计及柔性负荷的高比例风光渗透下配电网孤岛划分策略

分布式电源的渗透率逐步提高，为配电网的供电恢复带来了新的方案，利用分布式电源进行孤岛划分可以在停电故障后迅速恢复供电以减小损失。然而在故障后孤岛划分时分布式电源的出力通常采用预测数据，由于风电和光伏等间歇电源的出力具有不确定性，为了确保能稳定、可靠的供电，因此有必要考虑风电和光伏作电源时的备用需求。首先建立考虑风电和光伏不确定性的预测误差模型，采用机会约束规划处理模型中的随机变量并得到各孤岛的备用需求。在孤岛划分阶段，首先得到最大供电范围，通过粒子群遗传混合算法充分发挥柔性负荷的调控能力，使得得到的孤岛结果最优。最后，以修改后的IEEE33节点配电系统为例，对所提出的模型和方法进行了验证。

Islanding strategy for distribution network with high proportion of wind/photovoltaic penetration considering flexible load

The rapid development of distributed generation has brought new solutions for the distribution systems restoration. By using distributed generation can quickly restore power supply after power failure and reduce the loss. However, when the island is divided after a failure, the output of the distributed generation usually adopts the forecast data. Since the output of the intermittent power sources such as wind power and photovoltaics power is uncertain, in order to ensure stable and reliable power supply, it is necessary to consider reserve demands of wind power and photovoltaic power. Firstly, a prediction error model considering the uncertainties of wind power and photovoltaic was established, and using chance-constrained programming to deal with the random variables in the model. In the islanding partition stage, getting the maximum power supply range firstly. Particle swarm genetic hybrid algorithm is used to give full play to the control ability of the flexible load, so that the island result is optimal. Finally, the proposed restoration strategy is verified by a modified IEEE 33-bus test system.