A simple big data methodology and analysis of the specific yield of all PV power plants in a power system over a long time period

In this work, we have reviewed recent literature concerning the performance of PV plants in different power systems and detected that, usually, the raw data used is not complete in terms of the number of PV plants and also main parameters of poor quality need to be removed. Then, for the first time, a study of the specific yield of all PV plants in a power system with single-PV-plant resolution is presented. Thus, we have analyzed the official 237,588 monthly energy values obtained from 2005 to 2017 for the 1523 PV plants in the Canary Islands. This dataset is obtained from PV plants ranging from 0.53 kW(p) up to 9 MWp, and it has been supplied by the distribution system operator and main utility (ENDESA). Then, this dataset is compared to 153,120 irradiance and temperature data from a PVGIS database. Results show that the Spanish regulation has a direct effect not only on the development of the PV capacity in the Canary Islands, but also on the specific yields. Moreover, only combining meteorological data (irradiance, temperature and wind speed) from satellites, starting year of operation, and nameplate capacity we have developed a very simple theoretical model to predict the specific yield of a PV plant at any location in the Canary Islands, avoiding the requirement of any data from the owners of the PV plants. The simulation values obtained have been validated with the real specific yields for PV plants assumed to be well managed (multi-MW power plants placed in best locations) showing errors below a 3%. This theoretical model has also been used for detecting suboptimal PV plant designs and anomalous specific yield of PV plants above the clear sky limit. Recommendations to avoid anomalous specific yields in future are included.