| Direct current(DC)microgrid is a microgrid composed of interconnected DC distributed energy resources,loads and energy storage systems.It plays a significant role in advancing energy conservation,emissions reduction,and sustainable energy development as an integral component of future intelligent power distribution systems.Coordination of distributed generation units is es-sential for the stability of DC microgrids.Therefore,a mature control scheme is needed to achieve their design objectives.Distributed control strategies are one of the methods used in current sharing control in DC microgrids.However,the existing approaches mainly treat the DC microgrids as simple linear multi-agent systems,and the obtained voltage regulation and current sharing condi-tions are somewhat conservative.To gain a deeper understanding of current-sharing control in DC microgrids from a modern control perspective,this thesis proposes a more general nonlinear cascaded multi-agent system model based on the linearized DC microgrid model.Additionally,a distributed hierarchical hybrid control framework is established,which takes into account the adverse network-induced factors,such as the aperiodic sampling,the network-induced delays,and the mixed network attacks.Finally,the designed control algorithm-s are applied to DC microgrids,and simulations are conducted to validate the rationality and scientificity of the proposed theoretical framework.The main research contents are listed as follows:(1)For the scaled consensus problem of cascaded multi-agent systems,a hierarchical control framework including lower-layer continuous time con-trol and upper-layer distributed impulsive control is proposed.By applying the separation principle and the weighted discontinuous Lyapunov function,the consensus problem of heterogeneous multi-agent systems is decoupled into a lower-layer tracking problem for isolated subsystems and an upper-layer con-sensus problem for homogeneous multi-agent system.The sufficient conditions for scaled consensus in multi-agent systems are derived in the form of linear matrix inequalities.As an application,a relaxed distributed hybrid control al-gorithm is introduced for the current-sharing control in DC microgrids,which eliminates the balance requirements of the communication graphs and reduces the conservatism of current sharing conditions.(2)Considering the communication delays in the network,the multi-objective scaled consensus problem in cascaded multi-agent systems is investi-gated within the aforementioned hierarchical control framework.A novel sys-tem augmentation approach is presented to model the error system with impulse delays as a delay-free augmented error system.A switching-based time-varying Lyapunov function is employed to handle the switching impulses of the aug-mented system.As an application of the proposed distributed control scheme,a new hybrid current-sharing control algorithm of the DC microgrid against communication delays is designed.Through the various communication delays tests,load tests and limited information tests in an islanded DC microgrid,it is verified that the proposed delayed control scheme can effectively reduce the impact of communication delays on voltage regulation and current sharing.(3)For the cascaded multi-agent systems subject to replacement decep-tion attacks,a secure distributed hybrid controller is designed within the above hierarchical control framework to address the consensus control problem.The discrete-time deception signals are treated as bounded external disturbance se-quences,and the attacks are modeled by a set of Bernoulli processes.A nov-el mean-square exponential input-to-state stability(MSEISS)analysis method combined with the use of a weighted discontinuous Lyapunov function is de-veloped to to establish the MSEISS criteria of the consensus error system.The robustness performance against randomly occurring deception attacks with re-spect to mean-square quasi-consensus is quantified through MSEISS gains.The proposed resilient control algorithm is applied to the current-sharing control of the DC microgrids under deception attacks,and the excellent performance of the algorithm against deception attacks is verified by load tests.(4)For the secure consensus in cascaded heterogeneous multi-agent sys-tems subjected to mixed network attacks of denial-of-service and false data injection,a distributed hybrid control strategy is proposed within the afore-mentioned hierarchical control framework.Two sets of mutually independent Bernoulli probability distributions are employed to model denial-of-service at-tacks and false data injection attacks,respectively,while discrete-time false data is treated as bounded external disturbances.A weighted time-varying Lyapunov function based on MSEISS analysis is introduced for the stability analysis of error dynamics,and the MSEISS criteria of the error system are obtained.The proposed attack-resilient distributed hybrid control algorithm is applied to the current-sharing control of a DC microgrid,and the robustness performance of the algorithm against mixed-network attacks is verified by load tests and plug-and-play tests.(5)For the H_∞consensus problem of the multi-agent system subject to external disturbances and input saturation,a distributed hybrid controller is de-signed based on the above hierarchical control framework.By constructing a weighted time-varying Lyapunov function and exploiting the convex hull prop-erty combined with saturation characteristics,sufficient conditions for the H_∞consensus of the controlled error system are derived.Additionally,estimations of the attraction domain are provided.To estimate the attraction domain,appro-priately shaped ellipsoidal sets are introduced as reference sets,and a convex optimization-based algorithm for attraction domain computation is presented.The developed distributed control algorithm is applied to the current-sharing control in a saturated DC microgrid,and the effectiveness and feasibility are verified through controller saturation tests and load tests. |
PDF Full Text Request