Research On Microgrid Energy Management System Based On Predictive Control

In modern society,under the background of increasing environmental pollution,the depletion of non renewable energy and the rapid development of society,the power grid industry is constantly developing and exploring towards high efficiency,intelligence and sustainability.In the power grid industry,the development of microgrid technology not only opens up new fields,but also improves the stability and reliability of power supply of distributed generation system.Through the wide application of distributed new energy and the use of non renewable energy as fuel(i.e.fossil fuel energy),distributed generation and load complement each other,and energy can be used more effectively through integration.This is the current trend of sustainable development of power grid in the future,and also an important means to achieve efficient use of renewable energy and extensive access to the distribution network.However,due to a variety of technical problems,the traditional energy management strategy is inadequate,and the actual control requirements and power grid security can not be guaranteed.MPC(model predictive control)has many advantages,such as wide range of model application,strong compatibility,can effectively deal with a large number of constraints,realize the feedback mechanism of closed-loop control,these advantages make it can be applied to the energy management and control of microgrid system.The importance of economy is self-evident in the whole energy management system of microgrid.Therefore,this paper mainly studies how to improve the economy of microgrid through MPC.In MPC,the data obtained from load forecasting and generation forecasting have errors.Therefore,the energy optimal dispatch directly using the forecasting data in MPC is not the optimal solution.This paper first introduces the principle and structure model of MPC.By adding a feed-forward control to the rolling closed loop of MPC,the power values at both ends can be obtained at any time,and the real-time prediction error can be calculated with the power values.The prediction error data obtained is sent to each scheduling unit to help the system achieve the optimal power balance.The energy coordination management of microgrid is a nonlinear and variable optimization problem,which also has strong coupling.The controller must consider both continuous and discrete dynamic characteristics.The optimal operation of the system is affected by many factors,such as battery combination,economic dispatch,energy storage life and degradation factors,power purchase and sale of large power grid,load reduction scheme,etc.In this paper,the structure model of microgrid system based on model control strategy is established,including energy storage part and load cost model,interaction model with large power grid,micro source model of controllable generation,objective function aiming at total operation cost of the system.By introducing state space model prediction algorithm and hybrid logic dynamic control system,the original objective function and constraints are simplified The micro grid model and optimization constraints are analyzed comprehensively.The model is built by yalmip toolbox in Matlab platform and the objective function of minimum running cost is solved.In view of the shortcomings of the day ahead scheduling strategy,such as the poor system disturbance and the relative localization of the optimal solution,combined with the example,the MPC control strategy is compared with the control strategy of the same condition,and the total operating cost of the battery is converted to the total price.The total cost of the MPC strategy is 7.55% lower than that of the day ahead scheduling strategy.The effectiveness of the MPC control strategy of the positive link.Considering that the actual prediction error will affect and fluctuate the energy balance of the model control strategy system,the system will monitor the data of the load side and the generation side in real time,and allocate the total prediction error of the system to each energy dispatching unit(considering the constraints of the upper limit output power of the system inverter)through the feed-forward control method.The MPC strategy optimization is improved,and the multi-level control strategy is adopted The optimization steps and procedures of time scale predictive control strategy are solved by yalmip toolbox in Matlab platform.The total cost of time scale predictive control strategy in one day is 22.76% lower than that of model predictive control strategy.