One Year Submillisecond Fast Solar Database: Collection, Investigation, and Application

Big data is driving future renewable energy research and, in particular, has implications for fast dynamic control of power electronics. This paper discusses the collection of a full year of rooftop 5 kHz solar data. The database includes open-circuit voltage, short-circuit current, and maximum power information. Data reduction involved remapping of missing intervals, correction of unsynchronized metering, and regularization to go from more than eighteen months of raw data to a consistent full-year database. The results are made available at the full sampling rate. Data analysis shows how much downsampling can be employed without loss of useful information about variability. The downsampled databases are also made available. Given prior work that typically limits sampling rates to 1 Hz or takes fast samples over short intervals, the databases provided here support much more comprehensive examination of photovoltaic energy variability.