基于ESO与逆系统的非线性系统故障调节方法研究

发布时间：2018-06-18 08:22

本文选题：非线性系统 + 扩张状态观测器　；参考：《兰州理工大学》2017年硕士论文

【摘要】：随着当代科技的迅猛发展,实际工程中的各种控制系统变得日趋复杂,其规模变得越来越庞大,系统本身也表现出更强的非线性。一旦系统发生故障,轻则其影响性能,重则导致整个系统瘫痪,出现灾害性事故。因此,对非线性系统安全性与可靠性的研究变得尤为迫切。鉴于此,考虑实际系统中存在的非线性、执行器易发生故障以及外部扰动等因素,本文针对一类具有执行器定值和时变故障的非线性系统,提出了变增益扩张状态观测器(extended state observer,ESO)、新的基于扩张状态观测器的扩张状态滤波器(new extended state filter based on extended state observer,NESF-ESO)、新型ESF(extended state filter,ESF)三种方法;并结合逆系统理论进行了故障调节的研究;最后采用力学、物理学和电子学研究中的经典数学模型Van Der Pol振荡器进行了相应仿真实验研究,具体内容包括:1)基于变增益ESO与逆系统结合的连续系统故障调节方法研究针对一类非线性系统执行器时变故障的容错问题,研究了一种基于变增益扩张状态观测器的逆系统故障调节方法。该方法首先通过设计一种时变增益,来改进传统的恒增益扩张状态观测器;其次,在依据故障估计值对原系统进行补偿调节的基础上,借助于逆系统的引入,对原非线性系统线性化;进而又为其设计了鲁棒保性能控制;最后,采用典型非线性系统Van Der Pol振荡器,分别在发生恒值和时变故障情形下进行了仿真研究,验证了所提方法的有效性与优越性。2)基于NESF-ESO与逆系统结合的离散系统故障调节方法研究针对传统非线性系统故障诊断方法中存在的线性化误差、滤波发散等问题,首先构造出一种可用于故障诊断的扩张状态滤波器,并在此基础上推证出了增广系统渐近稳定的充分条件,并借助于原系统中已知的非线性动态进行故障的分离;其次,依据分离出的故障估计值对原系统进行补偿调节,再辅之以逆系统方法,对原非线性系统进行线性化;进而考虑逆建模误差的存在,设计了离散时间下的鲁棒保性能控制器;最后,通过仿真实验对比验证了该方法的有效性与优越性。3)基于新型ESF与逆系统结合的离散系统故障调节方法研究针对传统滤波算法ESF效果一般的问题,研究了一种新型ESF对含噪离散非线性系统故障滤波及调节方法。首先借鉴扩张状态滤波器思想,设计一种新型的ESF,采用递归学习方法来提高对含噪系统的状态及故障估计的准确性;其次,利用新型ESF滤波后的故障估计值作为故障调节项,并结合离散逆系统控制算法,对原非线性系统进行线性化;进而考虑逆系统控制方法未形成闭环以及存在逆建模误差的不足,设计了一种离散时间下的鲁棒保性能控制器;最后,仍采用经典非线性数学模型Van Der Pol振荡器对其进行有效性验证。
[Abstract]:With the rapid development of modern science and technology, various control systems in practical engineering become more and more complex, their scale becomes larger and larger, and the system itself shows stronger nonlinearity. Once the system fails, it will affect the performance of the system, and the whole system will be paralyzed and catastrophic accidents will occur. Therefore, it is urgent to study the safety and reliability of nonlinear systems. In view of this, considering the factors such as nonlinearity, actuator fault and external disturbance, this paper deals with a class of nonlinear systems with actuator fixed value and time-varying fault. This paper presents three methods of extended state observer, new extended state filter based on extended state observer, new extended state filter based on extended state observer, new extended state filter and new ESF, and studies the fault regulation of ESOs based on inverse system theory, and proposes three methods of extended state observer (ESO), new extended state filter (ESF) based on extended state observer (ESO), and new extended state filter (ESF) based on extended state observer (ESO), which is based on extended state observer (ESO). Finally, the Van Der Pol oscillator, a classical mathematical model in mechanics, physics and electronics, is used to carry out the corresponding simulation experiments. The specific contents include: (1) based on the combination of variable gain ESO and inverse system, the fault tolerance problem of a class of nonlinear systems with time-varying actuator faults is studied, which is based on the combination of variable gain ESO and inverse system. A fault regulation method for inverse system based on variable gain extended state observer is studied. Firstly, a time-varying gain is designed to improve the traditional constant gain extended state observer. Secondly, based on the compensation and adjustment of the original system based on the fault estimation, the inverse system is introduced. The linearization of the original nonlinear system and the design of robust guaranteed cost control for the nonlinear system are presented. Finally, the Van Der Pol oscillator of a typical nonlinear system is simulated in the case of constant value and time-varying faults, respectively. The effectiveness and superiority of the proposed method are verified. 2) based on the NESF-ESO and inverse system fault regulation method, the linearization error and filtering divergence in the traditional nonlinear system fault diagnosis method are studied. Firstly, an extended state filter for fault diagnosis is constructed, and a sufficient condition for the asymptotic stability of the augmented system is derived, and the fault separation is carried out with the help of the known nonlinear dynamics in the original system. The original system is linearized by the inverse system method, and the robust guaranteed cost controller under discrete time is designed, considering the existence of the inverse modeling error, according to the separated fault estimation value, the original system is compensated and adjusted, and the inverse system method is used to linearize the original nonlinear system. Finally, the effectiveness and superiority of this method are verified by the simulation experiments. 3) based on the combination of new ESF and inverse system, the fault adjustment method of discrete system is studied, aiming at the general problem of the traditional filtering algorithm ESF. A novel ESF fault filtering and regulating method for noisy discrete nonlinear systems is studied. First, a new ESFs is designed based on the idea of extended state filter. The recursive learning method is used to improve the accuracy of the state and fault estimation of noisy systems. Secondly, the fault estimation value of the new ESF filter is used as the fault adjustment term. Combining with the control algorithm of discrete inverse system, the original nonlinear system is linearized, and a robust guaranteed cost controller under discrete time is designed considering the lack of closed loop and inverse modeling error in inverse system control. Finally, the classical nonlinear mathematical model Van Der Pol oscillator is used to verify its validity.
【学位授予单位】：兰州理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP277

【参考文献】