【作者】 张坚；

【导师】 费敏锐；

【作者基本信息】 上海大学 ， 控制理论与控制工程， 2007， 博士

【摘要】 在过去的十几年时间里，非线性系统的跟踪控制问题受到众多专家学者的重视，并取得了许多重要的成果。然而到目前为止，大多数非线性系统的跟踪控制方案都无法对给定的轨迹进行渐近跟踪，而且很难综合考虑非线性模型中的不确定性及H₂和H_∞等性能指标。有少数方案通过采用自适应控制等算法可对给定轨迹进行渐近跟踪控制，但是这些方案往往使得控制系统的实时性变差。针对以上问题，本文以保证跟踪控制系统的综合性能指标为中心，在保证系统实时性的前提下提出多种基于T-S模糊模型的跟踪控制方案。所提出的控制方案可使连续非线性系统（简称非线性系统）在包含参数不确定性的情况下很好地跟踪给定参考轨迹，与此同时消除外界干扰对系统的影响并保证闭环系统的内部稳定性。以电站中速磨煤机（简称中速磨）为应用对象，采用所提出的控制方案设计其跟踪控制系统，达到改善控制性能的目标。本文的主要工作概括如下：1．提出非线性系统的模糊渐近跟踪控制方案。将需要跟踪的给定轨迹及被控对象所受外界干扰的内模嵌入到控制系统中，抵消跟踪轨迹与外界干扰的传递函数中包含的不稳定极点，从而使非线性系统的输出可以渐近跟踪给定轨迹并抑制扰干扰。为了使控制系统可应用于状态变量不可测量的情况，设计了模糊跟踪控制器对应的模糊观测器。接着分析了跟踪控制器及其观测器稳定的充分条件及跟踪控制器和观测器分步设计后系统整体稳定的充分条件。然后讨论了对跟踪控制系统进行简化设计的方法。2．提出非线性系统满足H_∞抗干扰性能指标的鲁棒模糊跟踪控制方案。与非线性系统的模糊渐近跟踪控制方案相比，此方案考虑了T-S模糊模型中的逼近误差及控制系统的H_∞抗干扰性能指标。使得受外界特性未知扰动影响的非线性系统的输出可以很好地跟踪给定轨迹。针对状态变量不可直接测量的情况，设计了H_∞鲁棒模糊控制器对应的模糊观测器，进一步地，给出了H_∞鲁棒模糊跟踪控制器和对应观测器稳定及整个H_∞鲁棒模糊跟踪控制系统稳定的充分条件。3．提出非线性系统的混合H₂／H_∞鲁棒模糊跟踪控制方案。与非线性系统的H_∞鲁棒模糊跟踪控制方案相比，该方案考虑了系统的H₂性能指标及T-S模糊模型中子系统输出方程的逼近误差，更完整地描述了实际的非线性系统。混合H₂／H_∞鲁棒模糊跟踪控制方案能很好地跟踪给定轨迹，而且满足对特性未知干扰抑制程度的H_∞性能指标与控制过程中的H₂优化性能指标。所证明的控制器及整个跟踪控制系统的稳定性定理则从理论上保证了所提控制方案的正确性。4．提出非线性系统的鲁棒模糊渐近跟踪控制方案。将非线性系统所包含的不确定性体现为T-S模型中的参数不确定性。将系统所需跟踪的轨迹及外界干扰的内模加入控制系统中，从而对给定轨迹进行渐近跟踪并近似完全抑制干扰对系统的影响，改善了系统的抗干扰的性能。接着，对跟踪控制器及其观测器的稳定性进行了分析并给出了保证鲁棒模糊跟踪控制系统整体稳定的充分条件。5．在对中速磨的运行机理分析的基础上，改进其已有的非线性数学模型。基于这个改进后的模型，提出了中速磨的T-S模糊模型，从而便于设计其控制器。所提出的T-S模糊模型由几个线性子模型所组成，采用遗传算法可以辨识得到子模型所包含的系数从而获得整个T-S模糊模型。该T-S模糊模型完全等价于改进后的中速磨的非线性数学模型。对一台实际运行的中速磨进行模糊建模，得到的T-S模糊模型真实地反映了中速磨的运行状况，验证了所提模型的有效性。6．对鲁棒模糊跟踪控制系统的设计过程进行了分析与归纳。在此基础上开发了中速磨的鲁棒模糊跟踪控制系统设计软件，方便现场工程师开发控制系统并对其进行仿真试验。然后在中速磨出口变量的参考轨迹为典型的阶跃信号和复杂信号的情况下分别对鲁棒模糊跟踪控制系统进行了仿真，结果表明控制系统具有良好的跟踪控制性能。更多还原

