine contouring using minimum curvature: Geophysics 1974; 39(1): 39–48.
Procedia Engineering
ProcediaProcedia Engineering 00 (2012) 000–000 Engineering 32 (2012) 399 – 406 www.elsevier.com/locate/procedia
I-SEEC2011
An approach for estimating average daily global solar radiation from cloud cover in Thailand P. Nimnuan∗ and S. Janjai Solar Energy Research Laboratory, Department of Physics, Faculty of Science, Silpakorn University,Nakhon Pathom, 73000, Thailand Elsevier use only: Received 30 September 2011; Revised 10 November 2011; Accepted 25 November 2011.
Abstract Global solar radiation is important for solar energy applications. In this work, an approach for estimating monthly average daily global solar radiation in Thailand was proposed. In developing this approach, cloud cover and global solar radiation measured at 24 sites in Thailand were analyzed. According to the analysis, values of normalized monthly average daily global radiation were correlated with those of monthly average cloud cover. Then linear equations relating the normalized global radiation with the cloud cover were formulated. In the next step, the values of the slope and intercept of the equations from all sites were determined. These values were interpolated to cover the entire areas of the country and they were presented as contour maps. In the final step, a general empirical equation for estimating the average daily global radiation from the cloud cover was proposed. The model coefficients for the entire areas of the country were obtained from the maps. For the validation, the general equation was used to calculate average daily global radiation at 4 sites whose data were not used in the model development process. Finally, the results were compared with the measurements and they were in good agreement.
© 2010 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of I-SEEC2011 Keywords: Global solar radiation; cloud cover; model
1. Introduction Information on the amount of solar global radiation at the earth’ surface is essential for designing solar energy systems such as solar photovoltaics and solar water heating systems Ideally, such information should be obtained from a dense network of pyranometer stations. However, in reality the number of stations in
* Corresponding author. Tel.: +66-34-270761; fax: +66-34-271189. E-mail address: [email protected].
1877-7058 © 2012 Published by Elsevier Ltd. doi:10.1016/j.proeng.2012.01.1285
400 2
P. Nimnuan S. Janjai / Procedia Engineering 32 (2012) 399 – 406 P. Nimnuan and and S. Janjai / Procedia Engineering 00 (2012) 000–000
the network of most countries is too sparse to provide sufficient solar radiation data for solar energy applications. An alternative solution to this problem is to estimate solar radiation from other meteorological parameters which routinely measured in most meteorological stations. Among these parameters, cloud cover is considered to be a good estimator of solar radiation since clouds affect strongly solar radiation at the earth’s surface. As a result, a number of models relating global solar radiation with cloud cover have been formulated [1-5]. For the case of Thailand, Exell and Salicali [6] developed an empirical model relating sunshine duration to cloud cover and used sunshine duration to estimate global solar radiation. Kirtikara et al. [7] proposed empirical model to predict global solar radiation from various meteorological parameters including cloud cover data. Janjai and Tosing [8] developed a model for the estimation of global solar radiation from cloud cover using a four-year period (1995-1998) global radiation and cloud cover measured at three meteorological stations in main regions of Thailand. Although a number of models for predicting global radiation from cloud cover have been proposed by using solar radiation from various parts of the world, most of these models were developed for specific areas. For the case of Thailand, the existing cloud-global radiation models were formulated from old and limited solar radiation data. Since 2002, the country has new solar radiation monitoring network with more than 25 pyranometer stations across the country. The objective of this study was to establish a new cloud-global radiation model from the new data set. In addition, an interpolation scheme was proposed to generalize the application of the model for the entire country.
