In august 2017 an article was published by the National Renewable Energy Laboratory (NREL). The article was 
written by I. Reda and A. Andreas. This study discusses the component summation method for pyranometer calibration. For this method, the pyranometer is placed under a clear sky with the measured reference global solar irradiance consisting of two components. The ‘subtended beam’ solar irradiance is measured by a pyrheliometer traceable to the World Radiometric Reference. The other component, known as the diffused horizontal solar irradiance, does not have an internationally recognized reference. The study proposes a method for calibrating pyranometers consistently for measuring the diffuse component of solar irradiance.
Pyranometer calibration method
The NREL has developed a method to automate the calibration of radiometers called, “the Broadband Outdoor Radiometer Calibration”, or ‘BORCAL’ process. The BORCAL process uses the summation technique, where the responsivity of a pyranometer is described by the following expression:
With:
V = Thermopile output voltage. [µV]
= Infrared Net responsivity. [µV/(W/m2)]
= Infrared net irradiance measured by a pyrgeometer nearby. [W/m2]
N = Direct beam irradiance. [W/m2]
= Solar zenith angle. [°]
D = Diffuse irradiance. [W/m2]
As mentioned before, there is no standard for diffuse irradiance (D) measurement. The study has provided a guide of 23 steps in which they explain how they calculate the diffuse irradiance by using multiple sets of data.
Using the Hukseflux SR25 pyranometer, calibrated according to their proposed diffuse irradiance reference, an uncertainty of 0,89% was found with respect to values of the World Radiometric Reference. As, under clear skies, the ratio of the diffuse irradiance to the global irradiance is 1:10, the error in the reference global irradiance is in the order of 0,09%. This uncertainty is more than adequate for the purpose of calibrating unshaded pyranometers.
Pyranometer calibration evaluation
The Hukseflux SR25 pyranometers will be calibrated every year with the proposed diffuse reference. The resulting responsivity from each calibration will be recorded in order to evaluate the stability of the diffuse reference over time.
Original Article:
“Calibration Procedure of a modified Hukseflux SR25 as an example to establish the diffuse reference for the outdoor broadband radiometer calibration.”