A study on the degradation rates and the linearity of the performance decline of various thin film PV technologies

In most available analyses of the annual degradation rate (R-D, in %/year) of thin film PV (TFPV), light-induced degradation is not separated from long-term degradation and degradation linearity is virtually not explored. A study on the performance loss of various TFPV technologies in Jaen (Spain) is presented with a double goal: (a) providing more reliable values of R-D by avoiding the approaches above and (b) assessing the suitability of some performance metrics and statistical techniques intended to estimate R-D, depending on PV degradation linearity. Thus, monthly time series were constructed with PR DC, temperature-corrected PR DC and experimental peak power for two grid-connected TFPV technologies -a-Si and a-Si/mu c-Si- and four TFPV modules-a-Si,a-Si/mu c-Si CdTe and CIGS- tested outdoors over a five-and-a-half-year and a six-and-a-half-year period, respectively. Classical seasonal decomposition (CSD) and Year-on-Year (YOY) statistical techniques were subsequently employed. The confidence intervals of R-D found for the a-Si and a-Si/mu c-Si PV fields overlap for CSD and YOY over the five-and-a-half-year measurement period and its sub-periods analysed, with R-D similar to 1.3%/year. This holds for the three performance metrics under scrutiny, a fact that is attributable to the closely linear degradation experienced by both PV fields. Mainly due to nonlinearities, CSD applied to any of the three performance metrics for the four TFPV modules fails to provide reliable values of R-D. However, the values of R-D obtained by applying YOY to such performance metrics for the aforementioned four TFPV modules over the five and-a-half-year measurement period and its sub-periods studied are within the uncertainty range.