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

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.