Solar radiation forecasting with non lineal statistical techniques and qualitative predictions from spanish national weather service

Apr 15, 2013Download as PPT, PDF0 likes592 views

1) The study uses neural network models to predict half-daily solar radiation values in Spain, aiming to improve predictions for solar energy feed-in tariffs. 2) The best neural network model reduced the mean root squared error of predictions compared to a persistence model by over 9%. 3) While the first neural network model had errors limited by nonlinear signal behavior, a second model further improved predictions, achieving the best error level possible with the presented methodology.

Solar radiation forecasting with non-lineal statistical techniques and 20 08
UN 008
qualitative predictions from Spanish National Weather Service OS r 2
EUR be
to BON
Oc IS
Martín L., Zarzalejo L.F., Polo J., Navarro A., Marchante R. L

1. INTRODUCTION 3. RESULTS
Solar energy feed-in tariff is regulated by (RD 436/2004, Errors of the models essayed are measured in terms of
661/2007) in Spain. Predictions must be given for next 72 mean root mean squared deviation (RMSD). The best
hours and deviations are strongly penalized. A new method NN(z) model is compared to persistence model (PER) in
to predict half daily values of solar radiation is presented. terms of improvement of RMDS.
2. METHODOLOGY 1 N 45.0 ° N

∑ ( xi − xi )
2
RMSD = ˆ 42.5 ° N

Solar radiation is transformed to a new gaussian and N i=1 40.0 ° N
• Madrid RRN AEMet

stationary variable. “Lost component” (LC) is the difference
 i − ierrorm  37.5 ° N

betwen extratrrestrial and ground measured solar radiation. improvement =  1 − ÷
 i − ierror ÷
35.0 ° N
°
15.0 ° W12 ° ° 0.0 E
.5 W10.0 ° W ° E 5.0 ° E 7.5 E 1

 p 
7.5 ° W 5.0 ° W 2.5 ° W 0.0 ° 2.5

6000
38

5000
Lost Component L
36
NN(1)
4000 C NN(2)
34

Halfday)
NN(3)
P NN(4)
3000 R NN(5)
32

2
E NN(6)
2000 D NN(7)

% RM SE Prediction G (W /m
I 30 NN(8)
C NN(9)
T NN(10)
1000 28 P ersistence
I
O
0
0
N 26
100 200 300 400 500 600 700 S
Half Day
24

Synoptic predictions of sky conditions (SYN) are used as 22
1 2 3 4 5 6
input to the neual network to test the improvement of the 40
Pre diction horizon (Halfdaily )

predictions. AEMET offers this predicitons in its web page W
I
for each location of Spain and 7 days in adavance. T
35

H NN(1)
Halfday)

NN(2)
30
NN(3)
S NN(4)
Y
2

NN(5)
%RMSE Prediction G (W/m

N 25
NN(6)
NN(7)
C 20
NN(8)
NN(9)
O
NN(10)
N Persistence
D 15
I
T
10
I
O
N 5
S 1 2 3 4 5 6
Prediction horizon (Halfdaily)

4. CONCLUSIONS
The error of the first model is limited by an upper level
which is due to deterministic nonlinear behaviour of the
signal which can’t be followed correctly by neural
network models. The second model improves
Neural Network (NN) is used to predict future values from considerably the prediction. The error has a lower level
observations. NN(z) índica el tamaño del vector patrón de of nine percent which is the best prediction error that can
entrada empleado z=1…10. be achieved with the methology presented.

División de Energías Renovables (Departamento de Energía), CIEMAT, Av. Complutense
nº22, Madrid, 28040, (Madrid) España, +34 913466048, luis.martin@ciemat.es

The document discusses the impact of polarization sensitivity on ocean color measurements by the Visible Infrared Imaging Radiometer Suite (VIIRS) instrument. Extensive testing characterized VIIRS's polarization sensitivity, which varies by band and detector. Left uncorrected, polarization effects could cause errors in retrieved water leaving radiance (nLw) used for ocean color products. Simulations show polarization correction in processing reduces nLw errors significantly, while uncertainty in sensor characterization has a small additional impact. Updating algorithms with polarization correction and detector-dependent calibration is needed to achieve ocean color performance comparable to legacy sensors.

Ispra 2007 luis martín2IrSOLaV Pomares

��

The document presents preliminary results on daily radiation forecasting using statistical methods. It explores using linear and non-linear prediction techniques like TAG(p) and discusses using historical piranometric data from 1996-2003 to predict daily global solar radiation values and analyze the data through partial autocorrelation. The goals are to characterize and predict solar radiation as an energy resource for applications like short and medium term forecasting.

