Suchergebnisse

From 25 May to 5 June 2015, the 10th regional intercomparison campaign of the Regional Brewer Calibration Center – Europe (RBCC-E) was held at El Arenosillo atmospheric sounding station of the Instituto Nacional de Técnica Aeroespacial (INTA). This campaign was jointly conducted by COST Action ES1207 EUBREWNET and the Area of Instrumentation and Atmospheric Research of INTA. A total of 21 Brewers, 11 single- and 10 double-monochromator instruments from 11 countries participated and were calibrated for total column ozone (TOC) and solar UV irradiance. In this 2015 campaign we have introduced a formal approach to the characterisation of the internal instrumental stray light, the filter non-linearity and the algorithm for correcting for its effects on the TOC calculations. This work shows a general overview of the ozone comparison and the evaluation of the correction of the spectral stray light effect for the single-monochromator Brewer spectrophotometer, derived from the comparison with a reference double-monochromator Brewer instrument. At the beginning of the campaign, 16 out of the 21 participating Brewer instruments agreed within better than ±1%, and 10 instruments agreed within better than ±0.5% considering data with ozone slant column between 100 and 900DU, which does not require instrumental stray light correction. ; This article is based upon work from COST Action 1207 EUBREWNET. This work has been supported by the European Metrology Research Programme within the joint research project ENV59 "Traceability for atmospheric total column ozone" (ATMOZ). The EMRP is jointly funded by the EMRP participating countries within EURAMET and the European Union. We also gratefully acknowledge further support by the Fundación General de la Universidad de La Laguna. This study and the campaigns were supported at large part by ESA project CEOS Intercalibration of ground-pectrometers and lidars (ESRIN contract 22202/09/I-EC).

Póster elaborado para el Quadrennial Ozone Symposium celebrado en Edimburgo del 4 al 9 de septiembre de 2016. ; This work has been supported by the European Metrology Research Programme (EMRP) within the joint research project ENV59 "Traceability for atmospheric total column ozone" (ATMOZ). The EMRP is jointly funded by the EMRP participating countries within EURAMET and the European Union.

Taking advantage of UV (295 385 nm) irradiance measurements is one of the objectives of this paper. A newindex termed kt″ is established for this band. This newindexworks as a zenith angle independent clearness index for band measurements and has similar applications to those of kt′ for broadband measurements. The new index may be applied to identify cloudless instants from UV band measurements. Both indexes were correlated throughout the period 1998 2004 with a R2 of 0.85. A selection criterion of kt″UV >1.1 classified cloudless sky conditions with a probability of 95% in comparison with a selection that two criteria-applying broadband measurements would make. This index may be of interest for classifying cloudless sky conditions when only UV band measurements are available. An estimation method from the literature was applied to the period 1998 2004. This method was previously validated for the UV band with a measurement campaign made in Valencia (Spain) in the summer season. ; This work was supported by the Spanish Government through MEC grant MAT2006-02279, and was a part of the activities of the Grup d'Optoelectronica i Semiconductors of the Universitat Politecnica de Valencia, Spain. ; Serrano, M.; Boscá Berga, JV. (2013). Selection of cloudless sky conditions by applying solar globalultraviolet irradiance measurements. Atmospheric Research. 132-133:291-298. https://doi.org/10.1016/j.atmosres.2013.05.020 ; S ; 291 ; 298 ; 132-133

In-situ measurements of ultraviolet (UV) and solar irradiance is very sparse in Nigeria because of cost; it is estimated using meteorological parameters. In this work, a low-cost UV and pyranometer device, using locally sourced materials, was developed. The instrument consists of a UV sensor (ML8511), a photodiode (BPW34) housed in a carefully sealed vacuumed glass bulb, the UV and solar irradiance sensor amplifiers, a 16-bit analog-to-digital converter (ADS1115), Arduino mega 2560, liquid crystal display (LCD) and microSD card for data logging. The designed amplifier has an offset voltage of 0.8676 mV. The sensitivity of the irradiance device is 86.819 Wm-2/mV with a correcting factor of 27.77 Wm-2 and a maximum range of 1200 Wm-2. The instrument validation error is 9.67% and a correlation coefficient of 0.89 when compared with a standard SRS100 pyranometer. The UV sensor showed a close response with a correlation of 0.99 in comparison with a standard Skye instrument. From 08:00 to 16:00 local time (LT), there is a very close agreement between the standard device and the developed counterpart, with marginal differences of about 9.6% observed at the two extremes.

