Affiliations: [a] Energy and Power Engineering College, Changsha University of Science and Technology, Changsha, Hunan, China
| [b] State Key Laboratory of Alternate Electrical Power system with Renewable energy Sources, North China Electric Power University, Beijing, China
| [c] School of Electronic Information and Electrical Engineering, Changsha University, Changsha, Hunan, P. R. China
Abstract: In order to overcome the serious errors of wind farm load abnormal fluctuation forecasting results caused by traditional forecasting methods, a wind farm load abnormal fluctuation forecasting method based on probabilistic neural network is proposed in this paper. The probabilistic density is screened out by probabilistic neural network, and the maximum posterior probability density neuron is used as the output to realize wind farm load forecasting. According to the prediction results, a comprehensive severity subordinate function is constructed based on fuzzy reasoning to classify the severity of wind farm anomalies. According to the fuzzy operation rules, the abnormal fluctuation of wind farm load can be warned. The experimental results show that the operation error of the proposed method is only 0.49, the accuracy of early warning is high, and the effective fitting index is up to 0.95, which shows that the proposed method has high practical application value.
