改进的粒子群算法及其在电力系统无功优化中的应用

发布时间：2018-06-25 23:09

本文选题：无功优化 + 粒子群算法　；参考：《东北大学》2014年硕士论文

【摘要】：电力系统无功优化是典型的非线性优化问题,具有多约束、多变量、离散性等特点,在利用目前的优化方法求解时存在很多问题,例如难以找到全局最优、求解时间长、求解精度低等。因此我们有必要研究新的方法或者新的改进策略。以粒子群算法为代表的群智能算法是求解无功优化这类复杂非线性优化问题的有力工具,但粒子群算法是一种相对新的优化技术,在理论分析和应用研究等方面还处于初级阶段,有很多问题值得研究,例如如何提升算法跳出局部最优解的能力,如何提升算法求解高维度、复杂、多峰问题的精度和速度。本文首先针对标准粒子群算法容易陷入局部最优解的缺陷进行改进,提出了两种改进策略——包含精英权重的全面学习策略和锦标赛邻域拓扑改进策略,并将两种改进策略与标准粒子群算法相结合,形成改进的粒子群算法并将其运用到了电力系统无功优化领域。然后本文进一步研究大规模复杂电力网络无功优化的模型及求解方法,在改进粒子群算法的基础上提出了基于协同进化理论框架的大规模复杂电力网络无功优化的求解方法——分布协作粒子群算法,在电网分区的基础之上,以一种分布协作的方式处理复杂系统的无功优化问题。相比传统的集中式的优化计算方式,该方法不仅能够降低问题的复杂程度,加快算法的计算速度,而且能处理各分区无功调节设备的协调配合问题,有利于实现电力系统综合无功电压控制。最后,本文将理论研究成果应用到某市分散式风电项目之中。在对项目调研的基础上,设计了一套综合无功电压控制系统,并利用分布协作粒子群算法完成其中的无功优化部分。通过对比引入综合无功电压控制系统前后风电场无功电压调节设备的动作次数、网损大小及优化计算的时间等关键指标,可以看出本文设计的基于分布协作粒子群算法的综合无功电压控制系统相比传统各风电场单独优化控制和集中式的优化计算方式有很大的优越性。
[Abstract]:Reactive power optimization in power system is a typical nonlinear optimization problem, which has the characteristics of multi-constraint, multi-variable, discreteness and so on. There are many problems in solving the problem by using the current optimization methods, such as the difficulty to find the global optimum and the long time to solve the problem. The accuracy of the solution is low. Therefore, we need to study new methods or new improvement strategies. The swarm intelligence algorithm, represented by particle swarm optimization, is a powerful tool for solving complex nonlinear optimization problems such as reactive power optimization, but particle swarm optimization is a relatively new optimization technique. The theoretical analysis and application research are still in the primary stage, there are many problems worth studying, such as how to improve the ability of the algorithm to jump out of the local optimal solution, how to improve the accuracy and speed of the algorithm to solve high-dimensional, complex and multi-peak problems. In this paper, we first improve the standard particle swarm optimization (PSO) algorithm because it is easy to fall into the local optimal solution, and propose two improved strategies: a comprehensive learning strategy with elite weights and a tournament neighborhood topology improvement strategy. Two improved strategies are combined with the standard particle swarm optimization algorithm to form the improved particle swarm optimization algorithm and apply it to the field of reactive power optimization in power system. Then this paper further studies the reactive power optimization model and solution method of large-scale complex power network. Based on the improved particle swarm optimization (PSO) algorithm, a new method of reactive power optimization for large-scale complex power networks based on co-evolution theory is proposed, which is distributed cooperative particle swarm optimization (DCPSO), which is based on the partition of power network. Reactive power optimization of complex systems is dealt with in a distributed and cooperative way. Compared with the traditional centralized optimization method, this method can not only reduce the complexity of the problem and speed up the calculation of the algorithm, but also can deal with the coordination problem of reactive power regulation equipment in different districts. It is propitious to realize the integrated reactive power and voltage control of power system. Finally, this paper applies the theoretical research results to a decentralized wind power project in a certain city. Based on the research of the project, a set of integrated reactive power and voltage control system is designed, and the reactive power optimization part is completed by using distributed cooperative particle swarm optimization algorithm. By comparing the operation times of reactive power and voltage regulating equipment, the network loss and the time of optimization calculation before and after the introduction of integrated reactive power and voltage control system, and so on, the key indexes of wind farm reactive power and voltage regulation equipment are compared. It can be seen that the integrated reactive power and voltage control system designed in this paper based on distributed cooperative particle swarm optimization algorithm has great superiority compared with the traditional wind farm single optimal control and centralized optimal calculation method.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM714.3

【相似文献】