Measurements and Uncertainties of Solar Irradiance

Measurements and Uncertainties of Solar Irradiance

High-quality long-term irradiance measurements are the criteria for judging whether a solar radiation dataset can benefit. However, apart from a handful of observatories in the Pacific Northwest that provide high-quality total and direct solar irradiance data for more than 30 years, there are few long-term high-quality data that can be used to accurately assess the variability of solar resources. Measured irradiance data is available from more than 1,400 observation sites in the U.S. and many observatories around the world, but these data vary widely in quality and duration. Few of these stations have well-preserved irradiance data and associated records. Only a few observatories are used to measure solar radiation, providing data for potential solar power plants. This subsection will discuss the data provided by the observatory containing GHI and DNI measurements, and will focus on the steps necessary to generate the most reliable data.

  1. High quality DNI, GHI and DHI measurements.

The GHI and DNI can be measured by a total pyranometer and a direct pyranometer, respectively, and the DHI measurement can be obtained by using a total pyranometer and a shading plate. The DNI value measured by the absolute cavity radiometer is the most accurate, with an international standard of 95% confidence level and an uncertainty of ±0.3%; and the uncertainty of the absolute cavity instrument after calibration according to the international standard is ±0.4%. However, cavity radiometers are very expensive and are not designed for continuous field measurements.

The uncertainty level for the thermopile pyranometer is 95% and varies from ±0.7% to ±2.0% (depending on the instrument used). If a cavity pyranometer is not available, a total pyranometer that meets the secondary standard with an absolute accuracy exceeding ±2% can be used. For the solar zenith angle less than 70°, a first-class thermopile pyranometer with an absolute accuracy of ±3% can be used. Scattered radiation is best measured with a secondary black and white pyranometer mounted on an automatic sun tracker and protected from direct sunlight. The Solar Radiation Instrumentation is discussed in detail in Vignola et al. (2012).

The above uncertainties represent the accuracy of the measurements during calibration. When operating normally and the equipment is well maintained, the measurement uncertainties for GHI, DNI and DHI are ±5%, ±3% and ±7%, respectively, at the 95% confidence level. In order not to exceed the above uncertainties, it is necessary to: regularly clean the shrouds and windows of the equipment, point the direct pyranometer measuring DNI at the sun; use a visor when measuring DHI; calibrate the instrument regularly. If equipment is regularly serviced and maintained, the quality and accuracy of the data measured by the measuring station can be greatly reduced. The response rate of a total pyranometer is generally reduced by 0.5% to 1.0% per year, so it is recommended that it be field calibrated annually.

High-quality DNI measurements enable a more accurate assessment of system performance. DNI is a solar component that mainly provides energy for the solar power generation system and provides the required energy for the concentrating system.
In the absence of DNI measurements, the DNI component must be obtained from satellite model estimates (15% uncertainty in annual mean RMSE for DNI) or from correlations with GHI values. As mentioned earlier, the errors in the DNI obtained from the correlation can be compensated for with the relative errors in the DHI, and the uncertainty in the total irradiance modeling (south-facing surface) is very close to the uncertainty in the GHI values sex. For DNI, both methods have high average uncertainty and bias. Therefore, the measured DNI value greatly enhances the confidence level in the performance estimation of the modeled system.

The three irradiance components (ie, GHI, DNI, and DHI) are interrelated and can therefore be used to check the accuracy of the data and identify problems with the data. NREL has a software project called SERIQC that can help assess the quality of solar data with two or three components.

  1. Rotating shading belt radiometer

A rotating shading strip radiometer (RSR for short) can also be used to collect the three main irradiance measurements. The instrument includes a pyranometer with a light-shielding strip that can be rotated to pass the pyranometer at regular intervals. Through a series of corrections and correlation analyses, GHI, DNI and DHI can be measured.

