摘要
准确的风电场功率预测对评估风电场运行状态、风电场运营维护以及保证电力系统的安全稳定可靠运行具有重要意义。为提高风电场短期功率预测精度,提出了一种基于小波变换和长短期记忆(Long Short-Term Memory,LSTM)神经网络的风电场短期功率预测方法。首先,运用小波理论对功率数据进行多分辨率分解以详细分析多频域尺度层次上的信息,把分解后的子序列分别重构,使得子序列尺度与原序列均一致。然后,分别对子序列建立LSTM神经网络预测模型,最终预测结果为包含不同信息的所有子模型的叠加值,结合某风场的实际运行数据进行多时间尺度的功率预测。实验证明,所提方法相比于LSTM和ELMAN预测方法具有更高的预测精度,验证了所提方法的有效性。
Accurate wind farm power forecasting is of great significance for assessing wind farm operating conditions,wind farm operation and maintenance,and ensuring safe,stable and reliable operation of power systems.In order to improve the wind power prediction accuracy,this paper proposes a short-term power prediction of wind farms based on the wavelet decomposition and long short-term memory(LSTM)neural network.First of all,the wavelet theory was applied to conduct the multi-resolution decomposition of wind power series.For analyzing the information at multiple frequency domain scales in details,the decomposed subsequences were reconstructed separately,so that the subsequence scale was consistent with the original sequence.Then,a prediction model for the power at each layer was established with the LSTM neural network,and the prediction results of each sub-model were superimposed as the final predicted value.Finally,the power prediction was performed on multiple time scales based on the actual operational data of a wind farm.Experimental results show that the proposed method generates higher forecasting accuracy than the LSTM or ELMAN method and verify the feasibility of the proposed method.
作者
林涛
赵参参
赵成林
刘航鹏
LIN Tao;ZHAO Shen-shen;ZHAO Cheng-lin;LIU Hang-peng(School of Artificial Intelligence and Data Science,Hebei University of Technology,Tianjin 300130,China)
出处
《计算机仿真》
北大核心
2020年第6期81-84,共4页
Computer Simulation
基金
河北省科技计划项目(17214304D)。
关键词
功率预测
小波分解
长短期记忆
多时间尺度
Wind power forecasting
Wavelet decomposition
Long short-term memory
Multiple time scales
作者简介
林涛(1970-),男(汉族),河北保定市人,博士,教授,主要研究领域为物联网技术、风电系统;赵参参(1992-),女(汉族),河南省周口市人,硕士研究生,主要研究领域为风电功率预测、性能评估;赵成林(1994-),男(汉族),天津市人,硕士研究生,主要研究领域为风电故障诊断、状态检测;刘航鹏(1994-),男(汉族),河北省衡水市人,硕士研究生,主要研究领域为风电功率预测。
