On the basis of the theory of consistence and model predictive control

this paper proposes a distributed optimization framework of local centralized optimization and global distributed iteration

which can solve the problem of large-scale optimization of microgrid cluster(MGC). At each iteration

each microgrid solves its own sub-problem in parallel

and Lagrangian multiplier is updated by the solution of interconnecting variables until the convergence is achieved. Meanwhile

local distributed energy levels consumed by the local microgrid can be controlled by adjusting the penalty factor

so that the MGC has a strong flexibility. Especially

the proposed distributed framework can achieve the function of plug and play of sub-microgrids by adjusting auxiliary convergence constants. The simulation results show that the proposed distributed framework can achieve the global optimal only by a small number of iterations

and prove the validity of plug and play functionality.