节点文献

改进粒子群优化算法在整流变压器设计中的应用研究

Research on Improved Particle Swarm Optimization in the Design of Rectifier Transformer

【作者】 李鹏

【导师】 白惠珍;

【作者基本信息】 河北工业大学 , 电气工程, 2011, 硕士

【摘要】 整流变压器在电化学工业、电镀、直流电源以及直流输电等领域有着十分广泛的应用。随着工业水平的不断提高,对整流变压器的容量、结构形式也有了许多新的要求,这就要求设计人员根据不同的情况设计出不同特点的整流变压器。通过对ZS-2000/10kV整流变压器电磁部分采用手工计算,能够得知计算时间长。为了提高计算的效率,论文采用Visual C++软件对其计算过程进行程序设计。设计过程中对电磁部分进行分块,采用分块的方法编写程序,降低编程难度,提高工作效率,节省工作时间。目前在整流变压器设计方面,许多的参数没有具体的国家规定。每个设计者的经验不同,选取的参数往往不同,计算结果也会有差异,故对其进行优化设计是十分必要的。本文通过查阅资料,分析研究各种优化算法的优劣,针对整流变压器计算参数的特点,选择了粒子群优化算法。粒子群优化算法是一种可以应用于多种优化模型的全局搜索算法,计算过程简单、快速、准确。但是该算法在初始参数不多的情况下容易陷入局部最优,故对其进行了改进。在基本的粒子群算法之中引进了控制因子α,以此来调整惯性权重w取值的优劣。控制因子将最大迭代次数分为若干段,每段迭代开始计算时每个粒子的惯性权重由其自身的适应度值和群体平均适应度值的比较结果来决定。适应值好的粒子趋向于局部搜索,惯性权重不作调整;适应值差的粒子将进行全局探测,调整惯性权重,以便发现最优解,使整个粒子群具有多样性和好的收敛性。最后,本文将改进的粒子群优化算法应用到ZS-2000/10kV整流变压器优化设计中,并且在优化计算中采用计算机编程,以人机对话界面显示最终的结果,将优化结果与手工计算的结果进行比较可以看出,改进后的算法具有较好的可行性和有效性。

【Abstract】 Rectifier transformer has a very wide range of applications in industrial field, including electrochemical industrial, electroplating, DC power and DC transmission. With the continuous improvement of industry standards, the new capacity and structure of the rectifier transformer have many new requirements, which require the designer to design different characteristics of the rectifier transformer according to different situations. It is too time-consuming to calculate the electromagnetic part of the ZS-2000/10 rectifier transformer manually. Therefore, in order to improve the calculation efficiency, the paper uses Visual C + + software for the calculating process, in which segmented partition method program is adopted for the design process of electromagnetic parts, as a result it reduces programming difficulty, improves efficiency and saves work time.The current design of the rectifier transformer, many parameters have no specific national regulations. Each designer’s experience and the selected parameters are often different, the results will be different, and so it is very necessary to optimize the design. This paper studies the advantages and disadvantages of various optimization algorithms through data access, aiming at the characteristics of the calculated parameters rectifier transformer, the particle swarm optimization algorithm is choused. PSO is an optimization model which can be applied to many of the global search algorithm; the calculation is simple, fast and accurate. However the algorithm is limited in the case of few initial parameters and it is easy to fall into local optimum, so it needs improvement. In the basic PSO, the control factorαis introduced, so as to adjust the value of inertia weight w . The control factors divide the maximum iterating times into several segments, each particle’s inertia weigh is decided by the comparison results of self-adapting and group average fitness value when each iteration begins. Good fitness particle tend to local search, inertia weight is not made adjust; poor fitness particles will carry on global surveyor, adjust inertia weight to find the optimal solution, so that the whole particle swarm diversity and good convergence.Finally, the improved particle swarm optimization algorithm is applied to ZS-2000/10 kV rectifier transformer design optimization, and the calculation of the electromagnetic part is calculated by Visual C + + programming, interactive interface to display the final results. By comparing the results with manual calculation shows that the improved algorithm is more feasible and effective.

节点文献中: 

本文链接的文献网络图示:

本文的引文网络