基于随机矩阵和历史场景匹配的配电网无功优化

配电网的无功优化是保证配电网供电可靠、经济运行的一项重要工作。将大数据理论引入配电网无功优化，提出一种基于随机矩阵的无功优化方法，它不依赖于配电网的模型和参数，直接利用配电网在运行过程中产生的运行大数据以及当地的环境数据构造7种高维随机矩阵，提取57种特征指标，再应用主成分分析法对这些特征指标进行降维处理，然后匹配历史数据库中已有的场景，快速找到特征指标与当前系统最相近的场景，直接采用匹配场景的控制策略作为当前系统的无功优化控制策略，以减小有功网损和节点电压偏移。最后，在改造的IEEE-37节点配电网仿真模型上进行算例验证，其中增加了光伏/风电等分布式发电和电动汽车充电站等随机负荷模型。结果表明，方法可有效地对配电网进行无功优化，摆脱了配电网模型的限制，可快速做出控制决策。

Reactive Power Optimization of Distribution Network Based on Random Matrix and Historical Scenario Matching

Reactive power optimization and voltage management of distribution networks is very important for the optimal operation of distribution networks. By introducing big data theory into reactive power optimization of distribution networks, a reactive power optimization method is proposed based on random matrix and historical scenario matching, which does not need the model and parameters of the distribution network, and can directly use the data generated during the operation of the distribution network to construct seven high-dimensional random matrices and extract 57 characteristic indicators. The extracted characteristic indicators are reduced in dimension to match the existing scenarios in the historical database, and the scenarios closest to the statistical characteristics of the current system are found. The control strategy under the matching scenarios is adopted as the reactive power optimization control strategy in the current period to reduce the active power loss and node voltage deviation. Finally, the method is verified on the modified IEEE-37 node distribution network model, where the model of random loads, including the distributed generations such as photovoltaic/wind power, and electric vehicles are added. The results show that the proposed method can effectively optimize the reactive power of the distribution network without the need of its model and parameters, and the online decision-making speed is fast.