摘要
在“双碳”政策下,亟须提出并计算碳流指标。首先基于碳排放流理论,建立了碳流率、支路碳流密度、节点碳势等指标,为后续的碳势预测奠定了基础。其次,提出了一种基于SConv-LSTM-Attention的配电网节点碳势预测模型,该模型采用基于筛选机制的卷积网络(SConv)作为特征提取模块,将长短期记忆(LSTM)神经网络作为时序预测模块,并加入注意力(Attention)机制计算权重,动态选择LSTM神经网络输出,从而增强了预测能力。最后,以我国实际配电网为例,验证了所提方法的可行性和有效性。
Under the“dual carbon”policy,there is an urgent need for the development and calculation methods of carbon flow indicators.First,based on the theory of carbon emission flow,indicators such as carbon flow rate,branch carbon flow density,and node carbon potential are established,laying the foundation for subsequent carbon potential prediction.Second,a prediction model for distribution network node carbon potential based on SConv-LSTM-Attention is proposed.This model employs a convolutional network with a filtering mechanism as the feature extraction module,an LSTM neural network as the time-series prediction module,and incorporates an attention mechanism to calculate the weight and dynamically select LSTM outputs,thereby enhancing predictive ability.Finally,the feasibility and effectiveness of the proposed method are verified using a practical distribution network example from China.
作者
杨冰芳
姜皓喆
刘一童
徐友刚
吴继健
汪龙召
YANG Bingfang;JIANG Haozhe;LIU Yitong;XU Yougang;WU Jijian;WANG Longzhao(State Grid Qingpu Electric Power Supply Company,SMEPC,Shanghai 201700,China;School of Electrical Engineering,Shanghai University of Electric Power,Shanghai 200090,China)
出处
《电力与能源》
2025年第1期20-26,共7页
Power & Energy
作者简介
杨冰芳(1994-),女,工程师,从事变电检修与电气试验方面的工作。
