Upload
m-shahid
View
212
Download
0
Embed Size (px)
Citation preview
Arab J Sci Eng (2014) 39:175–179DOI 10.1007/s13369-013-0819-3
RESEARCH ARTICLE - EARTH SCIENCES
Estimating Global, Diffuse Solar Radiation for Chhorand Validation with Satellite-Based Data
Asif Ali Abbasi · M. Shahid Qureshi
Received: 13 November 2012 / Accepted: 15 April 2013 / Published online: 26 October 2013© King Fahd University of Petroleum and Minerals 2013
Abstract The study of solar irradiation has been carriedout for the first time over Chhor (city of Sindh province inPakistan). In the present work, measured data of bright sun-shine hour of the regions have been used to estimate monthlyaverage daily global and diffuse solar radiation. Regressioncoefficients a and b have been calculated from the first-orderAngstrom type correlation for the city using relationshipgiven by Tiware and Sangeeta. The results obtained throughfour empirical models i.e., Angstrom, Liu and Jordon, Page,Hawas and Muneer were compared with values obtainedfrom NASA Satellite-based Global and Diffuse radiationdata. A good agreement was found between satellite-basedvalues with Muneer and Hawaas model. The global irradia-tion is found high from April to July for the period of study.The statistical error tests—mean bias error, root mean squareerrors are used to validate the estimates using satellite-baseddata. MBE has lowest values for diffuse solar radiation thanglobal radiation and shows over estimation. The value ofcoefficient of determination implies that 98.7 % of NASA canbe accounted by estimation of global solar radiation Chhor.Low values of variance suggest that the correlation is bestfitted.
Keywords Global and diffuse radiation · Angstromcoefficient · Mean bias error · Root mean square error
A. A. Abbasi (B)Institute of Space and Planetary Astrophysics,University of Karachi, Karachi, Pakistane-mail: [email protected]
M. S. QureshiDepartment of Mathematical Sciences, IBA, Karachi, Pakistan
1 Introduction
The sun radiant energy is the only source that influencesthe atmospheric motion which can be used as an alterna-tive energy resource, in view of future depletion of fossilfuel reservoirs. Information about global solar radiation isthe most important for wide variety of applications, e.g.solar energy system’s design, building’s design, crop dry-ing, photosynthesis etc. Solar radiation data are collected inthe major parts of the world but is unavailable in develop-ing countries like Pakistan which cannot afford expensiveinstruments Viorel Badescu [1]. Due to the fact, it is impor-tant to develop methods of estimating the solar radiation onthe basis of available meteorological data. Several formulashave also been developed by various authors to estimate theamount of global solar radiation on horizontal surfaces usingvarious parameters, such as sunshine duration, cloud cover,
123
176 Arab J Sci Eng (2014) 39:175–179
humidity, maximum and minimum temperatures, wind speedetc. Recently, Wu et al. [2], used the meteorological data ofNanchang Station (China) to predict daily global solar radia-tion from sunshine hours and dew points. An effective resultwas achieved by model which uses sunshine duration. Araset al. [3], developed empirical models to predict the monthlyaverage daily global solar radiation over twelve provinces inthe Central Anatolia Region (CAR) of Turkey and to com-pare calculated values obtained from developed models withdata measured by the Turkish State Meteorological Service(DMI) in the period from January 1990 to December 1996based on the various statistical methods. The measured dataof global and diffuse solar radiation on a horizontal surface,the number of bright sunshine hours, mean daily ambienttemperature, maximum and minimum ambient temperatures,relative humidity, and amount of cloud cover for Jeddah (lati-tude 21◦42′37′′N, longitude 39◦11′12′′E), Saudi Arabia, dur-ing the period (1996–2007) were analyzed by El-Sebaii et al.