CAJ | 학술논문

风电功率预测方法分为两类，即直接预测法与功率曲线转换法。因风电功率具有混沌特性，故将混沌时间序列的相关理论引入到风速和风电功率预测中。鉴于预测精度在很大程度上取决于模型参数的选择，为此先用C-C法联合优化了重构相空间的参数，再用径向基RBF神经网络模型直接预测风电功率，或者由该模型得到风速预测值后，根据对应的风电机组功率特性曲线而推算出风电功率预测值。实例分析结果表明：所提出的两种方法均有较高的预测精度，其中基于混沌径向基RBF神经网络的风电功率直接预测法效果更优。
풍전공솔예측방법분위량류，즉직접예측법여공솔곡선전환법。인풍전공솔구유혼돈특성，고장혼돈시간서렬적상관이론인입도풍속화풍전공솔예측중。감우예측정도재흔대정도상취결우모형삼수적선택，위차선용C-C법연합우화료중구상공간적삼수，재용경향기RBF신경망락모형직접예측풍전공솔，혹자유해모형득도풍속예측치후，근거대응적풍전궤조공솔특성곡선이추산출풍전공솔예측치。실례분석결과표명：소제출적량충방법균유교고적예측정도，기중기우혼돈경향기RBF신경망락적풍전공솔직접예측법효과경우。
There were two kinds of method to predict the wind power, which were the direct forecasting method and the power curve conversion forecasting method. The wind speed and the wind power were predicted with the theories related to the chaotic time series because the wind power was chaotic. First the parameters of the phase-space reconstruction were optimized by C-C method because the accuracy of the prediction largely depended on the parameters used; then the wind power was predicted by the RBF neural network. The predicted wind power could also be achieved based on the curve of wind turbines after the wind speed was predicted by the RBF neural network. The analysis of the example shows that both of the methods have good performances in accuracy and the direct way based on the chaotic RBF neural network is better than the other one.