Elke Lorenz

Irradiance Forecasting for the Power Prediction of Grid-Connected Photovoltaic Systems

The contribution of power production by photovoltaic (PV) systems to the electricity supply is constantly increasing. An efficient use of the fluctuating solar power production will highly benefit from forecast information on the expected power production. This forecast information is necessary for the management of the electricity grids and for solar energy trading. This paper presents an approach to predict regional PV power output based on forecasts up to three days ahead provided by the European Centre for Medium-Range Weather Forecasts (ECMWF). Focus of the paper is the description and evaluation of the approach of irradiance forecasting, which is the basis for PV power prediction. One day-ahead irradiance forecasts for single stations in Germany show a rRMSE of 36%. For regional forecasts, forecast accuracy is increasing in dependency on the size of the region. For the complete area of Germany, the rRMSE amounts to 13%. Besides the forecast accuracy, also the specification of the forecast uncertainty is an important issue for an effective application. We present and evaluate an approach to derive weather specific prediction intervals for irradiance forecasts. The accuracy of PV power prediction is investigated in a case study.

Solar energy assessment using remote sensing technologies

About 20% of the final energy consumed in Europe is used in buildings. The active and passive use of solar energy is an approach to reduce the fossil energy consumption and the greenhouse gas emissions originated by buildings. Consideration of solar energy technologies in urban planning demands accurate information of the available solar resources. This can be achieved by the use of remote sensing data from geostationary satellites which show a very high spatial and a sufficient temporal resolution compared to ground station data. This paper gives a brief introduction to the HELIOSAT method applied to derive surface solar irradiance from satellite images and shows examples of applications: The use of daylight in buildings, the generation of correlated time series of solar irradiance and temperature as input data for simulations of solar energy systems and a short-term forecast of solar irradiance which can be used in intelligent building control techniques. Finally an outlook is given on potential improvements expected from the next generation of European meteorological satellites Meteosat Second Generation (MSG).

Short-term forecasting of solar radiation: a statistical approach using satellite data

Short-term forecasting of solar irradiance is an important issue for many fields of solar energy applications. As the solar surface irradiance can be inferred from satellite measurements with a high temporal and spatial resolution, we use satellite images as a data source for forecasting. The satellite data provide information on cloudiness, the most important atmospheric parameter for surface irradiance. This paper describes the application of a statistical method to detect the motion of cloud structures from satellite images. Extrapolating the temporal development of the cloud situation, solar radiation can be predicted for time scales from 30 min up to 2 h. The forecasts are evaluated with respect to accuracy and an example for the application of the forecast algorithm to predict PV power output is presented.

Prediction of Solar Irradiance and Photovoltaic Power

Power generation from solar and wind energy systems is highly variable due to its dependence on meteorological conditions. An efficient use of these fluctuating energy sources requires reliable forecast information for management and operation strategies. We give an overview of different applications and state-of-the-art models for solar irradiance and photovoltaic power prediction, including time series models based on on-site measured data, models based on the detection of cloud motion in satellite images, and numerical weather prediction-based models. In the second part of this chapter, we show evaluation results for selected irradiance and power prediction schemes.

Short term fluctuations of wind and solar power systems

Wind and solar power are known to be highly influenced by weather events and may ramp up or down abruptly. Such events in the power production influence not only the availability of energy, but also the stability of the entire power grid. By analysing significant amounts of data from several regions around the world with resolutions of seconds to minutes, we provide strong evidence that renewable wind and solar sources exhibit multiple types of variability and nonlinearity in the time scale of {\it seconds} and characterise their stochastic properties. In contrast to previous findings, we show that only the jumpy characteristic of renewable sources decreases when increasing the spatial size over which the renewable energies are harvested. Otherwise, the strong non-Gaussian, intermittent behaviour in the cumulative power of the total field survives even for a country-wide distribution of the systems. The strong fluctuating behaviour of renewable wind and solar sources can be well characterised by Kolmogorov-like power spectra and

Statistical Learning for Short-Term Photovoltaic Power Predictions

q-

Short-Term Forecasting of Surface Solar Irradiance Based on Meteosat-SEVIRI Data Using a Nighttime Cloud Index

exponential probability density functions. Using the estimated potential shape of power time series, we quantify the jumpy or diffusive dynamic of the power. Finally we propose a time delayed feedback technique as a control algorithm to suppress the observed short term non-Gaussian statistics in spatially strong correlated and intermittent renewable sources.

