Decoupling Internal Model Control And Event-triggered Failure Compensation Control For Complex Systems

With the development of society and the progress of science and technology,the demand for good quality control in the process industry is getting higher and higher.At the same time,controller design with high quality performance becomes more challenging due to the increasing number of device components and system complexity.In view of the characteristics of complicated controlled objects in the process industry,such as multivariable,strong coupling,time-delays,Right Half Plane(RHP)zeros,unknown nonlinearity,limited system resources and actuator failures,this paper studies the decoupled internal model control(IMC)with inverted decoupling and the event-triggered control methods,in the presence of external disturbances and model uncertainties.The main work is as follows:1.The inverted decoupling method is extended to multidimensional complex systems with multiple time-delays and RHP zeros,in which the necessary conditions for the existence of the decoupling elements are theoretically analyzed and deduced for the first time.On the premise that these conditions are satisfied,the design procedure and realizability of the decoupling matrix elements are discussed,and an approximate compensation method for the RHP zeros is proposed.The proposed method is not limited by the system form,and is suitable for both square system and non-square system,and the proposed decoupler has the advantages of simple calculation and flexible form.2.The traditional Butterworth filter is modified as an improved Butterworth filter.The characteristics of the traditional and improved Butterworth filters are compared and analyzed,which shows that the improved Butterworth filter has better regulating ability to get a better trade-off between tracking performance and robustness.The influence of the change of the improved Butterworth filter design parameters on the system performance is analyzed.Then,using the classical internal model control structure,the improved Butterworth filter is used as the internal model filter to design the internal model controller;and the control parameter selection method is studied based on the closed-loop system performance.3.The fractional order theory is introduced to design the fractional order improved Butterworth filter.Considering the needs of practical computer control systems,the discretization method of the filter is studied,such that the discrete fractional order improved Butterworth filter is obtained.Then,based on the discrete transfer function matrix model of the controlled process,the controller is designed with the internal model control structure,and a discrete internal model controller based on the fractional order improved Butterworth filter is proposed.With the introduction of the fractional order filter,the order of internal model filter is more precisely satisfied.4.Considering the nonlinear characteristics of the process industry systems,the radial basis function(RBF)neural network is used to realize the approximation of the unknown nonlinear function,and the tracking error is always constrained within the boundary of the prescribed performance functions by introducing a prescribed performance error transformation.Combining with the switching threshold event-triggering mechanism(SWT-ETM),an adaptive neural network event-triggered controller(ETC)is designed.The system error can always converge within the prescribed performance function boundaries,and the communication pressure of the system and the burden of the actuator can be significantly reduced.5.For the unknown failures that may occur to the system actuator,the upper bounds of the sum of the relevant failure parameters are estimated by the adaptive method,which can avoid the system oscillation caused by the direct adaptation of the failure parameters.Under the condition of actuator redundancies,combining with the SWT-ETM,an adaptive neural network event-triggered failure compensation controller is designed.Since the introduction of the event-triggering mechanism can reduce the communication pressure of the system and avoid unnecessary operation of the actuator,therefore the proposed control method achieves successful failure compensation and reduces the incidence of potential actuator failures,which means the reliability of the actuator is greatly improved.