Frank Vignola

A new operational model for satellite-derived irradiances: description and validation

We present a new simple model capable of exploiting geostationary satellite visible images for the production of site/time-specific global and direct irradiances The new model features new clear sky global and direct irradiance functions, a new cloud-index-to-irradiance index function, and a new global-to-direct-irradiance conversion model. The model can also exploit operationally available snow cover resource data, while deriving local ground specular reflectance characteristics from the stream of incoming satellite data. Validation against 10 US locations representing a wide range of climatic environments indicates that model performance is systematically improved, compared to current visible-channel-based modeling practice.

Determination of atmospheric turbidity from the diffuse-beam broadband irradiance ratio

A semi-physical method is proposed to evaluate turbidity from broadband irradiance measurements and other atmospheric parameters. This method demonstrates the utility of diffuse data when estimating atmospheric composition with broadband irradiance data. An error analysis and various tests against measured data show that this method can predict accurate turbidities provided that the sky is perfectly cloudless and the diffuse irradiance data are very accurate. Yet, this method is insensitive to errors in input data such as precipitable water and ozone amount. Applications of this method to the quality control of radiation data are discussed. Tests with actual data from Florida and Oregon show good agreement with other methods. Evaluation of the model required a detailed discussion of the accuracy and cosine error of pyranometers, and the uncertainty in precipitable water estimates.

Beam-global correlations in the Pacific Northwest

Abstract Daily, 10-day, and monthly averaged beam-global correlations are derived for seven sites in the Pacific Northwest. The beam-global correlations are found to vary with time of year in a manner similar to the seasonal variations exhibited by diffuse-global correlations. A good description of the seasonal dependence is produced by introducing a term dependent on the year-day into the correlations. The correlations are compared with data from several sites in the National Oceanic and Atmospheric Administration (NOAA) network and to results from other studies.

Correlations between diffuse and global insolation for the Pacific Northwest

Abstract Correlations between diffuse and global insolation averaged over 1, 5, 10, 15, and 30 days are presented for six locations in Oregon and Idaho. Data collected over a 2-to 4-yr period was used to obtain the correlations. The statistical accuracy of the correlations for data averaged over 10, 15, and 30 days was improved through the use of moving averages. Linear regression fist appear adequate for all but the daily data. The correlations obtained do not appear to be influence by climatic conditions or altitude. However, a repeating semi-annual pattern is present.

Diffuse-global correlation: Seasonal variations

Abstract Residual differences between measured diffuse fractions and those determined from the best one-parameter ( K T ) empirical correlations are studied for six Pacific Northwest sites. A pronounced variation with time of the year is found. By including a simple sinusoidal term dependent only on the year day in the correlation, the variance is significantly reduced and the seasonal dependence of the residuals is removed. Clear day solar noon transmission values are shown to have seasonal variations similar to the correlations. This leads to the hypothesis that the atmospheric factors contributing to the transmission values (air mass, water vapor, and turbidity) are the same ones responsible for the observed seasonal variation.

Chapter 5 – Bankable Solar-Radiation Datasets

The characteristics of solar-irradiance datasets used to forecast the performance of solar electrical systems and to estimate the ability of the electricality generated to meet scheduled loan repayments are discussed in this chapter. Examples are given of datasets currently available, and their strengths and weaknesses are illustrated. Measured datasets, ground-based measurement models, and satellite-derived models are evaluated. Steps needed to create a sound, bankable solar-radiation database are described. A discussion of the use of such datasets to generate P50, P90, and P99 probabilities of exceedance is presented along with a description of the bootstrap method used to generate a larger dataset for a more robust analysis of these probabilities.

Evaluation of methods to correct for IR loss in Eppley PSP diffuse measurements

The IR loss in diffuse measurements made by thermopile pyranometers is examined. Diffuse measurements are used for the study of IR losses because diffuse irradiance is much smaller than the total irradiance and hence the IR effects can be more clearly seen. Specifically, diffuse measurements of an Eppley PSP pyranometer are compared to those made with a Schenk Star pyranometer. Pyranometers with black and white or star type junctions suffer minimal IR loss because the reference and receiving junctions of the thermopile are at the same thermal level. The difference between diffuse values can be attributed to calibration and cosine response errors as well as IR loss. This is a preliminary study over one month when pyrgeometer data are available. Examination of the differences at various times of the year and at more than one location is necessary to generalize the findings in this report. Several methods of correcting for IR loss are examined. First subtracting out the average nighttime offset during the day is tested. Next an extrapolation between early morning and late evening offsets is tested. This should help eliminate the IR offset in both the morning and evening hours, but underestimate the IR losses during the rest of the day. Next, correlations of IR losses calculated using pyrgeometer measurements with temperature, relative humidity, and irradiance are evaluated. Initial results show that it should be possible to use more commonly available measurements rather than prygeometer data to estimate IR loss for Eppley PSP pyranometers.

Uncertainty of rotating shadowband irradiometers and Si-pyranometers including the spectral irradiance error

Concentrating solar power projects require accurate direct normal irradiance (DNI) data including uncertainty specifications for plant layout and cost calculations. Ground measured data are necessary to obtain the required level of accuracy and are often obtained with Rotating Shadowband Irradiometers (RSI) that use photodiode pyranometers and correction functions to account for systematic effects. The uncertainty of Si-pyranometers has been investigated, but so far basically empirical studies were published or decisive uncertainty influences had to be estimated based on experience in analytical studies. One of the most crucial estimated influences is the spectral irradiance error because Si-photodiode-pyranometers only detect visible and color infrared radiation and have a spectral response that varies strongly within this wavelength interval. Furthermore, analytic studies did not discuss the role of correction functions and the uncertainty introduced by imperfect shading. In order to further improve the...

New data set for validating PV module performance models

A new publicly available data set was completed for use in validating models that estimate the performance of flat-plate photovoltaic (PV) modules. The data were collected for one-year periods at three climatically diverse locations (Cocoa, Florida; Eugene, Oregon; and Golden, Colorado) and for PV modules representing all technologies available in 2010 when the work began. The same makes and models of PV modules were tested at all locations and common data acquisition systems were used with calibrations performed at the National Renewable Energy Laboratory. For use in determining model parameters and coefficients, baseline and post-deployment measurements were performed indoors with solar simulators, including per the requirements of IEC 61853 Part 1: Irradiance and Temperature Performance Measurements and Power Ratings. Outdoors, the PV modules were characterized per the requirements of the Sandia array performance model. A users manual describes the contents of the data set and how to access the data.

Modeling IR Radiative Loss from Eppley PSP Pyranometers

A method has been developed to estimate IR radiative losses using solar radiation and meteorological data without the need for pyrgeometer data. The modeled IR radiative losses are not as accurate as that obtained using pyrgeometer information, but 95% of the modeled IR radiative losses are with a few W/m2 of the actual IR radiative losses. Currently this method is limited to having a least some period when pyrgeometers measurements are available. More testing and evaluations are needed at a number of locations to test the general applicability of the model developed.