[4], Recently Viorel Badescu et al. [5,6], computed globaland diffuse solar hourly irradiation on clear sky as well asdiscussed the accuracy and sensitivity analysis for 54 mod-els of computing hourly diffuse solar irradiation on clearsky. Mehmet Kaya [7] studied the models of global solarradiation on the horizontal surface in the literature that areinvestigated, and new empirical models based on the sun-shine hour data for Erzincan, Turkey are developed. Severalresearchers have used one or more meteorological data forthe estimation of global solar radiation on horizontal surfacein Pakistan. Firoz Ahmed et al. [8], studied the prospects ofsolar energy utilization in Karachi’s. Ilyas et al. [9], stud-ied global solar radiation over Pakistan. Akhlaque Ahmedet al. [10,11], estimated global and diffuse solar radiation ofHyderabad and Lahore in recent years. Chhor is located atlatitude 25.52◦ and longitude of 69.46◦ near to Thar Desertof Sindh, Pakistan. In this study, an attempt was made by cal-culating regression coefficient from the first-order Angstromtype correlation as well as to estimate monthly average dailyglobal and diffuse radiation for Chhor. The significance ofestimating the averaged daily diffuse radiation on a horizon-tal surface is that, when the sun is obscured by thick cloud, orthe sun is below the horizon, gives a direct measurement ofthe energy received on a horizontal solar panel. Dependingon the altitude and azimuth of the sun at any moment if thesolar panel can be adjusted according to the solar angle, thereis increase in the energy received. Such estimates then leadto the obtained estimates of the energy output from any solarinstallation, depending on the efficiency of the system. In theend, mean bias error (MBE) and root mean bias error (RMSE)were also used to evaluate the accuracy of data Karakoti andDatta [12]. The objective of this study is to develop a sin-gle correlation between global radiation and bright sunshinehours, which can be used to estimate monthly average globalsolar radiation with the input of only sunshine records. This
procedure will be implemented by using the measured datafor both of these parameters for Chhor station. It also com-pares statistically the performance of our correlation.
2 Methodology
The Angstrom [13], equation relates monthly averaged dailyirradiation Ho to clear day irradiation H and the number n ofhours of bright sunshine.
H = Ho[a + b(n/N )] (1)
Ho is the monthly-mean daily radiation on a horizontal sur-face in the absence of atmosphere. To compute coefficientsa and b, the relationship given by Tiware and Sangeeta[14], can be used. The values of Ho, the radiation receivedunder the absence of any atmosphere may be calculated byMuneer [15]:
Ho = (0.024)/Isc[1 + 0.033 cos(360 DN/365)]× [cos LAT cos DEC sin Ws + (2πWs/360)
× sin LAT sin DEC (2)
where Isc is the solar constant, LAT is the latitude, DEC isthe solar declination, and Ws is the sunset hour angles.
Yallop’s algorithm [16], enables a high precision com-putation of DEC. The routine in this algorithm is valid forthe period 1980–2050 and has an accuracy of 1 min of arcfor DEC. For longer period high precession calculation, thecomplete version of VSOP theory, Bretagnon and Francou[17], may be used.
3 Prediction of Diffuse Solar Radiation
A regression between monthly averaged values of diffuse andglobal irradiation was first developed by Liu and Jordon [18],which correlate diffuse fraction (ratio of diffuse radiation Hd
and global radiation H) with clearness index which is theratio of the global radiation to extraterrestrial radiation Ho
i.e., KT = H/Ho, where Hd is the monthly average dailydiffuse radiation incident on horizontal surface.
Hd/H = 1.390 − 4.027KT + 5.53(KT)2 − 3.108(KT)3
(3)
Hawas and Muneer [19], worked on the data which was basedon long-term measurement undertaken at 13 stations in Indiafor the period 1957–1975 and the model proposed for Indiansubcontinent is
Hd/H = 1.35 − 1.61KT (4)
For temperate climates and for location close to the tropics,the correlation developed by Page [20], is widely used and isgiven as follows
Hd/H = 1.00 − 1.13KT (5)
123
Arab J Sci Eng (2014) 39:175–179 177
4 The Validation Method
The ground measurement and satellite-derived solar radia-tion data complement each other and are required to builda complete solar radiation database. The model accuracy isdetermined by comparison of modeled data series againstground-estimated parameters as shown in the studies of Boyoand Adeyemi [21], Davies et al. [22], Djemaa and Delorme[23], Perez et al. [24], Pereira et al. [25], Argirion et al. [26],Schillings et al. [27], and Lefevre and Wald [28]. The sta-tistical indicator allows models to be compared and at thesame time indicate whether or not a model’s estimates arestatistically significant at a particular confidence level Stone[29]. Mean bias error and RMSE have been found out, andthe values of correlation coefficient of determination (R2),variance (S), were also determined for regression equationdeveloped for Chhor.
