文档名:基于EMDBiLSTMANFIS的负荷区间预测
摘要:考虑到新型电力负荷随机性增强,传统的准确预测方法已无法满足要求,提出一种EMD-BiLSTM-ANFIS(EmpiricalModeDecomposition-Bi-directionalLongShort-TermMemory-AdaptiveNetwork-basedFuzzyInferenceSystem)分位数预测负荷概率密度的方法,使用负荷预测区间取代点预测的准确数值,能为电力系统分析与决策提供更多数据,增强预测的可靠性.首先将原始负荷序列通过EMD(EmpiricalModeDecomposition)分解成若干分量,并通过计算样本熵分为3类分量.然后将重构后的3类分量与由相关性筛选的外界因素特征采用BiLSTM、ANFIS模型进行训练和分位数回归(QR:QuantileRegression),并将分量的预测区间结果累加得到最终负荷的预测区间.最后利用核密度估计输出任意时刻用户负荷概率密度预测结果.通过与CNN-BiLSTM(ConvolutionalNeuralNetwork-BidirectionalLongShort-TermMemory)、LSTM(LongShort-TermMemory)模型对比点预测及区间预测结果,证明了该方法的有效性.
Abstract:Consideringthattherandomnessofthenewpowerloadisenhanced,thetraditionalaccurateforecastingmethodscannotmeettherequirements,anEMD-BiLSTM-ANFIS(EmpiricalModeDecompositionBi-directionalLongShortTermMemoryAdaptiveNetworkisproposedbasedFuzzyInferenceSystem)quantilemethodtopredicttheloadprobabilitydensity.Itreplacestheaccuratevalueofpointpredictionwiththeloadpredictioninterval,whichcanprovidemoredataforpowerSystemanalysisanddecision-making,Thereliabilityofpredictionisenhanced.First,theoriginalloadsequenceisdecomposedintoseveralcomponentsbyEMD,andthendividedintothreetypesofcomponentsbycalculatingthesampleentropy.Then,thereconstructedthreetypesofcomponentsandthecharacteristicsofexternalfactorsscreenedbycorrelation.AndtheyareusedtogetherwiththeBilstmandANFISmodelsforpredictiontrainingandQR(QuantileRegression),andaccumulatetheresultsofthepredictionintervalofthecomponentstoobtainthepredictionintervalofthefinalload.Finally,thekerneldensityestimationisusedtooutputtheuserloadprobabilitydensitypredictionresultsatanytime.ThevalidityofthismethodisprovedbycomparingthepointpredictionandintervalpredictionresultswithCNN-BiLSTM(ConvolutionalNeuralNetwork-BidirectionalLongShort-TermMemory)andLSTM(LongShort-TermMemory)models.
作者:李宏玉 彭康 宋来鑫 李桐壮Author:LIHongyu PENGKang SONGLaixin LITongzhuang
作者单位:东北石油大学电气信息工程学院,黑龙江大庆163318
刊名:吉林大学学报(信息科学版)
Journal:JournalofJilinUniversity(InformationScienceEdition)
年,卷(期):2024, 42(1)
分类号:TP183TM715
关键词:经验模态分解 双向长短期神经网络 模糊推理系统 分位数回归 概率密度预测
Keywords:empiricalmodedecomposition twowaylongandshorttermneuralnetwork fuzzyinferencesystem quantileregression probabilitydensityprediction
机标分类号:TM715TV737TP391
在线出版日期:2024年4月3日
基金项目:基于EMD-BiLSTM-ANFIS的负荷区间预测[
期刊论文] 吉林大学学报(信息科学版)--2024, 42(1)李宏玉 彭康 宋来鑫 李桐壮考虑到新型电力负荷随机性增强,传统的准确预测方法已无法满足要求,提出一种EMD-BiLSTM-ANFIS(EmpiricalModeDecomposition-Bi-directionalLongShort-TermMemory-AdaptiveNetwork-basedFuzzyInferenceSystem)分...参考文献和引证文献
参考文献
引证文献
本文读者也读过
相似文献
相关博文
基于EMD-BiLSTM-ANFIS的负荷区间预测 Load Interval Forecast Based on EMD-BiLSTM-ANFIS
基于EMD-BiLSTM-ANFIS的负荷区间预测.pdf
- 文件大小:
- 5.55 MB
- 下载次数:
- 60
-
高速下载
|
|