Validating satellite-based solar radiation data with ground measurements

Validating satellite-based solar radiation data with ground measurements

SHARE

Introduction
The power obtained from solar radiation reaching the earth is many times greater than the power generated by man by other conventional sources. Solar radiation is a perpetual source of natural energy and has great potential for a wide variety of applications. It is abundant, accessible and pollution free and hence can be used as a supplement to the non-renewable sources of energy which have finite depleting supplies [1-3]. The renewable energy sector depends upon the assessment of resources for planning and marketing their energy production technology. For solar based renewable energy technologies, the basic source is the Sun [3].

India is located between 60N to 320N latitude and hence most of the locations in India receive abundant supply of solar energy. This energy can be utilised for the development of solar energy systems. For this purpose the solar radiation data are required in various forms depending upon the nature of application. Numerous researchers e.g. Riordan et al. [4], Aladas et al. [5], Hirata and Tani [6] and Antonio et al. [7] have predicted the importance of solar radiation data for design and efficient operation of solar energy systems. Proper design and performance of these solar appliances require accurate information on solar radiation availability [8]. Hence many studies [9-11] have been made on solar climates of specific regions for the development of solar technologies.

The ground measured and satellite derived solar radiation data complement each other and are required to build a comprehensive solar radiation database. It is difficult to have high capability solar radiation monitoring network and also accuracy in the interpolation of data decreases with the increase in distance between sites. However, satellite measurements have lack of accuracy and short time interval data are needed for the engineering and site-specific studies. Hence, combining these two, ground- based and satellite- derived measurements create a comprehensive solar radiation database.

Methodology
The analysis is carried out for the North West region (Ahmadabad, Bhopal, Jodhpur, Jaipur and New Delhi) of the country. The ground based data and satellite (Meteosat 5 & 6) data of horizontal and tilted global radiation are used in the study. The ground measured data are taken from Indian Meteorology Department (IMD) while satellite data are derived from the maps developed by National Renewable Energy Laboratory (NREL). Comparative plots of ground-based measurement and satellite derived data are drawn for global horizontal and global tilted radiation.

The three statistical parameters – mean percentage error (MPE), root mean square error (RMSE) and mean bias error (MBE) are used in the validation of data.

The mean percentage error is

RMSE provides information on the short-term performance of a model. The lower the RMSE, the more accurate is the estimation [12]. MBE provides information on the long-term performance of a model. Positive and negative MBE values show overestimation and underestimation respectively [12]. However, a zero value of RMSE or MBE is ideal.

The ground measurements and satellite derived data are plotted against months to test the systematic difference.

The statistical errors are computed for the each of the five stations viz. Ahmadabad, Bhopal, Jodhpur, Jaipur and New Delhi and then plots showing comparison among the different cities are drawn for each of the statistical parameter.

Validation

Comparative study on global radiation
The ground based and satellite-derived data on monthly mean daily global radiation for all the five station viz. Ahmedabad, Bhopal, Jodhpur, Jaipur and New Delhi are listed in Table 1 and Table 2 respectively. It can be seen that the satellite derived values are higher than the ground measured in all months of the year for Ahmedabad and Jodhpur. In the case of Jaipur and New Delhi, ground based values slightly exceed the satellite data in the months of April and December respectively.


Table 1: Ground measured global horizontal radiation (kWh/m2)

Table 2: Satellite derived global horizontal radiation (kWh/m2)

The statistical errors (MPE, MBE and RMSE) for global radiation are given in Tale 3. The extreme values of MPE are for New Delhi (-1.756%) and Ahmadabad (-7.621%). For other cities Bhopal, Jodhpur and Jaipur, mean percentage error is -7.559 %, -6.243 % and -5.742 % respectively.

MBE is lowest for Jaipur (0.278 kWh/m2) while highest is for Ahmadabad (0.401 kWh/m2). For Bhopal, Jodhpur and New Delhi, it is 0.385 kWh/m2, 0.351 kWh/m2 and 0.105 kWh/m2 respectively. MBE values show overestimation for all of the locations. RMSE ranges from 0.177 kWh/m2 to 0.416 kWh/m2being minimum for New Delhi and maximum for Ahmadabad as illustrated in listed in Table. The other stations Bhopal, Jodhpur and Jaipur have 0.416 kWh/m2, 0.394 kWh/m2 and 0.392 kWh/m2.

Comparative study on tilted solar radiation
The ground based and satellite-derived data on monthly mean daily global tilted radiation for all the five stations are tabulated in Table 4 and Table 5 respectively. It is obvious that for all five cities the tilted radiation derived from satellite image exceeds the ground measured values throughout the year.


