Estimation of the hourly solar radiation

: Hybrid facility investments from the renewable energy sources have been increased in recent years. In general, solar power is that the secondary energy source in hybrid systems. The reason why solar energy is most commonly used in hybrid power systems is that solar energy is cheaper than other types of renewable energy. Solar-hydroelectric (SHE) is one among the foremost compatible hybrid energy pairs, as solar and hydroelectric power generation profiles complement one another. Hybrid systems consisting of solar and hydropower have complementary characteristics due to the shared use of infrastructure systems and different periodicities in power generation. Energy management is important for SHE integrated facilities. Since only a limited amount of energy can be injected into the grid from transformer capacities, energy management in hybrid systems is of great importance. To manage energy in hybrid energy systems, the amount of energy that can be produced each hour must be determined. In hybrid energy plants, there is usually already another renewable energy plant in place, and solar energy is added on top and optimized. Since there are no pyrometers in the existing plants, the daily radiation data from National Aeronautics and Space Administration (NASA) is used, but the daily energy production amount may be insufficient for accurate energy management. To realize this, it’s necessary to reveal the energy generation on an hourly basis. During this study, the quantity of radiation on an hourly basis determined to calculate solar power generation. Empirical and econometric models utilized in radiation amount determination were performed, and also the most appropriate method was clarified by comparing with one another. Hourly based radiation is achieved with an empirical method by using National Aeronautics and Space Administration (NASA) daily radiation.


