Robust Fault-tolerant Control Research Of Nonlinear NCS

Compared with traditional systems where various components are usually connected via point-to-point cables, the NCS have large scale, complicated construction and more fault inducing factors. In the event of failure may cause the huge loss of personnel and property. Hence, NCS are increasingly received a extensive attention via design of fault-tolerant control to improve the safety and reliability of the system. Aiming at some problems for the existing achievements of NCS fault-tolerant control, such as controlled plant is mostly linear, system performance generally is integrity, state feedback control strategy is usually adopted, and the delay-independent conclusions are mostly large conservative, and so on, this paper addresses the problem of robust fault-tolerant control for nonlinear networked control systems (NNCS) with network-induced delay and packed dropout. Overall briefly as follows:1) For a class of uncertain NNCS with network-induced delay and packed dropout, a T-S fuzzy model is employed to represent the nonlinear controlled plant. Considered the state variables are measurable and immeasurable respectively, the packed dropout as a special type of time-delay, state feedback and observer-based dynamic output feedback control strategy are used respectively to set up the closed loop failure model of nonlinear NCS against actuator or sensor failure. Especially, in the modeling of observer-based dynamic output feedback control strategy, delay section processing method is adopted, which more accord with the engineering practice, and the upper and lower bound of the time-delay are considered respectively. These have laid the foundation for subsequent less conservative result.2) For a uncertain NNCS with time-varying delay and packet dropout, using the state feedback control strategy, by construct a proper Lyapunov-Krasovskii functional and employ the free weighting matrix, linear matrix inequalities, several integral inequality, etc., the less conservative delay-dependent sufficient conditions and controller design methods are reserved for system with robust integrity, robust H∞integrity, robust guaranteed cost and robust H∞guaranteed cost fault-tolerant performances against sensor or actuator failure. Moreover, effectiveness and conservative of all conclusions are comparatively studied. Due to considering all information of time delay property, especially the lower bound of the time-delay, less conservative results are obtained. In addition, Neither model transformation nor magnifying or ignoring useful items in the theoretical result derivation, but free weighting matrixes are introduced via Newton-Leibniz formula and integral inequality, which amplify degrees of freedom of solution, and conservative is further reduced.3) For a uncertain NNCS with time-varying delay and packet dropout, we adopt observer-based dynamic output feedback control strategies when state variables are immeasurable. Firstly, used research method similar to that of the state feedback control, the problem of robust integrity and robust H∞integrity fault-tolerant control is investigated against actuator failure. Then considering complexity of the calculation which is caused by excessive free-weighting matrices due to segmentation process of time delay, we research the problem of robust guaranteed cost and robust H∞guaranteed cost fault-tolerant control via introducing the intermediate variables α1and α2of delay information to decrease matrix dimension and using Jensen inequality to process cross terms. Because any free-weighting matrix is not introduced, computational burden is decreased significantly and conservative which may be produced by optimization of excessive decision variables is avoided. Besides that, conservative and complexity are studied comparatively for two kinds of research methods in this paper. The result of the comparison show it is better on reduce conservative and lower computational complexity to the second method. And then benefit is flowed to improve the fault-tolerant satisfaction of uncertain NNCS.4) Relying on PCS platform supplied by the advanced control laboratory of Lanzhou University of Technology, we develop successfully virtual NCS test platform based on OPC and other technologies. In this experimental platform, aforementioned theoretical achievements are carried on experiment research of engineering usability. Results show that it is reliable to build for the test platform and it is correct and effective for all theoretical results. Due to the flexibility of the platform, it provides the convenience for the study of the complex NCS.In short, the results with little conservative, low computational complexity and high fault-tolerant satisfaction from the paper, will become the important way to improve safety of NCS. Especially, observer, controller, network properties and fault are unified under the T-S fuzzy model framework to study robust fault-tolerant control when state variables are immeasurable, which first to explore the effective methods for the passive fault-tolerant control of NNCS.