Combination of Transformed-means Clustering and Neural Networks for Short-Term Solar Radiation Forecasting

In order to provide an efficient conversion and utilization of solar power, solar radiation datashould be measured continuously and accurately over the long-term period. However, the measurement ofsolar radiation is not available to all countries in the world due to some technical and fiscal limitations. Hence,several studies were proposed in the literature to find mathematical and physical models to estimate andforecast the amount of solar radiation such as stochastic prediction models based on time series methods. Thispaper proposes a hybridization framework, considering clustering, pre-processing, and training steps for shorttermsolar radiation forecasting. The proposed method is a combination of a novel data clustering method,time-series analysis, and multilayer perceptron neural network (MLPNN). The proposed Transformed-Means clustering method is based on inverse data transformation and K-means algorithm that presents moreaccurate clustering results when compared to the K-Means algorithm; its improved version and also otherpopular clustering algorithms. The performance of the proposed Transformed-Means is evaluated usingseveral types of datasets and compared with different variants of K-means algorithm. The proposed methodclusters the input solar radiation time-series data into an appropriate number of sub-datasets which are thenpreprocessed by the time-series analysis. The preprocessed time-series data provide the input for the trainingstage where MLPNN is used to forecast the solar radiation. Solar time-series data with different solar radiationcharacteristics are also used to determine the accuracy and the processing speed of the developed forecastingmethod with the proposed Transformed-Means and other clustering techniques.