基于分层设计方法的非线性多源干扰系统精细抗干扰控制

Recently, anti-disturbance control based on disturbance observer mostly focuses on single type of disturbance or integrates multiple disturbances into a new equivalent disturbance. The characteristic of multiple disturbances and the influence of disturbances to system performance are not given enough attention, which limits further improvement of system performance. Morover, the main part of research object are linear systems or especial nonlinear systems with well-defined disturbance relative degree. Elegant anti-disturbance control with a core of hierarchical design method based on classified modeling, hierarchical control and independent composite is proposed for nonlinear systems with multiple disturbances and arbitrary disturbance relative degree. Firstly, nonlinear modeling method for disturbance is presented based on nonlinear vibration theory and output regulation theory, which broadens descriptions range of linear disturbance mode and eliminates matched condition of disturbance. Secondly, iterative learning control is applied to design novel nonlinear disturbance observer which can estimate the disturbances online, as a result, the accuracy and response speed of disturbances observation are further improved. Then, by integrating disturbance observer based control (DOBC) with robust control and stochastic control, a novel type of elegant anti-disturbance control scheme with characteristic of decoupling and separability is presented to achieve high-accuracy anti-disturbance control. Finally, the hard disk drive system is given to demonstrate the effectiveness of the proposed methods. And it can provide an effective anti-disturbance control scheme for high-accuracy active vibration control.

目前，基于干扰观测器的抗干扰控制研究大多针对单一类型干扰或者将多源干扰整合为一个等价干扰，而对多源干扰的特征及对系统的影响考虑不足，因而制约了系统性能的提高。另外，研究对象的主体部分大多是线性系统或者是具有固定干扰相对阶的特殊非线性系统。本项目针对具有任意干扰相对阶的非线性多源干扰系统，提出基于分类建模、分层控制和独立复合的分层设计方法及以其为核心的精细抗干扰控制策略。首先，结合非线性振动理论及输出调节理论，提出干扰的非线性建模方法，拓展线性干扰模型的描述范围，消除干扰匹配条件的限制。其次，利用迭代学习控制方法，设计可以在线估计的新型非线性自适应干扰观测器，提高干扰观测的精度和响应速度。然后，将基于干扰观测器的控制与鲁棒控制、随机控制相结合，提出具有解耦和分离特性的精细抗干扰控制策略，以实现高精度控制。最后，将研究结果应用于硬盘驱动器系统，为高精度振动系统主动控制提供有效的抗干扰控制方法。

现有基于干扰观测器的抗干扰控制研究大多针对多源干扰的特征及系统的影响考虑不足，从而影响了系统性能的提高。本项目针对非线性多源干扰系统，提出了基于分类建模、分层控制和独立复合的分层设计方法及以其为核心的精细抗干扰控制策略。本项目从反作用飞轮系统入手，研究了多源干扰的形成机理，提取和建立了多源干扰的分类模型。本项目将多源干扰建模为两种类型，第一类干扰表示初值未知且含有摄动项的外源系统，代表已知频率但未知振幅、相位的干扰，涵盖常值、谐波、中立型噪声、随机等信号。研究将原有的匹配干扰发展至非匹配干扰，拓宽了干扰的描述范围，进而拓展了所提方法的适用范围；第二类干扰归结为随机变量或者范数有界变量。对于第一类干扰，设计具有自学习能力和分离特性的非线性干扰观测器进行在线估计和抵消，将干扰观测器由线性化设计发展为非线性设计，减少因外部干扰和未建模动态等相互耦合而产生的干扰估计的保守性。研究结果表明本课题所提出干扰观测器的设计方法与现有文献中的方法相比，提高了干扰观测的精度和响应速度。针对第二类干扰，将DOBC方法与鲁棒控制、自适应Backstepping方法结合，提出具有解耦和分离特性的非线性系统精细抗干扰控制策略。