【Abstract】 The tracking control problem of nonlinear systems has been extensively investigated by a number of researchers and many results have been made over the past decade. By far, most tracking controllers of nonlinear systems can’t track the desired trajectory asymptotically. They also can’t consider such performance indexes as robustness, H₂ performance, H_∞performance at the same time, cew tracking control schemes can make the output of a nonlinear system track the desired trajectory asymptotically by adopting the adaptive scheme, e owever, these schemes will make control systems slower. Aiming at the above problem, this paper will propose several fuzzy tracking control schemes which can consider important performance indexes of tracking control systems with guarantee of their real-time characteristic. These control schemes can make the output variables of continuous nonlinear systems （nonlinear systems for short） with uncertainties track the desired trajectories asymptotically. At the same time, the control schemes can attenuate the effect of external disturbance and maintain internal stability of the close-loop systems, curthermore, the tracking control system of a pulverizer of a power plant is designed according to the proposed fuzzy tracking control scheme. Control performance of the new tracking control system is better than the former control performance. The main work is generalized as follows:1. The fuzzy asymptotically tracking control scheme for nonlinear systems is presented. The internal model of the desired trajectory and the external disturbance is embedded in the control system to counteract the unstable poles of the disturbance and trajectory. Thus the output variable of the nonlinear system can track the desired trajectory asymptotically and the external disturbance can be attenuated without adaptive control scheme. The corresponding fuzzy observer is presented to deal with the condition that the state variables are not measurable directly. The sufficient conditions of stability of the fuzzy tracking controller, its corresponding observers and the fuzzy tracking system containing a controller and a observers are proven to guarantee the validity of the presented control schemes. Then the simplified design methods of fuzzy tracking control are introduced.2. The robust fuzzy tracking control scheme is proposed which can satisfy the H_∞disturbance attenuation performance. Compared to the fuzzy asymptotically tracking control scheme, the proposed scheme considers the approximation error in the T-S fuzzy model and the H_∞disturbance attenuation performance. It can make the output variables of nonlinear systems with external unknown disturbance track the desired trajectories very well. The corresponding H_∞fuzzy observer is also presented to deal with the condition that the state variables are not measurable directly. Furthermore, the sufficient conditions of H_∞fuzzy tracking controller, its corresponding observer and the H_∞fuzzy tracking control system are proven.3. The mixed H₂/H_∞robust fuzzy tracking control scheme is proposed. The approximation error between the output of nonlinear system and its corresponding T-S fuzzy model is considered. So the T-S fuzzy model is closer to the nonlinear system. The mixed H₂/H_∞robust fuzzy tracking control scheme can make the output variable track the desired trajectory very well while the H_∞disturbance attenuation performance and H₂ optimization performance are satisfied at the same time. The stability of the mixed H₂/H_∞, robust fuzzy tracking controller and control system containing a controller and a observer are proven to ensure the validity of proposed control scheme.4. The robust fuzzy asymptotically tracking control scheme of nonlinear systems is proposed. The uncertainties in nonlinear systems are expressed as the parametric uncertainties in T-S fuzzy model. Internal models of the desired trajectory and the external disturbance are embedded in the robust fuzzy tracking control system so that the control system can track the desired trajectory asymptotically and the effect of disturbance can be attenuated effectively. The disturbance attenuation performance is improved. Then the sufficient conditions of stability of robust fuzzy tracking controller, its corresponding observer and the robust fuzzy tracking control system are proven.5. Based on the analysis of operation mechanism of a medium speed pulverizer, its present nonlinear mathematical model is improved. Then a T-S fuzzy model of a medium speed pulverizer is proposed based on the improved model, enabling easy design of the control system for the pulverizer. The T-S model is composed of several linear models which can be obtained by identifying a group of constant coefficients with the aid of genetic algorithms. It is equivalent to the improved model of a pulverizer. The actual modeling results show a satisfactory agreement between the T-S fuzzy model response and the measured data, confirming the effectiveness of the proposed model.6. The design process of control system of a medium-speed pulverizer in the DCS hardware platform is introduced. The design procedure of fuzzy asymptotically control system is summarized. Then the robust fuzzy control system design software of a pulverizer is developed which makes it easy to develop the tracking control system and perform the simulation examination for an engineer. Simulation examples of a pulverizer under desired trajectories of a step signal and a complex signal are given to illustrate tracking performance of the proposed tracking control system.更多还原