Introduction
Energy management and hybrid energy became the foremost significant issue to make sure the sustainability of energy production.In general, solar energy is that the most renewable energy hybrid pair.Therefore, it's necessary to work out energy generation amounts on an hourly basis and to confirm energy management.Solar energy may be a style of energy which will be easily accessed everywhere the globe.Electricity generation from solar energy is within the sort of function of radiation and temperature.Meteorological measurements are made regionally in many parts of the planet.The radiation amount is obtained from satellites.The NASA official website has daily irradiance data for any coordinate everywhere the planet.However, these data must be obtained on an hourly basis for energy management.The quantity of radiation on an hourly basis is obtained by basically two methods.One in every of them is that the empirical method, and also the other is that the econometric statistical method.During this study, both approaches were run separately and compared with actual data.The literature reviews are given within the following paragraphs.With the widespread use of hybrid energy, energy management is becoming more important.The main reason is that the capacity that can be fed into the grid does not change, no matter how much the installed power increases.For example, if a wind turbine with an installed capacity of 80 MW, which has the right to inject energy into the grid with an installed capacity of 50 MW, is additionally supported by solar energy with an installed capacity of 100 MW, the amount of energy that the turbine can inject into the grid will not be 180 MW, which is the total installed capacity of the two production facilities, but the amount of electricity injected into the grid will still be 50 MW.Therefore, energy management in hybrid energy systems is critical to the continuity of the system.In hybrid energy, solar energy is mostly used.In hybrid projects using solar energy, the amount of radiation should be determined on an hourly basis to ensure natural and reliable energy management.The amount of radiation is one of the main factors that determine the amount of energy production, as it directly affects the amount of electricity produced by the PV panels.With this study, the solar radiation amounts are accurately and quickly determined every hour and will be used for both efficient power optimization and energy management.
Tırmıkçı used the mathematical models and equations to see the foremost suitable angle in Sakarya province in her doctoral thesis.During this study, Tırmıkçı revealed a real-time comparison of the twoaxis scheme positioned with the sun and therefore the fixed scheme.Tırmıkçı worked on position with the foremost appropriate annual angle [1] .Jacovides et al. stated versatile correlations in determining the hourly solar radiation.During this study, radiation measurements for Cyprus supported the experiment, which was briefly defined empirically, by making use of the mathematical models previously stated within the literature studies by other researchers.By calculating the coefficients, hourly radiation is estimated [2] .Iqbal correlated the mean beam and diffuse radiation.Monthly average radiation amounts with the knowledge obtained from the stations for various cities of Canada are calculated [3] .Alsafadi and Başaran Filik used a design algorithm like machine learning mechanism for hourly solar radiation.Estimated the quantity of radiation on an hourly basis by applying machine learning (ML-machine learning) to empirical models for Eskişehir [4] .Serttaş proposed a brand-new pattern scan-based approach to radiation forecasting [5] .Maleki et al. estimated the global solar radiation on disposed sides.Empirical radiation simulation developed especially within the last ten years for a neighborhood [6] .Chandel and Aggarwal have worked on a one-hour energy model that will provide reliable and accurate finishes up within the western Himalayas.This study has been cited as a reference in many other articles [7] .
Khatib and Elmenreich proposed a new approach for obtaining hourly solar radiation.In this study, an artificial neural network is used to generalize daily solar radiation [8] .Berrizbeitia et al. stated experimental simulation for determining solar radiation.During this study, a review and experimental analysis were emphasized that daily radiation amounts are often easily obtained from NASA sources, but it had been emphasized that a model should be developed for the hourly radiation amount obtained on an hourly basis, and empirical coefficients for 19 different regions were tested [9] .Gueymard stated a new estimation approach for solar radiation.During this study, the number of radiation on a monthly and average hourly basis using two new models in line with the information provided by an outsized nation-state and lots of information stations.The empirical coefficients proposed during this study were also utilized in studies conducted by other researchers [10] .Al-Rawahi et al. suggested a new estimation method for Oman.During this study, the amount of radiation is calculated on an hourly basis from the available daily radiation amounts for Oman [11] .Ishola et al. developed a territorial correction factor for the estimation of hourly based solar irradiance.A comparison was made specific to the estimation model that determined ten regions for radiation amount and also