Robust Output Feedback Control Design And Batch Operation Optimization For Sampled Systems With Input Delay

The input or output delay phenomenon pervasively exists in industrial process operations due to mass transportation,energy exchange,and signal processing etc.If the influence of time delay were not considered in the control design,the closed-loop system performance would likely be affected seriously,or even become unstable.Besides,in engineering applications there often exist various load disturbances involved with system operations.How to eliminate or attenuate load disturbance is an important issue to be solved for control system design and performance optimization.Although control designs for time-delay systems,especially for state-delay sys-tems,have been reported in many references,there still exist a lot of problems to be solved for systems with input delay and time-varying uncertainties,e.g.,robust predictive feedback control design and anti-disturbance control design et al.This dissertation is concerned with robust con-trol design methods based on only output measurement for sampled systems with input delay.The main content and contributions include:An observer-based disturbance rejection control design method is proposed.Firstly,an extended state observer(ESO)is constructed to simultaneously estimate the system state and load disturbance.Then,a filtered predictor based composite controller is designed using the estimated system state and load disturbance to overcome the adverse impact caused by the input delay and load disturbance,such that the system performance of disturbance rejection could be improved.For the presence of time-varying load disturbance such as the sinusoidal type,the system output error bounds are quantitatively analyzed in terms of using two ESO-based disturbance rejection control methods(i.e.,state predictor feedback and disturbance compensation based predictor feedback control),respectively.A novel predictor design method,named state and disturbance observer-predictor(SDOP),is developed to simultaneously predict system state and load disturbance for a class of sampled systems with input delay and time-varying uncertainties.Moreover,the proposed SDOP is fur-ther extended to the sequential SDOPs(SSDOPs)where each SDOP only estimates the future system state and load disturbance in terms of a specified step size,such that the adverse im-pact caused by large input delay could be effectively overcome.Correspondingly,an SDOP(or SSDOPs)based feedback control design is developed which can significantly improve the sys-tem performance of disturbance rejection.Meanwhile,the sufficient condition for guaranteeing robust stability of the closed-loop system is analyzed.A robust disturbance rejection control method is proposed based on the delay-free output prediction.The filtered Smith predictor is combined with an ESO to estimate the undelayed system state and load disturbance.The ESO and controller are designed by specifying the desired characteristic roots of ESO and the closed-loop system poles,respectively,such that the ESO gain vector and the predictor-based controller could be analytically derived.The application to the temperature control system for a 4-litre crystallization reactor is used to demonstrate the effectiveness and advantage of the proposed control method.For industrial batch processes with input delay subject to time-varying uncertainties,a ro-bust output feedback iterative learning control(ILC)design is proposed based on a two-dimension(2D)system description.By constructing a 2D system state predictor,an ILC controller design is developed for the corresponding 'delay-free' 2D system.Based on the 2D Lyapunov-Krasovskii stability analysis theory,delay-dependent stability conditions are established for the resulting 2D closed-loop system.The ILC control law is explicitly derived by solving the established ma-trix inequality conditions.An illustrative example of injection molding machine is adopted to demonstrate the effectiveness and merit of the proposed method.