摘要： It is important to secure for every solar photovoltaic energy installation the highest-possible average (mean) capacity factor, as well as the lowest-possible standard-deviation, this latter computed with high frequency. High-frequency data of solar photovoltaic energy facilities are, however, very hard to be found. There are no estimations in the literature of the standard deviation of the capacity factor of solar photovoltaic energy installations sampled with high frequency. Here we show the data sampled every 5 minutes for the solar photovoltaic energy facilities connected to the Australian National Electricity Market grid during the year 2018. The average capacity factors are about 27-28% flat panel, and 31-32% tracking flat panel. The standard deviations have larger values at about 37-38%, for coefficients of variation 120-130%. As solar photovoltaic energy facilities only produce electricity during the daylight time, that is predictable, the statistic analysis is also applied to the daylight times only. While average daylight-only capacity factors are about twice the previous values, standard deviations are slightly less, for coefficients of variations more than halved at 55-65%. The analysis highlights the extreme variability of solar energy helping the design of the energy storage needed for making possible an efficient, resilient, renewable energy-only grid. The collected and interpreted information and the provided analysis are filing a gap in the area of renewable energy and energy harvesting technology based on wind and solar photovoltaics where variability has been so far underevaluated.