A method based on the multi-task asynchronous processing is proposed to quickly and accurately identify the vulnerable lines of power grid.The protection vulnerability is defined with the comprehensive consideration of the cooperative correlation degree and the colligated consequence severity of line protections.The power-flow betweenness with the comprehensive consideration of the topological structure and the operating mode of power system is adopted to estimate the vulnerability of line structure.A line vulnerability index combining the protection vulnerability and the power-flow betweenness is thus given.The multi-task asynchronous processing mode is applied to divide the calculation of index into several task modules to be dealt with concurrently.Results of case calculation show that,with better identification performance and high identification accuracy,the proposed method shortens the required time of vulnerable line identification effectively,suitable for online application.

Since current researches have not fully utilized the information of comprehensive edge betweenness obtained by GN(Girvan-Newman) algorithm to achieve more feasible partitioning schemes,an automatic electromagnetic loop network partitioning algorithm based on the GN algorithm improved with the backtracking concept is proposed.The GN algorithm is perfected according to the partitioning principles of electromagnetic loop network and the open-loop cut set obtained by GN algorithm based on the weak line set of power grid is backtracked to extend the diversity of partitioning scheme.The verified schemes are evaluated by the principal component analysis to determine the optimal one.The proposed algorithm is applied to a 500 kV/200 kV electromagnetic loop network and results show that,it is automatically partitioned while the practical partitioning principles are satisfied.

The configuration scheme for distributed generation has an important effect on the reliability and economy of distribution network.Aiming at this problem, a mathematical model is founded in this paper, whose target is the minimum active power loss.Besides normal constraints such as the node voltage, the node current and the capacity, the maximum short-current level limit is also taken into account, which is a key to ensure that the protection device dose not present in the form of malfunction.Meanwhile, in order to solve the slow convergence and local optimization problem of genetic algorithms, the adaptive genetic algorithm (AGA) is used to optimize the location and sizing of the distributed generation.The necessity for considering relay protection constraint as well as the better convergence and adaptation for AGA are eventually verified by the simulation of IEEE33 bus system example.

近年来,多核处理及多线程处理技术发展迅速,充分利用CPU资源,采用异步处理技术是大幅度提高电力系统网络拓扑分析计算效率的有效途径.本文提出了基于多任务异步处理的序网络拓扑分析方法,将电气元件抽象为节点和支路类型,建立连接关系类保存连接信息,异步化处理正序、零序网拓扑分析,采用任务嵌套的模式将耗时较长的节点阻抗矩阵的计算异步化为四个子任务独立执行,最终生成拓扑网络图及分析结果数据.针对四川某电网的实例拓扑分析表明了该方法的可行性与高效性.

光电传感器是将光信号转换为电信号的传感器,在检测和控制系统中得到广泛应用。本文从几种光电传感器的基本原理入手,研究了各种元件的光电转换原理,通过matlab仿真软件拟合曲线分别分析了其伏安特性和光照特性,并与理论曲线进行对比后作出误差分析。最后通过比较各种光电元件的线性性能,表明了适合用于线性检测的光电元件类型。