MBE =[∑
(Hic − Hio)]/
n (6)
RMSE ={ [∑
(Hic − Hio)2] /
n
}1/2
(7)
In general, a low RMSE is desirable. The positive MBEshows overestimation, while a negative MBE indicatesunderestimation.
5 Results and Discussion
Table 1 gives the input parameters declination of the sun, sun-shine hour’s n, and day length N, for each month. The daylength N and Angstrom coefficients a and b were calculatedusing equation given by Tiware and Sangeet. Monthly meandaily extraterrestrial radiation Ho is estimated from Eq. (2)and values are listed in Table 2. The values of Angstrom coef-
Table 1 Input parameters forestimation of monthly averagedaily global solar Chhor
Months Declination(in deg)
Monthly meansunshine hour (n)
Monthly averageday length (N)
Coefficient“a”
Coefficient“b”
Jan −19.95 9.55 10.672 0.3911 0.329
Feb −10.63 9.88 11.316 0.384 0.344
Mar 0.15 10.15 12 0.375 0.362
Apr 11.78 10.91 12.75 0.378 0.356
May 20.13 10.91 13.33 0.366 0.382
June 23.43 9.61 13.58 0.33 0.455
July 20.13 7.31 13.34 0.266 0.565
Aug 12.18 7.62 12.76 0.294 0.536
Sep 0.78 9.87 12.04 0.366 0.382
Oct −10.61 10.53 11.31 0.402 0.304
Nov −19.86 10.2 10.67 0.410 0.287
Dec −23.43 9.40 10.41 0.393 0.324
Table 2 Monthly average dailyglobal solar radiation data Chhor Months Ho
(kWh/m2)
Hest(kWh/m2)
HNASA(kWh/m2)
KT(H/H0)
Hd/H(LJ)
Hd/H(MH)
Hd/H(Page)
Jan 6.57 4.52 4.12 0.687 0.218 0.244 0.224
Feb 7.748 5.34 4.89 0.689 0.22 0.241 0.222
Mar 9.115 6.25 5.61 0.686 0.228 0.246 0.225
Apr 10.295 7.06 6.30 0.686 0.228 0.246 0.225
May 10.948 7.45 6.51 0.681 0.229 0.254 0.231
June 11.145 7.27 6.56 0.652 0.281 0.301 0.264
July 11.012 6.36 5.85 0.577 0.306 0.422 0.348
Aug 10.511 6.42 5.57 0.611 0.282 0.367 0.31
Sep 9.508 6.43 5.55 0.677 0.231 0.261 0.235
Oct 8.129 5.55 4.95 0.683 0.228 0.251 0.229
Nov 6.835 4.66 4.20 0.683 0.228 0.251 0.229
Dec 6.222 4.26 3.90 0.685 0.228 0.248 0.226
123
178 Arab J Sci Eng (2014) 39:175–179
ficient are then used to estimate monthly average daily globalsolar radiation Hest using Eq. (1) and are given in Table 2.Measured data are taken from NASA satellite for the globalsolar radiation HNASA as shown in Table 2. The transparencyof the atmosphere is indicated by KT fraction of extraterres-trial radiation that reaches the earth surface as global solarradiation. It is a measure of the degree of clearness of thesky. KT is calculated in Table 2 and values show that sky isvery clear almost throughout the year (i.e., 60 %). The ratioof the monthly average daily diffuse radiation to the monthlyaverage daily global radiation (Hd/H ) is calculated usingEqs. (3), (4) and (5) from different diffuse solar radiationmodels. The diffuse fraction is calculated for Liu and Jor-don, Page, Hawas and Muneer models and given in Table 2.
Table 3 Monthly averaged daily diffuse solar radiation data Chhor
Months Hd, LJ(kWh/m2)
Hd, MH(kWh/m2)
Hd, Page(kWh/m2)
Hd, NASA datasatellite-based(kWh/m2)
Jan 0.98 1.102 1.012 1.04
Feb 1.17 1.286 1.185 1.26
Mar 1.425 1.537 1.406 1.62
Apr 1.609 1.736 1.588 1.93
May 1.706 1.892 1.72 2.19
Jun 2.04 2.188 1.919 2.27
Jul 1.946 2.683 2.21 2.37
Aug 1.81 2.356 1.99 2.21
Sep 1.48 1.678 1.51 1.81
Oct 1.265 1.393 1.27 1.43
Nov 1.06 1.169 1.067 1.13
Dec 0.97 1.056 0.962 0.96
Table 4 MBE and RMSE for global and diffuse radiation of Chhor
Radiation MBE (KWh/m2) RMSE (KWh/m2)
Global −0.63 0.65
Diffuse −0.12 0.15
There is remarkable agreement between the estimated andmeasured values for global solar radiation. The maximumof global radiation for the month of June and July is quiteappreciable. Table 3 gives the values of diffuse solar radi-ation. Diffuse solar radiation is not commonly measured inany meteorological station of Pakistan; therefore, the dif-fuse solar is estimated by Liu and Jordon, Page, Hawas andMuneer method using Eqs. (3), (4) and (5). The estimated val-ues of diffuse solar radiation by Liu and Jordon, Page, Hawasand Muneer models are listed in columns (1–3), respectively,of Table 3.
