Accurate short-term solar forecasting is challenging due to weather uncertainties associated with cloud movements. Typically, a solar station comprises a single prediction model irrespective of time and cloud condition, which often results in suboptimal performance. In the proposed model, different categories of cloud movement are discovered using K-medoid clustering. To ensure broader variation in cloud movements, neighboring stations were also used that were selected using a dynamic time warping (DTW)-based similarity score. Next, cluster-specific models were constructed. At the prediction time, the current weather condition is first matched with the different weather groups found through clustering, and a cluster-specific model is subsequently chosen. As a result, multiple models are dynamically used for a particular day and solar station, which improves performance over a single site-specific model. The proposed model achieved 19.74% and 59% less normalized root mean square error (NRMSE) and mean rank compared to the benchmarks, respectively, and was validated for nine solar stations across two regions and three climatic zones of India.

The current research objective is to explore changes in solar radiation variables across the Indian region during the lock-down period. Global Horizontal Irradiance (GHI), Direct Normal Irradiance (DNI) and Diffuse Horizontal Irradiance (DHI) were considered in order to determine changes in radiation. The reduction in DHI and the increase in DNI have been investigated. The radiation levels of GHI, DHI and DNI were compared to the previous month/year reference data for industrial and rural locations and observed the percentage difference in solar radiation during the lock-down period. Overall, the results of lock-down are evident in DHI and DNI industrial areas relative to rural/remote locations. Significantly, the DNI values raised from 1% to 12% in urban areas compared to previous years.