Characterization and calibration of solar radiation sensors (A)


There are two principal types of instruments that measure the incident solar radiation, the pyrheliometer to measure the direct normal irradiance (DNI) and the pyranometer to measure the global horizontal irradiance (GHI) and the diffuse horizontal irradiance (DfHI). There are many manufacturers of each type of instrument and there are several novel instruments that measure all three components at once. In addition, there are instruments that measure the circumsolar irradiance (scattered light surrounding the solar disk) and spectroradiometers that measure the spectral distribution of incident DNI, GHI, DfHI, and ground reflected irradiance. Spectroradiometers and other specialized instruments such as ultraviolet sensors and pyrgeometers that measure long-wavelength radiation will not be discussed here. For a more thorough treatment of instruments that measure solar radiation see (Vignola et al., 2019).

Discussion will start with pyrheliometers that measure DNI irradiance, followed by a discussion of instruments being tested that measure the circumsolar irradiance and then pyranometers are reviewed. Multifaceted pyranometers that produce all three irradiance values will then be covered. Not covered are spectroradiometers that monitor the spectral distribution on the incident radiation. Instead of measuring the total or broadband incident solar radiation, spectroradiometers measure the spectral composition of the incident solar radiation.

In the case of measurements characterization, the relevant features for calibration sensors are defined at each calibration standard. Thus, in this case the relevant features for the characterization and calibration of solar radiation sensors at each project stage are the need of one specific type of characterization service.

(A) Prefeasibility

On site measurements are not required for prefeasibility analysis. This analysis relies main on satellite-derived irradiance values or previously available measured data.

TABLE 1. LOOK UP TABLE OF CATEGORIE AND USESGrey cells in Table 1 are later described. Cells width N.A. legend, means that no relevant applications are showed later.
Proyect developers needs at diferent plant stages Other users
Category of product service (A) Pre-feasibility (B) Feasibility & Design (C) Due Diligence Financing (D) Plant Accep-tance Test (E) Systems or Plant Operations (F) Grid operators (G) Policy makers (H) Education / Outreach

Calibration of pyrheliometer

Pyrheliometers consist of a tube with an aperture at one end of the tube and the sensor is at the other end of the tube. They have baffles to limit the field of view to the solar disk and a small area around the disk and are similar in appearance to small telescopes. Common pyrheliometers have a field of view that range from about 5.0° to 5.7° degrees with a 5.0° field of view now becoming standard.

There are three types of pyrheliometers, and their performance will be compared and contrasted.

  • Absolute Cavity Radiometers:
  • Thermopile-based pyrheliometers:
  • Photodiode-based pyrheliometer:

Calibration of a pyrheliometer is done by comparing the output of the instrument under clear stable atmospheric conditions against the output of an absolute cavity radiometer that has its calibration traceable to the World Radiometric Reference (WRR). Alternatively, the instrument can be calibrated against another pyrheliometer of the same type that has its calibration traceable to the WRR (ASTM E816-05, 2015; ISO 9847:1992, 1992).

² The term beam normal irradiance (BNI) is being used more frequently to avoid confusion with direct and diffuse, especially when the components are on horizontal or tilted surface

The uncertainties determined for the instruments in this section are obtained using the Guide to the Expression of Uncertainty in Measurements (JCGM, 2008). Pyrheliometers are classified into types to help differentiate the uncertainties in irradiance measurements from these instruments.

Calibration of field pyranometers

There are three distinct methods for calibrating pyranometers:

  • Shade/unshade calibration: The most accurate method is the shade/unshade method. The DNI irradiance is calculated by subtracting the unshaded measurement from the shaded measurement and projecting this difference equal to the angle of incidence.
  • Summation method: The second method is to calculate the reference GHI from the combination of DNI projected onto a horizontal surface plus the measured DfHI irradiance (ASTM, 2015; ISO 9847:1992, 1992).
  • Side by side calibration: The third method is method is a side by side comparison with a reference pyranometer that has its calibration traceable to the WRR (ISO 9847:1992, 1992).

The percent uncertainties given for pyranometer are for a limited range of incident angles, typically from 30° to 60°. Therefore, under clear skies when the solar zenith angle is greater than 60° or less than 30° systematic biases in the data can be greater than the stated uncertainty. This is particularly true during the winter months when the latitude is 40° or more from the equator (Wilcox & Myers, 2008). During the middle of winter at these high latitudes, the SZA is always larger than 63° and this is outside the range specified for the pyranometer uncertainty (see Figures 6.a and 6.b).

Calibration of RSBR

Rotating Shadowband Irradiometer (RSI) and SPN1 Sunshine Pyranometers are two instruments that determine all three irradiance component, GHI, DfHI, and DNI irradiance, simultaneously and are often used in regions with where little maintenance is available. The generic name for these instruments is a rotating shadowband irradiometer (RSI), however some use the term Rotating Shadow Band Radiometer (RSBR) because irradiometers typically refer to instruments measuring infrared radiation. A brief description of these instruments is given along with a discussion of the strengths and weaknesses of each type.

