A Regime-Dependent Neural Network Approach to Short-Range Solar Irradiance Prediction Using Surface Observations and Satellite Data

An accurate short-range irradiance forecasting system is essential for utility companies to maintain reliable energy grids as the delivery of power from solar power increases. We test several cloud-regime dependent short-range solar irradiance forecasting systems (RD-ANN) to make 15-min average clearness index predictions for 15-Min, 60-Min, 120-Min and 180-Min lead-times. The RD-ANN system that shows the lowest forecast error on independent test data classifies cloud regimes with a k-means algorithm based on a combination of surface weather observations, irradiance observations and GOES-East satellite data. Then, this method trains Artificial Neural Networks (ANNs) on each cloud regime to predict the clearness index. This RD-ANN system improves over the mean absolute error of the prediction by the baseline clearness index persistence forecast by 1%, 21.0%, 26.4% and 27.4% at the 15-Min, 60-Min, 120-Min and 180-Min forecast lead times. When a version of this method was configured to predict the irradiance variability, this method showed more accurate irradiance variability predictions than a smart variability persistence technique.