Abstract：A microgrid optimization scheduling model considering dynamic network reconstruction was

proposed for the parallel and off grid scheduling problem of microgrids. The proposed model was based on Benders

decomposition and decomposes the optimal scheduling problem of microgrids into a grid-connected operation

master problem and an islanded operation subproblem. When the grid-connected operation scheduling plan cannot

guarantee the feasibility of island operation，the network topology structure，unit output，and scheduling plan of

energy storage and adjustable loads in the main problem can be dynamically adjusted to meet the relevant

constraints of the subproblem. A new linearized power flow model for distribution networks that considers network

reconstruction was proposed to support the solution of the proposed optimized scheduling model. A random

framework based on unscented transform（UT）was introduced to make the proposed model more in line with the

actual power grid situation. The effectiveness and superior performance of the proposed model were verified

through simulation analysis of the IEEE33 node test network.

      Key words：microgrid；optimization scheduling；network reconstruction；linear power flow model；unscented

transform（UT）