Research On Energy Management Strategy Of Micro-grid Based On Economical Optimal MPC

With the increasingly serious problems of environmental pollution and energy shortage,the demand for electric energy increases year by year,and the contradiction between environment and resources is becoming more and more prominent.It is one of the important ways to solve the above problems to develop the low carbon new energy represented by wind and light,and to improve the permeability of renewable energy.Recently,the popular microgrid technology improves the reliability of distributed power generation,and realizes the integrated operation of distributed power supply and load through large-scale access and application of distributed power supply,especially renewable energy.Governments around the world are also encouraging large-scale promotion of renewable energy generation and research and development of new energy generation technologies.Microgrid system is a hybrid system with hybrid dynamic characteristics.Due to the intermittence of new energy,the variety of load types,the complexity of network topology and the restriction of power market,the traditional energy management strategy is difficult to meet the actual control needs.In order to ensure the safety,reliability and economy of microgrid,it is necessary to study the energy management of microgrid.In this paper,there is a difference between the real value and the numerical value of the new energy generation forecast and load forecasting in the model predictive control system.The energy scheduling scheme calculated from the predictive input signal will not satisfy the energy balance of the real system,and the strategy of adding feedforward control to the rolling closed-loop model of model predictive control is adopted.The total predictive error is obtained by real-time monitoring the actual power values of the generator and the user,and then the predictive error is allocated to each energy dispatching unit to realize the real-time power balance of the system.In this paper,the day-ahead scheduling strategy is introduced firstly,and the optimization steps of the scheduling strategy and the simulation results are given by a numerical example.Then the disadvantages of day-ahead scheduling are analyzed,and the MPC strategy is introduced by adding rolling optimization and feedback correction to the scheduling strategy,and the basic principle and structure of model predictive control are introduced in detail.Then,the steps of MPC policy optimization and the simulation results of the code program are given.By comparing the results of the two strategies,it is found that the total cost per day of the MPC strategy is 6.21% lower than the total cost of the day-ahead scheduling economic strategy.Considering the prediction error problem of the actual system,the system balance of the MPC strategy will be affected.In this paper,a simple analysis and calculation of the cost of the compensation prediction error in the traditional distribution network regulation mode(power limit or load shedding)is presented in this paper.Then the strategy of real-time monitoring power generation and load data and allocating the system total prediction error to each energy dispatching unit by feedforward control(considering system constraints such as power threshold)is proposed.The strategy is improved on the basis of the optimization step of MPC strategy.The improved MPC strategy optimization steps and program codes are obtained,and the simulation results are obtained.By comparison,the results show that the total cost per day of the improved MPC strategy is 22.69% lower than that of the MPC strategy.