摘要
介绍了基于模糊聚类的风速短期组合预测方法。以模糊聚类技术进行训练样本的选择,可以大大提高训练样本的相似度。在时间序列模型、多元线性回归模型、灰色模型、神经网络模型基础上,根据风电场的风速(及气象)特性优化组合模型权重,得到适合本风电场的组合预测模型,应用实例表明该方法具有广泛的自适应性,应用范围更广,效果更佳。
The method of short term combined prediction of wind speed based on fuzzy clustering was presented.The prediction samples set were classified through fuzzy clustering analysis,which greatly improved the similarity of prediction samples.Based on the time-series models,multiple linear regression model,gray model,neural network model,the model weight of combined prediction was optimized according to wind speed characteristics of the wind electric field,and a suitable combined prediction model was obtained.Application examples show that this method has a wide range of self-adaptability,a broader range of applications and better prediction results.
出处
《华东电力》
北大核心
2010年第2期295-299,共5页
East China Electric Power
基金
国家科技部863项目(2007AA05Z458)
关键词
模糊聚类
样本选择
组合预测
自适应性
fuzzy clustering
sample selection
combination Forecast
self-adaptability
