As the name implies, the physical limits approach to solar data quality assessment compares measured data with estimated or defined limits. For instance, is the radiation component within the range of zero to the maximum possible expected value? Is the direct normal irradiance greater than zero and less than the extraterrestrial value? Is the global horizontal no greater than the vertical component of the extraterrestrial beam? Is the diffuse irradiance more than the expected Rayleigh diffuse sky? Note that while somewhat crude, and allowing the possibility of one or more components to pass such tests even when bad (such as direct beam of zero on a clear day, because the tracker was not pointing correctly, or an attenuated DNI due to a dirty pyrheliometer window), the last example above did lead to the identification of the thermal offset problem described in Sect. 5.
However, for the most part, the physical limits tests cannot provide the level of accuracy needed to assure that measurement instrumentation is indeed functioning properly, unless used with intensive human interaction. This interaction includes regular cleaning of windows, checking of tracker and shading disk alignment, etc. Other intensive human operations involve daily inspection of diurnal irradiance profiles, combined with knowledge of meteorological conditions on the day being inspected. Clearly, these are difficult procedures to automate and implement. They also involve a good deal of qualitative, rather than quantitative evaluation. Thus, the quality assessment may be more subjective than objective.