Application of Fuzzy Neural Network in Predicting Small Hydropower Short-term Generating Capacity

Abstract

This paper proposes a prediction method on short-term generating capacity of small hydropower projects based on fuzzy neural network, taking into consideration of the implications of factors such as season, climate, rainfall, temperature, and dates influencing small hydropower generation in a short-time. The research subject is regional small hydropower project groups as a collective. It screens out sectional flow of large hydropower projects as a correlation factor. We then conduct analysis on the large and small hydropower projects and test for the significant difference. Based on the regression models for these two, we utilize the predicted value of large hydropower for prediction of the small hydropower’s capacity, thus providing an effective solution to resolve the difficulty in capacity prediction of small hydropower in short term. The application results prove that the method produces satisfactory prediction results and it provides valuable reference to the predictability of short-term small hydropower generating capacity.