Luis martin oldenburg 7-2007IrSOLaV Pomares

��

The document summarizes research on forecasting solar radiation using the Weather Research and Forecasting (WRF) mesoscale simulation model. It presents results on hourly and daily forecasts of global solar radiation for locations in Europe with 27km resolution. Hourly forecasts had mean biases between -14.56% to 7.46% and root mean square errors of 34.39-75.45%. Daily forecasts showed mean biases of -16.01% to 9.14% and root mean square errors of 24.36-64.01%. The model showed better performance forecasting solar radiation in Granada, Spain compared to Oviedo, Spain.

Solar radiation and plasma diagnosticsUniversity of Glasgow

��

Solar radiation and plasma diagnostics outlines key concepts about solar radiation and how it is detected. It discusses (1) how solar radiation is formed through radiative transfer processes and depends on temperature, opacity, and optical depth; (2) the interaction of radiation with plasma is described by the radiative transfer equation; and (3) different spectral features form in different layers, with absorption lines seen in the photosphere and emission seen in the chromosphere.

Time series solar radiation forecastingIrSOLaV Pomares

��

1. The document discusses global solar radiation forecasting techniques used by the Spanish National Weather Service. 2. It outlines various statistical models for predicting clearness index, lost component, and qualitative predictions up to 6 hours in advance. 3. The best models were neural networks for clearness index and lost component predictions, achieving up to 28% improvement over persistence.

Global solar radiation forecasting with non lineal statistical techniques and...IrSOLaV Pomares

��

This document discusses methods for forecasting global solar radiation using statistical techniques and qualitative predictions from the Spanish National Weather Service. It presents various data preprocessing techniques to remove non-stationarity from time series data, including using clearness index, lost component variable, subtracting mean values, and removing annual harmonics. Graphs show improvements in predicting the lost component using differencing. Qualitative predictions from weather services are also combined with the lost component technique, showing further improved predictions over longer time horizons.

Programa jornada predicciónIrSOLaV Pomares

��

The document outlines the agenda for a workshop on applications of solar forecasting held at CIEMAT on June 11, 2013. The workshop covered topics such as forecasting needs for solar energy production, managing energy from solar power plants and photovoltaic systems, commercial solar forecasting services, and future guidelines on solar forecasting from research and meteorological perspectives. Presentations were given by experts from universities, research institutions, solar energy companies, and meteorological organizations on integrating solar forecasting into energy management, electricity markets, and power plant operations.

Solar radiation forecasting with wrf model in the iberian peninsulaIrSOLaV Pomares

��

This document summarizes research validating solar radiation forecasting models in the Iberian Peninsula. It evaluates the ECMWF ERA-40 global model and the WRF mesoscale model at hourly and daily resolutions against ground measurements. The ECMWF model underestimates daily solar radiation with errors up to 39.55%. The WRF model has hourly errors ranging from 30-98% and daily errors from 23-89%. While WRF with NCEP data fails to accurately reproduce synoptic situations, using ECMWF data inputs may improve forecasts. However, cloud movement remains challenging to predict deterministically. Further progress is needed to meet a 20% error requirement for hourly solar forecasts in Spain.

CSP, CPV or PV solar plant simulation reportIrSOLaV Pomares

��

Este informe técnico analiza los datos de radiación solar medidos en la estación de STATION entre diciembre de 2005 y julio de 2009. Se realiza un análisis de calidad de los datos mediante filtros físicos y comparaciones entre las componentes medidas para verificar su precisión. También se comparan los datos con los de otras 5 estaciones cercanas y con un modelo de cielo despejado para validar la calidad de las mediciones a lo largo de los años.

Solar radiation ground measured data quality assessment reportIrSOLaV Pomares

��

This technical report analyzes radiometric data measured at a site in XXX from June 2011 to May 2012. It assesses data quality using various filters and comparisons to clear sky models. Global, diffuse, and direct normal irradiance were measured. The methodology section describes transforming time to true solar time, calculating hourly/daily averages, and quality analysis including physical limits checks and cross-component relationship checks. Graphs of measured and clear sky data are presented and used to visually inspect data quality.