(
Fig. 1. Positions of the stations where solar radiation and cloud cover data were used in this work stations used for the development of the approach, stations employed for the validation of the approach)
P.P.Nimnuan Engineering 3200 (2012) 399 – 406 Nimnuanand andS.S.Janjai Janjai/ Procedia / Procedia Engineering (2012) 000–000
3
2. Measurement and data Global solar radiation data used in this study were obtained from 25 pyranometer stations of the Department of Alternative Energy Development (DEDE). The pyranometer stations were established by our research group during 2000-2001. The systematic measurements of solar radiation were started at the beginning of 2002. We have taken care of the measurements and processed the data since the beginning of the measurements. Most stations are located at the existing meteorological stations of the Thai Meteorological Department where cloud cover is routinely observed. Additionally, global radiation data from 3 stations of our laboratory in Chiang Mai, Ubon Ratchathani and Songkhla were also used in this work. For all stations of DEDE, global solar radiation was measured by using a Kipp&Zonen pyranometer (model CM11). Global radiation at 3 stations of our laboratory was measured by using pyranometer of Kipp&Zonen (model CM21). The output voltage of all pyranometers was captured every 1 second by a Yokogawa datalogger (model DC100). The values of the voltage were averaged over a 10minute period and the average value was recorded by the datalogger. The average voltage data were sent to our laboratory once a month. At the laboratory, the voltage data were converted into solar irradiance by using a conversion factor or pyranometer sensitivity of each station. The 10-minute average irradiance from each station was integrated over a day to obtain daily global radiation in MJ/m2. All pyranometers were calibrated once a year by using a traveling standard pyranometer recently calibrated by Kipp&Zonen. For the quality control of solar radiation data, the abnormal data or data which violate physical laws were discarded from the data set. For the cloud cover data from each station, they were obtained from visual observation of a welltrained meteorologist. The observation at each station was carried out at 3 hour time interval. In this work the data from cloud observation at 7:00h, 10:00h, 13:00h and 16:00h local time were used. The cloud cover was quantified as tenths of the sky covered by clouds. This means that values of cloud cover vary between 0 and 10. Cloud data 7:00h, 10:00h, 13:00h and 16:00h local time were averaged over the month to obtain monthly average cloud cover. Both solar radiation and cloud cover data for the period of 4-8 years (Table 1) from 28 stations (Fig. 1) were used in this work. They were separated into 2 data sets, one set from 24 stations for the model formulation, the other 4 stations for testing the model (see Table 1) 3. Development of the approach for estimating global radiation from cloud cover As shown in our previous work [8], global solar radiation correlated well with cloud cover for the long-term monthly average data. For this reason, in this work we put emphasis on the development of a general model for the estimation of the monthly average daily global