Adaptive neuro-fuzzy inference system for DC power forecasting for grid-connected PV system in Sharjah

Solar energy forecasting is essential to maintain PV system’s performance in uncertain environmental conditions. Factors such as module temperature, ambient temperature, solar irradiance, and wind speed contribute to the DC current generated by PV systems. In this study, an adaptive neuro-fuzzy inference system (ANFIS) is developed on MATLAB to study a 2.88 kW grid-connected PV system in the harsh weather conditions of Sharjah. Solar irradiance, ambient temperature, module temperature, and wind speed are considered as the input membership functions in the developed ANFIS model. The output parameter considered in this study is the current DC generation, which critically depends on the defined membership functions. The accuracy of the model was determined based on the comparison with the experimental dataset. The R2 value has shown that the proposed model can forecast the DC current with minimal error, with a value of 99.12% and 99.13% for training and testing, respectively. Moreover, the spatial 3-D surface has shown that the optimum DC current generation is achieved at the highest solar irradiance and ambient temperature while minimizing the module temperature for enhanced electrical efficiency.