The development of a new PV performance laboratory and PV outdoor research facility.
The output of a PV system under field conditions is not constant, but fluctuates throughout the year according to four key variables, the solar irradiance, the module temperature, the solar spectrum and the solar angle. With new infrastructure constructed as part of this project, we demonstrate that careful measurements of these parameters allows the output to be mathematically corrected back to the standard reference conditions used to nameplate the modules in the factory. This will work provided the irradiance is predominantly direct in nature, i.e. for times when the sky was clear.
We find that the error in this correction scales with the proportion of the sunlight that is diffuse in nature at the time of the measurement. This highlights the difficulty in accurately predicting PV output during cloudy periods, or for cloudy climates.
We find that by correcting the output of a PV system back to standard reference conditions we can compare the performance of the system with the rated performance of the installed modules, thus measuring the quality of system design as well as the health of the system. This relatively simple approach can be applied to acceptance testing for new PV systems, as well as output monitoring to ensure a rapid response to any faults during operation.
Since it is not practical to install expensive equipment for measuring the solar spectrum at every possible site, we show that a simple method exists for simulating the above correction based only on measurements of irradiance and temperature. The method is practical and reasonably accurate, with its only shortcoming being a small error arising from the difference between the mean local AM1.5 solar spectrum and the standard AM1.5 reference spectrum. We introduce the term site spectral offset to describe this error.