radiation from cloud cover. The formulation of the model started with the calculation of long-term monthly average daily global radiation ( H ) and cloud cover ( C ). To eliminate the effect of the variation of the sun-earth distance, the global radiation was divided by monthly average daily extraterrestrial radiation ( H 0 ). H 0 was calculated using the formula reported in Iqbal [9] with the new solar constant of 1366.1 W/m2. Then values of H / H 0 were plotted against C and the results are shown in Fig. 2.
401
402 4
P. Nimnuan and S. Janjai / Procedia Engineering 32 (2012) 399 – 406 P. Nimnuan and S. Janjai / Procedia Engineering 00 (2012) 000–000
0.5
cloud cover
8
10
0
2
4
6
cloud cover
8
0
10
Tak
0.3 R2 = 0.95 0
2
4
0.3 0.2 0.1
H/H0 = -0.0418C + 0.7497
0
0.4
R2 = 0.92
0 6
cloud cover
8
0
10
0.8
2
4
6
cloud cover
8
0.2 H/H0 = -0.0174C + 0.6351 0.1 R2 = 0.78 0 0 2 4 6
cloud cover
0.4 0.3 R2 = 0.95
0 0
0.8
2
4
6
cloud cover
8
0.2 0.1
0
2
4
0.3 R2 = 0.98
0
6
cloud cover
8
10
0
0.8
2
4
cloud cover
8
0.3 0.2 0.1
R2 = 0.87 0
2
4
0.3 0.2 0.1
H/H0 = -0.0231C + 0.6204
0
0.4
R2 = 0.93
0 6
cloud cover
8
10
0
Songkhla
0.7
2
4
6
cloud cover
8
H/H0
0.5 0.4 0.3 0.2
H/H0 = -0.0635C + 0.962
0.1
2
R = 0.93
0 0
2
4
0.3 H/H0 = -0.0496C + 0.8139 2
R = 0.89
0 6
cloud cover
8
10
0
2
4
cloud cover
8
10
10
8
10
8
10
8
10
0.4 0.3 H/H0 = -0.0578C + 0.8935 R2 = 0.93 2
4
6
cloud cover
2
Kanchanaburi
4
0.6 0.5 0.4 0.3 0.2 0.1
R2 = 0.92
H/H0 = -0.0191C + 0.6246 R2 = 0.80
0 6
cloud cover
8
10
0
2
4
6
cloud cover
0.8
Chumphon
R2 = 0.91 2
4
0.6 0.5 0.4 0.3 0.2 0.1
H/H0 = -0.0425C + 0.751
0
Loei
0.7
0.3
H/H0 = -0.0238C + 0.6122 R2 = 0.87
0 6
cloud cover
8
0
10
2
4
6
cloud cover
0.8
Chonburi
0.4 0.3 2
R = 0.77 2
4
0.6 0.5 0.4 0.3 0.2 0.1
H/H0 = -0.0166C + 0.5939
0
Trad
0.7
0.6 0.5
0 6
8
Ranong
0.6 0.5
0.7
0.4
0.2 0.1
10
0.3
0
H/H0 = -0.0658C + 0.9620
0.7
0.4
8
0.4
10
0.8
0.6 0.5
0.2 0.1
cloud cover
8
0.6 0.5
10
Suratthani (punpin)
0.7
0.6
6
0.3
0
0.8
0.8
4
0.4
0.2 0.1
H/H0 = -0.0449C + 0.8477
0.2 H/H0 = -0.0400C + 0.725 0.1 R2 = 0.96 0 0 2 4 6
Nakhonpanom
0.6 0.5
0.8
0.7
0.6 0.5
H/H0
H/H0
0.4
2
0.8
Phrae
0.7
0.6 0.5
10
0
Narathiwat
0
0.8
Chiang Mai
0.7
8
0.2 0.1
0.6 0.5
10
6
0.7
R2 = 0.86
0 6
4
cloud cover
cloud cover
H/H0 = -0.0210C + 0.6484
0.2 0.1
H/H0 = -0.0660C + 0.9598
0.1
R2 = 0.90
0
0.4 0.2
H/H0 = -0.0580C + 0.8859
10
0.3
0.7
H/H0
H/H0
0.3
8
0.8
0.5
2
0.8
0
0.6
0.4
6
0.4
10
Phuket
0.7
0.6 0.5
4
cloud cover
Nakhonratchasima
0
0.8
Koh Samui
0.7
2
0.6 0.5
0.2 0.1
H/H0 = -0.0518C + 0.8867
0.1
10
R2 = 0.93
0.7
0.5
0.2
8
H/H0 = -0.0269C + 0.6664
0.8
H/H0
H/H0
0.3
R2 = 0.96 0
0.7
0.3
0
0.6
0.4
10
0.4
10
Ubon Ratchathani
0.7
0.6 0.5
cloud cover
8
Khonkhan
0
0.8
Surin
0.7
6
0.6 0.5
0.2 0.1
H/H0 = -0.0269C + 0.6608
H/H0 = -0.0392C + 0.7839
0
0.8
0.7
0.6 0.5
H/H0
H/H0
H/H0
0.4 0.2 0.1
Phetchabun
0.7
0.6 0.5
4
0.3 0.1
0.8
0.8
0.7
2
0.4 0.2
R2 = 0.89
0
H/H0
6
0.8
H/H0
R2 = 0.95
0
H/H0= -0.0271C + 0.6448
H/H0
4
0.1
H/H0
2
0.2
H/H0 = -0.0523C + 0.8849
H/H0
0
H/H0
0.3
0.1
R2 = 0.76
0
H/H0
0.4
0.2
H/H0 = -0.0152C + 0.6042
H/H0
0.6
0.5
H/H0
0.6
0.5 0.3
Prachuabkhirikhan
0.7
0.6 0.4
0.3
0.8
Thongphaphum
0.7
H/H0
H/H0
0.4
H/H0
Bangkok
0.7
0.6 0.5
0.2 0.1
0.8
0.8
Nakhonsawan
H/H0
0.7
H/H0
0.8
H/H0 = -0.0509C + 0.795 R2 = 0.97
0 6
cloud cover
8
10
0
2
4
6
cloud cover
