Distributed Peer-to-Peer Control For Renewable Power Generation DC Microgrid

The global energy and environmental crisis is boosting the exploitation of renewable energy,such as wind and solar energy.Microgrids provide an effective framework for organizing the distributed renewable sources connected into power systems.DC microgrids may offer high conversion efficiency,power capacity and system flexibility,and therefore are receiving increasing attention from both the academic and industrial community.This dissertation focuses on the operation control of renewable DC microgrid systems.Based on peer-to-peer control methodology,a series of self-disciplined,self-adaptive,and self-organized control solutions are proposed,to enable plug-and-play operation of diversified sources in a varying structure and open boundary.Firstly,the concept of self-disciplined system stabilization is proposed.It adopts a decentralized stabilization methodology to tackle the stability challenge of plug and play.The basic idea is to implant a self-stabilization function into each distributed terminal,so that the stability of the entire microgrid could be ensured in a varying structure of plug-and-play interconnection.To realize this concept,the passivity-based control theory is employed due to its robustness in structural variation.A passive admittance region is set as a self-disciplined stability criterion,based on which a passive margin criterion is derived to further enhance stability margin.It is also revealed that the proposed passivity-based self-disciplined criterion is essentially a distributed version of the conventional impedance-ratio phase-margin criterion.It separates the overall stabilization duty corresponding to microgrids' decentralized nature.For practical application of the proposed concept,a voltage feed-forward technique is designed to render a terminal passive by damping injection in converter control.Secondly,a self-adaptive power allocation strategy is proposed.Mode-adaptive and dynamic-adaptive control schemes are designed for peer-to-peer allocation of both static and dynamic power among diversified sources with heterogeneous feature.Extensive information is extracted from DC microgrid voltage to achieve local adaptability without external data exchange.In the mode adaptive control,the static level of DC voltage is used to designate microgrid modes and trigger adaptive mode transitions.In the dynamic adaptive control,the dynamic variation of DC voltage is used to share transient power according to the intrinsic response speed of different sources.These methods enable self-adaptive power allocation to fully exploit the versatility of distributed resources without extra communication,and thereby ensure both reliable and flexible power balance in plug-and-play operation.Thirdly,a self-organized energy management strategy is introduced.A peer-to-peer broadcasting communication network is used as an information sharing pool.Based on this pool,distributed terminals use decentralized negotiation to make cooperative decisions for energy management in an open boundary under plug-and-play condition.According to different timescales,the energy management tasks are divided into three aspects,namely power quality,real-time dispatching,and day-ahead scheduling.The power quality management uses the consensus regulation of each terminal to refine voltage profiles within several seconds.The real-time dispatching management uses distributed self-dispatcher to maintain minute-hour-scale energy balance and adequacy by cooperation of energy storage devices and the utility grid.The day-ahead scheduling management utilizes a power-market-based multi-agent system to optimize the power schedule of the next day for economic benefits.Taking advantage of the decoupled relationship between the utility and a DC microgrid,the energy management pattern is unified for both grid-connected and islanded condition,so the complex problem of grid transition is avoided.Finally,the aforementioned strategies are integrated into a composite peer-to-peer control framework with a three-level architecture,including converter control level,voltage coordination level,and communication dispatching level.The self-disciplined system stabilization,self-adaptive power allocation,and self-organized energy management functions are embedded into the three levels respectively,which comprehensively ensures stable,reliable,and optimal operation of a DC microgrid under plug-and-play condition.A laboratory-scale DC microgrid operation test platform is built to verify and demonstrate the proposed solutions.