Event-triggered Distributed MPC For Nonlinear Systems

With the improvement of modern industrial level,model predictive control has gradually become the most common method to deal with optimization problems in modern industry because of its better ability to deal with constraints.However,it cannot take the increasing continuity and complexity of modern industrial processes into consideration.While a novel distributed model predictive control scheme can handle these problems by transforming the whole system into multiple subsystems,without neglecting the communication among those subsystems,thus improving the performance of the whole system better.In addition,an event triggering mechanism is employed in this thesis.The timetriggered scheme is based on the sampling period to execute control behavior,so that it will waste plenty of system resources.However,the event-triggered scheme is based on the actual state of the system to define the next triggering instant,which can reduce the calculation load of the system and save the system resources badly.On the basis of previous researches,the event-triggered scheme and model predictive control scheme are applied to continuous-time systems as follows:For some nonlinear system models,owing to the difficulty in dealing with nonlinear systems,the feedback linearization method is used to transform nonlinear systems into linear systems.Furthermore,the principle of feedback linearization and the advantages of the nonlinear system are introduced,and the feasibility of the feedback linearization method is proved by the calculation and simulation analysis of an example.For a class of uncertain guaranteed cost control problems of continuous-time systems,the event-triggered scheme is employed to analyze the stability of the continuous-time system with network-induced delay.When proving the stability of the system,based on delay system approach and the linear matrix inequality method,a sufficient condition for the existence of robust cost controller and triggering parameters is presented.The conclusion is that there exists an upper bound for all admissible uncertainties of a specific cost function and the upper bound value is given.For continuous-time nonlinear systems with constraints and communication delays,the dual-mode event-triggered distributed MPC scheme is proposed to solve the optimization problem,combined with the event-triggered scheme and robust constraint inequalities.Then the feasibility of the scheme and stability of this system are verified.The results show that the feasibility can be satified when the prediction horizon is proper and the external disturbance is bounded,while the stability is related to the prediction horizon,the bound of external disturbance and event-triggering condition.