The Research Of Distributed Power Grid Integration Based On Predictive Control

With the continuous development of new energy power generation technology,represented by wind energy and solar power,the proportion of which to the total power generation has continued to rise.Therefore,the demand for distributed generations connected in grid is also increasing,especially in some regions where multiple types of distributed power grids are connected.This phenomenon poses challenges to the safety,reliability,and stability of distribution network operations.So it is necessary to plan the grid connection of distributed generations rationally to ensure the safe and stable operation of the distribution network,and maximize the economic benefits generated by grid connection of distributed generations,so that policy of energy conservation and consumption reduction can be fully implemented.In this paper,wind power and photovoltaic power are used as research objects to optimize their access locations and grid-connected capacity.It is necessary to establish the mathematical model of DG and analyze its impact on power flow when it is connected to the grid.At the same time,under the premise of the constraints of actual operation,an objective function should be established for the optimization indicators of voltage stability and network loss minimization.However,generally,the establishments of the objective function base on a simple weight form to combine different objective functions.This method is more suitable for the cases with fewer objective functions.This paper adopts a two-level programming approach,which can set multiple optimization goals better for different optimization priorities,and achieve decoupling of constraints between different layers.Constraints can be selected according to actual needs.The focus of the bottom level is the distribution network stability when distributed generations are connected to the grid.Therefore,the minimum network loss and the minimum node voltage offset are selected as the optimization goals.The first-level is designed for economics,so the optimization indicators include DG investment and operating costs,renewable energy policy subsidies,and electricity purchase costs.And the purchase cost should use the net loss of bottom level,so there is a connection between the two layers.In the research of predictive control algorithms,the basic principles of predictive control are analyzed,and the disadvantages of generalized predictive control,such as the slow calculation speed and the tendency to fall into local optimal solutions,are explained.On this basis,the genetic algorithm is combined with the GPC algorithm,and the good global convergence of the GA algorithm is used to make up for the shortcomings of the GPC algorithm.In order to further improve the convergence speed of GA-GPC,GA algorithm is improved by introducing a genetic strategy based on the error classification of bisexual population mating.This strategy defines chromosomes as positive and negative.When performing crossover operations,a heterosexual chromosome is selected as the target of the crossover operation.Compared with the traditional random selection of two chromosomes for crossover,the GA algorithm based on this strategy has better search capabilities.The test function is used to verify the superiority of the improved GA algorithm performance,and then it is used to improve the rolling optimization of the GPC algorithm to solve the problem that the GPC rolling optimization may be sensitive to initial value and easily fall into local optimal solutions by adopting the gradient method.Simulation results show that the improved GA-GPC has better control accuracy,and the calculation speed is significantly improved.The planning of DG grid connection is solved based on IEEE33 node distribution network system.The simulation results show that the improved GA-GPC algorithm finds a better planning solution than traditional GA-GPC algorithm in terms of minimum network loss,node voltage offset,and economic cost.The results also indicate that the improved GA-GPC algorithm can obtain the global optimal solution,and the superiority and effectiveness of the method are also be verified.