Research On Multi-level Robust Control Method For Distribution Network With Microgrids

With an increasing number of distributed generations(DGs)connecting to power systems,active distribution network has become a significant type of regional power grids,and its reliability and efficiency exerts direct influences on the security and economy of power grids.Although DGs integrated into grids have made the operation of power systems more complicated,such as the fluctuation of power flow and the dispersion of connection points,many power electronic devices make control tools diversified,enhancing the flexibility of grids operation.In this context,rationally integrating advanced communication tools,power electronic devices and intelligent control methods to build smart control systems in regional power grids can provide an effective scheme of efficient utilization of DGs,which is meaningful to support low-carbon economic development.Active distribution network features high-level uncertainty,so establishing intelligent control systems with strong robustness in active distribution networks is a vital way to construct strong smart grid.This thesis conducts research from two aspects,namely microgrid distributed control and active distribution network centralized control in regional power grids with various DGs,main research results show as follows:1.In the second section,to address the nonlinearation of DG models in microgrid,state feedback linearation is applied based on Lie derivative.Then,an extended state observer is designed to reduce the negative effects caused by model uncertainty and measurement noise,which also simplifies the implementation of measurement devices.The effectiveness of this observer system is verified at the end.2.In the third section,the sparse communication network is built based on undirected graph,and microgrid distributed voltage robust control model is established based on fast terminal sliding mode(FTSM)model to improve the convergence rate.Then,the FTSM model of voltage control solved by adaptive super-twisting(AST)model enhances the convergence rate also and simplifies the selection of control algorithm.Moreover,the distributed frequency control is designed considering active power sharing.In the end,several cases are designed to demonstrate the effectiveness,robustness and scalability of proposed control method.3.In the fourth section,based on sensitivity method and multicellular model,a reactive-voltage control model in active distribution network is established to address the uncertainty of DGs and loads.Then,overvoltage problem is solved by optimal robust control law based on H_? robust performance,and the adaptive cubic exponential smoothing algorithm compensates for the uncertain dealy of system-level control caused by mode communication delay and closed-loop control.Simulation cases are designed to illustrate the validity and applicability of proposed voltage robust control in the active distribution network.