Fig. 2. Correlation between the ratio of monthly average daily global to monthly average daily extraterrestrial radiation ( H / H0 ) and monthly average daily cloud cover ( C ) for 24 stations
P.P.Nimnuan Engineering 3200 (2012) 399 – 406 Nimnuanand andS.S.Janjai Janjai/ Procedia / Procedia Engineering (2012) 000–000
5
From Fig. 2, for each station it was observed that H / H 0 correlated well to C with a straight line fit. Additionally, the slope and y-intercept of the graphs varied with locations of the stations (Table 1). As a result, the correlation between H / H 0 and C can be written in a general form as:
H = a 0 + a1C H0
(1)
where a0 and a1 are the y-intercept and slope, respectively. The values of a0, a1 and the square of the correlation coefficient (R2) for all stations are shown in Table 1. Table 1. Positions of solar monitoring stations, data periods, model coefficients (a0,a1) and values of the square of correlation coefficient (R2) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Station
Nakhon Sawan Bangkok Thong Phaphum Prachuabkhirikhun Tak Phetchaboon Khonkhan Nakhon Panom Surin Ubon Ratchathani Nakhonratchasima Ranong Koh Samui Phuket Narathiwat Kanchanaburi Chiang Mai Phrae Chumphon Loei Songkhla Surat Thani Chonburi Trad Mae Sariang* Lopburi* Ubon Ratcahthani(sub)* Trung*
Period of data Jan, 2006-Dec, 2009 Jan, 2003-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2006-Jan, 2008 Jan, 2002-Dec, 2009 Mar, 2006-Dec, 2009 Jun, 2006-Dec, 2009 Jan, 2006-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2005-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2005-Dec, 2009 Feb, 2005-Dec, 2009 Feb, 2006-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2006-Dec, 2009 Mar, 2005-Dec, 2009 Mar, 2006-Dec, 2009 Jan, 2005-Dec, 2009 Jan, 2002-Dec, 2009 Jan, 2009-Dec, 2009 Mar, 2005-Dec, 2009
Latitude (degree) 15.80 13.67 14.75 11.83 16.77 16.43 16.19 17.42 14.88 15.25 14.97 9.98 9.47 8.13 6.42 14.02 18.78 18.16 10.48 17.40 7.20 9.13 13.37 11.77 18.17 14.83 15.23 7.52
Longitude (degree) 100.17 100.62 98.63 99.83 98.93 101.15 102.83 104.78 103.50 104.87 102.08 98.62 100.03 98.30 101.82 99.53 98.98 100.17 99.18 101.73 100.60 99.35 100.98 102.88 97.93 100.67 105.03 99.53
a1
a0
R2
-0.0152 -0.0523 -0.0271 -0.0392 -0.0418 -0.0269 -0.0269 -0.0400 -0.0174 -0.0518 -0.0210 -0.0578 -0.0580 -0.0660 -0.0658 -0.0191 -0.0231 -0.0449 -0.0425 -0.0238 -0.0636 -0.0496 -0.0166 -0.0509 -0.0288 -0.0313 -0.0378 -0.0630
0.6042 0.8849 0.6448 0.7839 0.7497 0.6608 0.6664 0.7250 0.6351 0.8867 0.6484 0.8935 0.8859 0.9598 0.9620 0.6242 0.6204 0.8477 0.7510 0.6122 0.9642 0.8139 0.5939 0.7950 0.6586 0.7189 0.7843 0.9378
0.7643 0.9501 0.8910 0.9626 0.9466 0.9171 0.9313 0.9637 0.7835 0.9475 0.8630 0.9316 0.8990 0.9756 0.9177 0.7983 0.8701 0.9347 0.9121 0.8701 0.9303 0.8881 0.7670 0.9667 0.7849 0.8080 0.8160 0.9500
* Stations used for model validation
In order to generalize Eq (1), the values of the y-intercept (a0) and slope (a1) were plotted as contours in the maps of Thailand. The values of a0 and a1 in the maps were interpolated to cover the entire areas of the country by using the minimum curvature spline surface method (MCS) [10] (Fig. 3). These maps allow users to obtain the values of a0 and a1 for all locations in Thailand.