the coefficients determining the regional radiation [12] .Hussein estimated hourly global radiation in Egypt employing a mathematical model.A mathematical model prediction calculates the amount of radiation for Egypt on an hourly basis, using meteorological data [13] .Within the studies examined, general and conceptual models were formed by fitting and measurements made in numerous parts of the world and at different times into mathematical models [1][2][3][4][5][6][7][8][9][10][11][12][13] .Among these models, mathematical models and experimental coefficients, which are briefly called empirical, were employed within the studies created by Maleki et al. [6] , Chandel and Aggarwal [7] , Gueymard [10] .In addition to empirical coefficient and prediction models, statistical models are also used when determining the quantity of radiation.When using statistical models, forward-looking estimations are made by using mostly actual data.For the regions where data are going to be obtained on an hourly basis, high correlation results are often obtained due to statistical methods.Huang et al. forecasted solar radiation by using continuance employing some auto-regressive and space (CARDS) models.During this study, the authors emphasized that the amount of radiation is seriously tormented by weather changes like cloudiness, during this case, the only approaches today are a statistic.It's stated some methods exist in computing networks and random variables called stochastically.Within the study, a self-connected (AR-autoregressive) model is used together with a dynamic system model.As a result of the study, it was stated that the proposed system confirmed and improved the predictions by around 30% [14] .Mbaye et al. estimated the solar availability of the Dakar site by using the ARMA model.A self-correlated dynamic average (ARMA-autoregressive moving average) model was used for short-term radiation estimation.Data and validations were obtained from a station located in Dakar.Data represent between October 2016 and September 2017 and are presented on an hourly basis.As a result of the study, it was emphasized that the ARMA method is reliable [15] .Adejumo and Suleiman estimated solar irradiance by using the ARMA-GARCH approach.A combined self-correlated dynamic mean model (ARMA) and generalized self-correlated and variable conditions model (GARCH-generalized autoregressive conditional heteroscedasticity) were employed during this study.For the southern element of Nigeria, an estimate was made for the amount of radiation [16] .Ghofrani and Alolayan used statistics for renewable energy forecasting.They stated statistics and connections that will be used for renewable energy and gave information about their general framework [17] .Hassan et al. estimated the daily based solar radiation of the United Arab Emirates (Al-Ain) by using an econometric statistics model.During this study, statistic connection models (ARIMA) radiation amount estimates including autoregressive integrated moving average were used.10 years of knowledge were taken (1995-2004) and a two-year estimation (2005-2007)  was made within the sunshine of these data.Correlation studies were also performed for the results obtained and thus the results were compared [18] .Prajapati and Sahay worked on a survey paper on the solar irradiance forecasting method.They used different and reliable statistical methods for the estimation of radiation amounts and also the results were explained [19] .Huang et al. used the ARMA model for estimating the solar energy capability.Short-time radiation estimation was made using the MATLAB program and ARMA statistical model [20] .
Ferrari et al. used statistical models approach for radiation prediction.The climatic data obtained for the town of Milan estimated the amount of radiation within the framework of the numerous statistical models and compared them with each other [21] .Marchesoni-Acland et al. analyzed the performance of the ARMA models.During this study, the quantity of short-term radiation employing a self-repetitive algorithm using satellite data is estimated [22] .Colak et al. carried out ARMA and ARIMA models for determining global solar radiation.The statistic model was constructed comparisons of ARMA and ARIMA models were made to this model and radiation amount estimation was made with the best results [23] .Diagne et al. reviewed a global solar radiation prediction model.During this study, they stated that solar energy contains a fluctuating production trend because the number of radiation changes very rapidly over time, and thus the irradiance amount estimation isn't effective in solar power.Statistical models were used for the quantity of radiation and compared with each other [24] .Alsharif et al. stated a statistic ARIMA model is utilized for the estimation of solar irradiance.This effort is performed as a case study for Seoul, the Republic of Korea.They used seasonal statistical methods (SARIMA) for the town of Seoul.Connections were established and comparisons were made employing an information set of roughly 37 years (1981-2017) [25] .Li et al. emphasized that due to the economic and technical restrictions, daily values of solar radiation were taken into account in most feasibility studies [26] .They improved a graphical solar irridation model by using daily solar irridadiance values.In this study, they claimed that they got approximately 10% accuracy more than usual methods.Gupta et al. reviewed estimation