The installed capacity of grid-tied photovoltaic (PV) systems around the globe is increasing rapidly due to the need for clean, sustainable and fuel independent energy. The PV technology is already a major part of the energy mix in many developed and developing countries. The electricity generated from these systems heavily depends on the prevailing weather conditions and is directly related to the available solar irradiance. Therefore, the electricity injected into the grid from these systems is intermittent reducing the utility of that generated power. Any sudden irradiance changes, due to passing clouds, rain, dust storms, etc., directly affect the power response of solar systems, compromising the security of electricity supply and resulting in the need for operating reserves to stabilize the supply. The further deployment of grid connected PV systems will benefit greatly by developing the methodology to systematically study and analyse the variability and quantity of the power produced by installed PV systems. This is done in this study by identifying existing patterns of the solar irradiance at the particular location of interest using historical data. The results of this research will eventually lead to a highly reliable and accurate forecasting of the energy production from such PV systems. In this domain, accurate and reliable PV supply forecasting will significantly increase its utility and will reduce the need to rely on conventional (fuel based) supply. In particular, solar irradiance has a direct bearing on the performance of PV systems and the quality of the energy supplied to the grid. The power production at a given location can be characterised via the quantity and the quality index. The quantity index reflects the amount of power produced, which mainly affects the scheduling of the centralised electricity generation by the system operator. In cases where PV penetration is significant this has to be supported by an appropriate energy mix (conventional units, energy storage systems (ESS), etc.) [1–3]. On the other hand, the quality index defines the frequency and ramp power of the fluctuations of the PV energy produced, caused mainly by passing clouds. Frequent and large fluctuations cause potential problems to the grid compromising the security of supply. Consequently, the quality of the resource dictates the corrective action that should be implemented to avoid grid problems. These actions can be either grid integrated energy storage systems or the allocation of appropriate spinning reserve in order to fill the energy valleys [3]. In this work a method for classifying and characterising the solar irradiance based on real outdoor measurements is outlined by calculating a quantity and quality index for each day. The analysis was performed with 1-minute resolution global horizontal irradiance (GHI) measurements and validated with 4 years of recorded data. The data are extracted from weather stations located between the 23rd and 27th parallels north, as indicated in Table 1. Different locations around similar latitudes are chosen in order to examine the solar irradiance behaviour at locations sharing similar climatic conditions and daytime periods. Furthermore, 4 years of data are used from each weather station in order to evaluate the repeatability of patterns identified, thus improving solar systems highly predictability. Locations of weather stations.

The quantity and quality of solar irradiance is of great importance as these can determine the possibilities and shortcomings that solar systems have in a region. In the scope of characterising the solar irradiance at the earth's surface, two important parameters are necessary, the instantaneous sky clearness index, kd, and the probability of persistence, POP day. The instantaneous sky clearness index is the ratio of the received irradiance at the earth surface to the extraterrestrial radiation. This index captures the instantaneous fluctuations of the solar irradiance and indicates the quantitative amount of solar irradiance the surface of the earth receives [4–5]. The quantity index, kday, is defined as the ratio between the daily received solar radiation to the daily extraterrestrial radiation. Consequently, the higher the quantity index the higher the amount of daily solar radiation available. Additionally, the quality of solar irradiance during a day can be found using a probabilistic approach. Firstly, an array Δkd is calculated containing the difference between consecutive values of clearness index, kday, within a day. Accordingly, the quality index, POP day for the day can be estimated by finding the probability of the Δkd values being equal to zero. Therefore, the higher the value of POP day for a day the lower the fluctuations to appear during that day. As a result, each day a pair with the daily value of the clearness index, kday and the probability of persistence, POP day is defined. The daily solar irradiance can be represented on a two-dimensional plot, where the “x” and “y” axes are the daily values of kd and POP day respectively. The plot of kday against POP day is divided into 9 classes, as shown in Fig. 1. Plot of daily solar irradiance characterisation and classification classes.

With reference to Fig. 1, the x-axis is divided into three sections based on the quantity of solar irradiance. Particularly they are divided into high quantity (classes 1, 4 and 7 i.e. 0.6 < kd), medium quantity (classes 2, 5 and 8 i.e. 0.3 < kd < 0.6) and low quantity (classes 3, 6 and 9 i.e. kd < 0.3). The y-axis, depicts the quality of solar irradiance and is similarly divided into 3 sections based on the quality of the daily sky conditions. Those are: clear or totally overcast sky with no or few fluctuations (classes 1, 2 and 3 i.e. 0.9 < POPd), relatively small and infrequent fluctuations (classes 4, 5, and 6 i.e. 0.7 < POPd < 0.9) and large and frequent fluctuations (classes 7, 8 and 9 i.e. 0.5 < POPd < 0.7). Typical solar irradiance profiles of each class can be found in Fig. 2. Typical profiles of daily solar irradiance plots for each class.

