基于DTW-FCBF-LSTM模型的超短期风速预测  

Ultra-short term wind speed prediction based on DTW-FCBF-LSTM model

作　　者：董治强 Dong Zhiqiang(School of Energy and Electrical Engineering,Hohai University,Nanjing 211100,China)

机构地区：[1]河海大学能源与电气学院,南京211100

出　　处：《电测与仪表》2020年第4期93-98,共6页Electrical Measurement & Instrumentation

摘　　要：为了实时调整电网调度计划、提高电网消纳风电的能力,提出了一种基于动态时间规整(DTW)进行相似数据分析、快速相关过滤方法(FCBF)进行输入属性特征选择、以及基于长短期记忆神经网络(LSTM)的超短期风速预测方法。利用DTW方法筛选出与待预测数据相似性高的训练样本;运用FCBF算法得到优选的输入特征集;构建LSTM模型进行超短期风速预测。以风电场实测数据为算例,将文中方法与现有算法的预测精度进行了对比,验证了所提方法的有效性和先进性。In order to adjust the power grid dispatching plan in real time and improve the ability of absorbing wind power,a new ultra-short term wind speed prediction method is put forward based on dynamic time warping(DTW)-fast correlation-based filter(FCBF)and long-short term memory neural network(LSTM).Firstly,high similarity of historical wind speed data is analyzed by using DTW method,and high similarity data to the predicted day are selected.Secondly,the FCBF method is used to select the optimal input feature set.Finally,the LSTM model is applied to predict the ultra-short term wind speed.The measured data of wind farm is taken as an example and the effectiveness and superiority of the proposed method is verified by comparing the prediction error with existing methods.

关 键 词：风速预测 人工智能 动态时间规整 快速相关过滤式算法 长短期记忆神经网络 

分 类 号：TM933[电气工程—电力电子与电力传动]