【作者】 于少娟；

【导师】 熊诗波；

【作者基本信息】 太原理工大学 ， 机械电子工程， 2012， 博士

【摘要】 本文的工作是将迭代学习控制理论引入电液伺服力控系统,从而改善其控制性能。所提出的控制方法不仅拓宽了控制理论应用的研究范围,而且具有一定的实际工程意义。电液伺服力控系统具有输出功率大,可靠性高,反应速度快,操作简单方便等优点,因而广泛应用于各种工程实践中。特别是应用于要求反应迅速精确的大功率自动控制系统中。本文立足于提高电液伺服力控系统的动态响应特性,从而使其满足多方面的应用需求,具有更良好的工业应用前景。电液伺服力控系统是一种参数时变、较多非线性因素、参数和结构具有不确定性、高阶、强耦合且经常存在干扰等特性的系统,这些特性使得常规控制方法较难实现对电液伺服系统的高精度控制。近年来,多种智能控制理论方法与鲁棒控制理论方法被用于该系统,为解决具有上述特性的电液伺服系统的建模和控制问题提供了有效的途径。但大多都是针对将电液力伺服系统简化为线性、低阶系统予以研究,目前电液力控系统对控制系统的动态响应特性、跟踪的精度、系统鲁棒性及稳定性的要求愈来愈高,系统的控制需求促使人们不断寻求针对性较强、控制性能较好的控制方法。迭代学习控制能实现在被控系统有限时间区间上的完全跟踪任务。它尤其适用于具有重复运行性质的被控对象,利用对象过去运行的信息,通过数次迭代修正控制信号,实现零误差。迭代学习控制算法简单、易于微机实现、跟踪精确,对解决一些实际问题有相当的优点。将迭代学习控制用于电液伺服力控制系统解决系统非线性、高阶等对控制特性的影响具有很好的效果,但单纯的迭代学习控制也存在学习律参数选择比较盲目和算法抗干扰技术有待提高等问题。为此考虑结合诸如自适应控制、变结构控制等鲁棒控制,利用这些算法鲁棒整定系统的能力,尽量避免学习律参数选择的盲目性,在开闭环迭代学习控制结构的基础上综合采用具有反馈与前馈作用的环节,以及引入H∞理论,旨在进行鲁棒迭代学习控制律的设计,力求给出同时兼顾收敛速度和跟踪性能的方法,针对所提出的一系列迭代学习算法,进行理论分析和仿真试验,使每种迭代学习算法的有效性都得到验证。目的是拓宽迭代学习控制的应用范围,加强迭代学习控制的实用性。本文的主要工作有：1、综述了电液伺服力控制系统的特点和控制要求、控制策略以及应用等方面的研究现状,简要介绍了迭代学习基本概念、原理、算法结构、研究内容以及应用成果,总结了提高算法的收敛速度和跟踪精度的迭代学习控制技术,阐述了本课题的相关研究背景及意义,提出主要研究工作2、研究建立电液伺服力控制系统模型的方法。模型研究为分析系统本质特性和提出合理的校正策略提供理论基础。分析了电液伺服阀的非线性因素,建立了电液伺服力控系统几种传递函数模型,在此基础上推导出基于力变量的系统状态空间模型,奠定了理论分析的基础,最后进一步建立了电液伺服力控系统的混合仿真模型,使数字控制器模型与系统的物理模型相结合,这种仿真与建模方法有效地模拟了工程实际。3、研究了智能鲁棒迭代学习控制策略,针对系统高阶、初始状态偏移和不确定的未建模动态以及缓慢变化,单纯学习律参数选择比较盲目问题,在开闭环迭代学习控制算法基础上结合自适应控制综合采用同时具有反馈与前馈作用的环节,针对系统的各种干扰、不确定特性、期望轨迹变化的影响,单纯迭代学习算法抗干扰能力不强的缺陷,结合滑模变结构控制,并采用H∞理论,设计鲁棒迭代学习控制律,力求使控制过程更快地收敛于期望值。收敛速度和跟踪性能这两个问题的解决将会推进迭代学习真正成为实用的控制技术。4、进行了阀控对称缸电液驱动力伺服系统控制方法研究,分析了电液驱动力伺服系统特性,针对系统模型知之甚少,尤其许多系统参数未知,采用具有前馈与反馈作用的开闭环综合迭代学习自适应控制策略,根据理论分析给出收敛条件,设计结构合理的控制器。重点解决高阶特性、初始状态偏移、未建模动态、油源压力变化、油温变化等未知信息给迭代学习控制用于电液伺服驱动力系统控制所产生的学习参数选择比较盲目的问题。仿真结果表明所提控制策略跟踪性能良好的前提下,迭代学习参数选择趋于合理,学习速度有较大提高。5、对电液伺服被动力系统控制策略进行研究。分析阀控非对称缸电液伺服被动力系统特性,针对系统内部位置干扰和负载变化产生多余力、期望随时间变化等问题,以及系统运动加载时不但有位置控制干扰,还有摩擦力产生外部干扰的问题,采用所提出的基于H∞理论的迭代学习变结构控制方法,重点研究进一步提高迭代学习控制抗干扰能力的问题,以增强其鲁棒性,并设计结构合理的控制器,分析给出收敛条件。仿真结果表明该控制策略不但保持传统迭代学习高精度跟踪的特性,鲁棒性得到了提高,使收敛条件得以放宽。6、进行电液伺服力控制系统的实验研究,以DSPACE为平台,使用实际的液压一机械系统物理模型和数字控制器模型,搭建了电液伺服驱动力试验系统,以及双缸电液伺服被动力试验系统,设计实验平台装置的硬件和软件。针对电液伺服力控系统的高阶特性、非线性特性、参数不确定性、负载变化、期望变化、抗扰性等问题,采用所构造的各种控制器实现了每种控制策略的快速原型试验。通过实时控制试验,验证所提策略对提高系统快速性、解决不确定性、增强鲁棒性的有效性。更多还原

【Abstract】 The iterative learning control theory is used to improve the control characteristic of electro-hydraulic force servo control system in this paper, and the proposed control method not only widen the research scope of the control theory application but also which has the theory and actual engineering significanceThe electro-hydraulic force servo control system has some merits such as great output power, high reliability, strong anti-interference ability, rapid respond speed, and simple operation, etc. Hence the system has been widespread used in various engineering practices. Especially, the system is used in the rapid and accurate respond requirement for the big power automatic control system. The dynamic characteristic of electro-hydraulic force servo control system is improved in this paper, which can not only meet various applications, and according with the continuous efforts and long-term target in this technology domain both domestic and overseas. Certainly, the dynamic characteristic improvement has very strong actual significance.The electro-hydraulic force servo control system is a system which consists of some characteristics, such as high order, strong coupled, time varying, strong nonlinear, uncertain parameters and structure, and normal endure uncertain load interference, and so on. The high accuracy control of electro-hydraulic force servo control system is difficult to realize by using the conventional control method due to these