基于相似度与神经网络的协同短期负荷预测模型

来源 :华东电力 | 被引量 : 0次 | 上传用户:laohu_you
下载到本地 , 更方便阅读
声明 : 本文档内容版权归属内容提供方 , 如果您对本文有版权争议 , 可与客服联系进行内容授权或下架
论文部分内容阅读
为了考虑除负荷本身外的其他因素对短期负荷的影响,提出了基于相似度与神经网络的短期协同预测模型.该模型首先通过计算负荷曲线的相似度对历史数据进行排序,然后选择与预测时刻相似度较相近的数据对未来时刻的负荷利用相似度进行预测,对于出现的误差,通过神经网络结合其他因素进行预测纠正.实验结果证明,该协同预测模型较之单纯的BP神经网络预测模型具有较高的预测精度.
其他文献
“女王诉杜德利与斯蒂芬案(Her Majesty The Queen v.Tom Dudley and Edwin Stephens)”作为世界法理学教材中比较著名的十大案例之一,历来引得各国法律学者的极大关注和热烈
利用Orlicz空间和LBaM空间中的范数关系,将Stancu-Kantorovich算子在Orlicz空间中的结果推广到LBaM空间,得到该算子逼近阶的一种估计.
Nowadays high intensity proton accelerators are extensively applied,and this paper gives particular emphasis on CYCIAE-100,a 100 MeV high intensity compact cycl
The frequency domain electromagnetic method has already been widely used for tomographic imaging or electromagnetic well logging.However,different from open hol
We completely characterize commutativity of Sρ and Sψ on L2α (Dn)⊥ for bounded pluriharmonic symbols ρ and ψ on Dn, and prove that SρSψ=Sψρ if and onl
目的:分析新疆刺山柑叶子部位挥发油成分.方法:采用水蒸气蒸馏法提取刺山柑叶子挥发油,通过气相色谱-质谱(GC-MS)联用技术进行挥发油成分分析.结果:检出50个色谱峰,鉴定出34
The Isochronous Mass Spectrometry (IMS) developed at GSI is a very efficient method for direct mass measurements of short-lived nuclides. By taking a recent IMS
Properties of the ground states and the excited states of N=127 isotones are investigated by using the nonlinear relativistic mean field theory with the interac
Using a new phenomenological (K)N interaction which reproduces A(1405) as an I = 0 bound state of (K)N, we have investigated K- -3 He(T = 0) and K- -4 He(T = 1/
We adopt the groove guide as the cavity of the undulator to reduce the diffraction effect.The groove guide has the advantages of lower surface energy loss,large