Q-learning算法优化的多种LSTM的超短期风电功率预测  

Q-learning Algorithm-Optimized Multi-LSTM Ultra-Short-Term Wind Power Forecasting

作　　者：辛鹏[1] 李超然 张勋 刘培瑞 袁成磊 XIN Peng;LI Chaoran;ZHANG Xun;LI Peirui;YUAN Chenglei(School of Information and Control Engineering,Jilin Institute of Chemical Technology,Jilin City 132022,China)

机构地区：[1]吉林化工学院信息与控制工程学院,吉林吉林132022

出　　处：《吉林化工学院学报》2024年第9期1-8,共8页Journal of Jilin Institute of Chemical Technology

基　　金：吉林省自然科学基金(YDZJ202301ZYTS271)。

摘　　要：针对风电功率预测中存在的特征选择困难和单一模型不稳定问题,提出了一种融合Q-learning算法的多种LSTM网络(Q_L-L-C-A)的超短期风电功率预测方法。该方法利用最大信息系数(MIC)对风电数据进行特征筛选,采用变分模态分解(VMD)将风电场功率数据分解为多个频率模态作为额外特征,将筛选和分解后的数据作为模型输入,进行LSTM、CNN-LSTM、Attention-LSTM这3种网络模型预测。在此基础上,依据Q-learning算法对3种模型的预测结果进行动态权重分配,以获得更优的组合预测结果。为了验证所提出Q_L-L-C-A模型的预测效果,以某风电场实测数据为模型输入,与6种模型进行对比实验,实验结果表明:所提出的Q_L-L-C-A模型的均方根误差和平均绝对百分误差结果均优于LSTM、CNN-LSTM、Atten-LSTM等模型,Q_L-L-C-A模型在超短期风电功率预测中具有更高的准确性和稳定性。To address the challenges of feature selection and the instability of single models in wind power forecasting,a novel ultra-short-term wind power forecasting method that integrates multiple LSTM networks with the Q-learning algorithm(Q_L-L-C-A)was proposed.This method employs the Maximum Information Coefficient(MIC)to filter the features of wind power data and uses Variational Mode Decomposition(VMD)to decompose the wind farm power data into multiple frequency modes as additional features.The filtered and decomposed data were then used as inputs for the LSTM,CNN-LSTM,and Attention-LSTM network models to make predictions.Based on this,the Q-learning algorithm was applied to dynamically allocate weights to the prediction results of the three models to achieve a more optimized combined forecasting outcome.To validate the forecasting performance of the proposed Q_L-L-C-A model,actual measurement data from a wind farm was used as the model input,and comparative experiments were conducted with six other models.The experimental results indicate that the Root Mean Square Error(RMSE)and Mean Absolute Percentage Error(MAPE)of the Q_L-L-C-A model are superior to those of the LSTM,CNN-LSTM,and Attention-LSTM models,demonstrating that the Q_L-L-C-A model exhibits higher accuracy and stability in ultra-short-term wind power forecasting.

关 键 词：功率预测 组合模型 Q-learning算法 深度学习 最大信息系数 变分模态分解 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]