characteristics. In recent years, many intelligent control and robust control theory method is used in the system, which provides the effective way for the electro-hydraulic force servo control system with above characteristics. However, the electro-hydraulic force servo control system is simplified as linear and low order system in most method. With the increasing of tracking accuracy, dynamic respond ability, stability, and robust, the complicacy of system requires to find the more strong target and more excellent control performance of control method.The controlled object of iteration learning control method has rerun characteristic, and the control method revises the control signals by using the iteration method and before operation information, which can realizes zero error tracking task in limited time interval of controlled system. The conventional iteration learning control is well known due to the simple algorithm form and accurate tracking, which has certain merits in resolve some actual question. In order to resolve the effect of nonlinear and high order for the system characteristic, the iteration learning control is used in the electro-hydraulic force servo control system. But the simple iteration learning exists in some questions such as the parameters choice of study law is aimless and the anti-interference of arithmetic need improve. Hence many robust control methods are combined with the conventional iteration learning method, such as adaptive control and variable structure control, which uses the robust setting ability for the system to avoid the aimless of study law parameters choice, and the open and closed loop combined with iteration learning control structure with feedback and feed forward action and H∞theory is adopted in the proposed method to design a robust iteration learning control law along with astringency and tracking ability, and a series of iteration learning law design method is proposed in this paper. The effectiveness of the every proposed iteration learning law is confirmed by using simulation and theory, which broadens the application domain and improves the practical applicability of iteration learning control.The main work in this paper:I) The characteristic and control require and control strategy and current research situation of application of electro-hydraulic force servo control system is summarized, then, the basic concept and idea of iteration is presented and the research and application results of iteration learning control is classify introduced and reviewed, and the convergence speed and tracking accuracy of improved arithmetic of iteration learning control technology is summarized, finally, the relative research background and significance of this project is explored and the main research work is presented.Ⅱ) The various modeling method for electro-hydraulic force servo control system is researched. The model research provides the theory base to analyze the nature characteristic of system and propose reasonable revise strategy. Then, the nonlinear characteristic of electric-hydraulic servo valve is analyzed, and the transfer function of valve controlling cylinder electro-hydraulic force servo system is established. Next, the state space model based on force variables and bias variables for electro-hydraulic force servo system are deduced, which can be used as the base of follow-up analysis. Finally, the hybrid simulation model of valve controlling cylinder electro-hydraulic force servo system is introduced, namely, the physical model and digital controller model of system is simulated by using the most close to the actual engineering simulation and modeling method.