NASA measured values of diffuse solar radiation are alsotaken into account. There is a good agreement between theestimated and measured values and the NASA satellite datafor diffuse solar radiation and was best fitted with the valuesestimated from Muneer and Hawaas model. The statisticalerrors (MBE and RMSE) for global and diffuse solar radi-ation are given in Table 4. The RMSE and MBE values arevery low, indicating fairly good agreements. MBE has low-est values for diffuse solar radiation than global radiationand shows overestimation (the negative values indicate thatthe present correlations slightly overestimate H). RMSE alsohas smaller values for diffuse radiation than global radiationdata.
The value of coefficient of determination implies that98.7 % of NASA can be accounted by estimation of globalsolar radiation Chhor as shown in Fig. 1. Low values of vari-ance suggest that the correlation is best fitted.
6 Conclusion
The estimated values of global and diffuse solar radiationsuggest that solar radiation can be used as an alternate energyresource for this region. The difference in ground-based andsatellite-derived data was small on clear sky. NASA satellitedata of diffuse radiation show good agreement with the valuesfrom Muneer and Hawaas model. The result obtained showsthat the solar energy utilization has bright prospects in Chhor.The analysis of the estimated and measured values of H
Fig. 1 Fitted regression line forthe estimated and NASAsatellite data of Chhor
123
Arab J Sci Eng (2014) 39:175–179 179
shows that the maximum values of global solar radiation areobserved in June while the minimum appeared in December.It is further suggested that Hawaas and Muneer model maybe best for an estimation of monthly daily diffuse radiationover this region.
Acknowledgments The data of global and diffuse solar radiationwere obtained from the NASA (Langley Research Center AtmosphericSciences Data Center POWER Project) [30]. The Pakistan Meteorolog-ical Department, Karachi, Sindh has also kindly provided the data setsof sunshine duration utilized in this study.
References
1. Badescu, V.: Modeling solar radiation at the earth’s surface. In:Recent Advances, vol. 7. Springer, New York (2008) (ISBN: 978-3-540-77454-9)
2. Guofeng, W.; Liu, Y.: Methods and strategy for modeling dailyglobal solar radiation with measured metrological data—a casestudy in Nanchang station, China. Energy Convers. Manag. 48,2447–2452 (2007)
3. Aras, H.; Balli, O; Hepbasli, A.: Global solar potential. Part 1:model development. Energy Sources Part B 1(3), 303–315 (2006)
4. El-Sebaii, A.A.; Al-Ghamdi, F.S.; Al-Hazmi; Faidah, A.S.: Esti-mation of global solar radiation on horizontal surfaces in Jeddah,Saudi Arabia. Energy Policy 37, 3654–3649 (2009)
5. Badescu, V.; Gueymard, C.A.; Cheval, S.; Oprea, C.; Baciu, M.;Dumitrescu, A.; Iacobescu, F.; Milos, I.; Rada, C.: Computingglobal and diffuse solar hourly irradiation on clear sky. Review andtesting of 54 models, Renew. Sustain. Energy Rev. 16, 1636–1656(2011). doi:10.1016/j.rser
6. Badescu, V.; Gueymard, C.A.; Cheval, S.; Oprea, C.; Baciu, M.;Dumitrescu, A.; Iacobescu, F.; Milos, I.; Rada, C.: Accuracy andsensitivity analysis for fifty-four models of computing hourly dif-fuse solar irradiation on clear sky. Theor. Appl. Climatol. (2012).doi:10.1007/s00704-012-0667-1
7. Kaya, M.: Estimation of global solar radiation on horizontal surfacein Erzincan, Turkey. Int. J. Phys. Sci. 7(33), 5273–5280 (2012)
