A new statistical approach for deriving global solar radiation from satellite images

Solar radiation derived from geostationary satellite images has become an advantageous technique for solar resource characterisation over large areas. The simplest methods for estimate solar radiation from the satellite information rely on straight forward relationships between a normalised parameter of the solar irradiance (such as clearness or clear sky index) and the cloud index. This paper presents a statistical fit of this relationship (fitted and tested using data from 28 Spanish radiometric station) different from the approach used by Heliosat-2 method (Rigollier, C., Lefèvre, M., Wald, L., 2004. The method Heliosat-2 for deriving shortwave solar radiation from satellite images. Solar Energy 77, 159-169), that includes local statistical measures of the cloud index distribution and the air mass.In particular, the inclusion of the local cloud index percentiles (median, first and third quartile) estimated from the whole series on each pixel improves clearly the model response, and is a way to account for the local climatological aspects of any location. The inclusion of the new explicative variables yield to practically unbiased results and the relative RMSE decrease to about 17% from the 21% result of the expression applied in the Heliosat-2 model.     

A method for the determination of the global solar radiation from meteorological satellite data

A statistical method is presented for the determination of the global solar radiation at ground level. It makes use of data from the meteorological satellites, which provide extensive coverage as well as adequante ground resolution. In the first step, a reference map of ground albedo is deduced from the time-sequence of satellite images. Then, by comparing the satellite data with the computed albedo map, a cloud coverage index is determined for each ground point of 5 km × 5 km. This index is linearly correlated to the atmospheric transiission factor. The regression parameters are estimated using a training set provided by ground pyranometers. Tests for two different time periods show a good agreement between the actual and model-derived hourly global radiation.     

Analysis of the influences of uncertainties in input variables on the outcomes of the Heliosat-2 method

The Heliosat-2 method, which employs satellite images to assess solar irradiance at ground level, is one of the most accurate among the available operational methods. Its input variables have uncertainties which impact on the final result. The General Law of Uncertainty Propagation is employed to analyze the impact of these uncertainties on a single pixel with Meteosat-7 inputs in various stages, beginning with the sensitivity coefficients and the changes induced in the clear-sky index (K_C) by each independent variable. Once these coefficients are known, the partial combined standard uncertainty (CSU) is calculated for K_C from each independent variable and albedo. Finally, the total CSU of K_C is calculated. All of the results are in agreement and show that the most influential variables in the uncertainty of estimation of cloudy skies are, in this order, the Linke turbidity factor (54% of K_C value), terrain elevation (33%), the calibration coefficient of the satellite sensor (13%) and the ground albedo (5%). What causes the initial uncertainty in the ground albedo is its variation over time and the difficulty in assessing it from a reflectance time-series for mixed clear and cloudy skies. The Linke turbidity factor is the most influential variable on the width of the uncertainty interval, not only because of its own uncertainty (17% in this study), but because it is also used in numerous intermediate calculations. For clear skies, the partial CSUs are considerably lower, except for ground albedo (5% also).     

Assessment and recalibration of the Heliosat-2 method in global horizontal irradiance modeling over the desert environment of the UAE

Remote estimation of global horizontal irradiance (GHI) by Heliosat-2 model has been benchmarked against ground-based measurements in different locations. The obtained results have shown that the level of agreement between ground-based and model-based GHI values are location dependent. To our knowledge no similar studies have been carried out over the Arabian Peninsula. The unique climatic condition in the Peninsula, characterized by high concentrations of airborne dust particles and high humidity makes the region a particularly interesting case.In this study Heliosat-2 method was used to derive the ground surface GHI in the United Arab Emirates. Inputs to the model were monthly Linke turbidity factor normalized to an air mass of 2 and a cloud index derived from SEVIRI sensor onboard the European satellite Meteosat. The Linke turbidity factor was obtained from a ground network of seven stations distributed across the UAE. A SEVIRI-based technique was developed and used to derive cloud index from high-resolution visible channels. Ground surface GHI measurements were collected from four inland stations for a period ranging from mid-2007 to mid-2010. The obtained results show that the Heliosat-2 model underestimates the GHI. The obtained root mean square error (RMSE) and mean bias error (MBE) values ranged from 16.3% to 18.5% and ?13.6% to ?15.8%, respectively. A constant bias was observed between modeled and measured GHI throughout the four stations. To correct this bias, the empirical equation used in Heliosat-2 to estimate the clear sky diffuse horizontal irradiance (DHI_clear) was recalibrated. With the new DHI_clear empirical equation, the modified Heliosat-2 model becomes more adapted to desert and dusty environments such as that of the UAE. By applying the modified DHI_clear equation, the RMSE and MBE values dropped to 9.5–10.3% and ?1.2 to +0.8%, respectively.     

