微型电脑应用

为了提高规划方案经济性，提出基于多源数据融合及深度学习的含分布式电源配电网两阶段规划设计方法。标准化处理含分布式电源配电网多源数据，采用主成分分析的DS(Dempster-Shafter)方法构建多源数据融合模型。在分布式电源接入容量约束条件下，将分布式电源运营商的利润最大化作为目标，构建阶段一规划模型。以最低网损和电压偏差总和为目标，考虑智能软开关约束和分布式电源消纳约束，构建阶段二规划模型。采用长短期记忆网络（LSTM）求解规划模型，确定分布式电源的最优接入节点和容量，实现含分布式电源配电网的优化运行。实验结果表明，所提出的方法可实现多源数据融合，规划方案的综合利润可达到最大值，并可通过智能软开关有效规划分布式电源消纳功率。

To improve the economy efficiency of planning schemes,a two-stage planning and design method for distribution network with distributed power supply based on multi-source data fusion and deep learning is proposed.Standardize the multisource data of distribution network with distributed power supply,and construct a multi-source data fusion model using the Dempster-Shafter(DS)method of principal component analysis.Under the constraint of access capacity of distributed power supply,a stage-one planning model is constructed with the goal of maximizing the profit of distributed power supply operators.A stage-two planning model is constructed with the goal of the minimum network loss and the sum of voltage deviation,considering the constraints of intelligent soft switching and distributed power consumption.Long and short-term memory(LSTM)network is used to solve the planning model,and the optimal access node and capacity of distributed power supply are determined,so as to realize the optimal operation of distribution network with distributed power supply.The experimental results show that the proposed method can realize multi-source data fusion,the comprehensive profit of the planning scheme can reach maximun,and the distributed power supply can be effectively planned through intelligent soft switching.