Ⅲ) Many intelligent robust iterative learning algorithms are researched, which is combined with adaptive control. In order to resolve these question such as original state excursion, uncertain and slow changing of non-modeling dynamic, and the aimless of study law parameters choice, the comprehensive open and close loop for the iterative learning control structure is adapted which has the feedback and feed forward action. The alone iterative learning algorithm has the drawback of poor antijamming capability, in order to realize the faster convergence of expected value during learning control process and resolve the effect of astringency and tracking performance during iterative learning control process, which is combined with variable structure control, and uses the H∞theory, and robust iterative learning control law due to the excellent astringency and tracking performance. The iterative learning algorithm became the practical control technology because of the two question is resolved.IV) The electro-hydraulic force servo control system is researched and the electro-hydraulic force servo characteristic is analyzed, the comprehensive open and close loop for the iterative learning control structure is adapted which has the feedback and feed forward action due to system model knows a little and a lot of system parameter is unknown. The reasonable structure controller is designed by using the convergence condition of theory analyses. The main question is that the unknown information induces the aimless learning parameter choice of electro-hydraulic force servo control system by using iterative learning control such as high order characteristic, original state excursion, non-modeling dynamic, and the changing of oil sources pressure and oil temperature, which is resolved in the paper.V) The control strategy of electro-hydraulic passive force servo system is researched and the electro-hydraulic passive force servo characteristic is analyzed, and the iterative learning and variable structure based on H∞theory is adapted to resolve these questions such as location interference of electro-hydraulic load simulator, the redundant force of load change, and the excepted value changing with time variation, etc. There are the location control interference during opposite vertex synergy and.the redundant force of external interference come from frictional force. Here, the drawback of weak anti-interference ability of alone iterative learning is the important question to improve the robustness. A reasonable structure controller is designed and the convergence condition is analyzed. The simulation results shows that the control strategy can remains the high accuracy tracking performance and improves the robustness and relaxes the convergence condition.VI) The DSPACE platform is used in the experimental research of electro-hydraulic force servo system, and the electro-hydraulic force servo driving force and double-cylinder electro-hydraulic passive force pilot system is established by using the actual physical model of hydraulic pressure and mechanical system and digital controller, and the hardware and software for experimental platform is designed. The rapid prototype test of every control strategy is realized by using the previous various controllers, and these questions of electro-hydraulic force servo system is improved, such as high order performance, nonlinear characteristic, uncertain parameters, load changing, expected values changing, and jamming-interference, and so on. The proposed strategy is proved to improve the rapidness, uncertainty, and robustness by using the real-time control experimental.更多还原