Research On Control Strategy Of Networked Multi-agent Systems And Application To Power Systems

In recent years,networked multi-agent systems have become a focal research topic and attracted a great deal of attention in various fields.Many systems in nature can be modeled as networked representation,which are composed of highly interconnected dynamical units.Such as the internet net,power grid,traffic network,nerve networks,sensor networks and so on.With the fast development of modern science and technology,the control systems we are faced becoming more and more complicated,many of them are composed of interconnected dynamic systems.The dynamic behaviors of the systems will become more complicated and the traditional theory can not solve this problem.So it is necessary to develop a method to understand the dynamical behavior of networked systems,which can help to improve the stability,safety and application.Motivated by the above mentions,decentralized control strategy of multi-agent systems and multi-area multi-generator interconnected power systems were researched in the viewpoint of decentralized networked control in this dissertation.The main contributions are summarized as follow.(1)We investigated decentralized stabilization problem of uncertain multi-agent systems with mixed delays including discrete and distributed time-varying delays based on passivity stability.We designed a decentralized statefeedback stabilization scheme such that the family of closed-loop feedback subsystems enjoys the delay-dependent passivity stability for each subsystem.Then,by employing a new Lyapunov-Krasovskii function,a Linear Matrix Inequality(LMI)approach was developed to establish the delay-dependent criteria for the passivity stability of multi-agent systems.The sufficient condition was given for checking the passivity stability.The proposed LMI result is computationally efficient.An example was given to show the effectiveness of the method.(2)A novel controller with adaptive coordinator was designed in order to realize networked multi-agent systems coordination control.Firstly,networked multi-agent system was modeling based on dynamic graphs with respective to nonlinear interconnection and non-avoidable time-varying delay.Then,applying distribution control policy,coordinator with adaptive parameter estimation was designed to obtain stable desired state by adjusting interconnection level.Error control system asymptotic stability was proved.This method can complement parameter estimation and coordination control simultaneously under uncertainty and own advantages of simple and easy to complement.Simulation example was given to verify the effectiveness of the proposed method.(3)We investigated adaptive synchronization control between two networked multi-agent systems with non-identical topological structures and unknown parameters.A novel controller with adaptive coordinator was designed in order to realize synchronization control.Applying decentralized control policy,coordinator with adaptive parameter estimation was designed to obtain stable desired state by adjusting interconnection level.Error control system asymptotic stability was proved.The designed controller can easily be implemented in practice.An illustrative example was given to demonstrate the effectiveness of the present method.(4)A pair-wise decomposition and decentralized coordinated control method for multi-area interconnected power system Automatic Generation Control(AGC)under structure perturbation was presented.Based on the permuted inclusion principle,AGC model was expanded into an expanded space,from which a recurrent reverse order of pair-wise subsystems was extracted.By a standard LQG control,pair-wise controllers can be designed to stabilize and coordinate the pair-wise subsystems.The cases of load perturbation and structure perturbation that is some area disconnected from or reconnected to the system have been considered.The effectiveness of the proposed methodology was demonstrated by an application of automatic generation control to two-area.three-area and four-area electric power system respectively.(5)Decentralized networked feedback control laws were designed for multi-area multi-generator interconnected power system AGC which controlled over networks in the power market.The dynamic model of interconnected power system AGC was treated as an interconnected system with overlapping states.Using the mathematic framework of the inclusion principle,the interconnected system was expanded into a higher dimensional space in which the subsystems appear to be disjoint.Then,for the locally extracted subsystems with respect to network-induced time-delay and packet-dropout,by modeling the NCS as ADS with rate constraints on events,the state feedback controllers were designed and the sufficient exponential stability criterion was derived.The design procedure was based on LMI.As a final step,the decentralized controllers were contracted back to the original space for implementation.A 2-generator simulation example was given to show the effectiveness of the method.(6)Considering a model of an n-machine interconnected power system,we refer the system motion to the motion of one of the machines as the comparision machine,and then perform a pairwise decomposition of the model into n-1 interconnected second-order subsystems.Each subsystem is of the Lure-Postnikov type with one nonlinearity.A distributed coordination controller is put forwarded to guarantee the local stability of subsystems.A simple analysis produces estimates of the stability regions for each disconnected subsystems.Then,we construct an aggregate model involving a vector lyapunov function and using the subsystem-region estimates to determine the stability of the overall system.This judging method simplifies the complexity of stability analysis and reduces the computation load.