摘要
由于不同因素对新能源场站集中式功率的影响不同,导致预测效果难以得到保障,为此,提出了基于组合神经网络的新能源场站集中式功率预测研究。利用肯德尔等级相关系数(KRCC)、皮尔逊相关系数(PCC)、斯皮尔曼等级相关系数(SRCC)以及最大互信息系数(MIC)筛选功率影响因素后,构建了卷积神经网络(CNN)和长短期记忆(LSTM)网络融合互轭模型,利用其对功率影响因素对应数据进行综合分析,实现功率预测。测试结果表明,所提方法的平均绝对误差始终稳定在0.1以内,与对照组相比,在稳定性和精确性方面均表现出显著优势。
Due to different factors affecting the centralized power of new energy stations,it is difficult to guarantee the prediction effect.Therefore,a research on centralized power prediction of new energy stations based on composite neural networks is proposed.After using Kendall Rank Correlation Coefficient(KRCC),Pearson Correlation Coefficient(PCC),Spearman Rank Correlation Coefficient(SRCC),and Maximal Information Coefficient(MIC)to screen power influencing factors,a Convolutional Neural Network(CNN)and Long Short Term Memory(LSTM)Network fusion yoke model was constructed to compre⁃hensively analyze the corresponding data of power influencing factors and achieve power prediction.The test results indicate that the average absolute error corresponding to the design method remains stable within 0.1,showing significant advantages in stability and accuracy compared to the control group.
作者
潘熙
PAN Xi(Marketing Service Center(Metering Center),State Grid Jiangsu Electric Power Co.,Ltd.,Nanjing 210000,China)
出处
《电子设计工程》
2025年第23期77-81,88,共6页
Electronic Design Engineering
关键词
组合神经网络
新能源场站
集中式
功率预测
功率影响因素
融合互轭模型
composite neural networks
new energy stations
centralized
power prediction
power inf⁃luencing factors
fusion yoke model
