A Distributed Predictive Control Strategy Based on State Estimator for islanded microgrid

The traditional droop control is usually used to share real and reactive power without communication for fully decentralization power system. Because of that the traditional droop control has its inherent limitations, this paper presents a modified droop controller based on Kalman filter to reduce load effect, and then utilizes a distributed consensus algorithm to transmit error voltage to other distributed generation (DG) with sparse communication network, and last make use of distributed model predictive control (MPC) to achieve accurately load power sharing between multi-parallel inverters. Simulation results is verified the reliability of the proposed algorithm in realizing load sharing and the voltage-frequency restoration.