【作者】 管春；

【导师】 吕厚余；

【作者基本信息】 重庆大学 ， 电气工程， 2004， 硕士

【摘要】 在电力系统中，由一端向多端供电的情况并不少见，如果要对其进行综合潮流控制，可以用以下2种方法：① 在供电端与各受电端间分别装设统一潮流控制器(UPFC)，假设受电端有N个，则共需N台IJPFC：② 在供电端与受电端间装设一台多端统一潮流控制器(M-UPFC)。为了研究上的方便，论文主要以二端统一潮流控制器(T-UPFC)(当N=2时)为研究对象，就其模型的建立以及控制器的设计等问题进行了较为深入细致的分析讨论，提出了相应的实用模型和控制方案，并通过仿真进行了研究。由于M-UPFC一般用于高压输电线，其功率等级要求很高。在综合研究大功率变换器的基础上，本文选择了一种适合于M-UPFC的变换器--基于二极管箝位的多电平变换器，并对其拓朴进行了一些改进。针对目前输出建模方法和拓扑建模方法的不足，通过引入开关函数的概念，本文建立了T-UPFC的可反映装置内部开关特性和运行机理的开关函数数学模型，较原有输出模型更具有一般性。构建了T-UPFC的动态模型，充分考虑了T-UPFC内部的调节过程，计及了T-UPFC逆变器的调制过程以及直流电容的充放电过程。本文将模糊神经网络(FNN)与改进遗传算法(IGA)结合起来用于T-UPFC的控制设计。由于模糊神经网络不需要对象的准确模型，它以分布的方式存储信息，利用网络的拓朴结构和权值分布实现非线性映射，在神经网络框架下引入模糊规则，使网络中的权值有明显的意义，且保留了神经网络的学习机制。对权值的学习采用改进的遗传算法，可避免BP算法极易陷入局部最优值的缺点，以及传统遗传算法存在的未成熟收敛的缺点。系统仿真表明，在频率波动较大情况下，本文所设计的T-UPFC能很好地抑制后续摆的振荡，有效、快速地抑制系统后续摆的振荡，从而提高了系统的动态稳定性。更多还原

【Abstract】 To control the complex powers emanating from a bus in power system, there are two solutions. The obvious solution is to equip each of the N lines radiating from the node with its own Unified Power Flow Controller (UPFC). The next evolution of the idea is to equip with a Multi-terminal UPFC (M-UPFC). To simplify the research, this thesis treats the Two-terminal UPFC (M-UPFC) (when N=2) as the investigative target mainly. A throughout analysis and study is conducted to modeling and designing of the controller with the T-UPFC. The relative model and control projects are presented and examined by simulation.The M-UPFC needs high power rate because it is generally used in high voltage transmission lines. Based on the study of the high-power converters, a suited inverter (diode-clamping multilevel inverter) is selected. The topology of the diode-clamping multilevel inverter is also inproved in the thesis.A switching function mathematical model of this device is constructed by introducing the concept of switching function. It has more generality than the output and the topological model with the consideration of the internal switching character and the physical course of the T-UPFC.The dynamic model of the T-UPFC was constructed which considers the process of the T-UPFC’s internal modulation and converters, the process of charge and discharge of DC capacitance.This paper applies Fuzzy Neural Network(FNN) based on the Improved Genetic Algorithm(IGA) to the T-UPFC.Because FNN can be used without the specific model of the object, and it stores useful information as distributed manner, we usually make use of the topological structure and weights of neural network to realize nonlinear mapping, which make those weights full of meaning, also reserve the training algorithm of the IGA. As it is known, we hope to get the universal best answer,so in the processing of weight training, we adopt IGA to avoid the defect of BP algorithm which is very easy to get into the local best answer and the defect of the Traditional Genetic Algorithm(TGA) which is very easy to get into premature convergence.The simulation results verify the efficiency of the T-UPFC designed in this paper. It can damp the subsequent swing quickly under serious power oscillation circumstance to enhance the system dynamic stability apparently.更多还原