This paper evaluates the relationship between the cloud modification factor (CMF) in the ultraviolet erythemal range and the cloud optical depth (COD) retrieved from the Aerosol Robotic Network (AERONET) "cloud mode" algorithm under overcast cloudy conditions (confirmed with sky images) at Granada, Spain, mainly for non-precipitating, overcast and relatively homogenous water clouds. Empirical CMF showed a clear exponential dependence on experimental COD values, decreasing approximately from 0.7 for COD = 10 to 0.25 for COD = 50. In addition, these COD measurements were used as input in the LibRadtran radiative transfer code allowing the simulation of CMF values for the selected overcast cases. The modeled CMF exhibited a dependence on COD similar to the empirical CMF, but modeled values present a strong underestimation with respect to the empirical factors (mean bias of 22%). To explain this high bias, an exhaustive comparison between modeled and experimental UV erythemal irradiance (UVER) data was performed. The comparison revealed that the radiative transfer simulations were 8% higher than the observations for clear-sky conditions. The rest of the bias (~14%) may be attributed to the substantial underestimation of modeled UVER with respect to experimental UVER under overcast conditions, although the correlation between both dataset was high (R2 ~ 0.93). A sensitive test showed that the main reason responsible for that underestimation is the experimental AERONET COD used as input in the simulations, which has been retrieved from zenith radiances in the visible range. In this sense, effective COD in the erythemal interval were derived from an iteration procedure based on searching the best match between modeled and experimental UVER values for each selected overcast case. These effective COD values were smaller than AERONET COD data in about 80% of the overcast cases with a mean relative difference of 22%. ; Manuel Antón thanks Ministerio de Ciencia e Innovación and Fondo Social Europeo for the award of a postdoctoral grant (Ramon y Cajal). C. Chiu was supported by the Office of Science (BER, US Department of Energy, Interagency agreement DE-SC0006001) as part of the ASR program. We also thank the AERONET team for providing instrument calibration and data processing. MODIS data were obtained from the Level 1 and Atmosphere Archive and Distribution System (LAADS, http://ladsweb.nascom.nasa.gov ). This work was partially supported by the Andalusian Regional Government through projects P08-RNM-3568 and P10-RNM-6299, the the Ministerio de Ciencia e Innovación through projects CGL2008-05939-C03-03/CLI, CGL2010-18782, CGL-2011-2992-1-C02-01 and CSD2007-00067, and by European Union through ACTRIS project (EU INFRA-2010-1.1.16-262254). This work is co-financed through FEDER (Programa Operacional Factores de Competitividade – COMPETE) and National funding through FCT – Fundaçaõ para Ciencia e a Tecnologia in the framework of project FCOMP-01-0124-FEDER-009303 (PTDC/CTE-ATM/102142/2008).

International audience ; Between 2006 and 2011, the French government set up an incentive policy in order to develop the electricity production from photovoltaics. In the overseas territories, like Reunion Island, the feedin tariffs proposed for the next 20 years were specifically high. It resulted an exponential increase of the installed PV systems. For these small non-interconnected grids, an important penetration rate of such an intermittent energy can destabilize the supply-demand balance and may lead to an electricity outage risk. As a consequence, a regulatory limit of 30% of the instantaneous power produced from intermittent renewable (solar, wind and waves) was defined in order to avoid this risk. For instance, this legal constraint was reached in 2012 in Reunion. In this context, the forecasting of solar irradiance is essential in order to increase the penetration rate of PV output power into the grid. More precisely, accurate solar forecasts will help the grid operator to better manage the means of production. Forecasts are also needed in order to optimize the operation of grid connected storage energy systems. In order to cope with specific plant operations, forecasts must be provided with different granularities and horizons. In this work, we will focus on forecasts from 10 minutes to to 4 hours ahead with a 10 minutes granularity. The chosen forecasting time horizon will permit to monitor the production and to adjust the scheduling. In this paper, we introduce a novel approach based on econometrics models to forecast the global solar irradiance. The use of econometrics methods is justified by the similar behavior exhibited by the clear sky index and log return time series Our approach combine 2 models: a ARMA to forecast the mean of solar irradiance and and heteroskedastic model to forecast the volatility.