Solar Forecasting in a Challenging Insular Context

A reliable prediction of photovoltaic (PV) power plays an important part as basis for operation and management strategies for a efficient and economical integration into the power grid. Due to changing weather conditions, e.g., clouds and fog, a precise forecast in a few hour range can be a difficult task. The growing IT infrastructure allows a fine screening of PV power. On the basis of big data sets of PV measurements, we apply methods from statistical learning for one- to six-hour ahead predictions based on data with hourly resolution. In this work, we employ nearest neighbor regression and support vector regression for PV power predictions based on measurements and numerical weather predictions. We put an emphasis on the analysis of feature combinations based on these two data sources. After optimizing the settings and comparing the employed statistical learning models, we build a hybrid predictor that uses forecasts of both employed models.

Overview of Irradiance and Photovoltaic Power Prediction

The cloud index is a key parameter of the Heliosat method. This method is widely used to calculate solar irradiance on the Earth’s surface from Meteosat visible channel images. Moreover, cloud index images are the basis of short-term forecasting of solar irradiance and photovoltaic power production. For this purpose, cloud motion vectors are derived from consecutive images, and the motion of clouds is extrapolated to obtain forecasted cloud index images. The cloud index calculation is restricted to the daylight hours, as long as SEVIRI HR-VIS images are used. Hence, this forecast method cannot be used before sunrise. In this paper, a method is introduced that can be utilized a few hours before sunrise. The cloud information is gained from the brightness temperature difference (BTD) of the 10.8 µm and 3.9 µm SEVIRI infrared channels. A statistical relation is developed to assign a cloud index value to either the BTD or the brightness temperature T10:8, depending on the cloud class to which the pixel belongs (fog and low stratus, clouds with temperatures less than 232 K, other clouds). Images are composed of regular HR-VIS cloud index values that are used to the east of the terminator and of nighttime BTD-derived cloud index values used to the west of the terminator, where the Sun has not yet risen. The motion vector algorithm is applied to the images and delivers a forecast of irradiance at sunrise and in the morning. The forecasted irradiance is validated with ground measurements of global horizontal irradiance, and the advantage of the new approach is shown. The RMSE of forecasted irradiance based on the presented nighttime cloud index for the morning hours is between 3 and 70 W/m2, depending on the time of day. This is an improvement against the previous precision range of the forecast based on the daytime cloud index between 70 and 85 W/m2.

Spectrally resolved DNI measurements: Results of a field comparison of spectroradiometers, component cells and the SOLIS satellite model

This paper aims at assessing the accuracy of different solar forecasting methods in the case of an insular context. Two sites of La Reunion Island, Le Tampon and Saint-Pierre, are chosen to do the benchmarking exercise. Reunion Island is a tropical island with a complex orography where cloud processes are mainly governed by local dynamics. As a consequence, Reunion Island exhibits numerous micro-climates. The two aforementioned sites are quite representative of the challenging character of solar forecasting in the case of a tropical island with complex orography. Hence, although distant from only 10 km, these two sites exhibit very different sky conditions. This work focuses on day-ahead and intra-day solar forecasting. Day-ahead solar forecasts are provided by the European Center for Medium-Range Weather Forecast (ECMWF). This organization maintains and runs the Numerical Weather Prediction (NWP) model named Integrated Forecast System (IFS). In this work, post-processing techniques are applied to refine the output of the IFS model for day-ahead forecasting. Statistical models like a recursive linear model or a nonlinear model such as an artificial neural network are used to produce the intra-day solar forecasts. It is shown that a combination of the IFS model and the neural network model further improves the accuracy of the forecasts.