Table 3: MPE, MBE and RMSE for global horizontal radiation

Table 5: Ground measured data of global tilted radiation (kWh/m2)

Table 7 lists the statistical parameters for tilted radiation. MPE has extreme values of -8.097% (Ahmadabad) and -10.482% (Jodhpur). It is -8.397%, -9.7528% and -10.237% for the other stations Bhopal, Jaipur and New Delhi respectively. The MBE being a true measure of difference between the ground based and satellite derived data is lowest for Ahmadabad (0.274 kWh/m2) and highest for Jodhpur (0.611 kWh/m2). The remaining locations have 0.44 kWh/m2 (Bhopal), 0.213 kWh/m2 (Jaipur) and 0.538 kWh/m2 (New Delhi) mean bias error.


Table 6: Satellite derived of global tilted radiation (kWh/m2)

Table 7: MPE, MBE and RMSE for global tilted radiation

Again Ahmadabad has the lowest RMSE (0.422 kWh/m2) for tilted radiation and Jodhpur has the highest one (0.628 kWh/m2). The other cities have values of 0.526 kWh/m2 (Bhopal), 0.628 kWh/m2 (Jaipur) and 0.606 kWh/m2 (New Delhi). It is obvious that Ahmadabad has smallest while Jodhpur has highest values of all the three statistical indicators. This may be due the different atmospheric conditions of the stations as in the clean atmosphere and clear sky the errors will be less and vice versa.

Conclusions
The difference in ground based and satellite derived data should be small on clear sky or totally overcast days. When examining global horizontal radiation, it is found that this dissimilarity is comparatively less mainly in February and March which is obvious from the fact that in these months sky is relatively clear. The variation between the data is high in July for all the stations. New Delhi has the smallest values of mean percentage error and mean bias error while Ahmadabad has the largest. The root mean square is high between the ground measured and satellite based data which is because of the fact that ground –based sample one small area of the sky while satellite are imaging larger area. It is minimum in New Delhi and maximum in Ahmadabad.

In the case of global tilted radiation, the high difference is observed in the winter (mainly in October). The errors are comparatively high and all the three statistical parameters used in the study have smallest value for Jodhpur and largest for Ahmedabad.

Applications

High resolution GIS maps
The validated solar radiation data are used in three mapping application examples that exploit the main strengths of the model:

  1. A high resolution (10 km) geostationary satellite archive for the North – West regions of India which has an ability to account for site-specific average resource is shown in Figure 1.
  2. Whole year is divided into four seasons- first (Jan, Feb & Mar), second (Apr, May & Jun), third (July, Aug & Sep) and fourth (Oct, Nov & Dec). The seasonal Irradiance maps of Global & Tilted Solar Radiation showing the seasonal variation are illustrated in Figures 2 and 3.
  3. The Peak Summer Irradiance maps of Global & Tilted Solar Radiation as shown in Figure 4, are the distribution of solar radiation in peak summer month May.

Figure 1: Annual Distribution of global horizontal and tilted radiation

Figure 2: Seasonal variation of global horizontal radiation for four seasons

Figure 3: Seasonal variation of global tilted radiation for four seasons

Figure 4: Global horizontal and tilted radiation for the peak summer month (May)

References:

  1. Ra IGD. Solar energy utilization. Naisarak, Delhi:Khanna Publisher, 1984, p. 24-25.
  2. Herman JR. Ultraviolet radiation at the Earth’s surface, 1999; 951-956.
  3. Khalil Samy A. Parameterization models for solar radiation and solar technology applications.Energy Conversion and Management 2008,doi:10.1016/j.enconman.2008.01.023.
  4. Riordan C et al. Solar radiation research for photovoltaic applications. Solar Cells 1991; 30: 489-500.
  5. Aladas-Arboledas L, Batlles FJ and Olmo F. Solar radiation resource assessment by means of silicon cells. Solar Energy 1995; 5(3): 183-192.
  6. Hirata Y and Tani T. Output variation of photovoltaic modules with environmental factors. Part one: the effect of spectral solar radiation on photovoltaic module output. Solar Energy 1995; 55(6): 463-468.
  7. Antonio P, Angelo S and Luciano R. Effect of solar irradiation conditions on the outdoor performance of photovoltaic modules. Opt Commun 1998; 153: 153-163.
  8. Jacovides CP et al. Comparative study of various correlations in estimating hourly diffuse fraction of global solar radiation. Renewable Energy 2006; 31: 2492-2504.
  9. Al-Aruri SD. The empirical relationship between global radiation and global ultraviolet (0.290-0.385) µm solar radiation components. Solar Energy 1990; 45: 61-64.
  10. Kudish AI, Ianetz A. Analysis of daily clearness index, global and beam radiation for Beer Sheva, Israel: partition according to day type and statistical analysis. Energy Conversion and Management 1996; 37: 405-416.
  11. Ianetz A et al. A method for characterization and inter-comparison of sites with regard to solar energy utilization by statistical analysis of their solar radiation data as performed for three sites in the Israel Negev region. Solar Energy 2000; 69: 283-293.
  12. Jayanta Deb Mondel et al. Solar radiation modeling for the simulation of photovoltaic systems, Renewable Energy, 33 (2008), 1109-20.

Note: This paper was presented at IGF 2012 held in Gurgaon