International audience ; Between 2006 and 2011, the French government set up an incentive policy in order to develop the electricity production from photovoltaics. In the overseas territories, like Reunion Island, the feedin tariffs proposed for the next 20 years were specifically high. It resulted an exponential increase of the installed PV systems. For these small non-interconnected grids, an important penetration rate of such an intermittent energy can destabilize the supply-demand balance and may lead to an electricity outage risk. As a consequence, a regulatory limit of 30% of the instantaneous power produced from intermittent renewable (solar, wind and waves) was defined in order to avoid this risk. For instance, this legal constraint was reached in 2012 in Reunion. In this context, the forecasting of solar irradiance is essential in order to increase the penetration rate of PV output power into the grid. More precisely, accurate solar forecasts will help the grid operator to better manage the means of production. Forecasts are also needed in order to optimize the operation of grid connected storage energy systems. In order to cope with specific plant operations, forecasts must be provided with different granularities and horizons. In this work, we will focus on forecasts from 10 minutes to to 4 hours ahead with a 10 minutes granularity. The chosen forecasting time horizon will permit to monitor the production and to adjust the scheduling. In this paper, we introduce a novel approach based on econometrics models to forecast the global solar irradiance. The use of econometrics methods is justified by the similar behavior exhibited by the clear sky index and log return time series Our approach combine 2 models: a ARMA to forecast the mean of solar irradiance and and heteroskedastic model to forecast the volatility.