A modified algorithm for calculating the cloud index

The cloud index is an important component of the Heliosat algorithm, which estimates solar radiation components from Meteosat High Resolution Visible images. The cloud index quantifies the reflective properties of the atmosphere, and varies from 0 at clear conditions to 1 at overcast. The algorithm is semi-empirical in the way that it includes several constants that need to be tuned. Some of these were removed in the Heliosat-II algorithm where the Meteosat calibration constant was introduced to replace the “pseudo-reflectivity” with a “real reflectivity”. This approach is followed here, and two additional changes are made: (1) An analytical expression is introduced to correct for backscattered radiation from air molecules. (2) A correction is made for non-lambertian reflectivity, removing the time consuming need for determining the ground reflectivity for each month and each slot. The new cloud index is used to calculate global irradiances which are validated against hourly measurements from five European ground stations. The average root mean square deviation is 15.5% for a six-month spring/summer period, of comparable accuracy as using the more time consuming traditional algorithm used in the EU project Satel-Light.     

The method Heliosat-2 for deriving shortwave solar radiation from satellite images

This article presents the method Heliosat-2 that converts observations made by geostationary meteorological satellites into estimates of the global irradiation at ground level. This new version integrates the knowledge gained by various exploitations of the original method Heliosat and its varieties in a coherent and thorough way. It is based upon the same physical principles but the inputs to the method are calibrated radiances, instead of the digital counts output from the sensor. This change opens the possibilities of using known models of the physical processes in atmospheric optics, thus removing the need for empirically defined parameters and of pyranometric measurements to tune them. The ESRA models are used for modeling the clear-sky irradiation. The assessment of the ground albedo and the cloud albedo is based upon explicit formulations of the path radiance and the transmittance of the atmosphere. The method Heliosat-2 is applied to Meteosat images of Europe for the months of January 1995, April 1995 and July 1994. Pyranometric measurements performed by thirty-five meteorological stations are used to assess the performances that are close to those of Heliosat-1 found in the literature. Possible improvements are discussed.     

A model for calculating hourly global solar radiation from satellite data in the tropics

A model for calculating global solar radiation from geostationary satellite data is presented. The model is designed to calculate the monthly average hourly global radiation in the tropics with high aerosol load. This model represents a physical relation between the earth-atmospheric albedo derived from GMS5 satellite data and the absorption and scattering coefficients of various atmospheric constituents. The absorption of solar radiation by water vapour which is important for the tropics, was calculated from ambient temperature and relative humidity. The relationship between the visibility and solar radiation depletion due to aerosols was developed for a high aerosol load environment. This relationship was used to calculate solar radiation depletion by aerosols in the model. The total column ozone from TOMS/EP satellite was employed for the determination of solar radiation absorbed by ozone. Solar radiation from four pyranometer stations was used to formulate the relationship between the satellite band earth-atmospheric albedo and broadband earth-atmospheric albedo required by the model. To test its performance, the model was used to compute the monthly average hourly global radiation at 25 solar radiation monitoring stations in tropical areas in Thailand. It was found that the values of monthly average of hourly global radiations calculated from the model were in good agreement with those obtained from the measurements, with the root mean square difference of 10%. After the validation the model was employed to generate hourly solar radiation maps of Thailand. These maps reveal the diurnal and season variation of solar radiation over the country.     