models of global solar radiation [27] .They emphasized that the solar energy is the most important renewable energy for obtaining net zero target by 2050.Hissou et al. examined the machine learning methodology for the predictiction of global solar irridation [28] .They emphasized that this new approach results were satiably in accordance with regression models.Guermoui et al. harmonized methodologies like time varying filter and empirical approach [29] .They stated that results of this new approach is better than classical estimation methods for all studied regions.Fan et al. analyzed empirical and machine learning methodology for estimating of solar irrdation in the various states of China [30] .12 machine learning models were used in the study.ANFIS, MARS and XGBoost models were indicated as the most recommended machine learning models.Gürel et al. reviewed various methods for the prediction of solar irradiation [31] .They compared empirical, time series, artificial intelligence and hybrid models.They emphasized that each model has advantages and disadvantages in accordance with targeted zone.They claimed that hybrid model is more applicable approach for the prediction of global solar radiation.Yarar et al. emphasized the importance of hourly based global solar radiation for harmonics analysis of the microgrid connection [32] .Kaysal and Hocaoğlu modelled a new approach, using by artificial neural network for the prediction of solar irradiation [33] .They used two stage forecasting model and discrete wavelet transform for their study.Mukhtar et al. integrated conventional and artificial neural networks for the prediction of hourly based solar irradiation [34] .They emphasized that, in general developing countries have not got solar pironometers.So, they claimed that the designated method can be used in developing countries, especially African countries.Geetha et al. offered an artificial neural network model for the prediction of global solar radiation [35] .They claimed that, due to this improved model, designing and evaluating stage of PV installation can be easier without meteorological data collection.In these studies, researchers claimed that hourly solar irradiance data are important for evaluating power plants which are using photovoltaic tools and equipments.They emphasized that, the most solar irradiance studies and actual measurements are not obtained by daily values due to the economic and technical restriction reasons, especially developing countries.In their study, they used data of identification, classification, clustering and regression of climate indicators and meteorological data.They compared estimation models and tried to get best availability.
In the studies conducted by researchers, it was mentioned that more effective results are obtained by using empirical and hybrid methods.In conducting the study, econometric models and empirical modeling approaches were used.As a result of these studies, it was found that empirical models are more useful.Comparing the studies conducted by the researchers, it is found that similar results were obtained, and it is evaluated that the researchers have developed models that contain complex calculations and relationships.The main objective of this study is to provide data for the algorithm to optimize the installed power and energy management of hybrid power plants.The developed algorithm determines the size of the hybrid structure with renewable energy pairs.This algorithm structure consists of 4 main sections.In the 1st section, the project characteristics of the renewable energy to be used in the hybrid structure are determined, in the 2nd part the project characteristics of solar energy, in the 3rd part the costs and in the 4th part the decision mechanism including the iterative functions are included.The study carried out, it is aimed to use solar radiation on an hourly basis, which is intended as a data input for the calculation of the amount of energy to be extracted from the solar energy sector.It is expected that the hourly solar radiation is practical and flexible so that it can be used in the algorithm.To meet the requirements of the algorithm, it was preferred to compare models based only on empirical and econometric approaches rather than the complex and mixed structure developed by other researchers.As a result of the comparison, it is found that it is more appropriate to use empirical models because they have high coefficients and can meet the requirements of the algorithm.It is found that the amount of solar radiation that provides the data input for the developed algorithm can be improved with hybrid or more advanced models.It is assumed that these integrated developed structures can be used, but in terms of practicality, the use of empirical models is considered sufficient.The designed algorithm is given in Figure 1.Thanks to the developed algorithm, the installed capacity of solar energy to be engaged in hybrid energy systems is determined and optimized.The algorithm has a simple but effective decision mechanism and basically works considering the generation profiles and characteristics, energy management of renewable energy plants in the hybrid structure.In this study, the amount of solar radiation data was estimated to feed into the algorithm and determine the amount of solar energy production.Since the energy management is done on an hourly basis, the predictions aim to determine the amount of solar radiation on an hourly basis.

