Rapid changes in insolation in southern England due to variation in cloud cover: influence on operation and testing of solar modules

Analysis of insolation data recorded in southern England during June 2009 reveals that on an active day, changes in irradiance (ramp rate) exceeding 1 % per second may occur up to 5 % of the time. This leads to uncertainties in outdoor PV performance figures, since current-voltage scans are taking place under non-constant light levels. We investigate the relationship between fill factor and ramp rate by an equivalent-circuit simulation, which provides an indication of the underlying noise levels to be expected from this source when compiling fill factor trends from raw outdoor test data. We have also computed the insolation amplitude spectrum and sought to assign the spectral features to diurnal, meteorological and stochastic effects. Comparison with experimental measurements suggests that ramps make a minor contribution to the noise level under typical measurement conditions. However the pronounced non-gaussian nature of the statistical frequency distribution of ramp rates suggests this may be responsible for ‘outlier’ data points.