A novel pattern search based approach to solar radiation forecasting

In this thesis, pattern-scanning based method Mycielski predictor which mostly used in signal processing applications such as compression and encryption is used to estimate the solar radiation. Solar radiation data is used as input and output parameters as the predicted for the predictor algorithm. Although there are many studies about solar radiation prediction and Mycielski applications in the literature, solar radiation prediction using Mycielski algorithm is firstly used in this study. Afyonkarahisar and Antalya region 2 and 4 yearlong solar radiation is used as solar radiation data. In this study, there are some developments on Mycielski predictor algorithm to prevent the errors while pattern-scanning or repeating the history. In the first level of the study, there is information about Mycielski method and upgrading of Mycielski in this thesis, in addition to this, in the second level, Markov and Mycielski methods are used hybrid ,n MATLAB software to increase the accuracy of the prediction. The proposed pattern search based forecasting approach have been applied on a known 2D model in the literature and better results have been gotten. In the last level of the study, the obtained solar radiation data is used in PVsyst and PV SOL wihch are solar energy design and planning software to do feasibility of 1 MW solar energy power plant.