1.广东电网有限责任公司广州黄埔供电局 广东 广州 510700;2.广东电网有限责任公司广州黄埔供电局
1.Guangzhou Huangpu Power Supply Bureau of Guangdong Power Grid Co.,Ltd,Guangdong,Guangzhou,510700;2.Guangzhou Huangpu Power Supply Bureau of Guangdong Power Grid Co., Ltd;3.Guangzhou Huangpu Power Supply Bureau of Guangdong Power Grid Co., Ltd,
随着分布式电源、储能及智能环网柜等技术在配电网中的快速发展,配电网的有源化和智能化为其孤岛运行提供了有力支撑。面向上级电网故障后有源配电网(Active distribution network, ADN)的孤岛划分问题,建立了负荷停电综合损失最小的鲁棒优化划分模型。其中,停电综合损失计及了切除可中断负荷的补偿成本和非协议停电所带来的负荷损失成本。损失成本采用分段线性化建模,并通过添加逻辑变量构建统一的线性约束式。为避免孤岛运行状态下任一负荷的频繁投退,提出了负荷停电连续状态约束及其线性化方法。进一步考虑到ADN中分布式能源出力的不确定性,将确定性模型转为两阶段鲁棒优化模型,引入鲁棒调节系数灵活调整方案的保守性,并基于对偶理论和列生成算法迭代实现模型求解。算例仿真验证了所提模型和方法的有效性。
With the rapid development of distributed power, energy storage and automation control technologies in the distribution network, the active and intelligent distribution network provides strong support for its island operation. Facing the problem of islanding of active distribution network (ADN) after a fault in the upper-level power grid, a robust optimal partition model with the least comprehensive loss of load outage is established. Among them, the comprehensive loss of power outage takes into account the compensation cost of cutting off the interruptible load and the load loss cost caused by non-agreement blackout. The loss cost is modeled by piecewise linearization, and a unified linear constraint is constructed by adding logical variables. In order to avoid the frequent switching on and off of any load in the island operation scheme, the continuous state constraint of load power failure and its linearization method are proposed. Further, considering the uncertainty of the output characteristics of distributed energy in ADN, the deterministic model is transformed into a two-stage robust optimization model, and the conservativeness of the flexible adjustment scheme of robust adjustment coefficients is introduced. At the same time, the model is solved based on duality theory and column generation algorithm. Cases studies verify the effectiveness of the proposed model and method.