构建电力系统状态矩阵的并行计算研究

发布时间：2018-04-07 17:14

本文选题：状态矩阵　切入点：Hager算法　出处：《郑州大学》2017年硕士论文

【摘要】：随着电网的互联,区域电网功率的相互支援提高了系统的经济性、安全性以及可靠性,然而低频振荡情况在系统中影响的范围也越来越大。目前分析小干扰稳定的基本方法是李雅普诺夫线性化处理方法。电网规模越大,状态矩阵维数越高,构建状态矩阵所需要的时间就越长。对于系统遭受的各种小干扰,如何能够快速、准确地判断出系统的稳定性是我们需要研究的问题。本文通过分析状态矩阵的形成过程,加入Hager缩减矩阵的外形法和OpenMP并行计算技术,来提高构建状态矩阵的计算速度。首先,通过分析系统状态矩阵的构建过程,发现在构建状态矩阵的过程中耗时较长的是矩阵求逆和矩阵相乘的运算,特别是大系统高阶数的状态矩阵,运算量大,内存占用多,计算速度缓慢。为了提升矩阵的计算速度,本文引入Hager算法和OpenMP并行处理技术。Hager算法是一种缩减矩阵外形的算法,该算法的基本思想是利用行列对换逐渐缩减矩阵的外形,从而减小矩阵存储及运算次数,提高计算速度。并行计算已在潮流计算和暂态稳定方面得到应用,通过分析状态矩阵的构建过程也能利用并行计算加快形成速度,OpenMP并行处理技术具备易于移植、串行等价以及能够增量并行等优点。然后,在Hager算法预处理后对状态矩阵构建过程进行并行算法设计,同时分析该方法并行计算的性能。最后,用两个实际电网的算例验证了在构建电力系统状态矩阵时加入Hager缩减矩阵外形法和OpenMP并行处理技术的方法有效提高了状态矩阵形成的速度。并行程序是在操作系统Windows7,Intel Parallel Studio XE 2011集成的Visual Studio 2010平台下,编译器Visual Fortran Compiler XE 12.0编译执行的。算例运行结果表明:该方法有效缩减了构建状态矩阵的时间,是切实可行的好方法。
[Abstract]:With the interconnection of the power grid, the mutual support of the regional power network improves the economy, security and reliability of the system. However, the influence of the low frequency oscillation in the system is increasing.At present, the basic method to analyze the small disturbance stability is the Lyapunov linearization method.The larger the scale of power grid, the higher the dimension of state matrix and the longer the time required to construct state matrix.How to judge the stability of the system quickly and accurately is a problem we need to study.In this paper, by analyzing the forming process of state matrix and adding the shape method of Hager reduced matrix and OpenMP parallel computing technology, the calculation speed of state matrix construction is improved.First of all, by analyzing the process of constructing the state matrix of the system, it is found that in the process of constructing the state matrix, the computation of matrix inverse and matrix multiplication is time-consuming, especially the state matrix of large scale system with high order number, which has a large amount of computation and a large amount of memory.The speed of calculation is slow.In order to improve the computation speed of matrix, this paper introduces Hager algorithm and OpenMP parallel processing technology. Hager algorithm is an algorithm of reducing matrix shape. The basic idea of this algorithm is to use column and column exchange to gradually reduce the shape of matrix.In order to reduce the number of matrix storage and operation, improve the speed of calculation.Parallel computing has been applied in power flow calculation and transient stability. By analyzing the construction process of state matrix, parallel computing can also be used to accelerate the formation speed. OpenMP parallel processing technology is easy to transplant.Serial equivalence and the ability to incrementally parallel, and so on.Then, after preprocessing the Hager algorithm, the state matrix construction process is designed in parallel algorithm, and the performance of the parallel computing method is analyzed at the same time.Finally, two practical power grid examples are given to verify that the addition of Hager reduction matrix contour method and OpenMP parallel processing technology to power system state matrix can effectively improve the speed of state matrix formation.The parallel program is compiled and executed by the compiler Visual Fortran Compiler XE 12.0 under the integrated Visual Studio 2010 platform of Windows 7 / Intel Parallel Studio XE 2011.The simulation results show that this method can effectively reduce the time of constructing state matrix and is a feasible and good method.
【学位授予单位】：郑州大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM74

【相似文献】