A TAKAGI-SUGENO fuzzy algorithm for solar irradiation forecasting

A new model for predicting daily solar irradiation based on a Takagi-Sugeno fuzzy algorithm is reported. The study was conducted with data measured at the meteorological station of Timisoara, Romania. The model construction and its operation are presented in detail. The proposed model is based on a simple algorithm with two input variables and one output variable. The daily clearness index, calculated from the measured daily global solar irradiation represents the physical quantity processed by the fuzzy algorithm. Basically, the model forecasts the clearness index at time t on the basis of two previous values, measured at time t-1 and t-2. An assessment of the accuracy of the algorithm was performed. Overall results show that the new forecasting model achieves an acceptable accuracy. The simplicity of the model makes it conveient for potential users. The algorithm can be applied successfully in other locations only by re-computing the model coefficients.