基于泛在电力物联网下双向长短时记忆的线损预测计算方法

Line Loss Prediction Method Based on Bidirectional Long Short Time Memory in Ubiquitous Power Internet of Things

随着风、光等可再生能源发电及柔性负荷的并网,使泛在电力物联网的运行调控异常复杂,导致输电线路损耗不断增大,对此提出基于双向长短期记忆的线损预测方法以指导运行调控决策。首先,在电力物联网的全时空量测环境下,提出了B系数法的线损计算模型;其次,以B系数法线损计算模型为基础,建立了泛在电力物联网的线损大数据;第三,基于人工智能中的双向长短时记忆网络方法,以线损数据为输入和约束条件,提出了线损预测计算方法;最后,通过实际电网的仿真验证,表明了所提方法的有效性。

With the integration of wind,solar and other renewable energy generation as well as flexible load,the operation and control of ubiquitous power Internet ofthings is very complex,which leads tothe increase of transmission line loss.A line loss prediction method based on bidirectional long-term memory is proposed in this paper so to guide the operation regulation decision-making.Firstly,the line loss calculation model of the least square B-cofficient method is proposed under the full space-time measurement environment of the power Internet of things.Then,the line loss big data of the ubiquitous power Internet of things is set up based on the line loss calculation model of the least square B-coefficient.After that,based on the bidirectional long-term memory network method in artificial intelligence and taking the line loss data as the input and constraint conditions,a new line loss calculation model is proposed.Finally,the effectiveness of the proposed method is verified by the simulation of the real power network.