节点文献
电力系统暂态稳定自适应变结构控制研究
Study on Adaptive Variable Structure Control for Power System Transient Stability
【作者】 马和邦;
【导师】 周建中;
【作者基本信息】 华中科技大学 , 系统分析与集成, 2012, 博士
【摘要】 我国电力系统正步入大机组、大电网、特高压、交直流混合输送的电力能源时代,电力系统复杂化程度越来越高,其运行与控制的难度越来越大,有效驾驭特大型电力系统正面临严峻挑战。电力系统作为典型的高维、强非线性、不确定性动力学系统,尚有许多不为人所知的复杂现象与运动规律,需要进一步加强理论和实践研究,以获得更加切实可行的控制策略和控制手段,保证电力系统长期稳定可靠运行。面对日益复杂多变的电力系统,传统的基于固定结构的精确化方法对于提高电力系统暂态稳定的控制能力受到很大限制。近几十年来,电力系统暂态稳定控制,尤其是励磁系统控制得到了世界各国电力系统专家广泛而深入的研究,提出了许多工程实践中行之有效的控制策略与控制方法,但仍然存在不少问题有待完善。本文以电力系统暂态稳定控制为研究对象,尤其针对励磁控制系统中出现的不确定性,在前人大量研究成果基础上,进一步研究非线性不确定电力系统暂态稳定自适应变结构控制策略,主要完成了以下几方面研究成果:(1)概述了电力系统暂态稳定的基本内涵,简要描述了系统故障发生后发电机内部机械与电磁之间的制约关系,回顾了电力系统暂态稳定控制的发展历程,综述了电力系统自适应变结构暂态稳定控制的研究现状。(2)电力系统是典型的强非线性、非自治、高维动力系统,其稳定性分析与综合比较复杂,研究工作需要合适理论指导。因此,引入了电力系统暂态稳定控制研究工作所需要的主要理论背景及其分析与设计工具。简要介绍了Lyapunov稳定性理论中常用的稳定性定理以及Barbalat引理;以一类线性化参数的不确定非线性系统控制为例,给出了反推自适应控制器的迭代设计步骤;解析了滑模变结构控制不变性的基本原理,并计算了快速终端滑模收敛时间;说明了变结构控制抖振现象产生原因以及消除措施;阐述了径向基神经网络的概念以及神经网络自适应控制基本模式。(3)将快速终端滑模变结构、反推控制以及自适应控制等控制策略进行有机结合,提出了单机无穷大电力系统暂态稳定自适应快速终端变结构励磁控制器,数字仿真实验表明,该控制器具有较好的适应性与较强的抗干扰能力,具有优良的控制品质。(4)针对多机电力系统励磁控制,提出了自适应多级滑模励磁控制器,结合干扰界观测器,解决了多机电力系统模型结构与参数非匹配,参数不确定等技术难题。为克服变结构控制输入信号抖振问题,引入了具有无穷可导的光滑切换函数。数字仿真实验表明,设计的基于多级滑模的自适应多机励磁控制器具有很强的抗干扰能力,能有效提高多机电力系统暂态稳定性。(5)考虑到电力系统暂态稳定励磁控制系统难以精确建模的特点,引入径向基函数(RBF)神经网络逼近系统不确定性。首先针对一类非线性不确定性系统,导出了一般意义上的神经网络自适应控制器的控制律、网络权重调整律以及参数调整律。然后结合单机无穷大电力系统励磁系统模型特点,构造了神经网络自适应励磁控制器,经数字仿真检验,控制器具有良好适应性与抗扰动能力。
【Abstract】 The state power systems are now entering a new era of power energy which with large power unit and with special high voltage, and power energy be transported by hybrid alternative current and direct current. The running and controlling of power system are becoming more and more complicated. How to control special large power systems is facing a serious challenge. Power system is of a typical nonlinear dynamics of multi-dimension and with strong nonlinearity and uncertainty and in which much intricate phenomena and movement law still unknown. In order to ensure power systems running credible stable for long period and obtain suitable and effective control approach, the study on the control theory and control practice must be further intensified. Facing on the fast change and complicated power systems, the ability of traditional exacting control method with fixed structure will be limited for improving power system stability. For the last decades, the control of power systems transient stability especially excitation control methods had be investigated widely and deeply by electricity specialists of world. Many effective control strategies and methods of the practices and theory be proposed, but still a lot of problems remain to unresolved. This thesis focused on the study of the power system transient control and especially on the excitation control with uncertainty based on prior knowledge and investigating the modern adaptive variable structure control strategy for uncertain power system transient stability. The following control methods and control strategies are presented in this thesis.(1) Summarizing the basic definition of power system transient stability and describing the relationship between the mechanical and electromagnetism briefly, and then review the developments of control strategies of power system transient stability. Finally, the progress of control methods of power system transient control is introduced.(2) Power system is of a typical dynamics systems with strong nonlinearity, non-autonomous and high-dimensions. It is a difficult task for analyzing and controlling and the study on it should be guided by suitable theory and necessary control tools. So it is necessary for providing background of stability theory. Lyapunov stability theory and tools of stability analyzing is introduced into and at the same time Barbalat lemma is provided. For a class of uncertain nonlinear systems with parameter can linearization, the design procedure of the controller by using backstepping methods is presented and also explaining the principle of control invariable of the sliding mode control. The time of the sliding to the balance point of the fast terminal system is calculated. The conception and control mode of neural network control is introduced.(3) A adaptive fast terminal sliding excitation controller for SMIB power system transient stability is proposed based on combining with fast terminal sliding mode control, backstepping and adaptive control. Simulation performed on SMIB shows that the proposed controller have good properties of adaptive and suppress disturbances and also shows good control performances.(4) A multi-sliding mode adaptive excitation controller is proposed for multi-machine power systems. The mismatched problem of parameter and disturbances for multi-machines power systems is solved by using disturbance observers. In order to overcome the chattering of input signal of sliding mode controller, a smooth S-type function which can infinitive differential is introduced to substitute for switching function. Simulation performed on 6-machine 22-nodes systems shows that the proposed controller have good properties of attenuating disturbances and can improve power systems transient stability effectively.(5) Considering the difficulty for exacting modeling of power system transient stability control, radial base function neural network is introduced to approximate the uncertainty and external disturbances. For a class of nonlinear uncertain system, adaptive control law, weight update law, parameter update law of the system in general meaning are given. By applying the above conclusion into SMIB, a adaptive excitation controller based on neural network control is proposed. Simulation studies are included to illustrate the effectiveness of the proposed controller.