Relevant literature shows that, if properly calibrated and maintained, a rotating shading strip radiometer can measure DNI with an uncertainty of ±5% (Myers et al., 2005), and GHI has a similar uncertainty (Stoffel et al. People, 2010; Wells et al., 1992).
The advantage of the rotating shading strip radiometer is that the remote application of the rotating shading strip radiometer is more stable in locations where routine maintenance cannot be performed (Meyers et al., September 2009). Considering the stability and price difference of the rotating shade strip radiometer system, many developers choose this device for evaluating solar resource development sites. Most rotating shading strip radiometers use photodiode-based pyranometers. The most commonly used model is 11-CORL1-200.

Similar to solar cells, photodiodes are sensitive to the spectral distribution of incident solar radiation. GHI and DHI have different spectral distributions during clear sky periods. Since the instrument needs to be calibrated frequently to provide accurate GHI, the recorded DHI must be adjusted to account for the different responsivity due to different GHI and DHI spectral distributions. In some places where the relevant instruments were built and tested, the calibration of the DHI went smoothly. However, the question of whether the correction is applicable to observations with different aerosol concentrations is still under investigation. Aerosols are particles distributed in the atmosphere that affect the spectral distribution of incident solar radiation. Adjustments made to the spectra mainly affect the DHI data and calculated DNI values. The correction factor for DHI may depend on the concentration and composition of aerosols in the air at the above observation points.

The calibration of the RSP instrument is very important because it is impossible to check all three solar components. DNI is calculated from GHI and DHI components. Early RSPs used a factory calibrated LI-COR Li-200 total pyranometer. The error of factory calibration can reach 8%. Therefore, it is very important to carry out a rigorous calibration of the instrument before it is installed in the field, as well as to carry out regular calibration checks.

For example, the RSP data for Corpus Christi, Texas is plotted in Figure 1. The instrument is factory calibrated with Ll-COR, and the observatory has no other calibration records for the instrument. Using climatic aerosol optical depth data, land-based measurements were compared with modeled irradiance under eye-air conditions generated by NREL. During clear sky periods when the clear sky index is high, the data and model should match.
While the eye-air model is only occasionally free of aerosol data input, it cannot be free of data input for the entire year. There is evidence that clear-sky models with high-quality aerosol input data have small errors (about ±2%). Therefore, the RSP data error in Corpus Christi is likely to be about 6% lower. This is the case if the instrument has been calibrated with high quality before or after field application.

Figure 1 - 2000 RSP data for Corpus Christi, Texas
Figure 1 – 2000 RSP data for Corpus Christi, Texas
  1. The importance of maintenance and calibration

The maintenance and calibration of instruments is of great significance in determining the usefulness of ground-based measured data.
If a solar monitoring station is established but not maintained in any way, the uncertainty in the data measured by the station increases significantly. Long-term data can be difficult to find as funding fluctuates over time. However, in order to obtain better monitoring results, it is necessary to maintain vigilance at all times during the monitoring work.

Long-term trends are often difficult to verify because they vary little and must be calibrated over the entire database period. All total pyranometers tend to change over time, so the response rate or calibration of the instrument needs to be monitored and updated appropriately to eliminate instrument-induced trends.

  1. Fusion value of satellite data and ground-based data

While obtaining ground-based irradiance data is difficult, combining parallel satellite data and historical datasets with 1-2 years of data can greatly improve the overall confidence in the solar resource. The uncertainty in high-quality measured data is much smaller than that in model data, and the comparison between measured data and satellite data can be used to determine the size and characteristics of systematic errors or systematic deviations in model data.

  1. Other important meteorological measurements

Auxiliary meteorological measurements are also significant and have an impact on projected system performance. Meteorological measurements of solar systems are usually related to ambient temperature, wind speed, wind direction, precipitation, relative humidity and air pressure. Models used to estimate PV system performance use at least measurements of ambient temperature (2m) and wind speed/direction (3m). This type of information removes system bias, so temperature and wind speed can be incorporated into system performance. For CPV and CSP systems, 10m wind speed measurements can be used to evaluate load loss estimates. Relative humidity affects solar panels slightly, while precipitation affects the accumulation of dirt and dust on solar panels. Barometric pressure measurements can be used for short-term predictions during system operation.

Read more: Calculation and scaling of solar irradiance

Related Posts