The solar irradiance data for this work were obtained from the “World Radiation Monitoring Center” and the “National Renewable Energy Laboratory (NREL)” [6]. The weather stations are located at latitudes between the 23rd and 27th parallel north, as in Table 1, covering the range of latitudes Qatar lies within. The resolution of the data is 1 minute and 4 years of data are used for the analysis. The extraterrestrial irradiance data are extracted from the online “Solar Calculator SOLPOS” of NREL. The methodology described here is global and to illustrate its functionality we will use Cyprus as a case study. The weather station in Cyprus is located at the southernmost part of the island (34.597N, 32.987E) and is operated by the PV Laboratory of the University of Cyprus. The characterisation of the daily sky conditions in Cyprus for 4 years showed that the highest concentration of days in Cyprus is found in classes 1, 4 and 5, where 321 days (88%) of a year experience high or medium quantity solar irradiance with rare and infrequent solar fluctuations. Moreover, from Table 2, it can be noted that 74.8%, are days experiencing high solar irradiation compared to 2.7% of days with low quantity solar irradiance. Also the quality of solar irradiance in Cyprus is very high with 52.5% of the examined days experiencing solar irradiance with small and very infrequent fluctuations (classes 1, 2 and 3). Daily solar irradiance percentile distribution into the 9 classes for Cyprus.

Additionally, examining Fig. 3 it can be clearly noted that the distribution of days over the evaluation period in Cyprus exhibit high periodicity. This can also be seen from centroid of the distribution points for all 4 years. The centroid values of kday and POPday are very similar throughout the years, located around 0.87 for POPday and 0.64 for kday. Daily solar irradiance distribution in Cyprus for the years 2010–2014.

Moreover, Table 3 shows the distribution of the solar irradiation measured at the SOV weather station located in Saudi Arabia (24.91N, 46.41E) for 4 years. As in the case of Cyprus the highest concentration of days in Saudi Arabia is found in classes 1, 4 and 5 representing 92% of a year or 336 days. However, the concentrations of days in SOV is higher for class 1 compared to Cyprus (66.3 and 48.4% respectively). This reveals that the prevailing weather condition in that region is clear sky thus high quantity and quality of solar irradiance. Daily solar irradiance percentile distribution into the 9 classes for SOV Weather Station.

Additionally, a close look at Fig. 4 reveals the repeatability and periodicity of patterns existing in the region. From a visual inspection it can be clearly seen that the distribution of data points is very similar for each year and also many data points through the years are overlapping. This can also be noticed from the centroids of the distributed data. The centroids for each year are very closely located on the plot and their mean value is 0.91 POP day and 0.68 for kday. Daily solar irradiance distribution at SOV weather station for the years 1999–2002.

These facts show that grid connected solar systems in Cyprus and the SOV region (24th parallel north) produce and deliver quality energy to the grid, without compromising the efficiency and quality of energy. Finally, the analysis reveals that the solar irradiance in both cases is highly predictable as repeating patterns are identified comparing all 4 years of data. This leads to the conclusion that the grid operators can rely on solar systems without compromising the quality and security of the supplied electricity in situations with higher penetration of grid-connected solar systems. The authors gratefully acknowledge the “World Radiation Monitoring Centre-Baseline Surface Radiation Network” for supplying the weather data from Saudi Arabia.


[1] M. A. Ortega-Vazquez and D. S. Kirschen, “Estimating the spinning reserve requirements in systems with significant wind power generation penetration,” IEEE Trans. Power Syst., vol. 24, no. 1, pp. 114–124, 2009.

[2] N. Amjady and F. Keynia, “A new spinning reserve requirement forecast method for deregulated electricity markets,” Appl. Energy, vol. 87, no. 6, pp. 1870–1879, 2010.

[3] M. Black and G. Strbac, “Value of bulk energy storage for managing wind power fluctuations,” IEEE Trans. Energy Convers., vol. 22, no. 1, pp. 197–205, 2007.

[4] B. O. Kang and K. Tam, “New and improved methods to estimate day-ahead quantity and quality of solar irradiance,” Appl. Energy, vol. 137, pp. 240–249, 2015.

[5] B. O. Kang and K.-S. Tam, “A new characterization and classification method for daily sky conditions based on ground-based solar irradiance measurement data,” Sol. Energy, vol. 94, pp. 102–118, Aug. 2013.

[6] National Renewable Energy Laboratory, “SOLAR and LUNAR POSITION CALCULATORS.” [Online]. Available: http://www.nrel.gov/midc/solpos/.


Article metrics loading...

Loading full text...

Full text loading...

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error