There are two types of RSI instruments. One uses a thermopile-based pyranometer and the other uses a photodiode-based pyranometer. In general, the RSBR measures GHI and then periodically a shadow band rotates in front of the pyranometer to measure the DfHI irradiance. The direct horizontal irradiance (DrHI) is calculated by subtracting the DfHI from the GHI and the DNI is determined by projecting the DrHI onto the normal surface.

Both instruments require adjustment algorithms to remove systematic effects associated with the problems of the pyranometer used to measure the irradiance (Vignola et al., 2016; Vuilleumier et al., 2017).

Several methods have been used to calibrate RSI instruments. Calibration methods have to consider the adjustment methodology used to remove the systematic biases associated with the instrument. A detailed description on the different methodologies can be found in (Jessen et al., 2016).

The algorithms used to remove the bias from the RSI measurements are different for each calibration methodology. The RSBR instruments are evolving as are the algorithms used to remove the systematic biases.

  1. Deviation from true cosine response, also called Lambertian response. The flux of energy incident on a surface is proportional to the intensity of the energy from a given direction multiplied by the cosine of the angle of incidence. This is called Lambert Law. In practice when one measures the intensity of light on the thermopile, the measurement intensity closely follows Lambert’s Law, but the match is not perfect. There can be many causes for this deviation from a true cosine response. These sources of deviation can range from the level of the sensor to the effects of the light passing through the glass dome(s). All pyranometers exhibit some deviation from the Lambert Law, especially at large incident angles.
  1. Thermal offsets

Thermal offsets are caused by radiative losses from the black disc to the sky with the radiation first heating the glass dome(s) before being radiated to the sky. The amount of thermal offset is dependent on the difference between the black disc and the sky temperature. When the sky is cloudy, the sky temperature is close to ambient temperature and the radiative offset is small. When the sky is clear, the sky temperature is typically much less than ambient temperature and the thermal offset can range from a few watt hours/m2 per hour to 10 or even 20 Whr/m2 per hour. Pyranometers that are on tilted surfaces exhibit smaller thermal offsets because the instrument ‘sees’ some ground that is close to ambient temperature.

  1. Dust or frost on the optics

Pyranometers need to be cleaned periodically otherwise dust can build up on the optics and this reduces the light reaching the sensor. When a pyranometer is mounted in a ventilator, buildup of dust, ice, or snow is greatly reduced. This helps maintain the quality of the measurements because dust or other contaminates reduce the amount of light reaching the sensor. Proper cleaning and maintenance of the instruments is important for a quality dataset.

  1. Change in responsivity over time

Thermopile-based pyranometer have a responsivity that degrades with time. The absorptivity of the black coating on the sensor disk degrades with exposure to UV radiation. The degradation of the responsivity can vary between 0.5% and 1.0% per year (see Figs. 5.a and 5.b). Therefore, it is important to calibrate the instruments every year or two. Since the calibration result typically has an uncertainty larger than the degradation rate, it is useful to record the change in responsivity and determine the trend to get a better idea of the calibration value. If the instrument is mounted in a ventilator, then the calibration should also be done in a ventilator because the ventilator affects the thermal offset, and this affects the calibration result.

  1. Spectral response:

Thermopile pyranometers respond to a wide range of wavelengths and usually have a glass dome that allows wavelengths up to 2500 nm to pass through without hindrance. Some pyranometers have special domes that have transmissivity that is approximately constant up to 4000 nm.

  1. Non-linearity

A perfect pyranometer would have a uniform responsivity over all intensities of light. For the better pyranometers, this responsivity doesn’t vary more than 0.2% to 0.5% from the responsivity at 500 W/m2 over a range from 100 W/m2 to 1000 W/m2.

  1. Temperature dependence

Thermopile-based pyranometers measure the temperature difference between the body of the pyranometer and the central disk glued to the thermopile. The body of the pyranometer has a large thermal mass (heat sink) that is minimally affected by the ambient temperature. The current from the thermopile is a measure of the heat flow from the sensor to the heat sink. However, there are other paths through which heat can flow such as by radiation and convection and these alternate paths reduce measured irradiance and the characteristics of the thermopile result in a temperature dependence of the measured irradiance. Factory specification for the pyranometers usually come with a plot of responsivity verses temperature and it is possible to adjust the measured value based on the ambient temperature measurement, but this is rarely done. However, by recording the ambient temperature along with the GHI or DfHI measurements, it is always possible at some future date to adjust the measured values for the effect of temperature. The temperature adjustment is also affected if the pyranometer is mounted in a ventilator that keeps air flowing over the dome to keep dust and/or frost from settling on the pyranometer dome. Sometimes ventilators warm the air and this affects the thermal offset and the temperature dependence of the instrument. For example, pyranometers mounted in ventilators can have smaller thermal offsets, especially if the ventilator uses a high speed DC fan (Dutton et al., 2001).