Ground reflectance estimation by means of horizontal and vertical radiation measurements

The importance of an accurate estimation of the ground reflectance is well-recognized in energy balances involving solar radiation. Despite there are methods to estimate it directly it is still a hard task in highly reflective environment or in urban areas. In those areas it is also hard to find measurements of the ground-reflected component of the solar radiation. However, data coming from solar energy installations could be employed to estimate the local ground reflectance. To achieve this goal a method to estimate the ground reflectance from global radiation measurements in several orientations is needed.In this work the question of the estimation of the ground reflectivity from measurements of the global solar radiation on orientations typical from the energy installations is addressed.Two different approaches are presented to solve the same problem. First, by direct solution of global radiation models the ground reflectivity is deduced from the measurements (known as direct problem approach). In second place, from physical hypothesis about the behavior of the solar radiation such as isotropy and diffuse correlations the ground reflectance is approximated from experimental data (known as inverse problem approach) at the points where reality is closer to these conditions. In this framework, it is shown the need of mathematical tools such as Tikhonov regularization.The two approaches have been tested by means of the data produced at the CIEMAT’s test facilities from the Energy Efficiency in Building R&D Unit (UiE3) at Plataforma Solar de Almería (PSA) located at the South-East of Spain. Approaches are tested on data coming from several pyranometers on the rooftop of a building. Cross-validation is performed with measurements coming from an albedometer.It is shown that the inverse technique can solve the problem by taking near two weeks of measurements of global solar radiation on the horizontal and vertical surfaces. With the same set of experimental data the direct technique is unable to estimate the ground reflectance even with the inclusion of a third measurement on other surface.     

Influence of three-dimensional cloud effects on satellite derived solar irradiance estimation—First approaches to improve the Heliosat method

Three-dimensional (3D) cloud effects which influence the surface solar irradiance are investigated by using the radiative transfer model SHDOM (spherical harmonics discrete ordinate method). The calculations are motivated by improving the Heliosat method which is used for the estimation of surface solar irradiance from satellite data. The modelled results lead to the assumption that it is possible to find a correction for moderate viewing and solar geometries for the current Heliosat method. Based on these results an empirical study has been done, comparing ground measurements with Heliosat derived data. It will be shown that a correction of the current Heliosat method for 3D cloud effects is possible. For two out of four sites the proposed corrections decrease the relative root mean square error (rRMSE) of the irradiance derived by the Heliosat method in a range of about 2%. The use of a site specific correction for different locations shows that the adaptability of the corrections depend on the dominant local cloud regimes, which governs e.g. the existence of certain cloud types.     

利用我国气象卫星遥感资料估算黄河源区净辐射辐照度的研究

以黄河源区为研究区域,利用我国静止气象卫星(FY-2D)及极轨气象卫星(FY-3A)资料,估算黄河源区一天中不同时刻的净辐射辐照度的空间分布。在晴天条件下,利用计算的向下太阳辐射照度,结合卫星遥感资料估算的地表反照率,得出地表太阳辐射收支。利用卫星遥感估算地表发射率和地表温度资料,结合地面观测站的气象观测资料,得到大气向下的长波辐射和地面向上的长波辐射值。在阴天条件下,借助FY-2D云顶的反照率资料,根据太阳辐射在大气中的衰减过程,推求透射比系数,进而得出阴天条件下地表净短波辐射照度。     

基于CERES数据分析中国云对地表太阳辐射影响特征

用于地面光伏发电的太阳辐射资源受到云影响所产生的变化,由于受到观测方式限制而对其特征及关联性探究不足。针对该问题,利用CERES卫星观测时空数据,对2001—2018年中国大陆地区云对地表太阳辐射量影响特征进行分析,利用不同高度云量与云太阳辐射强迫时空分布特点,讨论云影响的太阳辐射差异对光伏出力预计影响,并通过各高度云量趋势变化与云太阳辐射强迫关联关系给出相应太阳辐射资源与光伏出力的趋势变化估计,分析云主导的太阳辐射趋势变化成因,以期为未来新能源资源评估、光伏建设规划和电力优化调度等提供支撑与参考。     

Solar radiation estimations over India using Meteosat satellite images

India is endowed with good solar energy resource due to its geographic position in the equatorial Sun Belt of the earth, but its atmosphere can have eventual large presence of aerosols with the subsequent negative feedback to the solar radiation available. Therefore, solar resource assessment studies over India are of high interest for potential solar energy applications. In this work daily estimations of global horizontal and direct normal irradiation are presented for six locations in India covering the years from 2000 till 2007. These computations have been performed with IrSOLaV/CIEMAT method for computing solar radiation components from Meteosat images with a spatial resolution of less than 5 × 5 km. A brief assessment exercise of the model output have been made with ground measurements available from the World Radiation Data Centre database, which consists only of daily sums of global horizontal irradiation. The daily global horizontal irradiation estimated by the model has shown a general positive bias with the ground measurements in the range of 5%, being the root mean square deviation around 12%, excepting for Trivandrum location where there are evidences of higher uncertainty in the ground measurements. In addition, the dynamical behavior of daily global irradiation is quite well reproduced by the model as a consequence of adding to the input the daily turbidity values estimated from MODIS Terra satellite information. Estimations of direct normal irradiation are also exposed but no assessment was made due to the lack of ground measurements. The importance and need of accurate daily aerosol data with high spatial resolution for solar radiation estimations is pointed out in this work.     

