基于多因素的短期风电功率组合预测研究  

Research on Short-term Wind Power Combination Prediction Based on Multiple Factors

作　　者：屈伯阳 付立思[2] QU Bo-yang;FU Li-si(School of Electrical Engineering,Shenyang University of Technology,Shenyang 110870,China;College of Information and Electrical Engineering,Shenyang Agricultural University,Shenyang 110866,China)

机构地区：[1]沈阳工业大学电气工程学院,辽宁沈阳110870 [2]沈阳农业大学信息与电气工程学院,辽宁沈阳110866

出　　处：《水电能源科学》2024年第10期216-220,共5页Water Resources and Power

基　　金：兴辽英才计划(XLYC2008005)。

摘　　要：为了提高风电功率模型的预测精度,采用卷积神经网络(CNN)、双向长短期记忆(BI-LSTM)和注意力机制(AM)组合预测模型。首先,考虑气象因素(不同高度的风速与风向、温度、湿度、气压)全面反映当时天气条件对风电预测精度的影响;然后在气象因素基础上,探究历史风电功率时间序列数据与变分模态分解(WMD)信号作为特征量进行预测建模;最后,仅考虑时间序列数据与VMD信号作为特征量进行深度学习预测建模,发现基于气象因素和VMD组合信号作为特征输入的CNN-BI-LSTM-AM模型预测的准确率达到97.66%,而仅考虑VMD组合信号作为CNN-BI-LSTM-AM模型输入的预测准确率达到97.71%,在风电场功率预测的精度和稳定性方面均取得令人满意的结果。In order to improve the prediction accuracy of wind power model,this paper adopts a combined prediction model consisting of Convolutional Neural Networks(CNN),Bidirectional Long Short Term Memory(BI-LSTM),Network and Attention Mechanism(AM).Firstly,consider meteorological factors(wind speed and direction at different heights,temperature,humidity,and pressure)was considered to comprehensively reflect the impact of weather conditions on the accuracy of wind power prediction at that time.Then,based on meteorological factors,explore historical wind power time series data and Variational Mode Decomposition(VMD)signals were taken as feature variables for predictive modeling.Finally,only time series data and VMD signals were considered as feature variables for deep learning prediction modeling.It was found that the prediction accuracy of the CNN-BI-LSTM-AM model based on meteorological factors and VMD combined signals as feature inputs reached 97.66%,while the prediction accuracy of the CNN-BI-LSTM-AM model only considering VMD combined signals as input reached 97.71%.Satisfactory results were achieved in the accuracy and stability of wind farm power prediction.

关 键 词：风电预测 双向长短期记忆 变分模式分解 卷积神经网络 注意力机制 

分 类 号：TM614[电气工程—电力系统及自动化]