Consensus Algorithms Of Multi-Agent Systems And Its Application On Micro-Grid

Coordination control of multi-agent systems has attracted great attention recently in both theoretical and application areas.Consensus problem is one of the most important concerns in control of multi-agent systems(MAS).System consensus means agents' states reach to the common value by following certain consensus protocols which utilize local and neighboring information through network communication.Consensus algorithms has been utilized in many practical application,one of which is microgrid control.This thesis investigates consensus problems of multi-agent systems with different dynamics under different conditions,based on the matrix theory,Lyapunov stability theory and frequency domain analysis method.The content of the research involves the theoretical research of the consensus algorithm,and the application of the algorithm in microgrid systems.Chapter 3,4 and 5 study the consensus algorithm under different system models.Chapter 6 and 7 study secondary frequency regulation and energy management of microgrid under hierarchical control framework,by applying consensus protocols of MAS.The theoretical analysis of the Chapter 3,4 and 5 lay a theoretical foundation for the analysis of the stability of microgrid system under different network conditions.The main contents and contributions of this thesis are listed as follows.1)The thesis investigates the consensus problem of multi-agent systems subject to time delays arisen from communication and actuation,together with switching topologies.It is shown that by utilizing delayed local and neighboring information only,the proposed protocol is capable of ensuring system consensus in the presence of time delays in control input and state feedback,even when delays are nonuniform and time varying.The sufficient condition to guarantee the consensus is provided.The proposed algorithm is validated by a Lyapunov stability approach.2)This thesis studies the consensus problem for multiagent systems with nonlinear dynamics and time delays.A distributed adaptive consensus protocol is proposed in which the time delays are explicitly included in the adaptive algorithm.It is shown that the resultant closed loop system involves doubly larger time delays,making the stability analysis nontrivial.Stability condition on maximum tolerable time delay is established and controlled by the proposed two-hop adaptive algorithm.The explicit expression of the delay margin is derived and analyzed using a frequency domain method.3)The thesis investigates the consensus problem of nonlinear second-order multiagent systems with unknown disturbance.An adaptive consensus protocol is proposed using only delayed information for multi-agent systems with bounded disturbance and varying velocities.In this protocol,by estimating the control gains,the agents' velocities follows the variation of the leader agent or the average velocity in network with or without leaders respectively.The upper bound of the disturbance signals are estimated as well for system stability.Both the control input and the adaptive laws use only local and neighboring information.The stability of the system is validated by a LaSalle's method.4)The thesis explores the classic primary droop controllers proposed for inverterbased power grids,and analyzes closed loop frequency dynamics using Kuramoto model.A fully distributed adaptive control scheme is proposed in the secondary level of microgrid control to regulate the frequency and maintain the real power sharing.By estimating the parameters on power lines adaptively,the distributed generators do not need prior knowledge of the microgrid system.The adaptive nature of the proposed controller yields a stable performance of the paralleled inverter system.5)The thesis studies the energy management of microgrid with distributed generators.A simple and distributed ratio consensus scheme based on average consensus is presented for microgrid with both undirected graphs and unbalanced directed graphs.Average consensus is guaranteed in unbalanced digraphs by updating the weight matrix with both its row sums and column sums being 1.With this algorithm,real power is shared proportional among distributed generators according to their nominal powers.Finally,conclusion remarks are made and further research worth studying are presented.