CAJ | 학술논문

　　风的演变具有混沌特性，论文以混沌相空间重构理论为基础，应用神经网络技术进行风速短期预测。采用互信息法和饱和关联维数法分别计算最佳的延迟时间和嵌入维数，将原风速时间序列数据进行相空间重构，并用wolf法计算最大Lyapunov指数对时间序列的混沌特性进行判别。论文利用神经网络建立风速预测模型，并进行多步滚动式短期风速预测。论文还进行了仿真实验，实验结果初步验证了基于混沌理论的神经网络预测模型在风速短步长预测方面的效果。
　　풍적연변구유혼돈특성，논문이혼돈상공간중구이론위기출，응용신경망락기술진행풍속단기예측。채용호신식법화포화관련유수법분별계산최가적연지시간화감입유수，장원풍속시간서렬수거진행상공간중구，병용wolf법계산최대Lyapunov지수대시간서렬적혼돈특성진행판별。논문이용신경망락건립풍속예측모형，병진행다보곤동식단기풍속예측。논문환진행료방진실험，실험결과초보험증료기우혼돈이론적신경망락예측모형재풍속단보장예측방면적효과。
This paper is involved in the wind speed forecast model based on chaos theory and artificial neural network. The delay time and the embedding dimension are obtained with adopting the mutual information method and Grassberger-Procaccia algorithm. When the phase space of the original wind speed time series data is reconstructed, the new space can reflect the characteristics of wind speed with which the largest Lyapunov exponent is calculated with the wolf method. So the multi-step rolling forecast model using neural network is gotten in the paper. In the end, the experiment result of the simulation demonstrates that this model is helpful for improving accurate results.