Solar power output has an intermittent character due to atmospheric effects such as insolation variability resulting from cloud movements. Variations also occur due to diurnal patterns and physical shading. Accurate solar forecasting is therefore a highly beneficial information source for the successful integration of large amounts of intermittent solar generation, especially in stand-alone microgrids. Accurate solar forecasting enables variability to be managed and enables optimal matching of solar power supply and electrical demand in applications employing energy storage or load control.

This presentation discusses cloud motion forecasting research performed by CSIRO where a prototype system has been developed that incorporates algorithms for sun position tracking, cloud identification, classification and edge detection, cloud movement tracking, distortion correction, motion projection and a simple scheme for predicting solar irradiance. The discussion also includes the deployment of the prototype system in a remote hybrid solar-diesel microgrid in Australia for optimal control of the power system to enable a higher penetration of solar power.

Speaker