Distributed Predictive Secondary Control And Moving Horizon State Estimation Of Microgrid

The microgrid can improve power quality,reduce line losses,promote renewable energy consumption and has gained a lot of attention and research.The widely used droop control in the primary control of microgrid will cause the voltage and frequency of the system to deviate from the rated value,so it requires the secondary control to adjust the voltage and frequency.At the same time,the control problems in microgrid are mostly based on state feedback.However,there are noises and disturbances in system models and measurements,and sometimes some states cannot be directly measured.Therefore,we need to estimate the state of microgrid.In this paper,the principle of hierarchical control strategy of microgrid is firstly introduced.Then,the nonlinear system model of each distributed generation unit in the microgrid is established.Then the principle of input and output feedback linearization is introduced,and we apply it to process the dynamic model of DGs.A distributed model predictive controller is designed for the linearized system.The convergence and stability of the global system are analyzed,and the convergence and stability conditions are obtained.The proposed IOFL-DMPC based for the secondary control is fully distributed.It has good flexibility compared with traditional centralized control structure,which satisfies the plug and play characteristics of distributed generation in microgrid,and requires less communication.Through the simulation of microgrid,the results show that the algorithm can achieve the regulation of voltage and frequency well,and make the distributed generation units return to the rated value synchronously and uniformly,which verifies the effectiveness of the algorithm.In addition,we compare it with the common distributed coordinated control method.The simulation results show that the algorithm proposed in this paper has faster dynamic response speed,smaller variance and better control performance.As for the state estimation problem in microgrid,this paper establishes a dynamic model of the system,and introduces the traditional centralized control algorithm based on kalman filter.However,it is limited in dealing with the constraint problem,and the centralized control has large amount of computation and is sensitive to single point failure.In this paper,a distributed moving horizon estimation algorithm based on neighbor optimization is proposed to realize the distributed estimation of microgrid state.Each DG can be viewed as a subsystem and interacts with neighbor subsystem.The neighbor optimization based algorithm requires that each subsystem not only consider their own performance optimization,also consider the performance of adjacent subsystem,and obtain the local optimal solution as far as possible to converge to the global optimal solution,thus enhance the performance of distributed estimation.Finally,the effectiveness of the algorithm is verified by a microgrid system with 5 DGs.