申洪涛, 岳凡丁, 史轮, 刘林青, 李梦宇, 段子荷, 袁欢. 考虑DG及负荷时序性的多目标配电网重构与DG调控综合优化规划[J]. 现代电力, 2022, 39(2): 182-192. DOI: 10.19725/j.cnki.1007-2322.2021.0039
引用本文: 申洪涛, 岳凡丁, 史轮, 刘林青, 李梦宇, 段子荷, 袁欢. 考虑DG及负荷时序性的多目标配电网重构与DG调控综合优化规划[J]. 现代电力, 2022, 39(2): 182-192. DOI: 10.19725/j.cnki.1007-2322.2021.0039
SHEN Hongtao, YUE Fanding, SHI Lun, LIU Linqing, LI Mengyu, DUAN Zihe, YUAN Huan. Comprehensive Optimal Planning of Multi-objective Distribution Network Reconfiguration and DG Regulation Considering DG and Load Sequence[J]. Modern Electric Power, 2022, 39(2): 182-192. DOI: 10.19725/j.cnki.1007-2322.2021.0039
Citation: SHEN Hongtao, YUE Fanding, SHI Lun, LIU Linqing, LI Mengyu, DUAN Zihe, YUAN Huan. Comprehensive Optimal Planning of Multi-objective Distribution Network Reconfiguration and DG Regulation Considering DG and Load Sequence[J]. Modern Electric Power, 2022, 39(2): 182-192. DOI: 10.19725/j.cnki.1007-2322.2021.0039

考虑DG及负荷时序性的多目标配电网重构与DG调控综合优化规划

Comprehensive Optimal Planning of Multi-objective Distribution Network Reconfiguration and DG Regulation Considering DG and Load Sequence

  • 摘要: 为解决分布式电源(distributed generation,DG)出力及负荷的时变性给实际配电网调度所造成的不利影响,使配电网的优化规划方案更加切实可行,提出了一种基于配电网重构和DG选址定容结合的多目标粒子群动态优化模型,该模型以配电网有功损耗、电压偏差及经济成本为优化目标,考虑负荷及DG出力的时变性,对配电网络重构和DG调度进行综合优化求解。通过基于随机森林模型(random forest,RF)及长短期记忆神经网络(long short-term memory,LSTM)模型的混合预测模型对配电网负荷及DG出力进行预测。采用经帕累托最优理论改进的粒子群算法得到配电网重构及DG调控的帕累托最优解集并利用模糊隶属度函数法来确定帕累托最优解集中的最佳配电网调度方案。基于IEEE 33标准测试系统设计多个算例进行仿真分析,结果表明,所提考虑负荷及DG出力时序性的配电网重构和DG调度联合优化模型可显著改善配电网络运行的经济性和稳定性。

     

    Abstract: To mitigate the adverse effects due to time-variance property of distributed generation (DG) output and load on the dispatching of practical distribution network and to make the optimal planning scheme of distribution network more feasible, a particle swarm-based multi-objective dynamic optimization model, which was based on distribution network reconfiguration and the combination of site selection with capacity determination, was proposed. Taking active loss, voltage deviation and economic cost of distribution network as optimization objectives and considering the time-variance property of loads and DG output, the comprehensive optimization solution of distribution reconfiguration and dispatching of DG were conducted. By means of the hybrid forecasting model based on random forest model and long short-term memory neural network the output of distribution network and its load were predicted. The particle swarm algorithm, which was improved by Paleto optimal theory, was utilized to obtain the Paleto optimal solution set for distribution network reconfiguration and the regulation and control of DG, and by use of fuzzy membership function method the optimal distribution network scheduling scheme within Paleto optimal solution set was determined. Based on IEEE 33 bus standard test system several numerical examples were designed to perform simulation analysis. Simulation results show that using the proposed joint optimization model, in which the time sequence of DG output and the load are considered, both economy and stability of distribution network operation can be evidently improved.

     

/

返回文章
返回