Multi-time Scale Optimization Scheduling Of AC-DC Distribution Network Based On Model Predictive Control

With the emergence of energy shortage,people begin to look for new alternative energy sources,and the flexibility of distributed power source has attracted great attention.Distributed power supply is mostly DC,so DC grid can absorb distributed power supply well.The actual operating load is mostly AC load,in order to give consideration to both,AC-DC hybrid distribution network emerges.The running state of the system needs to be calculated by power flow.Accurately evaluating the running state of the system can better optimize the scheduling of the system.The study of optimal dispatching will play a positive role in the safe and economic operation of the system,reduce the operation cost of distribution network and improve the reliability of distribution network operation.This paper focuses on the study of power flow calculation and optimal scheduling of AC/DC hybrid distribution network.In this paper,the power output models of various distributed power supply and AC and DC distribution network power flow models are established,and the loss and control mode of voltage source converter(VSC)are studied.In order to deal with the nonlinearity of the power flow equation,the nonlinear power flow equation is transformed into linear power flow equation by a second-order cone,and the alternating iteration method is applied to calculate the power flow of AC-DC hybrid distribution network.The effectiveness of the alternating iteration method is verified.In order to better solve the optimal scheduling problem of AC-DC distribution network,this paper proposes a multi-time scale optimal scheduling strategy for AC-DC hybrid distribution network.Day-ahead scheduling considers the influence of uncertainty,uses the scenario analysis method to analyze the uncertainty quantitatively,takes the lowest operating cost of the system as the objective function,and solves the planned output of each unit of day-ahead AC-DC hybrid distribution network.The intra-day model predictive control strategy was used,and the day-before output plan was taken as the benchmark,and the minimum deviation between day-before optimal scheduling solution value and day-ahead optimal scheduling plan value was taken as the objective function,and the intra-day optimal scheduling plan value of each unit was obtained by rolling optimization,so as to ensure the validity of day-ahead optimal scheduling plan value.In the real-time stage,the distributed model predictive control method is adopted to realize the optimal control of the whole system through the cooperative control of each sub-region.The improved IEEE33 node is simulated by MATLAB,and the effectiveness of the proposed method is verified.