403
404 6
P. Nimnuan S. Janjai / Procedia Engineering 32 (2012) 399 – 406 P. Nimnuan and and S. Janjai / Procedia Engineering 00 (2012) 000–000
(a)
(b)
Fig. 3. Maps showing the values of the slope (a) and the y-intercept (b)
4. Performance evaluation of the approach To evaluate its performance, this approach was used to estimate monthly average daily global radiation at 4 stations, namely Mae Sariang, Ubon Ratchathani(sub), Lopburi and Trung (Fig. 1). As the data from these stations were not involved in the development of the approach, they are independent data set. Cloud cover observed at these stations were used for the estimation. The values of a0 and a1 were retrieved from the maps in Fig. 3. With these values of input parameter, the values of monthly average daily global radiation at these stations were calculated from Eq (1). Then, they were compared with those obtained from solar radiation measured at these stations and the results are shown in Fig. 4. The discrepancy in term of root mean square difference (RMSD) and mean bias difference (MBD) are shown in Table 2. 30
4 stations 25
Mae Sariang Ubon Ratchathani (sub)
Hmodel
20
Lopburi Trung
15 10 RMSD = 7.5% MBD = 0.1% N = 48
5 0 0
5
10
15
20
25
30
H m eas
Fig. 4. Comparisons between measured ( H meas ) and calculated ( H mod el ) monthly average daily global radiation at four stations (N is total number of data)
P.P.Nimnuan Engineering 3200 (2012) 399 – 406 Nimnuanand andS.S.Janjai Janjai/ Procedia / Procedia Engineering (2012) 000–000
7
Table 2. Root mean square difference (RMSD) and mean bias difference (MBD) between solar radiation calculated from the model and the measurements Station Mae Sariang Ubon Ratchathani Lopburi Trung Combined data from four stations
MBD (%) 1.5 -4.5 -1.5 5.0 0.1
RMSD (%) 8.5 6.7 6.0 8.6 7.5
From Fig. 4 and Table 2, it was observed that the values of RMSD and MBD from 4 stations are in the range of 6.0-8.6% and -1.5-5.0%, respectively. For the combined data from four stations, RMSD was 7.5% and MBD was 0.1%. These results indicated that values of monthly average global radiation estimated from this approach agreed well with those obtained from the measurement. 5. Conclusion An approach for estimating monthly average global solar radiation from cloud cover in Thailand has been proposed. Global radiation and cloud cover at 24 station across the country were used to formulated a model relating global radiation to cloud cover. The model expressed normalized global radiation ( H / H 0 ) as a linear empirical function of cloud cover ( C ). Maps of the values of the model coefficients covering the entire areas of the country were constructed. From the values of cloud cover together with the model coefficients from the maps, monthly average global radiation could be estimated from the model. The performance of this approach was tested against an independent data set. It was found that the estimated and measured global radiation were in good agreement. Acknowledgements We would like to thank Department of Alternative Energy Development and Efficiency (DEDE) for inviting Silpakorn University to carry out the project on the establishment of a new solar monitoring network of Thailand. The authors are grateful to the Thai Meteorological Department for providing cloud cover data. References [1] Kimball H.H. Variations in the total and luminous solar radiation with geographical position in the United States. Mon. Weather Rev. 1919;47(11):769-793. [2] Black J.N. The distribution of solar radiation over the earth’s surface. Arch. Meteorol. Geophys. Bioklimatol 1956;7:165189. [3] Kasten F. and Czeplack G. Solar and terrestrial radiation dependent on the amount and type of cloud. Solar Energy 1979; 24: 177-189. [4] Gul, M., Muneer, T., Kambezidis, H. Models for obtaining solar radiation form other meteorological data. Solar Energy 1998;64:99–108. [5] Zhang Q Y., Joe H., Yang H. and Lou C. Development of Models to Estimate Solar Radiation for Chinese Locations. Journal of Asian Architecture and Building Engineering 2003; 2(2): 35-41. [6] Exell R H. B. and Salicali K. The Availability of solar energy in Thailand. Research Report 1976; 63. Asean Institude of Technology. Bangkok. Thailand. [7] Kirtikara K, Siriprayuk T, Namprakai P. The use of regression equation for estimating solar radiation from meteorological data. Proceedings of The Third Conference on Renewable Energy and Applications KMITT. 4-6 November 1981. Bangkok, Thailand.
405
406 8
P. Nimnuan and S. Janjai / Procedia Engineering 32 (2012) 399 – 406 P. Nimnuan and S. Janjai / Procedia Engineering 00 (2012) 000–000
[8] Janjai S. and Tosing K. A new model for calculating global radiation from cloud cover data for Thailand. Proceedings of The First Regional Conferrence on Energy Technology Towards a Clean Environment. JGSEE. 1-2 December 2000. Chiang Mai, Thailand. [9] Iqbal, M. An Introduction to Solar Radiation. New York. Academic Press; 1983. [10] Briggs, I. C. Machine contouring using minimum curvature: Geophysics 1974; 39(1): 39–48.