Mathematical model
The ratio of the amount of solar energy on an empirical hourly basis to the amount of solar energy daily is calculated by Equation (1) given below.
In Equation ( 1),  ℎ represents the hourly based solar irradiance (kWh/m 2 ),   represents the daily based solar irradiance (kWh/m 2 ),   represents the hour angle at sunrise (° degrees). represents the hour angle (° degrees) which the displacement of the sun concerning a define zone on the world and can be found by Equation ( 2) below; The  (apparent solar time) specified in Equation (2) represents the observed or actual solar time. is calculated by Equation (3) given below; The GMT (Greenwich Mean Time Zone) specified in Equation ( 4) represents the time intervals arranged according to the city of Greenwich. is a coefficient and is found as stated in Equation ( 6) below; The  specified in Equation ( 6) represents the number corresponding to the day in the calendar year.For example, while the number of  for January 1 is 1, the number  for February 1 is 32.
specified in Equation ( 1) is found by Equation (7); The   specified in Equation ( 7) represents the sunrise angle.In this equation ∅ represents the angle of latitude (℃), and  represents the slope (℃). is calculated by Equation (8) The angle  varies between +23.45°/−23.45°.The following Figure 2 shows the change in the angle during the year;  [6] .
In the study carried out by Chandel and Aggarwal [7] the relationship between daily and hourly solar radiation is given in Equation ( 9) below; The   term specified in Equation ( 9) represents the conversion rate between hourly and daily solar radiation.The   ratio is obtained by Equation (10) The   given in Equation (10) represents the ratio of the earth outside hourly based solar irradiance to the daily irradiance.This ratio is calculated by Equation (11) given below.The coefficients, , (  ) and (  ) are given in Equations ( 12)- (14).
Given in Equation (10),  is the incidence angle of the beam, ℎ 0 is the sun height of the atmosphere,  0 is the daytime duration.The corresponding Equations ( 16)-( 18) are given below to calculate the relevant parameters;  = cos∅.cos (16)  sinℎ 0 = .(  )   (17)    0 = .  (18)   In addition to empirical methods, the amount of solar radiation can be determined by using econometric statistical methods.If hourly data is available for the studied site, hourly solar radiation data can be obtained, which is likely to occur prospectively.The most commonly used models in the literature and their explanations are given below.Time series lead to statistical data.These models, which are especially used for sun and wind, produce new predictions in line with the actual data.One of the most used models is the self-connection (AR-autoregressive) model.The AR model is calculated as given in Equation (19); In Equation (19),   represents time series value,   represents noise amounts, ɸ = (ɸ 1 , ɸ 2 , … , ɸ  ) vectors of model coefficients,  represents positive integer.
Another model is the moving average (MA-moving average) model.In this model, unlike the AR model, time series are created by using weighted total values in this model.The MA model is calculated as given in Equation (20); Equation ( 20),   represents the time series value,   represents the noise amounts,  = ( 1 ,  2 , … ,   ) is the vector coefficients of the model,  represents the positive integer. 0 is accepted as  0 = 1.
Another model is the self-correlated moving average (ARMA) model.This model has emerged by using AR and MA models together.It is calculated by Equation ( 21 In Equation (21),   is time series value,   is amount of noise,  = ( 1 ,  2 , … ,   ) is vectors of moving average model coefficients, ɸ = (ɸ 1 , ɸ 2 , … , ɸ  ) is vectors of self-correlated model coefficients,  and  represent a positive integer. 0 is accepted as  0 = 1.
Another model is the self-related and external variable moving average (ARMAX).This model has emerged by using AR and MA models together, as well as expanding the overall scope by using variables.It is calculated by Equation ( 22) given below.
Another model is the self-correlated and integrated moving average (ARIMA) model.This model is achieved by using AR and MA models together, as well as evaluating non-stationary variables.It is calculated by Equation (23) given below; In Equation ( 23),   is the time series value,   is the amount of noise,  = ( 1 ,  2 , … ,   ) is the vectors of moving average model coefficients, ɸ = (ɸ 1 , ɸ 2 , … , ɸ  ) is the vectors of self-correlated model coefficients,  and  represent a positive integer. 0 is accepted as  0 = 1.The expressions S = 1 -q − 1 and ɸ  (q), represent the stationarity and translatability expressions of the AR (m) operator.