8. Ahmad, F.; Naqvi, S.A.: Characteristic distribution of total, directand diffuse solar radiation at Karachi. Pak. J. Sci. Res. 24(5–6),171 (1981)
9. Ilyas, S.Z.; Nasir, Sh.M.: Estimation and comparison of diffusesolar radiation over Pakistan. Int. Sci. J. Altern. Energy Ecol. 3, 47(2007)
10. Akhlaque, M.; Ahmad, F.: Estimation of global and diffuse solarradiation for Hyderabad, Sindh, Pakistan. JBAAS 5(2), 73–77(2009)
11. Akhlaque, M.; Ahmad, F.: Distribution of total diffuse solar radia-tion at Lahore, Pakistan. J. Sci. Res. XXXX. 1, 37–42 (2010)
12. Karakoti, I.; Datta, A.: Validation of satellite based solar radiationdata with ground measurements. In: 11th ESRI India user confer-ence (2010)
13. Angstrom, A.: Solar and terrestrial radiation. Q. J. R. Met. Soc. 50,121–126 (1924)
14. Tiwari, G.; Suleja, N.: Solar thermal engineering system. NarosaPublishing House, India (1997)
15. Muneer, T.: Solar radiation and day light modeling, vol. 2. Elsevier,London (2004) (ISBN: 075065974)
16. Yallop, B.D.: In: Technical note. Royal Greenwich Observatory,Cambridge (1992)
17. Bretagnon, P.; Francou, G.: Planetary theories in rectangular andspherical variables. VSOP87 solutions. Astron. Astrophys. 202,309–315 (1988)
18. Liu, Y.; Jordan, H.: The inter relationship and characteristic distri-bution of direct, diffuse and total solar radiation from metrologicaldata. Solar Energy 4, 1 (1960)
19. Hawas, M.; Muneer, T.: Study of diffuse and global radiation char-acteristics in India. Energy Convers. 24, 143 (1984)
20. Page, J.K.: The estimation of monthly mean value of daily shortwave irradiation on vertical and inclined surfaces from sunshinerecords for latitudes 60◦N to 40◦S. In: BS32. Department of Build-ing Science, University of Sheffield, UK (1977)
21. Boyo, A.O.; Adeyemi, K.A.: Analysis of solar radiation data fromsatellite and Nigeria Meteorological Station. IJRER 1(4), 314–322(2011)
22. Davies, J.A.; McKay, D.C.; Luciani, G.; Wahab, M.A.: Validationsof models estimating solar radiation on horizontal surfaces. In:Task IX final report to the solar heating and cooling program of theInternational Energy Agency (1988)
23. Djemma, B.; Delorme, C.: A comparison of one year daily globalirradiation from ground-based measurements versus meteo satimages from seven locations in Tunisia. Solar Energy 48, 325–333(1992)
24. Perez, R.; Seals, R.; Zelenka, A.: Comparing satellite remote sens-ing and ground network measurements for the production forsite/time specific irradiance data. Solar Energy 60, 89–96 (1997)
25. Pereira, E.B.; Martins, F.R.; Abreu, S.L.; Beyer, H.G.; Colle, S.et al.: Cross validation of satellite radiation mode during SWERAproject in Brazil. In: Proceedings of the ISES solar world congress,Goteborg, CD-ROM, paper 06 23 (2003)
26. Argiriou, A.; Lykoudis, S.; Kontoyiannidis, S.; Balaras, C.A.;Asimakopoulos, D.; Petradis, M.; Kassomenos: Comparison ofmethodologies for TMY generation using 20 years data for Athens,Greece. Solar Energy 66, 33–45 (1999)
27. Schillings, C.; Meyer, R.; Mannstein, H.: Validation of a methodfor deriving high resolution direct normal irradiance from satellitedata and application for the Arabian Peninsula. Solar Energy 76,485–497 (2003)
28. Lefevre, M.; Wald, L.: Using reduced data sets ISCCPB2 from themeteosat satellites to assess surface solar irradiance. Solar Energy81, 240–253 (2007)
29. Stone, R.J.: Improved statistical procedure for the evaluation ofsolar radiation estimation models. Solar Energy 51, 289 (1993)
30. NASA: Surface meteorology and solar energy data and informa-tion. Accessed October (2010)
123