Mención Internacional en el título de doctor ; Renewable energies are the leading alternative to fossil fuels, facing the constant threat of climate change. The development of these new resources has grown in the latest years, especially in the field of solar and wind energy. These renewable power sources have gathered a series of research challenges that, to this date, are still to be solved, with many contributions to this end in the last decade. The role of estimation and forecasting of solar energy is key to the development of the solar energy market, because it cheapens instrumentation costs and improve the efficiency of solar energy market participation in the power grid. The forecast of solar energy is fundamental to estimate costs and operational regulations of a solar plant, although the intermittence of solar energy makes this a difficult task. On the other hand, the estimation of solar irradiance can replace expensive measuring devices such as pyranometers or pyrheliometers; or the need of expert supervision on meteorological stations for cloud type classification. In order to improve estimation, two proposals are studied. The first approach to estimation is the automatic classification of clouds by including ceilometer information. This is a device capable of measuring height and thickness of a cloud, information that has never been applied to cloud classification. The next proposal is the estimation of irradiance by directly analyzing images with Convolutional Networks and multiple perspectives, a never before used technique for solar energy estimation. To improve forecasting the integration of prediction models is proposed. This technique compares and combines existing predictive models to obtain a final, more accurate, prediction. Although this is not a new approach, it has never been applied to various prediction models specialized in different horizons, or for short-term forecasting. Given that clouds produce the greatest interference between extraterrestial and surface irradiance, whole-sky cloud images are a valuable source of data for radiation estimation. To study the cloud type classification problem a Random Forest algorithm is employed. The algorithm is trained using information from cloud height and thickness, which is combined with camera im- 3 age features. Including cloud height and width proves to noticeably improve accuracy even when difficult to classify cloud types are included. Results for 10-class cloud classification, including multiple clouds in a single image, show 71.12%, an improvement over the 50.6% achieved without ceilometer information. This study shows the positive impact of ceilometer information in the cloud classification problem. Irradiance estimation can also be estimated directly from camera images. To face this problem various models have been created using convolutional neural networks, a Machine Learning technique fit for image recognition. Two approaches are proposed, a model with information from a single camera and a model with multiple sky perspectives. In addition to the common RGB colour channels used in image processing, two new channels are included: the distance from a pixel to the sun and the cloudy pixels of an image. Multiple perspectives improve noticeably all alternatives proposed, proving the contribution of the multi-view convolutional network proposed. There are many predictive models that predict with diverse capabilities at different predictive horizons. In this thesis, this process is called forecast integration (or blending). An integration model is proposed to blend four physical models from four meteorological stations at the south of the Iberian peninsula. Using support vector regression these are combined in a linear and non-linear way using the four predictors as inputs to machine learning. Two approaches are presented: a horizon approach that builds a model for each predictive horizon, and a general approach that builds a single prediction model for all horizons. In addition, a regional model is proposed, capable of of making predictions at a regional level instead of a station level. Results from integration are very positive compared with the baseline models for global and direct irradiance. Some absolute improvements reach 15% when comparing integration models to any predictor model when rRMSE and rMAE are evaluated on global and direct irradiance. At a regional level, there are also improvements, at an absolute 5% on global radiation over the predictor models and 10% for direct irradiance. The general approach is specially remarkable because, using a single model, it can obtain the best results on rMAE and match the results of other integration models on rRMSE. ; Las energías renovables son una importante alternativa a los combustibles fósiles ante el constante avance del cambio climático. El desarrollo de estos nuevos recursos se ha acelerado en los últimos años, especialmente en el campo de energía eólica y solar. Estas fuentes energéticas han atraído una serie de desafíos de investigación que siguen en progreso de ser resueltos, con numerosas contribuciones en la última década. La labor de estimación y predicción de energía solar es integral para el desarrollo del mercado energético, ya que permite abaratar costes instrumentales y mejorar la eficiencia de la penetración de la energía solar en la mezcla energética. La predicción de energía es fundamental en el mercado energético para estimar costes y regulaciones operativas de plantas solares, aunque la intermitencia de la energía solar hace que sea una tarea difícil. Por otro lado, la estimación de radiación solar permite reemplazar herramientas de alto coste como piranómetros y pirheliómetros; o la necesidad de expertos para detectar tipos de nube. Para la mejora de estimación se estudian dos propuestas diferentes. En primer lugar se trata de abordar el problema de clasificación de nubes, incluyendo información de ceilómetro. Esta es una herramienta que mide altura y anchura de una nube, cuyo uso nunca ha sido aplicado en la clasificación de nubes. La siguiente propuesta es la estimación de radiación directa a partir de imágenes, usando Redes Convolucionales y múltiples perspectivas, una técnica que nunca ha sido empleada para la estimación de energía solar. Para la mejora de la predicción de energía solar se propone la integración de modelos predictivos. Esta técnica consiste en la combinación de modelos predictivos existentes para obtener una predicción final mucho más precisa que las iniciales. Aunque esta no es una aproximación nueva, su exploración ha sido insuficiente para varios modelos especializados en distintos horizontes, o para predicción a corto plazo. Dado que las nubes producen el mayor impacto entre la radiación extraterrestre y la radiación que alcanza la superficie, las imágenes de nubes son una fuente de datos valiosa para la estimación de radiación. Para estudiar la clasificación del tipo de nube se emplea un algoritmo Random Forest entrenado con información sobre la altura y ancho de la nube, que se combina con estadísticos obtenidos a partir de imágenes. La información del ceilómetro permite mejorar notablemente los resultados incluso cuando se incluyen ejemplos de nube difíciles para expertos. Se logra predecir 10 tipos de nube con un 71.1% de precisión frente al 50.6% obtenido sin ceilómetro. Este estudio prueba que la inclusión de información del ceilómetro tiene un impacto muy positivo en los resultados. La estimación de radiación también se puede afrontar directamente a partir de las imágenes. Para tratar este problema se han creado varios modelos usando redes convolucionales apropiadas para el análisis de imágenes. Se proponen modelos que utilizan información proveniente de una sola cámara y otro modelo con múltiples perspectivas del cielo. Además de los canales habituales utilizados en el proceso de imágenes con redes convolucionales (RGB) se incluyen varios canales adicionales: la lejanía de los píxeles al sol y los píxeles que representan nubes. Las múltiples perspectivas y canales de información adicionales mejoran notablemente las alternativas propuestas, demostrando el aporte de la red convolucional multi-perspectiva propuesta. Existen multitud de modelos predictivos que ofrecen predicciones con capacidades diversas a distintos horizontes de predicción. En esta tesis, se propone un modelo integrador de cuatro modelos predictivos. Usando Maquinas de Vectores de Soporte para regresión se combinan de manera lineal y nolineal los cuatro predictores, utilizando como entradas al modelo las predicciones de los cuatro predictores. Se proponen dos aproximaciones, una por horizontes, construyendo un modelo para cada horizonte de predicción, y otra general, construyendo un modelo único para todos los horizontes. Los modelos han sido evaluados con datos procedentes de cuatro localizaciones al sur de la península ibérica. También se propone un modelo integrador regional, capaz de aportar predicciones a nivel regional en lugar de a nivel de estación. Los resultados de integración son muy positivos tanto para radiación global como directa, mostrando mejoras absolutas hasta del 15% frente a cualquier predictor tanto en rRMSE como en rMAE. A nivel regional también se obtienen mejoras del 5% para radiación global y del 10% para radiación directa. La aproximación general es especialmente destacable, haciendo uso de un único modelo, es capaz de obtener los mejores resultados en rMAE e igualar al resto de modelos de integración en rRMSE. ; This dissertation has been developed under the project PROSOL ENE2014-56126-C2 (Towards an integrated model for solar energy forecasting) in collaboration with the research group MATRAS (University of Jaen) and funded by the Ministry of Science and Innovation (Spanish Government). All the data shown in this text has been provided by MATRAS and has been used with their permission. ; Programa de Doctorado en Ciencia y Tecnología Informática por la Universidad Carlos III de Madrid ; Presidente: Pedro Isasi Viñuela.- Secretario: Esteban García Cuesta.- Vocal: Ricardo Simón Carbajo

