Distributed Cooperation Of Multi-agent And Applications On Power Optimization Of Power Grid

Distributed cooperation of multi-agent system is a hot topic in control com-munity recently,and it can be used in numerous fields,such as multi mobile robot formation,sensor networks,information fusion,distributed computing and so on.As one type of fundamental problem for cooperation control,the main idea of dis-tributed consensus problem is to design appropriate algorithm such that all agents in the controlled systems agree upon convergence by information exchange with their locally neighbors.The working time of Pulse-Modulated intermittent con-trol is discontinuous,which is greatly reduces the burden on the implementation of the controller.Moreover,the control input function can be adjusted according to the control constraints.This dissertation,based on the leader-following problem of multi-agent and considered the limitations of sensors bandwidth,measuring capac-ity and communication delay,systemically and deeply investigates the consensus-tracking and multi-tracking problem of second order multi-agent systems.There-fore,the optimization problem of secondary frequency control is also investigated by using the properties of distributed consensus in the multi-agents.The distributed tracking problem is proposed in the second order multi-agent systems,and the stochastic quantization is introduced to overcome the bandwidth limitation of sensors.At the same time,two distributed pulse-modulated intermit-tent control protocol are proposed for the leader-following consensus problem under measurable and unmeasurable velocity,respectively.Based on stochastic quanti-zation,Hurwitz stability,discretization method and algebraic graph theory,some necessary and sufficient conditions are obtained to ensure consensus tracking of the controlled system,where the communication topology is a directed graph with a s-panning tree or a connected undirected graph.Specifically,some simply restrictions are derived under sample and impulse control.The achievement of consensus track-ing of second order multi-agent systems depends on not only the pulse function,connection topology and sampling interval,but also with the control duration.For multi-tracking problem with velocity unmeasurable,the stochastic quanti-zation is also introduced to overcome the bandwidth limitation of sensors.A quan-tization intermittent control protocol based on only sampled position information is proposed,and the multi-tracking problem of second order multi-agent systems with directed communication topology is investigated.Through the theoretical analysis,some necessary and sufficient conditions are derived to ensure multi-tracking of the controlled system.The convergence performance of multi-tracking can effected by communication delay,which may naturally arise in practical applications.A distributed control protocol is proposed for the multi-tracking problems of second order multi-agent systems,which is a extension problem of last question.Then,under different interval of communication delay,some necessary and sufficient conditions are obtained for achievement of multi-tracking problem.Assume the communication topology of the controlled systems contain a spanning tree,when the sampling interval,the control gain and the eigenvalues of the communication topology are satisfied certain conditions,the multi-tracking of second order multi-agent systems can be reached.Demand side frequency control is introduced for the power system frequency optimization control.Due to the capacity limitation of traditional AGC units,re-sponsive load aggregators as excellent spanning reserves can engage the frequency control when a large disturbance occurs in the system.In order to keep the stable and rated system frequency,two-step optimal model that considered with the cost of load aggregators reductions and AGC regulations is proposed.In the first step,Lambda-Iteration method is used to yield the total active power command for load group and AGC group.In the second step,?—logarithmic barrier is introduced to reconstruction of the cost function for load aggregators and AGC units.A dis-tributed optimization allocation scheme is proposed to achieve the optimal active power reductions of load aggregators and AGC unit regulations.The effectiveness,robustness and flexibility of the two-step allocation scheme are verified.More im-portantly,under the optimal control conditions,the stability and improvement of power system frequency control is achieved.Finally,the research works of this dissertation are summarized,and the further researches on distributed cooperative control and its applications are prospected.