| A large-scale nonlinear interconnected system is comprised of several subsystems with obvious interconnections, which can be found in engineering areas, such as power systems, communication networks, and industrial processes. Due to nonlinearity, different locations, high dimensionality and interconnections, large-scale systems bring to the increasing difficulty for solving their control problems. On the other hand, there exist some facts in practical applications, such as modeling inaccuracies, parameter uncertainties, unmeasured disturbances and time delays. These conditions generally lead to unsatisfactory performance and even cause instability. Therefore, it is quite important to study the problems of robust static-output-feedback(SOF) H_∞control and H_∞filter for large-scale nonlinear interconnected systems from theoretical and practical perspectives.This dissertation will focus on the problem of robust static-output-feedback(SOF)H_∞control for a class of large-scale nonlinear interconnected systems, and the problem of H_∞filtering for a class of large-scale nonlinear interconnected systems with time-varying delay. The considered system is comprised of several nonlinear subsystems, each nonlinear system is represented by a Takagi-Sugeno(T-S) model. Since centralized control brings to higher computation and cost, this dissertation proposes decentralized control approach, and addresses two schemes: decentralized robust SOF H_∞controller design and decentralized H_∞filtering design. Since the considered system is characterized by several subsystems, fuzzy formulations and interconnections, most of existing results are nonlinear matrix inequalities, or more conservative. Some new approaches is proposed to derive the decentralized robust SOF H_∞controller and decentralized H_∞filtering design results in terms of linear matrix inequalities(LMIs), and the obtained results are less conservative. These approaches are piecewise Lyapunov function, piecewise Lyapunov–Krasovskii function, description system approach, small scale gain(SSG) theorem, model transformation with two-term approximation, and some matrix inequality linearization techniques.First, the problem of decentralized robust SOF H_∞control is examined for continuoustime large-scale T-S fuzzy interconnected systems. By making some assumptions on the output and input of the considered system, and based on some matrix inequality linearization techniques, sufficient conditions for existing a decentralized robust fuzzy SOF H_∞controller are derived in terms of LMIs. In order to release these assumptions on the considered system, by introducing some virtual state variables, the closed-loop fuzzy control system is transformed into two different descriptor representations. Based on the descriptor system approach, the results to decentralized robust fuzzy SOF H_∞controller design are also derived in terms of LMIs. Furthermore, based on the piecewise Lyapunov function, the attention is focused on the problem of decentralized robust SOF H_∞control for discrete-time large-scale T-S fuzzy interconnected systems. Based on some matrix inequality linearization techniques and the descriptor system approach, three different results of decentralized robust piecewise SOF H_∞controller design are derived in terms of LMIs, and are less conservative.Then, the problem of decentralized H_∞filtering design is studied for continuoustime large-scale systems with time-varying delay. Assume that time delay is of an intervallike form, and the bounds of time-varying delay are available a priori, and introduce a decentralized memory non-parallel distributed compensation(non-PDC) fuzzy filter. Instead of direct Lyapunov function approach, an input-output stable theorem is proposed.A two-term approximation method is firstly presented to transform the filtering error system into an interconnected formulation, such that the decentralized H_∞filtering problem is reformulated in the context of input-output(IO) stability. Then, based on a LyapunovKrasovskii functional(LKF) combined with the SSG theorem and some matrix inequality linearization techniques, the results to decentralized memory non-PDC fuzzy H_∞filtering design will be derived for the considered system in terms of LMIs. In addition, a piecewise Lyapunov-Krasovskii functional will be introduced to study the decentralized H_∞filtering problem for discrete-time large-scale systems with time-varying delay. Based on the model transformation with two-term approximation, the SSG theorem, and some matrix inequality linearization techniques, the results to decentralized memory piecewise H_∞filtering design will be derived for the considered system in terms of LMIs. In this part, the model transformation with two-term approximation, the SSG theorem, memory non-PDC fuzzy filter, memory piecewise filter, piecewise LKF, these methods guarantee that the design results are less conservative.In the end, the methods proposed in this dissertation will be applied to the motion control of chip mounter machines. The background for the design of chip mounter machines is first introduced, and the key techniques in the control of chip mounter machines is analysed. Due to these two requirements of high speed and high precision in chip mounter machines, a driving system with two motors is proposed, and the motion systems of x-axis and y-axis are modeled into mathematic formulation, respectively, which belongs to an interconnected system with two nonlinear subsystems. Then, based on T-S model approach and the theoretical results given in the third chapter, the decentralized robust piecewise SOF H_∞tracking controller with the full closed-loop control technique is designed. It will be shown that the closed-loop fuzzy control system is asymptotically stable while the tracking error in loading position is guaranteed with an H_∞disturbance attenuation level. This part is a test work from the methods proposed in this dissertation to the practical application, which not only strengthens the control methods of lager-scale nonlinear interconnected systems, but also supports for the design of chip mounters.
|
PDF Full Text Request