基于聚类的超闭球算法短期风速预测研究

Research on Short Term Wind Speed Forecasting of Super Closed Ball Algorithm Based on Clustering

中文关键词:  风速预测  聚类  信度非平均分配

英文关键词:wind speed forecasting  clustering  non average reliability

中文摘要:

      具有较高精度的超短期风速预测有着重要的作用，它对建立和保障并网运行风电场风电功率预测预报系统有着举足轻重的作用。但是，由于风速的影响因素较多，且存在着巨大的波动性、随机性，以及较高的自相关性。这些因素，极大地影响了传统的风速预测方法。因此，探究一种短期风速预测方法是十分必要的，此方法以聚类的小脑超闭球算法为基础，此超闭球方法，对减少数据输入的地址碰撞有着很好的作用，提高了学习速度，另通过模糊聚类对输入数据确定节点数和节点值，提高了学习精度。仿真结果证明基于聚类的小脑超闭球网络相比应用较为成熟的BP神经网络等能很好地预测未来1 h风速。

英文摘要:

      Ultra short term wind speed forecasting with high accuracy is a prerequisite and guarantee for the establishment and operation of wind power prediction and forecasting system for grid connected wind farm.Due to many factors affecting wind speed,there are huge volatility,randomness,and high self-correlation,which brings a great challenge to the traditional wind speed forecasting method.A short-term wind speed forecasting method based on the clustering of the cerebellar hyper sphere algorithm is proposed.The algorithm is based on the fuzzy clustering to determine the node number and node value of the input data to improve the learning accuracy.The simulation results show that the neural network based on the clustering of the cerebellum is more mature than that of BP neural network.