Abstract
The interest in methodologies for the estimation of the performance loss rate (PLR) using data from PV systems arises with the aging of an increasing amount of PV installations at different climates. These methodologies must be simple, i.e. involve few quantities as input. Furthermore, the associated uncertainty must be kept low in order to be a reliable tool for the verification of module manufacturers’ warranties. This study compares the PLR of seven PV systems based on different technologies using four different methodologies that are presented and discussed. Results are compared on the basis of the PLR and the respective regression uncertainty. We find that, especially for technologies with a strong seasonal response, the error of the estimated PLR can be significantly reduced using seasonal decomposition methods such as a simple moving average. Other more physical approaches like the PVUSA method to reduce seasonality reduce the error as well, however remaining oscillations that are not captured by the physical description enlarge the error.