A case study
The second method employed in obtaining the quantity of radiation is that the approach that features econometric and statistical methods.The foremost widely used econometric models within the literature are autoregressive integrated moving average (ARIMA), autoregressive moving average (ARMA), autoregressive moving average model with exogenous variables (ARMAX), autoregressive (AR), moving average (MA), models.Additionally, artificial neural network (ANN) model was used as a synthetic intelligence element.Econometric models were created using the "R" modeling program and MATLAB.The models used were compared with one another and therefore the best suited one was chosen.Within the first place, a one-day forecast was made for the AR and MA models, which are econometric models.Econometric models are fed from the data set from a solar energy plant which is found in Denizli province, Acıpayam district of Türkiye.A seven-day real data set was accustomed predict one-day radiation.MATLAB program is employed for this prediction.The demonstration of model is given in Figure 3.The study was distributed with the "R" modeling program for ARIMA and ANN models.Six days of information acquisition is employed for 1 day and 6 days following estimations.The study results and existing condition comparisons are demonstrated within the graphs below.The detail of study is given in Figures 4 and 5   Looking at the comparative figure, it is clear that the econometric models differ from the actual data.Taken on its own, it can be said that the ARIMA model produces results that are relatively similar to the original data.Although the ARIMA model consists of the integration of other AR and MA models, it is a more advanced econometric model.Finally, using the data within the first 2 months of the year-January and February-radiation amount predictions were made for the primary six days of March.The subsequent chart shows the relevant studies and their results.Estimates obtained with AR and MA models produced very different results from the particular data.Within the more advanced models, ARIMA and ANN models, estimations were made using only two days data set.Additionally, "R" modeling, which is an econometric program that has more precise results for these models, was used.First, the ARIMA model was run.The quantity of estimation that offers the foremost optimum solution in step with the determined number of steps has been revealed.The blue drawing within the graph below gives the optimum ARIMA line.Other gray areas reflect the framework of other and sensibility (Figure 6).Using econometric models can provide reasonable results if the info is as detailed and long as possible.It's not always easy to achieve this detail supported coordinates in radiation estimations.An empirical model comes into play in these circumstances.The daily radiation amounts are often reached supported coordinates on the official website of NASA.When the empirical model compares with artificial neural network and ARIMA models; • It doesn't need one-to-one hourly radiation.It is often obtained on an hourly basis using daily radiation.
• While the number of radiation at any point on the world daily is obtained from public sources, it's too difficult to access hourly radiation amount data supported coordinates.• High correlation with real values.
• Does not need long-term data entry like ARIMA and ANN.
• Obtaining realistic predictions not only in short-term forecasts like ARIMA and ANN but also in long-term forecasts.
Due to its prominent features, it's been decided to use an empirical model in radiation predictions.The results obtained, using the MATLAB GUI program, an interface has been created to display the quantity of radiation within the day, month, year, or any desired period.Changes within the amount of radiation throughout the year, start and end times are often determined within the interface, and may be obtained at desired month, day, and hour intervals.So, empirical model is chosen for hourly based estimation.The radiation amounts of the coordinates of the solar energy plant located in Denizli province, Acıpayam district, which are used for empirical models, were obtained daily from the NASA website [36] .By making use of the equations described within the mathematical model section, hourly radiation amounts were obtained in daily radiation amounts.Additionally, the study was enriched with the MATLAB GUI program.By using the MATLAB GUI, it's possible to draw graphs of the quantity of radiation on an hourly basis at any time during the year.While designing the MATLAB GUI graphic area, radiation amounts were graphed in line with the beginning and end times.The timeline relies on month, day and hour.Desired graphs are drawn by entering the periods to be seen in these boxes.8760 hours-based radiation dataset was created to hide the entire year.The MATLAB GUI chart below shows the 24 h radiation amounts for 31 March, 30 June, 30 September, and 31 December.This demonstration is given in Figure 7.The daily radiation amounts were taken from the NASA website for the solar energy plant located within the Acıpayam district of Denizli province.This original dataset is given in Figure 8.
Figure 8.The daily solar radiation obtained from the NASA website (365 days) [26] .
The amount of solar energy produced may be calculated by using the panel data.To check the operation of the approach, dataset of a SPP which is located at a distance of 2 km were taken, daily NASA data for the identical region were correlated hourly within the light of original data using the equations above.The parametric statistic obtained as a result of 86% gives a robust concept the acceptance is usable for solar radiation.The daily NASA data are converted to the hourly irradiance shown within the figure below.The correlation graph of the study is given in Figure 9.The energy generation amount was estimated by using the empirically obtained radiation amount and panel/inverter data of a solar energy plant located within the Acıpayam district of Denizli province.The particular production of this existing facility has been compared with the energy generation amounts prepared using the radiation amounts obtained on an hourly basis.X variable references the time, Y is that the radiation.The coefficient of correlation is obtained as 0.8625.The result is given in Figure 10.With the assistance of the designed interface, analyzes may be made for specified periods, and changes in radiation amount may be observed.After determining the quantity of radiation, the quantity of energy to be obtained from the sun determined.After obtaining the number of radiation, the number of energy which will be obtained from the sun was calculated following the mathematical model and algorithm explained within the above sections.Before calculating the quantity of radiation, the energy production figures of an SPP facility located in Denizli province/Acpayam district were compared with the energy amount obtained by the proposed method, the technical specifications of the panels utilized in the prevailing SPP facility were taken as a basis.Within the current SPP facility, Yingli Solar branded 250 watt polycrystalline PV panels are used.Within the manuals (data sheet) containing the technical specifications of the PV panels, the measurement results under the quality test (STC-standard test condition) and nominal operating cell (NOCT) conditions are included.
The number of power that can be obtained from PV panels depends on radiation and temperature.The graph below shows the I-V and power curves of the PV panels embedded in MATLAB.Technical specifications also include changes in current and voltage against a unit natural process.After entering the NOCT values of the chosen PV panel, the values to be obtained by this and voltage were determined by using the hourly data obtained from the meteorology station for Acıpayam district.The quantity of energy that may be obtained from the solar power system was calculated by multiplying the determined current and voltage values with one another.Since a high production forecast of roughly 91% can is obtained in comparison to the particular electricity generation.It's been accepted that the said approach is employed in other regions similarly.Accordingly, the calculation was made by using the tactic described above within the determination of the number of radiation for the related HEPP facilities area.

