Analyzing the mechanisms behind temporal correlation between power sources using frequency separated time scales: A Swedish case study on PV and wind
Oskar Lindberg, David Lingfors, Johan Arnqvist
Abstract
The temporal correlation between different power generation sources is important for quantifying the reduction in variability when constructing co-located hybrid power parks (HPPs) that combine multiple power sources. This study investigates the physical mechanisms behind correlation on time scales relevant for the power system using frequency separated time scales. The methodology is universally applicable to any data set consisting of at least two power sources and could be adjusted accordingly. The methodology is demonstrated and validated in a case study across Sweden for wind and PV power generation, using the meteorological reanalysis dataset CosmoREA-6. All studied time-scales (seasonal, mid-term, synoptic and diurnal) showed anti-correlated characteristics, although the magnitude of temporal correlation is highly dependent on the time-scale considered. The highest potential for useful anti-correlation is found on the seasonal cycle, followed by the diurnal cycle where existing wind turbine sites, on average, have stronger anti-correlation than the average site. The validation showed good correspondence with measurements for all time-scales. However, an underestimations of the results were found for the diurnal and seasonal cycle while this was shown to have a minor effect when analyzing the correlation on different time scales. The methodology of the case study should be generally valid for all similar climates.