Development of a method for generating operational solar radiation maps from satellite data for a tropical environment

Despite a considerable number of publications which use satellite data to map solar radiation, relatively few studies have been undertaken in a tropical environment. In this study, we have developed a method to produce operational solar radiation maps from satellite data for this environment. The method is based on a physical model which relates the satellite-derived earth–atmospheric reflectivity from visible channel of GMS-4 and GMS-5 to the transmissivity of the atmosphere. Cloud reflectivity was determined from satellite data, while radiation absorbed by water vapour, ozone and aerosols and radiation scattered by aerosols were determined from ground-based meteorological data. Techniques for determining the radiation depleted by these atmospheric constituents over a whole country were also presented. Satellite data of a six-year period (1993–1998) with approximately ten thousand satellite images were used to construct the maps. When tested against an independent data set, monthly average of daily global irradiation calculated from this method agree with that obtained from the measurements with the relative root mean square difference of 6.8% with respect to the mean values. Solar radiation is presented as twelve maps showing the monthly average of global irradiation and one map showing the yearly average of global irradiation. Radiation patterns from the maps show a strong influence of the tropical monsoons.     

双面光伏组件辐照度模型的研究

针对双面光伏组件正面和背面获均能吸收太阳光的特点,通过光线跟踪辐照度模型分析,构建区分阴影区和无阴影区的热传输理论视觉因子太阳辐照度模型。模拟结果表明:当双面光伏组件倾斜角比单面组件增加约4°,在地面反射率为30%和50%的情况下,年辐照度增益可提高17.41%和28.79%;且随着离地高度与行间距增加,年辐照度可进一步提高。双面光伏组件辐照度模型为双面光伏组件电站安装时的地面反射率、最佳倾斜角、离地高度及行间距的设置提供了理论支撑。     

Photovoltaic power generation using albedo and thermal radiations in the satellite orbits around planetary bodies

In this paper, we have studied photovoltaic power generation in satellite orbits around different planetary objects using albedo and thermal radiations. We have addressed the power generation in the darkside of satellite orbits using solar cells in particular. LILT solar cells can be used for photovoltaic power generation in the darkside of satellite orbits around moon, mars etc. receiving the weak albedo radiation from the nearby brightest celestial object. We can also make use of the emerging IR photovoltaic technology to generate power using planetary thermal radiation from the nightside of planets like earth.     

Solar radiation mapping from satellite image using a low cost system

An operational low cost system is developed for mapping daily and monthly mean daily global solar radiation reaching the ground over a big geographical area from APT (Automatic Picture Transmission) images of NOAA polar orbiting satellite. The system is suitable for use in south and southeast Asia and is said to be acceptable for engineering, agricultural, and climatological applications requiring the knowledge of geographical and temporal variation of solar radiation. It is alterable for use in other locations. The system comprises a personal computer attached with a small APT receiver along with other peripherals such as monochrome monitors and simple printer. An improved empirical bispectral method which applies visible and far infrared APT images is used to calculate the cloudiness of the sky. The empirical coefficients obtained by correlating cloudiness with ground measured solar radiation are used to estimate solar radiation at a target on the ground knowing the cloudiness of its sky from a satellite image. The geographical coverage of the map can be as big as 15° latitude by 20° longitude near the Equator. The spatial resolution of the radiation map is 111 km. The mean standard error of estimation of daily global radiation is 12.9% with a range of 6–19% of the measured mean. This accuracy is acceptable and comparable to other similar methods developed so far.     

