主动配电网可靠性评估源荷模型改进及并行处理

随着主动配电网节点数及分布式电源种类与数量的增多,基于蒙特卡洛模拟法的主动配电网可靠性评估结果愈加依赖于分布式电源与负荷的可靠性模型,且计算复杂度会急剧增加。针对此问题,文中首先建立了考虑更多随机因素的光伏发电系统与风电机组随机输出模型、计及负荷反弹效应的多状态负荷模型,并就主动配电网协调控制优化特性建立了用于系统运行状态判定的储能—负荷协调控制模型。然后,为降低上述模型引入与时序蒙特卡洛模拟法对计算耗时的影响,提出和设计了结合多机与多核并行的可靠性评估并行处理方案,建立了并行计算任务分配环节的数学模型与一维搜索求解算法。最后,通过算例验证了基于并行计算的主动配电网可靠性评估方法的有效性与优越性,在极大减少计算时间的同时,能与串行计算持有一致的收敛速度与评估精度。

Source-load Model Improvement and Parallel Computing for Reliability Evaluation of Active Distributed Networks

Along with the increasing active distributed network(ADN)nodes, distributed generator(DG)types and quantities, the reliability evaluation results of ADN based on Monte Carlo simulation are more and more dependent on the reliability models of DG and load. And the computational complexity has also increased dramatically. To solve these problems, this paper proposes random output models for photovoltaic system and wind turbines that involve more random factors. Furthermore, a multi-state load model considering the load rebound effect is developed. According to the optimization characteristics of ADN coordinated control, an energy storage and load coordinated control model for judging the system operating status is also developed. In order to reduce the influence of the above reliability model and the sequential Monte Carlo simulation approach on the computing time, a reliability evaluation method based on multi-machine and multi-core parallel computing is designed. And a mathematical model of task allocation in parallel computing and one-dimensional search algorithm are proposed for further enhancement. Finally, case simulation results show that the reliability evaluation method based on parallel computing is highly effective and superior. While greatly reducing the computing time, the reliability evaluation method proposed maintains a convergence speed and evaluation accuracy consistent with that of serial computing. This work is supported by National Natural Science Foundation of China(No. 51207051)and Fundamental Research Funds for the Central Universities(No. 2016XS02).