In this work, the impact of component reliability on large scale photovoltaic (PV) systems' performance is demonstrated. The analysis is largely based on an extensive field-derived dataset of failure rates of operation ranging from three to five years, derived from different large-scale PV systems. Major system components, such as transformers, are also included, which are shown to have a significant impact on the overall energy lost due to failures. A Fault Tree Analysis (FTA) is used to estimate the impact on reliability and availability for two inverter configurations. A Failure Mode and Effects Analysis (FMEA) is employed to rank failures in different subsystems with regards to occurrence and severity. Estimation of energy losses (EL) is realised based on actual failure probabilities. It is found that the key contributions to reduced energy yield are the extended repair periods of the transformer and the inverter. The very small number of transformer issues (less than 1%) causes disproportionate EL due to the long lead times for a replacement device. Transformer and inverter issues account for about 2/3 of total EL in large scale PV systems (LSPVSs). An optimised monitoring strategy is proposed in order to reduce repair times for the transformer and its contribution to EL.

The uncertainty analysis of irradiance and temperature measurements in relation to the energy yield

No abstract

The uncertainty analysis of irradiance and temperature measurements in relation to the energy yield prediction of the photovoltaic (PV) modules are presented. A Monte Carlo simulation approach is demonstrated separately to propagate the monthly and annual measurement uncertainties of irradiance and temperature to annual energy yield prediction uncertainty for two commercially available PV modules. The annual irradiation uncertainty as measured with a thermopile pyranometer is calculated as ±1.56%. Uncertainty of the annual average of ambient temperature measurement is calculated as ±0.08 O C. Finally, the uncertainty in the energy yield estimation of the PV devices is determined as 2.8% and 15.5% for crystalline silicon (c-Si) and copper indium gallium (di)selenide (CIGS) modules, respectively.