Conclusion
Summarized results of this study are given in below.
In recent years, worldwide applications and research disbursed for SHE plants, which is one among the hybrid renewable energy systems.One in all the foremost issues for these integrated facilities is that the management of the number of energy generation from solar and hydroelectricity.To confirm reliable energy management, hourly electricity generation must be obtained.Electricity generation from solar energy may be a function that depends on the quantity of radiation and temperature.To see the number of solar energy on an hourly basis, the number of radiation must be determined.While it's easy to access the daily radiation data on a coordinated basis, this data isn't disclosed to the public on an hourly basis.
There are two basic methods employed in the literature to get the quantity of radiation on an hourly basis.One among them is that the empirical method and also the other is that the econometric method.In this study, the quantity of radiation determined by using both methods.As a result of the studies, the quantity of radiation on an hourly basis was obtained from the daily radiation data, while more realistic results were obtained with the empirical method.The most difference during this comparison is that the requirement of longer-term and more detailed data for econometric statistical models to get an affordable result.By using the empirical model, the quantity of radiation on an hourly basis was obtained from the daily radiation.These data are often easily employed in the energy management of renewable hybrid energy systems like solar hydroelectric facilities.The suggested hourly based global solar radiation which is obtained by empirical model can be used in developed algorithm of hybrid renewable energy system.It is simple and practical way for the estimation of solar energy amount.The existing results are adequate in addition to this, study can be improved by using advanced artificial intelligence methods.

Figure 1 .
Figure 1.The algorithm of installed capacity for solar energy in integrated hybrid renewable structure.

Figure 2 .
Figure2.Angle of inclination and change over the year[6] .

Figure 3 .
Figure 3. Estimation of the hourly solar radiation by (a) AR model; and (b) MA model. .

Figure 5 .
Figure 5.Comparison of the econometric models.

Figure 6 .
Figure 6.Prediction of the solar radiation of 1-6 March by using (a) ARIMA; and (b) ANN.

Figure 7 .
Figure 7. Estimation of the hourly based solar radiation by using the empirical model in MATLAB GUI (a) 31 March; (b) 30 June; (c) 30 September; and (d) 31 December.

Figure 9 .
Figure 9.The hourly based solar irradiance obtained from the daily solar irradiance dataset (8760 h = 24 h/day × 365 day).

Figure 10 .
Figure 10.Predicted and real solar radiation correlation plot.
given below.
given below;