Application of extreme learning machine for estimating solar radiation from satellite data

In this paper, a simple and fast method based on extreme learning machine (ELM) for the estimation of solar radiation in Turkey was presented. To design the ELM model, satellite data of the National Oceanic and Atmospheric Administration advanced very high‐resolution radiometer from 20 locations spread over Turkey were used. The satellite‐based land surface temperature, altitude, latitude, longitude, month, and city were applied as input to the ELM, and the output variable is the solar radiation. To show the applicability of the ELM model, a performance comparison in terms of the estimation capability and the learning speed was made between the ELM model and conventional artificial neural network (ANN) model with backpropagation. The comparison results showed that the ELM model gave better estimation than the ANN model for the overall test locations. Moreover, the ELM model was about 23.5 times faster than the ANN model. The method could be used by researchers or scientists to design high‐efficiency solar devices such as solar power plant and photovoltaic cell. Copyright © 2013 John Wiley & Sons, Ltd.     

Estimation of global solar radiation in arid and semi-arid climates of East and West Iran

In Iran, most of the models used so far, have provided solar estimation for a few specific locations based on the short-term solar observations. Using different radiation models, (e.g. Sabbagh, Paltridge–Proctor, Daneshyar) and various input parameters (e.g. cloud cover, sunshine duration, relative humidity, temperature, and altitude) we developed a general height-depended formula for the prediction of the direct and diffuse monthly average daily solar radiation for 64 mountainous arid and semi-arid locations in West and East Iran. The models mentioned are modified and new coefficients are defined for the diffuse component based on the long-term observed diffuse data. Model results are validated against up to 13-year daily solar observations at 10 solar radiation sites. In comparison with the previous studies, the newly developed method performs more accurate estimation (less than 3% MPE error) in the arid and semi-arid regions. Comparison of the model results indicates that calibration of the coefficients made to the diffuse formula against the longer period experimental data can improve the estimations of global solar radiation.     

Solar radiation assessment using Meteosat 4-VIS imagery

Meteosat 4-VIS high-definition imagery was used for testing solar radiation assessment over northeast Brazil during January 1991 and 1994. Actinographs of solarimetric stations in Paraíba State provided proper ground reference. A simple physical model was developed, based on two basic hypotheses: visible and infrared intervals are decoupled, and radiation transfer in the visible interval is essentially conservative. Some ideas are borrowed from the well known model of Gautier et al. for cloudiness assessment. The resulting GL model can run on a personal microcomputer, processing targets of 5 × 5 pixels for all the region. It exhibits rather good behaviour on a monthly scale, reproducing the mean diurnal cycle and assessing mean irradiation with bias and mean standard deviation lower than 10 W m⁻². Fifteen-day means exhibit similar accuracy. Ground albedo is a highly important parameter of the model. As expected, the main sources of error seem to be planetary reflectance and cloudiness assessment, because of the non-lambertian reflection and the different count thresholds for each kind of cloud cover. Further studies considering longer time series of images (VIS and IR channels) as well as the anisotropic properties of finite cloud fields may improve the model.     

Solar resources estimation combining digital terrain models and satellite images techniques

One of the most important steps to make use of any renewable energy is to perform an accurate estimation of the resource that has to be exploited. In the designing process of both active and passive solar energy systems, radiation data is required for the site, with proper spatial resolution. Generally, a radiometric stations network is used in this evaluation, but when they are too dispersed or not available for the study area, satellite images can be utilized as indirect solar radiation measurements. Although satellite images cover wide areas with a good acquisition frequency they usually have a poor spatial resolution limited by the size of the image pixel, and irradiation must be interpolated to evaluate solar irradiation at a sub-pixel scale. When pixels are located in flat and homogeneous areas, correlation of solar irradiation is relatively high, and classic interpolation can provide a good estimation. However, in complex topography zones, data interpolation is not adequate and the use of Digital Terrain Model (DTM) information can be helpful. In this work, daily solar irradiation is estimated for a wide mountainous area using a combination of Meteosat satellite images and a DTM, with the advantage of avoiding the necessity of ground measurements. This methodology utilizes a modified Heliosat-2 model, and applies for all sky conditions; it also introduces a horizon calculation of the DTM points and accounts for the effect of snow covers. Model performance has been evaluated against data measured in 12 radiometric stations, with results in terms of the Root Mean Square Error (RMSE) of 10%, and a Mean Bias Error (MBE) of +2%, both expressed as a percentage of the mean value measured.