Typical Meteorological Year Report for CSP, CPV and PV solar plantsIrSOLaV Pomares

��

This technical report analyzes the solar resource available at a site in Northern Cape, South Africa selected to host a solar thermal power plant. It presents a typical meteorological year (TMY) developed using 12 years of hourly solar radiation data for the site. The TMY is generated using a methodology that selects the most representative month from each year for key meteorological variables. It is comprised of months from 2007 to 2010 that best match the long-term averages for global horizontal and direct normal solar radiation at the site. The TMY and long-term averages are presented and show a close match in monthly and daily solar radiation patterns for use in modeling solar power production at the site.

Technical report site assessment of solar resource for a csp plant. correctio...IrSOLaV Pomares

��

This technical report summarizes a solar resource assessment for a CSP solar project site in Morocco. It details the methodology used by IrSOLaV to estimate solar radiation from satellite images, which has been validated against ground measurements. IrSOLaV corrects initial satellite estimates using on-site meteorological station data, achieving hourly and daily GHI and DNI estimation uncertainties of 12-18% and 5-10%, respectively. The report presents satellite-derived solar radiation data for the project site location from 2011-2013 and compares it to on-site pyranometer and pyrheliometer measurements.

Forecasting commercial services from S2M - Juan Liria (Sun2Market)IrSOLaV Pomares

��

Sun to Market Solutions was founded to provide independent technical support and engineering services for the growing solar power sector. It has over 25 engineers located across five continents with experience conducting studies in over 30 countries. Sun to Market Solutions has become a leading global advisor for the solar power industry. It developed a nowcasting system that provides short-term weather forecasts up to 24 hours in advance to help optimize performance of solar power plants and their integration with the electric grid. The nowcasting system collects and processes data from sky cameras, weather stations, satellites and numerical models to predict direct normal irradiation levels with an accuracy of within ±150 W/m2.

Future guidelines the meteorological view - Isabel Martínez (AEMet)IrSOLaV Pomares

��

This document discusses nowcasting and forecasting of solar irradiance using meteorological data. Nowcasting uses observations from the past 6 hours to predict clouds and irradiance up to 2 hours ahead for a specific site. Forecasting uses numerical weather prediction models to predict clouds and irradiance out to days or weeks ahead on regional to global scales. The document outlines various nowcasting techniques including the use of sky cameras, satellites, and neural networks. It also describes several forecast models run operationally at ECMWF and AEMET including HIRLAM, HARMONIE, and the ECMWF model. Prognostic aerosols are also modeled to improve irradiance forecasts.

Future guidelines on solar forecasting the research view - David Pozo (Univer...IrSOLaV Pomares

��

The document discusses solar radiation forecasting research conducted by the MATRAS solar radiation and atmosphere modelling group at the University of Jaen. The group has developed facilities for measuring and forecasting direct normal irradiance (DNI) using sky cameras, ceilometers and numerical weather prediction models. Their research aims to improve short-term DNI nowcasting and forecasting up to 72 hours ahead for applications such as solar power plant operation and electricity market participation. They are also investigating how to optimally balance solar and wind power resources to reduce production variability.

Forecasting commercial services from IrSOLaV - Luis Martin (IRSOLAV)IrSOLaV Pomares

��

IrSOLaV is a spin-off company founded in 2007 that provides solar radiation data and forecasting services using satellite imagery. It has participated in over 3.4GW of solar energy projects worldwide. IrSOLaV's products include time series data, reports, and maps of solar radiation. It also provides short-term (72 hours) and nowcasting (6 hours) forecasts of solar radiation. IrSOLaV conducts research partnerships and has offices in Spain, India, and Chile.

Managing the energy purchasing - Jorge gonzalez (Gesternova)IrSOLaV Pomares

��

Gesternova is an energy company that supports renewable energy development in Spain. They offer renewable energy to customers and help renewable energy producers connect to the grid and register with agencies to receive payments. There are several steps and agencies involved in connecting a solar plant to the grid and ensuring the plant owner receives payment for the electricity generated. Gesternova assists plant owners with registration, forecasting energy output, and interacting with grid operators and energy market agencies to ensure payments are processed accurately and on time.

Forecasting energy fo pv system - Miguel Martínez (Wenner Solar)IrSOLaV Pomares

��

Este documento presenta un taller sobre la predicción de la energía de sistemas fotovoltaicos. Incluye información sobre la predicción de la producción eléctrica de plantas fotovoltaicas, la influencia del mantenimiento en la producción y una visión general del sector fotovoltaico. El documento también proporciona detalles sobre el análisis inicial, el estudio de radiación y el rendimiento energético necesarios para predecir la producción de energía de una planta fotovoltaica.