A detailed knowledge of the solar resource is a critical point in the design and control of Concentrating Solar Power (CSP) plants. In particular, accurate forecasting of solar irradiance is essential for the efficient operation of solar thermal power plants, the management of energy markets, and the widespread implementation of this technology. Numerical weather prediction (NWP) models are commonly used for solar radiation forecasting. In the ECMWF deterministic forecasting system, all forecast parameters are commercially available worldwide at 3-hourly intervals. Unfortunately, as Direct Normal solar Irradiance (DNI) exhibits a great variability due to the dynamic effects of passing clouds, 3-h time resolution is insufficient for accurate simulations of CSP plants due to their nonlinear response to DNI, governed by various thermal inertias due to their complex response characteristics. DNI series of hourly or sub-hourly frequency resolution are normally used for an accurate modeling and analysis of transient processes in CSP technologies. In this context, the objective of this study is to propose a methodology for generating synthetic DNI time series at 1-h (or higher) temporal resolution from 3-h DNI series. The methodology is based upon patterns as being defined with help of the clear-sky envelope approach together with a forecast of maximum DNI value, and it has been validated with high quality measured DNI data. ; This project has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No 654984.

Recent research on solar irradiance forecasting has attracted considerable attention, as governments worldwide are displaying a keenness to harness green energy. The goal of this study is to build forecasting methods using deep learning (DL) approach to estimate daily solar irradiance in three sites in Kuwait over 12 years (2008–2020). Solar irradiance data are used to extract and understand the symmetrical hidden data pattern and correlations, which are then used to predict future solar irradiance. A DL model based on the attention mechanism applied to bidirectional long short-term memory (BiLSTM) is developed for accurate solar irradiation forecasting. The proposed model is designed for two different conditions (sunny and cloudy days) to ensure greater accuracy in different weather scenarios. Simulation results are presented which depict that the attention based BiLSTM model outperforms the other deep learning networks in the prediction analysis of solar irradiance. The attention based BiLSTM model was able to predict variations in solar irradiance over short intervals in continental climate zones (Kuwait) more efficiently with an RMSE of 4.24 and 20.95 for sunny and cloudy days, respectively.