Solar radiation forecast based on fuzzy logic and neural networks

This paper presents a solar radiation forecast technique based on fuzzy and neural networks, which aims to achieve a good accuracy at different weather conditions. The accuracy of forecasted solar radiation will affect the power output forecast of grid-connected photovoltaic systems which is important for power system operation and planning. The future sky conditions and temperature information is obtained from National Environment Agency (NEA) and the sky and temperature information will be classified as different fuzzy sets based on fuzzy rules. By using fuzzy logic and neural network together, the forecast results can follow the real values very well under different sky and temperature conditions. The effectiveness of the approach is validated by a case study where four different scenarios are tested. The Mean Absolute Percentage Error (MAPE) is much smaller compared with that of the other solar radiation method. •This paper presents a solar radiation forecast technique based on fuzzy logic and neural network.•The proposed technique can tell the difference of the solar radiations between the different sky conditions.•The forecast results can follow the real values very well under different weather conditions.•Compared with the other solar radiation method, the Mean Absolute Percentage Error (MAPE) is much smaller.