Gemasolar a thermal solar power plant with 15 hours, Ignacio Burgaleta (Torre...IrSOLaV Pomares

��

The Gemasolar plant is a 19.9 MW solar thermal power plant in Spain that uses a central tower receiver with molten salt storage. It has over 2,600 heliostat mirrors that focus sunlight onto the receiver to heat molten salt to 565°C. The heated salt is then stored in tanks for up to 15 hours. This allows electricity production both during sunlight and after sunset. The plant began commercial operation in 2011 and has consistently exceeded performance guarantees, producing electricity continuously for over 12 days. Future, larger plants are planned to achieve even greater economies of scale.

Solar Thermal Power Plant with Thermal Storage - Ignacio Burgaleta (Torresol ...IrSOLaV Pomares

��

Torresol Energy operates several solar thermal power plants in Spain that use parabolic trough collectors and central tower technology. These plants include molten salt storage systems to allow electricity production when the sun is not shining. The document discusses Torresol Energy's experience with molten salt storage, including the advantages it provides in improving plant efficiency and enabling dispatchable solar power. It also describes the components and operation of the company's 50 MW parabolic trough plant with 7 hours of thermal storage. Accurate forecasting of solar irradiance and clouds is important for optimizing plant operations and grid integration of the solar power.

General situation of solar thermal energy - Eduardo Iglesias (Protermosolar)IrSOLaV Pomares

��

1) Solar thermal energy has potential for large-scale deployment as a carbon-free electricity source but currently only has 3 GW installed globally. 2) In Spain, solar thermal provided over 500 GWh in July 2012 and its economic impact in 2012 included over 1.8 billion euros in GDP contribution and nearly 18,000 jobs. 3) For solar thermal to reach its full potential, costs must continue to decline as deployment increases, with projections of 14-25 euro cents/kWh by 2030 as the industry matures and capacity grows towards 300 GW.

The CECRE: Making renewable energy technologies compatible with the security ...IrSOLaV Pomares

��

Red Eléctrica de España is the Spanish transmission system operator. It operates and maintains the transmission grid to ensure security of electricity supply. Renewable energy sources now account for 39% of Spain's installed capacity, including 22.6% from wind and 4.4% from solar PV. Integrating high levels of renewable energy presents challenges like variability in generation, lack of observability and controllability, and impacts on grid constraints and voltage control. The CECRE control center helps address these issues by providing centralized monitoring and control of renewable facilities.

Workshop on Applications of Solar Radiation Forecasting - Introduction - Jesú...IrSOLaV Pomares

��

This document provides an overview of solar radiation forecasting techniques and related activities. It discusses (1) the classification of forecasting methods, including very short-term nowcasting using sky imagers and satellite images, and forecasting with numerical weather prediction models, (2) Ciemat's involvement in solar forecasting research through IEA Tasks 36 and 46, benchmarking different methods, and (3) Ciemat's role in the DNICast and COST Wire projects which aim to improve techniques for direct normal irradiance forecasting and integrate forecasts with power plant and grid models.

Assessment and evaluation of solar resources adb courseIrSOLaV Pomares

��

This document discusses the assessment and evaluation of solar resources. It covers topics like solar geometry, interaction of solar radiation with the atmosphere, measurement of solar radiation, and databases of solar radiation. Key points include how solar position varies daily and yearly due to Earth's orbit and rotation, atmospheric effects on solar irradiance like scattering, absorption and reflection, and parameters used to characterize solar resources like extraterrestrial and global irradiance. The goal is to provide tools to help evaluate solar power projects.

Solar resource assessment luis martinIrSOLaV Pomares

��

The document provides an overview of solar resource evaluation methodology. It discusses the need to evaluate solar radiation for energy system studies and compares classical evaluation using measurements to evaluation from satellite images. The proposed procedure involves determining what type of data is needed like hourly or monthly time series data or maps. It then describes whether satellite information or measurements will be used. The document also covers topics like solar radiation characteristics, components, clear sky models, and methods for measuring solar radiation including typical monitoring station setups.

Review solar prediction iea 07-06IrSOLaV Pomares

��

This document reviews different techniques for predicting solar irradiance levels including: 1) Numerical weather prediction models and statistical prediction for short to long term forecasting. 2) MOS (Model Output Statistics) techniques using sky cover products from weather centers for short term prediction. 3) Satellite-based methods comparing different approaches and error analysis for short term forecasting. 4) Signal analysis techniques including wavelet transforms and artificial neural networks combined with recurrent networks for improved prediction. Future work proposed includes combining these methods along with normalized data and forecasts from weather centers to improve prediction accuracy across timescales.

Quality of ground data for assessment and benchmarkingIrSOLaV Pomares

��

This document discusses the importance of assessing the quality of ground-based solar radiation data used for model development, benchmarking, and assessment. It outlines several existing quality control procedures from organizations like BSRN, ARM, and NREL that check for physically realistic values and consistency between radiation components. Common errors found in some databases are also described, such as errors in the recorded time reference affecting clearness index calculations and erroneous beam radiation near sunrise/sunset. The document raises questions about whether Task 36 should propose a general quality control procedure and which criteria should be included in a solar radiation data guide.

Luis martin jornada becarios 2006IrSOLaV Pomares

��

Este documento presenta una revisión de técnicas de predicción de la radiación solar diaria como modelos estadísticos, redes neuronales y predicción a partir de imágenes de satélite. Luego describe la metodología utilizada, que incluye datos experimentales, modelos de predicción como perceptrones multicapa y wavelets, y la evaluación de resultados. Finalmente concluye la necesidad de predecir la radiación solar y que los datos terrestres previos producen un menor error de predicción.

Solar radiation forecasting with non lineal statistical techniques and qualitative predictions from spanish national weather service

1. Solar radiation forecasting with non-lineal statistical techniques and 20 08 UN 008 qualitative predictions from Spanish National Weather Service OS r 2 EUR be to BON Oc IS Martín L., Zarzalejo L.F., Polo J., Navarro A., Marchante R. L 1. INTRODUCTION 3. RESULTS Solar energy feed-in tariff is regulated by (RD 436/2004, Errors of the models essayed are measured in terms of 661/2007) in Spain. Predictions must be given for next 72 mean root mean squared deviation (RMSD). The best hours and deviations are strongly penalized. A new method NN(z) model is compared to persistence model (PER) in to predict half daily values of solar radiation is presented. terms of improvement of RMDS. 2. METHODOLOGY 1 N 45.0 ° N ∑ ( xi − xi ) 2 RMSD = ˆ 42.5 ° N Solar radiation is transformed to a new gaussian and N i=1 40.0 ° N • Madrid RRN AEMet stationary variable. “Lost component” (LC) is the difference  i − ierrorm  37.5 ° N betwen extratrrestrial and ground measured solar radiation. improvement =  1 − ÷  i − ierror ÷ 35.0 ° N ° 15.0 ° W12 ° ° 0.0 E .5 W10.0 ° W ° E 5.0 ° E 7.5 E 1  p  7.5 ° W 5.0 ° W 2.5 ° W 0.0 ° 2.5 6000 38 5000 Lost Component L 36 NN(1) 4000 C NN(2) 34 Halfday) NN(3) P NN(4) 3000 R NN(5) 32 2 E NN(6) 2000 D NN(7) % RM SE Prediction G (W /m I 30 NN(8) C NN(9) T NN(10) 1000 28 P ersistence I O 0 0 N 26 100 200 300 400 500 600 700 S Half Day 24 Synoptic predictions of sky conditions (SYN) are used as 22 1 2 3 4 5 6 input to the neual network to test the improvement of the 40 Pre diction horizon (Halfdaily ) predictions. AEMET offers this predicitons in its web page W I for each location of Spain and 7 days in adavance. T 35 H NN(1) Halfday) NN(2) 30 NN(3) S NN(4) Y 2 NN(5) %RMSE Prediction G (W/m N 25 NN(6) NN(7) C 20 NN(8) NN(9) O NN(10) N Persistence D 15 I T 10 I O N 5 S 1 2 3 4 5 6 Prediction horizon (Halfdaily) 4. CONCLUSIONS The error of the first model is limited by an upper level which is due to deterministic nonlinear behaviour of the signal which can’t be followed correctly by neural network models. The second model improves Neural Network (NN) is used to predict future values from considerably the prediction. The error has a lower level observations. NN(z) índica el tamaño del vector patrón de of nine percent which is the best prediction error that can entrada empleado z=1…10. be achieved with the methology presented. División de Energías Renovables (Departamento de Energía), CIEMAT, Av. Complutense nº22, Madrid, 28040, (Madrid) España, +34 913466048, luis.martin@ciemat.es

�ݺ�ߣ

Solar radiation forecasting with non lineal statistical techniques and qualitative predictions from spanish national weather service

Recommended

More Related Content

More from IrSOLaV Pomares (20)

Solar radiation forecasting with non lineal statistical techniques and qualitative predictions from spanish national weather service